Background
Early, sensitive, and reproducible evaluation of left ventricular function is imperative for the diagnosis of cardiac dysfunction in patients with Duchene muscular dystrophy. The aim of this study was to test the hypothesis that combining two-dimensional strain analysis with catecholamine stress could be a sensitive method for detecting early cardiac dysfunction.
Methods
Mdx (C57BL/10ScSn-Dmd mdx /J, a mouse model of DMD) and control (C57BL/10ScSn) mice were studied with conventional M-mode and high-frequency ultrasound-based two-dimensional speckle-tracking echocardiography using long- and short-axis images of the left ventricle at baseline and after intraperitoneal isoprenaline (ISO) administration (2 μg/g body weight).
Results
Conventional M-mode analysis showed no differences in left ventricular fractional shortening, wall thickness, or internal diameter at diastole between mdx and control mice before the age of 6 months. ISO increased left ventricular ejection fraction and fractional shortening to the same extent in mdx and control mice at young ages (3, 4, and 5 months). No differences in basal peak systolic strain (PSS) but increased SDs of times to PSS between young mdx and control mice were found. After ISO, PSS and percentile changes of PSS were significantly diminished in mdx mice compared with control mice at young ages. ISO increased the normalized maximum difference of times to PSS in young mdx mice but not in young control mice, suggesting that ISO reduces cardiac contractile synchrony in young mdx mice.
Conclusions
This study suggests that catecholamine stress coupled with two-dimensional strain analysis is a feasible and sensitive approach for detecting early onset of cardiac dysfunction, which is instrumental for early diagnosis of cardiac dysfunction and early treatment.
Highlights
- •
Cardiac dysfunction becomes an important contributing factor to mortality and morbidity in patients with muscular dystrophy.
- •
There is still a lack of reliable approaches to diagnose early cardiac dysfunction in patients with DMD.
- •
The combination of 2D strain analysis with ISO stress showed reduced β-adrenergic reserve and increased contractile dyssynchrony in mdx mice at a very early stage.
- •
Two-dimensional strain analysis with stress testing is a feasible and sensitive approach to diagnose early cardiac dysfunction and to evaluate treatments in patients with DMD.
- •
The treatment of cardiac dysfunction should be started at a young age in patients with DMD.
Duchenne muscular dystrophy (DMD) is a hereditary X-linked recessive disease resulting from the mutation of the dystrophin gene, which affects 1 in 3,500 male births. An early manifestation of DMD is weakness in the skeletal muscle. A weakened diaphragm may cause death in patients with DMD at young ages because of respiratory failure, but with the support of mechanical ventilators, patients with DMD can now survive much longer than they did a decade ago. Cardiomyopathy is becoming a more important cause of mortality and morbidity in patients with DMD. At rest, cardiac dysfunction can be detected during the early teenage years in patients with DMD. With more advanced technologies, such as three-dimensional strain, distortion, and dispersion analyses, cardiac dysfunction at rest can be detected as early as at the age of 8 or 9 years. Reduced cardiac reserve is an early indication of cardiac dysfunction and is evaluated with a treadmill or dobutamine stress test in the clinic. However, there have been few reports regarding cardiac reserve in patients with DMD.
Mouse models, such as x-linked muscular dystrophy ( mdx ) mice, have been used to investigate the pathophysiology of and to test new therapies for DMD. To develop a safe and effective therapy for cardiomyopathy in DMD, early, sensitive, and reproducible measurements of left ventricular (LV) remodeling and dysfunction are needed. High-frequency ultrasound has become a valuable tool to phenotype cardiac structure, hemodynamics, and global function in small animals. Furthermore, high-frequency ultrasound-based two-dimensional (2D) speckle-tracking echocardiographic (STE) imaging has been used to quantitatively measure myocardial strain and strain rate as potentially sensitive markers for early regional myocardial dysfunction in mdx mice, showing that peak systolic strain (PSS) analysis is feasible and reproducible for detecting LV dysfunction in mdx mice.
The catecholamine (e.g., dobutamine in the clinic or isoprenaline [ISO] in animal experiments) stress test has been used to evaluate cardiac reserve in many studies. Previously, using invasive in vivo intra-LV hemodynamic measurement and an ex vivo Langendorff system, we found that the cardiac β-adrenergic response was reduced in 4-month-old but not in 2-month-old mdx mice. In a 16-year-old patient with DMD, it was found that there was a reduced cardiac response to dobutamine stress, although his LV systolic function was normal at rest. However, the use of the combination of catecholamine stress testing and high-frequency ultrasound-based 2D STE imaging to detect early cardiac dysfunction in patients with DMD or animal models has not been reported. Therefore, we hypothesized that reduced abnormal β-adrenergic response in mdx mice at young ages could be detected by the combination of catecholamine stress and LV strain analysis but not by conventional echocardiography or 2D STE imaging at rest alone.
Methods
Animal Models
All mice were handled in compliance with the “Principles of Laboratory Animal Care” of the National Society for Medical Research and the “Guide for the Care and Use of Laboratory Animals” published by the National Institutes of Health (publication 85-23, revised 1996). The research protocol was approved by the Institutional Animal Care and Use Committee of Temple University. Mdx (C57BL/10ScSn-Dmdmdx/J) and control mice (C57BL/10ScSn) (The Jackson Laboratory, Bar Harbor, ME) were used at the ages indicated in the figures to study cardiac morphology and function at rest. Both male and female animals were used. Three sets of animals were used for the study. We used the first set of animals (control, n = 10; mdx , n = 8) to detect cardiac function alterations for 1 year with conventional echocardiography. The second set of animals (control, n = 7; mdx , n = 7) were used to detect changes in cardiac strain at the ages of 6, 9, and 12 months with high-frequency ultrasound 2D imaging followed by 2D STE analysis. The third set of mice included a total of 21 mdx and 21 control mice; separate sets of seven mdx and seven control mice each were studied at three separate time points (3, 4, and 5 months) for combined high-frequency ultrasound analysis and ISO stress to avoid potential carryover effect of previous ISO injection. Because the mdx mouse model is a model with mild disease progression, none of the mice died during the 1-year study period.
Echocardiography
A high-frequency ultrasound system (Vevo 770; VisualSonics, Toronto, ON, Canada) was used to acquire ultrasound images. A single-crystal mechanical transducer (RMV 707B) with a central frequency of 30 MHz (15–45 MHz) and focal length of 12.7 mm was used. The Vevo 770 machine with the RMV 707B scan head has lateral resolution of 115 μm, axial resolution of 55 μm, and up to 240 frames/sec, allowing us to acquire high-resolution images for conventional echocardiographic and speckle-tracking analysis when combined with proprietary software, TomTec Image Arena 4.0 (TomTec Imaging Systems, Unterschleissheim, Germany).
Mice were anesthetized using 3% isoflurane initially and then maintained at 1.5%. The animals were laid in a supine position on a heated platform with all legs taped to electrocardiographic electrodes for recording. Body temperature was monitored using a rectal thermometer and maintained at 36°C to 38°C. Hair was removed from the chest using chemical hair remover before imaging.
M-Mode and 2D High-Frequency Ultrasound
Eight mdx and 10 control mice were followed for conventional M-mode analysis every 2 months until the age of 12 months to evaluate cardiac morphology and function. M-mode images were obtained in the LV short-axis view at the midpapillary muscle level. The M-mode gate was set in the middle between the two papillary muscles for ejection fraction and fractional shortening analysis.
For 2D STE studies, short-axis B-mode cine loops were obtained at the mid-LV level as described above, and long-axis B-mode cine loops were recorded with the scan head placed at the mid-LV level (i.e., along the left sternal border with clear view of both the apex and the LV outflow tract). At the ages of 3, 4, and 5 months, seven mdx and seven control mice at each age were studied without reuse because they were also used for ISO stress testing. For ages of 6, 9, and 12 months, seven mdx and seven control mice were followed serially. The field of view was adjusted to maximize the display resolution of the region of interest and the frame rates (up to 160 frames/sec for short-axis images and 120 frames/sec for long-axis images). Image gain and other parameters were adjusted carefully to delineate all myocardial segments. All images were acquired for multiple cardiac cycles (CCs) and stored digitally in the hard drive for offline analysis.
LV remodeling and global functional measurements, including LV end-diastolic dimension, LV end-systolic dimension, LV ejection fraction (LVEF), and LV fractional shortening (LVFS), were derived from M-mode echocardiography using high-frequency ultrasound by averaging three consecutive and stable CCs.
High-Frequency Ultrasound with ISO Stress
High-frequency ultrasound with ISO stress was performed for mdx and control mice at the ages of 3, 4, and 5 months. At each age, seven control and seven mdx mice were used. Mice were not reused after ISO stress, to avoid potential complications. After long- and short-axis M-mode images and B-mode cine loops of the left ventricle were acquired at baseline, ISO was injected intraperitoneally (2 μg/g body weight). The animals were fully anesthetized to avoid pain response during ISO injection, and their heart rates (HRs) ranged from 400 to 450 beats/min. The relatively low HRs also allowed us to obtain more images in each CC and thus perform better strain analysis. Images were acquired every minute for up to 14 min after ISO injection. Long-axis cine loops and M-mode images of the left ventricle were acquired at 1, 3, 5, 7, 9, 11, and 13 min after injection. Short-axis cine loops and conventional M-mode images of the LV were acquired at 2, 4, 6, 8, 10, 12, and 14 min after injection. The images and cine loops showing maximum responses to ISO were used for stress response analysis.
Offline Strain Analysis
Because cardiac muscle degeneration and cardiac dysfunction are heterogeneous in different regions of the heart in patients with DMD, and there has been thus far no report on epicardial and endocardial strains in mdx mice, we analyzed endocardial and epicardial strains in different regions of the heart, including six segments of short-axis cine loops and six segments of long-axis cine loops. B-mode cine loops of the short and long axes recorded with a 707B scan head (15–45 MHz, 100–240 frames/sec) of a Vevo 770 System were converted to Digital Imaging and Communications in Medicine format without compression and imported into TomTec Image Arena 4.0 for offline strain analysis. TomTec Image Arena has been widely used for vendor-independent cardiac strain and time-to-peak analysis. All analyses were based on at least three successive and stable CCs. Longitudinal, circumferential, and radial global PSS values of the LV endocardium and epicardium were obtained by manual tracing, editing, and semiautomated tracking of LV myocardial deformation. The software automatically divides the long-axis LV images into six segments: basal anterior, mid anterior, apical anterior, apical posterior, mid posterior, and basal posterior ( Supplemental Figure 1 A, available at www.onlinejase.com ). The software also divides the short-axis images into six segments: anterior, anterior septum, posterior septum, inferior, posterior, and lateral ( Supplemental Figure 1 B, available at www.onlinejase.com ). Strain measurements in each segment were averaged, and the PSS values of all segments were further averaged to obtain PSS. Five PSS measurements were determined: longitudinal PSS of the endocardium (L-Endo), longitudinal PSS of the epicardium (L-Epi), circumferential PSS of the endocardium (C-Endo), circumferential PSS of the epicardium (C-Epi), and radial PSS ( Supplemental Figures 2 and 3 , available at www.onlinejase.com ). L-Endo and L-Epi were analyzed with long-axis cine loops and C-Endo, C-Epi, and radial PSS were obtained from short-axis cine loops obtained at the mid-LV level. LVEF, LVFS, and PSS measurements were made under baseline and ISO stress conditions. Maximum values for LVEF, LVFS, and PSS after ISO injection were selected for statistical analysis. Percentage increase in PSS was calculated as (difference between the measurements after and before ISO injection/measurement before ISO injection) × 100%. To determine dyssynchrony of LV segments, the maximum time difference between times to PSS of segments (ΔT peak,max ) was determined. The SD of times to peaks of six segments of each heart were normalized to the CC (i.e., RR interval) to further characterize the dyssynchrony of contractions. The person performing the echocardiographic analysis was blinded to the mice population.
Intraobserver and Interobserver Variability
A total of 20 data sets at rest and 20 data sets during stress were selected, and measurements were made by two blinded, experienced observers to determine intra- and interobserver variability. To determine intraobserver variability, one observer analyzed and measured the strain data (PSS) from rest and stress echocardiographic results twice at a 4-week interval, while blinded to the results of the first measurements. A second observer, blinded to the results of the first observer, also analyzed and measured the strain data independently to determine interobserver variability. The intra- and interobserver variabilities are expressed as percentile differences: (absolute difference between the measurements/mean of the measurements) × 100%.
Statistical Analysis
Continuous variables are expressed as mean ± SEM. The difference between two continuous variables was tested using a t test with a two-tailed distribution or one-way or two-way analysis of variance with or without repeated measurements. Bonferroni adjustment was done for comparison between pairs for analyses of variance. Statistical significance was defined as P < .05. All statistical analyses were done using SPSS version 17.0 (SPSS, Chicago, IL).
Results
Conventional M-Mode High-Frequency Ultrasound Analysis in Combination with ISO Stress Testing Cannot Detect Cardiac Dysfunction in Young mdx Mice
We first used conventional M-mode analysis to determine cardiac function in mdx and control mice every 2 months from the age of 2 months until the age of 12 months ( Figure 1 ). This technique only detected reduced cardiac dysfunction in mdx mice at rest after the age of 8 months ( Figure 1 B), while diastolic LV posterior and septal wall thickness increased ( Figures 1 C and 1D). Diastolic LV internal diameter decreased slightly because of the increases in LV wall thickness in mdx mice seen at the age of 8 months ( Figure 1 ). These results agree with those of previous studies.
Because altered β-adrenergic reserve is an early indication of cardiac dysfunction, we tested if young mdx mice (ages 3, 4, and 5 months) had reduced cardiac reserve with a catecholamine (intraperitoneal ISO 2 μg/g body weight) stress test assessed by conventional M-mode echocardiography. During stress, LVEF and LVFS increased gradually after injection and reached peak values within 3 to 7 min (4.25 ± 1.16 min) for all mdx and control mice. ISO significantly increased HR, LVEF, and LVFS in both control and mdx mice at the ages of 3, 4, and 5 months. After ISO, there were no differences in HR, LVEF, and LVFS between control and mdx mice of the same age ( Figure 2 ).
High-Frequency Ultrasound STE Imaging Does Not Detect Cardiac Dysfunction in Young mdx Mice at Baseline
It has been reported that cardiac strain analysis is a more sensitive method for detecting local and global cardiac dysfunction. We used this method to determine if it could reveal a difference between young mdx and control mice. At the ages of 3 and 4 months, there was no difference in PSS between mdx and control mice. PSS started to decrease in mdx mice at the age of 5 months, and the differences in L-Endo and radial PSS reached statistical significance at the age of 6 months, progressing with age. Differences in L-Epi and C-Endo between mdx and control mice were significant at the ages of 9 and 12 months ( Figure 3 ). The impairment of strain in mdx mice appeared earlier and more severe in the endocardium than in the epicardium, indicating that endocardial PSS values were more sensitive measurements of regional cardiac dysfunction in mdx mice. These data show that strain analysis is able to detect cardiac contraction differences between control and mdx mice only after the age of 6 months.
PSS in Young mdx Mice Had Reduced Response to ISO
During stress, LVEF, LVFS, and PSS increased gradually after injection, and all reached peak values between 3 to 7 min (4.25 ± 1.16 min) for all mdx and control mice. Endocardial and epicardial PSS in longitudinal, circumferential, and radial directions increased significantly in control and mdx mice with ISO stimulation ( Figure 4 ). In control mice aged 3 to 5 months, age did not significantly change PSS at baseline or after ISO administration. ISO also significantly increased PSS in young mdx mice, but the increases in most PSS were less in mdx mice than in control mice except for L-Epi and C-Epi at the age of 3 months. From the age of 3 to 5 months, PSS responses to ISO in mdx mice tended to decrease with age. For L-Endo and radial PSS during stress, significant differences between mdx and control mice appeared in all age groups from 3 to 5 months ( Figure 4 ). However, for L-Epi and C-Endo, the differences between mdx and control mice after ISO were not significant at the age of 3 months but reached significance at the age of 5 months. For C-Epi, differences between mdx and control mice could not be detected in the 3-, 4-, and 5-month groups.
Percentage Increases in LV PSS Were Less in Young mdx Mice Compared with Young Control Mice
Because there were subtle differences in baseline PSS between mdx and control mice, we determined the percentage increase in LV PSS to better compare the effects of ISO ( Figure 5 ). There were significantly lower percentage increases in L-Endo, C-Endo, and radial PSS in mdx mice than in control mice aged 3, 4, and 5 months ( Figures 5 A, 5C, and 5E). The percentage increases in L-Epi and C-Epi were not significantly different between mdx and control mice at the ages of 3 and 4 months. At the age of 5 months, the increases of these two PSS values reached statistical significance between mdx and control mice ( Figures 5 B and 5D). These data suggest that the combination of 2D strain analysis and ISO stress can detect reduced cardiac functional reserve in mdx mice starting at a very young age (3 months). L-Endo, C-Endo, and radial PSS were more sensitive than L-Epi and C-Epi for distinguishing between control and mdx mice.
β-Adrenergic Stimulation Decreases Myocardial Contraction Synchrony in Young mdx Mice
To determine if there is dyssynchrony of myocardial contractions at baseline or after β-adrenergic stimulation in young mdx mice, SDs and maximum differences of times to peak of strains (ΔT peak,max ) were analyzed for the epicardium and endocardium in the longitudinal, circumferential, and radial directions. At baseline, no differences in ΔT peak,max in all directions were observed between 3-, 4-, and 5-month-old control and mdx mice. ISO stimulation did not significantly alter ΔT peak,max in control mice in all three age groups or in 3-month-old mdx mice. However, ISO significantly increased L-Endo and C-Endo ΔT peak,max in 5-month-old mdx mice, suggesting that ISO increases dyssynchrony in 5-month-old mdx mice ( Supplemental Figure 5 , available at www.onlinejase.com ). Because ISO increased the HR and reduced CC duration, we normalized ΔT peak,max to CC duration. Again, ISO did not significantly change ΔT peak,max /CC in control mice but significantly increased L-Endo, C-Endo, and radial ΔT peak,max /CC in 4-month-old mdx mice and ΔT peak,max /CC in all directions (L-Endo, L-Epi, C-Endo, C-Epi, and radial) in 5-month-old mdx mice ( Figure 6 ). It seems that the normalization of ΔT peak,max to CC is a more sensitive parameter for evaluating myocardium dyssynchrony, and ISO increases myocardium dyssynchrony in mdx mice. The early contraction or delayed contraction was not restricted to a specific region in the midlevel short-axis and long-axis videos ( Supplemental Tables 1–3 , available at www.onlinejase.com ).
We also studied the dyssynchrony of PSS by analyzing the SD of time to peak (SD Tpeak ) of PSS. Baseline SD Tpeak was not significantly changed in control mice but increased in mdx mice during the growth from 3 to 5 months, reaching statistical significance between these two groups of mice at the age of 5 months for all PSS values and at the age of 4 months for L-Endo. ISO decreased SD Tpeak in all mice but did not further enhance the difference in SD Tpeak between control and mdx mice ( Supplemental Figure 6 , available at www.onlinejase.com ). When SD Tpeak was normalized to CC (SD Tpeak /CC), there were significant differences in the SD Tpeak /CC of L-Endo and L-Epi PSS at the age of 4 months and of all PSS at the age of 5 months between control and mdx mice at baseline or after ISO ( Figure 7 ).
Intraobserver and Interobserver Variability in STE Measurements
Intraobserver and interobserver variabilities for specific PSS values were <10%, as shown in Table 1 and the Bland-Altman plots in Supplemental Figures 7 and 8 (available at www.onlinejase.com) . The intraclass correlation coefficients for intraobserver and interobserver observations were all >0.90.
| Strains | Rest ( n = 20) | Stress ( n = 20) | ||||||
|---|---|---|---|---|---|---|---|---|
| Intraobserver | Interobserver | Intraobserver | Interobserver | |||||
| Variation (%) | ICC | Variation (%) | ICC | Variation (%) | ICC | Variation (%) | ICC | |
| L-Endo | 5.30 ± 3.16 | 0.965 | 6.58 ± 4.43 | 0.947 | 5.57 ± 3.95 | 0.958 | 6.79 ± 4.73 | 0.921 |
| L-Epi | 6.09 ± 4.41 | 0.932 | 7.08 ± 3.85 | 0.970 | 6.21 ± 4.62 | 0.927 | 7.52 ± 4.74 | 0.960 |
| C-Endo | 5.74 ± 3.26 | 0.936 | 6.64 ± 3.33 | 0.958 | 6.13 ± 4.12 | 0.931 | 6.91 ± 4.02 | 0.951 |
| C-Epi | 6.82 ± 4.07 | 0.900 | 7.06 ± 3.91 | 0.960 | 7.12 ± 4.62 | 0.920 | 8.16 ± 4.59 | 0.934 |
| Radial PSS | 5.73 ± 3.84 | 0.929 | 6.54 ± 3.71 | 0.956 | 5.94 ± 3.80 | 0.932 | 7.88 ± 4.68 | 0.917 |
Results
Conventional M-Mode High-Frequency Ultrasound Analysis in Combination with ISO Stress Testing Cannot Detect Cardiac Dysfunction in Young mdx Mice
We first used conventional M-mode analysis to determine cardiac function in mdx and control mice every 2 months from the age of 2 months until the age of 12 months ( Figure 1 ). This technique only detected reduced cardiac dysfunction in mdx mice at rest after the age of 8 months ( Figure 1 B), while diastolic LV posterior and septal wall thickness increased ( Figures 1 C and 1D). Diastolic LV internal diameter decreased slightly because of the increases in LV wall thickness in mdx mice seen at the age of 8 months ( Figure 1 ). These results agree with those of previous studies.
Because altered β-adrenergic reserve is an early indication of cardiac dysfunction, we tested if young mdx mice (ages 3, 4, and 5 months) had reduced cardiac reserve with a catecholamine (intraperitoneal ISO 2 μg/g body weight) stress test assessed by conventional M-mode echocardiography. During stress, LVEF and LVFS increased gradually after injection and reached peak values within 3 to 7 min (4.25 ± 1.16 min) for all mdx and control mice. ISO significantly increased HR, LVEF, and LVFS in both control and mdx mice at the ages of 3, 4, and 5 months. After ISO, there were no differences in HR, LVEF, and LVFS between control and mdx mice of the same age ( Figure 2 ).
High-Frequency Ultrasound STE Imaging Does Not Detect Cardiac Dysfunction in Young mdx Mice at Baseline
It has been reported that cardiac strain analysis is a more sensitive method for detecting local and global cardiac dysfunction. We used this method to determine if it could reveal a difference between young mdx and control mice. At the ages of 3 and 4 months, there was no difference in PSS between mdx and control mice. PSS started to decrease in mdx mice at the age of 5 months, and the differences in L-Endo and radial PSS reached statistical significance at the age of 6 months, progressing with age. Differences in L-Epi and C-Endo between mdx and control mice were significant at the ages of 9 and 12 months ( Figure 3 ). The impairment of strain in mdx mice appeared earlier and more severe in the endocardium than in the epicardium, indicating that endocardial PSS values were more sensitive measurements of regional cardiac dysfunction in mdx mice. These data show that strain analysis is able to detect cardiac contraction differences between control and mdx mice only after the age of 6 months.
PSS in Young mdx Mice Had Reduced Response to ISO
During stress, LVEF, LVFS, and PSS increased gradually after injection, and all reached peak values between 3 to 7 min (4.25 ± 1.16 min) for all mdx and control mice. Endocardial and epicardial PSS in longitudinal, circumferential, and radial directions increased significantly in control and mdx mice with ISO stimulation ( Figure 4 ). In control mice aged 3 to 5 months, age did not significantly change PSS at baseline or after ISO administration. ISO also significantly increased PSS in young mdx mice, but the increases in most PSS were less in mdx mice than in control mice except for L-Epi and C-Epi at the age of 3 months. From the age of 3 to 5 months, PSS responses to ISO in mdx mice tended to decrease with age. For L-Endo and radial PSS during stress, significant differences between mdx and control mice appeared in all age groups from 3 to 5 months ( Figure 4 ). However, for L-Epi and C-Endo, the differences between mdx and control mice after ISO were not significant at the age of 3 months but reached significance at the age of 5 months. For C-Epi, differences between mdx and control mice could not be detected in the 3-, 4-, and 5-month groups.
