Division of Pediatric Cardiology, Cardiovascular Development Research Program, Children's Hospital of Pittsburgh, University of Pittsburgh, Pittsburgh, Pennsylvania
Division of Pediatric Cardiology, Cardiovascular Development Research Program, Children's Hospital of Pittsburgh, University of Pittsburgh, Rangos Research Center, Room 3320D, 3460 Fifth Avenue, Pittsburgh, PA 15213
Mechanical load is one of the major exogenous factors that regulate embryonic ventricular function and structure during cardiac morphogenesis (Keller, 1998). Following onset of the heart beat, the embryonic heart rapidly transforms from a single straight tube to a three-dimensional (3D) complex four-chamber heart. The operating pressure within the embryonic cardiovascular system is only 0.1–10 mm Hg during early cardiac morphogenesis and thus small changes in mechanical load affect normal pump function and structural phenotype (Clark et al., 1989; Tobita and Keller, 2000; Tobita et al., 2002).
Experimental paradigms that alter regional embryonic left ventricular (LV) mechanical loads produce changes in local cardiomyocyte proliferation patterns, protein expression, myocardial architecture, and myocardial function (Clark et al., 1989; Saiki et al., 1997; Sedmera et al., 1999, 2002; Tobita and Keller, 2000; Schroder et al., 2002; Tobita et al., 2002). Reduced LV volume load produced by left atrial ligation (LAL) results in LV hypoplasia with decreased LV myocardial volume and accelerated trabecular compaction (Sedmera et al., 1999). Increased LV pressure load produced by conotruncal banding (CTB) induces LV chamber dilatation, thickening of the compact myocardium, and precocious trabecular spiraling indicating acceleration of morphogenesis (Sedmera et al., 1999). Mechanical load also regulates the maturation of embryonic myocardial myofiber architecture at the subcellular level (Shiraishi et al., 1992; Price et al., 1996; Alford and Taber, 2003). However, little is known regarding how the embryonic myocardium acquires a mature 3D myofiber architecture and how altered mechanical load influences myofiber architecture and maturation.
In the present study, we tested the hypothesis that chronically altered mechanical load changes the local 3D fiber architecture of the developing embryonic LV myocardium. We measured local transmural myofiber angle distribution of the LV compact myocardium in normal development and the development of altered mechanical loads in Hamburger-Hamilton stages 21–36 chick embryos. In normal developing chick embryos, we found that transmural myofiber angles in the LV compact myocardium were oriented in a circumferential direction until stage 27 (−10 to 10°). Myofibers in the outer side of the compact myocardium shifted to more longitudinal direction by stage 36 (10 to 40°), producing a transmural gradient in myofiber orientation. Developmental changes in transmural myofiber angle distribution were accelerated following CTB, while the changes in angle distribution were significantly delayed following LAL. Thus, our results are consistent with a global paradigm that mechanical load modulates the maturation process of myofiber architecture distribution in the developing LV compact myocardium.
MATERIALS AND METHODS
Fertile white Leghorn chicken eggs were incubated in a forced-draft constant-humidity incubator and studied at Hamburger-Hamilton stages 21 (3.5 days), 27 (5 days), 31 (7 days), and 36 (10 days) of a 46-stage (21-day) incubation period as previously described (Hamburger and Hamilton, 1951). Embryos that were dysmorphic at stage 21, such as lacking retina development or displaying axial twisting, were excluded in the present study. Our research protocols conform to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (85-23, revised 1985).
Left Atrial Ligation to Reduce LV Volume Load
Embryos were initially incubated to stage 21. A 1 cm2 hole was made in the shell and the inner shell and extraembryonic membranes were removed to expose the developing embryo. The embryo was then gently positioned with the left side up, and microforceps were used to make a slit-like opening in the thoracic wall above the primitive left atrium. A loop of monofilament 10-0 nylon suture was placed across the primitive left atrium and tightened, which decreased the effective volume of the left atrium (Harh et al., 1973; Rychter et al., 1979; Tobita and Keller, 2000). Each embryo was then repositioned to its original right-side-up orientation. The eggshell opening was sealed with Parafilm M (Pechiney Plastic Packaging, Menasha, WI) and reincubated until stages 27, 31, or 36.
Conotruncal Banding to Increase LV Pressure Load
Embryos were initially incubated to stage 21. A monofilament nylon suture was passed underneath the midportion of the conotruncus and then tied snugly in an overhand knot until ventricular dimensions obviously increased without blood flow cessation or ventricular arrhythmia. The eggshell opening was sealed with Parafilm M and embryos were reincubated until stages 27, 31, or 36 (Clark et al., 1989; Tobita et al., 2002). Normal embryos were unoperated and incubated until stages 21, 27, 31, or 36.
Confocal Laser Scanning Microscopy
The heart was arrested at end-diastolic phase by injecting chick ringer solution containing 60 mM KCl, 0.5 mM verapamil, and 0.5 mM EGTA (Miller et al., 1997; Tobita et al., 2002). The embryo was fixed in 4% paraformaldehyde with PBS (pH 7.2) for 12 hr. The heart was excised from the embryo and rinsed with a PBS buffer. The heart was then placed in a dorsoventral orientation within a 13% polyacrylamide gel that was immediately polymerized with 2% ammonium persulfate (Fig. 1A) (Germroth et al., 1995; Nakagawa et al., 1997). When the heart was embedded in the polyacrylamide gel, we positioned the tangential plane of the LV free wall to be approximately perpendicular to the sectioning plane (viewed from dorsoventral orientation of the heart, Fig. 1). We defined the LV coronal plane as the plane including the LV longitudinal axis, which was the line between LV apex and midpoint of the LV atrioventricular groove (Figs. 1B and 2A). The angle (α) between LV coronal plane and the sectioning plane was measured when the heart was mounted on the vibratory microtome for further angle adjustment (Vibratome-1000, Vibratome, St Louis, MO; Fig. 1B). The angle α was expressed as ± 0° with the LV coronal plane parallel to the sectioning plane. A positive angle (α) represented that a section was made from ventral (LV base) to dorsal (LV apex) orientation with respect to the LV coronal plane.
Polyacrylamide-embedded hearts were then sectioned at a 200 μm thickness with a vibratory microtome (Fig. 2). Four to five serial sections were made in each heart and then stained for f-actin with FITC-conjugated phalloidin (Molecular Probes, Eugene, OR). The stained sections were mounted on a glass slide and the equator level of the LV free wall, the middle point of the LV free wall between LV apex and atrioventricular groove, was examined using a standard laser scanning confocal microscope (model TCS SP2, Leica Microsystems, Mannheim, Germany; Fig. 2A). Laser power, aperture, and gain of the confocal system were selected to minimize background fluorescence and to optimize the fluorescence intensity of the inner trabecular myocardium. We chose this approach because when we set the confocal system to optimize fluorescent signals in compact myocardium, the signals from trabecular myocardium tended to oversaturate and interfered with further image processing. Using a 10× objective, z-serial optical sectioning was performed on the middle 100 μm thickness of each section. We then optically sectioned the middle 50–60 μm with a 63× water-immersion objective. We set a z-axis depth and a section interval at 1 μm. Each image was saved in eight-bit grayscale TIFF format. Image resolution was set at 1.02 pixels/μm (10× objective, 512 × 512 pixels2) or 6.45 pixels/μm (63× objective, 1,024 × 1,024 pixels2).
Reconstruction of 3D Fiber Architecture
Individual sections were stacked according to the order of optical sectioning in which the top section represented the ventral side of the LV. We generated a 3D animation sequence from the stacked images by projecting 3D data sets through a 360° rotating plane at 1° intervals (see Supplemental Video clips 1 and 2 available at: The Supplementary Material referred to in this article can be found at the Anatomical Record website (http://www.interscience.wiley.com/jpages/0003-276X/suppmat/2005/v283A.html). Each frame in the animation sequence represented the result of projecting from a different viewing angle. All 3D image processing was performed using Scion Image software (Scion, Frederick, MD).
Myofiber Orientation Measurement
We digitally reconstructed either transverse or sagittal serial LV myocardial sections at a size of 50 μm × 50 μm × (compact myocardial thickness) with 1 μm depth at 1 μm intervals from the higher-resolution stacked images (6.45 pixels/μm; Fig. 2A).
We first measured the angle between the horizontal line and local epicardial tangential line at the middle level of the reconstructed transverse section and chose sections in which angle was within ± 15°. For transverse sections, we described myofiber angle with reference to the local epicardial tangential line as viewed from above (± 0°). A positive angle represents a myofiber angle that deviates in a counterclockwise direction from the reference axis. For sagittal sections, we described myofibril angles with ± 0° assigned to the dorsoventral LV circumferential direction and 90° assigned to the LV apicobasal longitudinal direction. Therefore, a positive sagittal plane angle represented a myofiber angle oriented in a counterclockwise direction from ± 0°. To determine the myofiber orientation of the compact myocardium, we then analyzed subregions of the reconstructed sections. The local myofiber angle in either transverse or sagittal section was determined by local intensity gradients technique (Karlon et al., 1998). In the present study, we used three-by-three horizontal and vertical Sobel filters to calculate intensity gradients in each section. Each reconstructed section was divided into small regions of interest (ROIs) with a size of 20–30 pixels2 (150–300 ROIs per section). Dominant local myofiber angle in each ROI was determined by the largest accumulator bin value and converted to range from −89 to 90°. The angle (α) between the LV coronal plane and sectioning plane influenced projected myofiber angles in both transverse (θt) and sagittal (θs) planes. To adjust the projected myofiber angles, the following equations were used: adjθs = θs − α, and adjθt = arctan [sin(θt)] [cos(α)], where adjθ is the adjusted myofiber angle.
The experimental numbers of transmural myofiber angle measurement were stage 21 normal (n = 6), stage 27 normal (6), stage 27 LAL (6), stage 27 CTB (6), stage 31 normal (8), stage 31 LAL (6), stage 31 CTB (6), stage 36 normal (6), stage 36 LAL (5), and stage 36 CTB (5) embryos. All calculations were performed using LabVIEW-based custom-made programs (LabVIEW 5.0; National Instruments, Austin, TX). Calculated local fiber angles were then inspected visually in each image (Fig. 2B).
The LV compact myocardium was divided into 10 segments (layers) from endocardium (0%) to epicardium (100%). Mean fiber orientation and circular standard deviation in each segment were calculated using circular statistics (Fisher, 1993). The mean orientation was computed by treating each measurement as a unit vector and averaging the vector components as follows: C = Σ(i)cos(2θi) and S = Σ(i)sin(2θi), where θi is the individual fiber angle; R2 = C2 + S2; and cos(2θmean) = C/R and sin(2θmean) = S/R, where θmean is the mean fiber angle. The circular standard deviation (δ) is expressed as δ = 0.5[−2 log(Rmean)]0.5, where Rmean = R/n and n is the number of measurement.
To test myofiber orientation directionality, the Rayleigh test was performed in each section. We performed a nonparametric ranking test to determine differences in mean fiber angle throughout the compact myocardium in normal, LAL, and CTB hearts. We defined statistical significance when P < 0.05.
Myocardial Architecture of LV Free Wall
Figure 3 shows representative changes in 3D LV free wall myocardial architecture during development (see also Supplemental Video Clip 1 available at: http://www.interscience.wiley.com/jpages/0003-276X/suppmat/2005/v283A.html). At stage 27, the LV myocardium consists of abundant inner trabecular myocardium and outer thin compact myocardium in all experimental groups. Following LAL, individual trabecula tended to approximate each other and intertrabecular space was reduced versus normal hearts at stage 27. The LV cavity size was reduced, as was intertrabecular spacing, particularly at the LV apex at stage 31. For normal and CTB ventricles, LV compact myocardium thickness increased by stage 31 and the angle between the long axis of major trabeculae and epicardial tangential line became shallow, consistent with the onset of myocardial compaction. By stage 36, the thickness of the compact myocardial layer increased in all experimental groups. Following LAL, the LV cavity was obviously smaller than normal, while myocardial compaction was accelerated following CTB. These morphological changes were consistent with a previously published report (Sedmera et al., 1999).
Transmural Myofiber Angle Distribution
Mean fiber angles in LV transverse sections were distributed uniformly through the entire compact myocardium in normal and LAL groups during all developmental stages (−10 to 10°; P < 0.001, Rayleigh test). Following CTB, mean fiber angles were initially distributed uniformly, similar to stage-matched normal controls; however, by stage 36, mean fiber angles became more negative than normal controls (−10 to −16°; P < 0.05 vs. normal, nonparametric ranking test; Fig. 4). In sagittal sections, mean fiber angles were distributed uniformly at stage 27 in all experimental groups (−5 to 15°; P < 0.001, Rayleigh test).
During normal development, mean fiber angles at the outer 30% of the compact myocardium significantly increased at stage 31 and extended to 40% by stage 36 (30 to 60°; P < 0.05 by nonparametric ranking test within a group; Fig. 5). Following LAL, mean fiber angles were distributed uniformly at stage 31, similar to stage 27 embryos (−5 to 5°; P < 0.001 by Rayleigh test; P < 0.05 vs. normal controls). Myofiber angle of the outer 20% of the compact myocardium increased by stage 36 following LAL (15 to 20°; P < 0.05 within a group). Following CTB, mean fiber angles of the outer 30% of the compact layer increased by stage 31 (20 to 30°; P < 0.05) and extended to the outer 70% of the wall by stage 36 (Figs. 4 and 5; P < 0.05).
The present study showed that transmural myofiber angle distribution in the LV compact myocardium changed from a uniform transmural circumferential orientation to a gradual transmural shift in orientation from circumferential (endocardium) to longitudinal (epicardium) during normal development, and that reduced LV mechanical load (LAL) induced LV hypoplasia and was associated with an immature transmural variation of myofiber angle, while increased LV mechanical load (CTB) induced LV hyperplasia associated with an acceleration of compact myocardial myofiber maturation.
Normal Transmural Myofiber Distribution During Development
Studies of myofibrillogenesis in the embryonic heart show that changes in myofibril organization and cell shape occur in regions with rapid shape change and proliferation (Manasek et al., 1984; Itasaki et al., 1989; Shiraishi et al., 1992; Price et al., 1996). Shiraishi et al. (1992) reported the 3D arrangement of actin filaments in stage 7–13 chick embryonic myocardium, in which cardiomyocytes in the outer layer were round while those of the inner layer were spindle-shaped with their long axes oriented in a circumferential direction. They speculated that transmural variation of myofiber arrangement maintained tubular heart structure and played an important role in generating effective blood flow. Itasaki et al. (1989) reported regional variations in myofiber arrangement in the looped chick embryonic heart with myofibers in the inner curvature arranged in a circumferential direction and some myofibers in the outer curvature arranged longitudinally. Recently, Alford and Taber (2003) measured regional epicardial wall strain in the looping embryonic heart and noted that the epicardial strain of the outer curvature was relatively isotropic, consistent with a lack of myofiber alignment, while the wall strains of the central and inner curvature regions were orthotropic, consistent with regional myofiber alignment. These studies provide evidence of the developmental relationship between regional cardiac architecture, myofiber alignment, and cardiac function.
To our knowledge, the present study is the first investigation of embryonic LV myofiber architecture at stages corresponding to rapid ventricular growth and secondary and tertiary ventricular trabeculation (Sedmera et al., 1997, 1998). During the secondary trabeculation stage, the embryonic ventricle is composed of a thick inner trabecular myocardium and a thin layer of outer compact myocardium. Results in the present study showed that transmural myofiber distribution in LV compact myocardium was uniform from stage 21 to 27, in which myofibers oriented parallel to the epicardial tangential line in the transverse plane, and those in sagittal plane oriented in the circumferential direction. Our previous study of epicardial wall strains (Tobita and Keller, 2000) showed that embryonic LV contraction pattern changed from uniform at stage 21 to a longitudinal dominant nonuniform pattern by stage 27. Developmental changes in the LV contraction pattern coincide with a dorsoventral orientation of secondary trabecular sheets within the LV (Sedmera et al., 1997) and the trabecular myocardium represents the majority of myocardial wall volume at this stage. Our current results suggest that myofibers in the outer thin compact myocardium contribute mainly to the circumferential myocardial contraction seen at this stage. By stage 31, transmural myofiber alignment has shifted (in the sagittal plane) from a uniform to a nonuniform pattern, with myofiber orientation shifting from a circumferential (endocardial side) to a longitudinal (epicardial side) pattern. By stage 36, further shifts in transmural myofiber alignment occurred coincident with tertiary trabeculation, intramural coronary vascular formation, maturation of Purkinje fibers within the conduction system, and myocardial compaction (Sedmera et al., 2000; Reckova et al., 2003).
During tertiary ventricular trabeculation, the arrangement of LV trabeculae shifts from a dorsoventral orientation to a counterclockwise spiral arrangement when viewed from the cardiac base. The trabecular sheets become larger and interconnected, similar to a wicker basket, stretching from the LV apex to the mitral valve orifice. The present study showed that transmural myofiber angle distribution changed from a uniform to a nonuniform pattern in which myofiber angles developed a gradient from circumferential (endocardial) to longitudinal (epicardial). The nonuniform transmural myofiber distribution indicates that the embryonic LV chamber has acquired the characteristic twist that is present during contraction of the mature LV (Taber et al., 1996). The change in transmural myofiber distribution occurs coincident with the shortened cycle length and increased cardiac function noted in the rapidly growing embryonic heart (Girard, 1973; Akiyama et al., 1999). The myofiber angle distribution in the stage 36 embryonic chick LV was similar to the myofiber angle distribution in the outer layers of second-trimester human fetal LV (Jouk et al., 2000). Our results suggest that the LV myocardium acquires a mature three-layer spiral muscle fiber system by fusing tertiary trabeculae and developing compact myocardium to form the composite myocardial wall (Streeter and Hanna, 1973; Jouk et al., 2000). Further studies are necessary to define the impact of mid- and late-gestation growth and remodeling on final myofiber phenotype.
Transmural Myofiber Distribution Following Altered Mechanical Loads
Following LAL, reduced LV chamber dimensions and myocardial volume (LV hypoplasia) become apparent between stages 29 and 31 when the embryonic LV myocardium makes the shift from secondary to tertiary trabeculation stages (Sedmera et al., 1999; Tobita and Keller, 2000). These changes in LV chamber dimensions are associated with reduced LV cell proliferation rate (Sedmera et al., 2002). The LV pressure, blood flow patterns, regional contraction patterns, passive material properties, and cytoskeletal protein expression all change following LAL and these changes in cardiac function precede LV chamber remodeling (Tobita and Keller, 2000; Schroder et al., 2002; Tobita et al., 2002). Reckova et al. (2003) recently reported a significant delay in conduction system maturation during tertiary trabeculation following LAL. In the present study, transmural myofiber angle distribution in LAL compact myocardium was similar to normal until stage 27. However, myofiber distribution patterns were significantly delayed at stages 31 and 36, coincident with delayed conduction system maturation.
Increased LV pressure load (CTB) induces LV chamber dilatation followed by thickening of compact myocardium and acceleration of tertiary trabeculation (Sedmera et al., 1999). Following CTB, myocardial cell proliferation rate increases (Clark et al., 1989; Saiki et al., 1997). Both peak and end-diastolic LV pressures increased and eventually resulted in altered passive material properties (Tobita et al., 2002; Miller et al., 2003). The process of LV conduction system maturation is accelerated following CTB (Reckova et al., 2003; Hall et al., 2004). In the present study, transmural myofiber angle distribution following CTB was similar to normal embryos at stage 27 and then the maturation patterns of myofiber distribution were accelerated at stage 36. In addition, transverse myofiber angles following CTB were more negative at stage 36, implying greater LV twist during contraction compared with normal embryos. As with LAL, these results suggest that altered mechanical loads change compact myocardial fiber architecture maturation process, conduction system maturation, and coronary artery development in the tertiary trabeculation stages.
In conclusion, mechanical load plays a critical role as an epigenetic factor in determining both local myofiber orientation and maturation process in the developing compact myocardium during cardiac morphogenesis.
The authors appreciate the recommendations made by Dr. Robert Price at the University of South Carolina regarding myocardial imaging techniques.