SEARCH

SEARCH BY CITATION

Keywords:

  • zebrafish;
  • cardiac morphogenesis;
  • ventricular trabeculation;
  • myocardium;
  • endocardium;
  • cloche;
  • AG1478;
  • weak atrium

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Maturation of the developing heart requires the structural elaboration of the embryonic ventricle through the process of trabeculation. Trabeculae form as the ventricular myocardium protrudes into the lumen of the chamber, thereby increasing muscle mass and altering functional output. Little is understood about the cellular basis for trabeculation and its genetic regulation. Here, we establish the utility of the zebrafish embryo for the analysis of the mechanisms driving trabeculation. In zebrafish, we can follow trabeculation in four dimensions and define morphologically discrete stages for the initiation, propagation, and network elaboration that form the ventricular trabeculae. We find that Neuregulin/ErbB signaling is required for the initial protrusion of the myocardium into the ventricular lumen. Additionally, we demonstrate that optimal blood flow through the ventricle is important for the advancement of trabeculation. Thus, our results indicate that the zebrafish provides a valuable model for investigating possible causes of congenital defects in trabeculation. Developmental Dynamics 240:446–456, 2011. © 2011 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Functional maturation of the embryonic heart is crucial for effective circulation of blood throughout the developing organism. One key aspect of cardiac maturation is the structural elaboration of the embryonic ventricle through trabeculation (Sedmera et al.,2000; Moorman and Christoffels,2003). Initially, the primitive ventricular chamber resembles a simple balloon with a smooth inner surface. Trabeculation transforms the interior terrain of the ventricle, creating numerous myocardial projections that protrude inward from the chamber wall. Ultimately, these projections, called trabeculae, form a complex network within the ventricular lumen.

Ventricular trabeculation has been proposed to influence cardiac function in multiple ways. Trabeculae increase the amount of myocardial mass in the chamber and also create elaborate lacunae through which blood flows (Sedmera et al.,2000). Both aspects of ventricular topography contribute to this chamber's characteristic contractile strength and compliant filling, and these properties determine a ventricle's overall workload efficacy (Icardo and Fernandez-Teran,1987; Sanchez-Quintana and Hurle,1987; Sedmera et al.,2000; Hu et al.,2001). Of clinical importance, infants born with a hypotrabeculated ventricle exhibit compromised hemodynamics; conversely, infants born with a hypertrabeculated ventricle can demonstrate impaired diastolic function, secondary to lack of compliance and ultimately leading to inadequate filling of the ventricle (Weiford et al.,2004; Breckenridge et al.,2007). Despite the clinical relevance of deficient or increased trabeculation, the cellular mechanisms governing trabeculation and their genetic regulation are relatively understudied.

Our general understanding of the cellular basis for trabeculation comes primarily from studies of fixed tissues from chick and mouse embryos at various stages (Sedmera et al.,2000). In these contexts, trabeculation is first apparent after the linear heart tube starts to loop, and the initial changes occur at the ventricular outer curvature. Trabeculae appear as ridges that converge near the atrioventricular (AV) canal and radiate out along the ventricular wall. The precise cell behaviors that initiate trabeculation are unknown. Appealing hypotheses include the ideas that cardiomyocytes actively invaginate into the lumen, that the endocardium actively evaginates into the myocardial wall, or that a passive buckling of the chamber wall produces ridges (Icardo and Fernandez-Teran,1987; Marchionni,1995; Sedmera and Thomas,1996).

The molecular regulation of trabeculation is also poorly understood. However, several compelling studies in mouse highlight Neuregulin signaling as an important influence on trabeculation (Gassmann et al.,1995; Lee et al.,1995; Meyer and Birchmeier,1995; Lai et al.,2010). Neuregulins are growth factors that serve as ligands of ErbB receptor tyrosine kinases and induce cell survival, proliferation, and differentiation in several different settings (Mei and Xiong,2008; Pentassuglia and Sawyer,2009). In mouse, Neuregulin 1 (Nrg1) is expressed by the endocardium and ErbB2 and ErbB4 are expressed by the adjacent myocardium (Meyer and Birchmeier,1995). Loss-of-function of Nrg1, ErbB2, or ErbB4 in mouse results in reduced or absent trabeculation (Gassmann et al.,1995; Lee et al.,1995; Meyer and Birchmeier,1995; Lai et al.,2010). These data strongly suggest that Neuregulin signals, delivered from the endocardium to the myocardium, are necessary for the successful completion of trabeculation. However, it is less clear whether Neuregulin signaling is essential for the initiation, propagation, or maintenance of trabeculae and which pathways work in concert with Neuregulin signaling in this context.

To facilitate a deeper understanding of trabeculation, it would be ideal to use a system in which the process can be visualized in live embryos that are amenable to genetic and pharmacological manipulation. Here, we establish the utility of the zebrafish embryo for the analysis of trabeculation. The optical clarity of the zebrafish embryo allows detailed visualization of this temporally and spatially dynamic process (Schoenebeck and Yelon,2007). This approach facilitates a four-dimensional analysis of trabecular morphogenesis in live specimens, thereby avoiding the limitations of single-timepoint analyses and potential artifacts derived from tissue fixatives.

Our study provides a unique characterization of the initiation of trabeculation and defines three progressive phases of trabeculation in the zebrafish embryo. First, trabeculation initiates as cardiomyocytes are displaced into the lumen from a specific location along the ventricular outer curvature. Second, lumenal ridges appear and propagate along the outer curvature wall in a radial pattern. Third, ridges mature into thicker bundles that partially detach from the ventricular wall and create an elaborate network. Additionally, we show that the presence of the endocardium and the reception of Neuregulin signaling are required for the initiation of trabeculation in zebrafish. Furthermore, we find that blood flow has a potent influence on the progression of trabeculation. Together, our results suggest multiple possible etiologies for trabecular defects in patients.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Zebrafish Transgenes Facilitate Live Imaging of the Ventricular Myocardium

To follow the process of trabeculation in live embryos, we took advantage of existing zebrafish reporter transgenes that drive expression of fluorescent proteins in cardiomyocytes. The transgenes Tg(myl7:egfp) (Huang et al.,2003) and Tg(myl7:dsredt4) (Garavito-Aguilar et al.,2010) drive expression of cytoplasmic green fluorescent protein (GFP) and DsRed, respectively, thereby allowing rendering of cardiomyocyte volume. Additionally, the transgene Tg(myl7:egfp-hshras) (Chi et al.,2008) drives expression of a membrane-tethered form of GFP and thereby highlights cardiomyocyte boundaries. Together, this combination of reagents allowed us to visualize the lumenal contours of the ventricular myocardium during trabeculation.

Trabeculation Initiates at a Characteristic Position Along the Outer Curvature

Prior analyses of zebrafish embryos have defined a broad timeframe for ventricular trabeculation: no trabeculae are evident at 48 hours postfertilization (hpf; Hu et al.,2000), some trabeculae are present by 100 hpf (Chi et al.,2008), and a network of trabeculae is present by 5 days postfertilization (dpf; Hu et al.,2000). We, therefore, began our investigation by using confocal microscopy to inspect live embryos after 48 hpf for the earliest signs of trabecular development. At 55 hpf, the ventricular wall has a generally uniform thickness, and the ventricular lumen has a correspondingly smooth inner contour (Fig. 1A–C; Supp. Movie S1, which is available online). However, by 60 hpf, characteristic features appear on the ventricle's lumenal surface (Fig. 1D–F; Supp. Movie S2). Most notably, the first displacement of cardiomyocytes into the ventricular lumen (Fig. 1E, red arrowheads) occurs at a reproducible location. In 19 of 20 wild-type embryos observed at this stage, this protrusion is readily discernible at a point in the outer curvature (Fig. 1E, top white arrowhead, and Fig. 1F, white arrowheads) that is positioned ventrally across from the superior side of the AV canal (Fig. 1E, bottom white arrowhead). Additional, but more subtle, protrusions are evident in adjacent regions (Fig. 1E, yellow arrowhead). All of the developing protrusions appear to extend in an orientation circumferential to the long axis of the ventricle.

thumbnail image

Figure 1. Trabeculation initiates between 55 and 60 hours postfertilization (hpf) in a characteristic location. A–F: Confocal reconstructions of the ventricular myocardium in wild-type (wt) embryos expressing Tg(myl7:dsredt4) and Tg(myl7:egfp-hshras). Volume reconstruction (A,D), lumenal surface reconstruction (B,E), and optical sections (C,F) are shown as described in the Experimental Procedures section. In A and D, the atrioventricular (AV) canal is outlined with a dotted white line. White asterisks indicate areas where the acquisition of fluorescent signal was blocked by overlying melanocytes. Yellow asterisks indicate examples of cells displaying epigenetic silencing of Tg(myl7:dsredt4). G,H: Diagrams depict the orientations of the hearts and optical sections analyzed, as described in the Experimental Procedures section. The axes indicate the orientation of the embryo: V, ventral; D, dorsal; P, posterior; A, anterior; L, left; R, right. Additional abbreviations used are A for atrium and V for ventricle. A–C: At 55 hpf, the lumenal surface of the ventricle is generally smooth with no prominent protrusions. See also Supp. Movie S1. D–F: By 60 hpf, lumenal protrusion becomes evident (E, red arrowheads), most notably at a characteristic position on the ventricular outer curvature (E, top white arrowhead, and F, white arrowheads). Additional portions of the lumenal surface of the chamber exhibit more subtle and variable protrusions (E, yellow arrowhead). See also Supp. Movie S2. Scale bars = 50 μm.

Download figure to PowerPoint

Initial Lumenal Ridge Becomes Prominent as a Radial Array of Ridges Forms

By 72 hpf, the inner contours of the ventricle become considerably more complex (Fig. 2A–C; Supp. Movie S3). The initially most prominent protrusion forms a more distinct ridge connecting the AV canal to the outer curvature (Fig. 2B, red arrowheads). Coincident with ridge formation, an inward folding of the ventricular wall (Fig. 2A–C, white arrowhead) aligns with the ridge, tethered opposite the superior portion of the AV canal. Additionally, this fold corresponds to an invagination of the outer surface of the chamber (data not shown). In addition to the continuing prominence of the initial ridge, an array of several other ridges radiates from a similar origin at the AV canal, elongating toward the outer curvature (Fig. 2B, yellow arrowheads).

thumbnail image

Figure 2. Lumenal ridges become prominent by 3 days postfertilization (dpf). A–F: Confocal reconstructions of the ventricular myocardium in wild-type (wt) embryos expressing Tg(myl7:egfp). All reconstructions and sections are as shown in Figure 1. Yellow asterisks indicate examples of cells displaying epigenetic silencing of Tg(myl7:egfp). White arrowheads indicate the position of the invagination in the ventricular wall. A–C: By 72 hours postfertilization (hpf), ridges are present on a broad portion of the lumenal surface of the ventricle (red and yellow arrowheads). The most prominent ridge (red arrowheads) extends from the atrioventricular (AV) canal to the outer curvature and appears tethered to an invagination in the ventricular wall (white arrowhead). See also Supp. Movie S3. D–F: By 91 hpf, the ridges (red and yellow arrowheads) become more complex, although most ridges form a radial array that coalesce at a common origin near the AV canal. See also Supp. Movie S4.

Download figure to PowerPoint

The ridged lumenal surface of the ventricle becomes further defined by 91 hpf (Fig. 2D–F; Supp. Movie S4). Propagation of the radial array of ridges remains evident (Fig. 2E,F, red and yellow arrowheads), and the most prominent ridge (Fig. 2E,F, red arrowheads) is still tethered to an invagination in the ventricular wall (Fig. 2E,F, white arrowhead). A larger number of ridges become apparent as the heart continues to mature, and most of these structures appear to converge at a common position near the AV canal.

Trabeculae Protrude Into the Lumen While Staying Connected to the Ventricular Wall

As chamber maturation continues, the ventricular ridges appear to transform into protruding trabeculae (Fig. 3). By 96 hpf, the ridges are clearly larger bundles of cells (Fig. 3A–C; Supp. Movie S5). The most prominent ridge is still in the same position as the initial ridge, and, in its widest portion, forms a stalk that is five to six cells thick (Fig. 3B,C, red arrowheads). The widest portion of the most prominent ridge is found on either the ventral or dorsal side of the outer curvature, in equal frequency (n = 30, Fig. 3B shows a ventral example), and is attached to additional ridges that span the chamber (Fig. 3B,C″, yellow arrowheads). In 30% of the wild-type embryos observed at this stage, one of these additional ridges extends from the main stalk toward the outflow tract (Fig. 3B, right yellow arrowhead).

thumbnail image

Figure 3. The main trabecular stalk separates from the ventricular wall between 96 hours postfertilization (hpf) and 120 hpf. A–F: Confocal reconstructions of the ventricular myocardium in wild-type (wt) embryos expressing Tg(myl7:dsredt4) and Tg(myl7:egfp-hshras) (A–C) or expressing only Tg(myl7:dsredt4) (D–F). All reconstructions and sections are as shown in Figure 1. A–C: By 96 hpf, the main ridge (red arrowheads) on the lumenal surface becomes a thick stalk that is connected to numerous other ridges crossing the chamber (yellow arrowheads). See also Supp. Movie S5. D–F: By 120 hpf, the main stalk (red arrowheads) separates from the ventricular wall, but is still connected to the wall by thin connecting rods (green arrowheads). Additionally, the interconnections between the other ridges (yellow arrowheads) become increasingly more elaborate. See also Supp. Movie S6.

Download figure to PowerPoint

By 120 hpf, the main stalk protrudes further into the lumen (Fig. 3D–F; Supp. Movie S6). It seems to detach almost completely from the ventricular wall (Fig. 3E,F, red arrowheads), while still maintaining fine rods of myocardium that connect the stalk to the wall (Fig. 3E,F, green arrowheads). The rest of the network of ventricular ridges becomes thicker, more elaborate, and further interconnected with each other and the main stalk (Fig. 3E, yellow arrowheads). The complex pattern of these interconnections varies between embryos, whereas the morphology of the main stalk remains relatively consistent (n = 30). Thus, by 4 dpf, the inner surface of the ventricle has a clear trabeculated morphology.

Trabeculation Fails to Initiate in the Absence of Endocardium

The morphological similarities between the trabecular networks in zebrafish and amniotes suggest that the genetic pathways controlling trabeculation could be highly conserved (Sedmera et al.,2000; Moorman and Christoffels,2003). Because endocardial–myocardial signaling is required for trabeculation in amniotes (Toyofuku et al.,2004; Stankunas et al.,2008; Pentassuglia and Sawyer,2009), we chose to investigate the impact of the endocardium on trabeculation in zebrafish. The zebrafish cloche (clo) mutation inhibits the specification of endothelial lineages; as a result, the endocardium does not form in clo mutants (Stainier et al.,1995; Liao et al.,1997). The clo mutant heart is dysmorphic, with an enlarged atrium and a malformed ventricle (Fig. 4A; Stainier et al.,1995; Schoenebeck et al.,2007). We carefully examined clo mutants for morphological signs of trabeculation; however, it appears that the trabeculation process never initiates in the clo mutant ventricle (Fig. 4; Supp. Movie S7). In clo mutants, the ventricular wall retains a uniform thickness, no myocardial protrusions are evident in the ventricular lumen, and no lumenal ridges form (Fig. 4B,C). Overall, the smooth lumenal surface of the clo mutant ventricle (Fig. 4B,C) resembles the lumenal surface of the wild-type ventricle before the initiation of trabeculation (Fig. 1B,C). This phenotype does not simply represent a developmental delay, because we do not observe progression of trabeculation in clo mutants even at 96 hpf (data not shown). Therefore, initiation of myocardial trabeculation in zebrafish seems to depend upon the presence of the endocardium, presumably due to important signals that emanate from this tissue.

thumbnail image

Figure 4. In clo mutant embryos, trabeculation fails to initiate. A–C: Confocal reconstructions of the ventricular myocardium in clo mutant embryos expressing Tg(myl7:egfp). All reconstructions and sections are as shown in Figure 1. A similarly staged wild-type (wt) embryo is shown in Figure 2A–C. A: The clo mutant ventricle is dysmorphic and has a large AV canal. B,C: The lumenal surface of the clo mutant ventricle is smooth with no significant protrusions or ridges, and the ventricular wall retains a simple, uniform thickness. See also Supp. Movie S7.

Download figure to PowerPoint

Neuregulin Signaling Is Required for the Initiation of Trabeculation

Due to the important role of Neuregulin signaling in promoting trabeculation in mouse (Gassmann et al.,1995; Lee et al.,1995; Meyer and Birchmeier,1995; Lai et al.,2010), we hypothesized that Neuregulin signals generated by the zebrafish endocardium could have a potent impact on the morphogenesis of the myocardium. In zebrafish, neuregulin 1 is expressed in the endocardium (Milan et al.,2006), and multiple ErbB receptor genes, such as erbb2 and erbb1a, are expressed broadly, including in the myocardium (Goishi et al.,2003; Thisse and Thisse,2004). To test the importance of Neuregulin/ErbB signaling during trabeculation stages, we used AG1478, an inhibitor of the ErbB1 receptor and the ErbB2/ErbB3 heterodimer (Levitzki and Gazit,1995) that has been used to phenocopy loss of Neuregulin signaling in other developmental contexts in zebrafish (Lyons et al.,2005; Budi et al.,2008; Honjo et al.,2008; Scherz et al.,2008).

In embryos treated with AG1478 beginning at 27 hpf (Fig. 5D,J), the ventricle appears slightly rounded and dilated, in comparison to the ventricular morphology of dimethyl sulfoxide (DMSO) -treated control embryos (Fig. 5A,G). Additionally, whereas DMSO does not disrupt trabecular morphogenesis (Fig. 5B,C,H,I; Supp. Movies S8 and S10), AG1478 has a potent inhibitory effect on the initiation of trabeculation (Fig. 5E,F,K,L; Supp. Movies S9 and S11). Similar to clo mutant embryos (Fig. 4B,C), AG1478-treated embryos exhibit a uniform thickness of the ventricular wall, a smooth lumenal surface of the ventricle, no significantly protruding cardiomyocytes, and no evident myocardial ridges (Fig. 5E,F,K,L). By 4 dpf, some cells in AG1478-treated ventricles exhibit subtle signs of lumenal protrusion (Fig. 5K, yellow arrowheads), but they do not proceed to form defined ridges. Therefore, Neuregulin signaling plays an essential role during the initiation of trabeculation in zebrafish.

thumbnail image

Figure 5. Neuregulin signaling is required for the initiation of trabeculation. A–L: Confocal reconstructions of the ventricular myocardium in wild-type (wt) embryos expressing Tg(myl7:dsredt4) and Tg(myl7:egfp-hshras) and treated either with dimethyl sulfoxide (DMSO; A–C,G–I) or AG1478 (D–F,J–L). All reconstructions and sections are as shown in Figure 1, except that the atrioventricular (AV) canal is not visible in G and J. A white asterisk indicates area where the acquisition of fluorescent signal was blocked by overlying melanocytes. Yellow asterisks indicate examples of cells displaying epigenetic silencing of Tg(myl7:dsredt4). A–C: At 75 hours postfertilization (hpf), DMSO-treated embryos exhibit lumenal protrusions and primitive ridges (yellow arrowheads), similar to untreated wt embryos at a comparable stage (Fig. 2A–C). See also Supp. Movie S8. D–F: In contrast, AG1478-treated embryos do not initiate trabeculation and instead exhibit a smooth lumenal surface and a uniform thickness of the chamber wall at 75 hpf. See also Supp. Movie S9. G–I: At 99 hpf, DMSO-treated embryos exhibit normal ventricular morphology, including defined trabecular ridges (yellow arrowheads). See also Supp. Movie S10. J–L: In contrast, AG1478-treated embryos fail to form defined myocardial protrusions or ridges, although some subtle and variable displacements of lumenal cardiomyocytes do emerge (K, yellow arrowheads). See also Supp. Movie S11.

Download figure to PowerPoint

Normal Blood Flow Is Required for the Progression of Trabeculation

Normal blood flow is essential for multiple facets of cardiac development (Bartman and Hove,2005), including the formation of AV endocardial cushions (Hove et al.,2003; Vermot et al.,2009), the enlargement and elongation of outer curvature cardiomyocytes (Auman et al.,2007), and the maturation of the conduction system (Sedmera et al.,2005; Chi et al.,2008). These processes, like trabeculation, take place after cardiac function begins and, therefore, occur in the presence of the biomechanical impact of blood flow. We therefore wondered whether blood flow influences any aspects of trabeculation.

To reduce blood flow through the ventricle without directly altering the intrinsic character of the ventricular myocardium, we used zebrafish weak atrium (wea) mutant embryos. Our prior studies have established that the wea locus encodes atrial myosin heavy chain (amhc), an atrium-specific gene that is required for atrial contractility but is not needed for ventricular contractility (Berdougo et al.,2003). As a consequence, wea mutant embryos have a noncontractile atrium and exhibit substantially inhibited blood flow through the ventricle, even though the specification and contractility of the ventricular myocardium are unaffected (Berdougo et al.,2003; Auman et al.,2007). Our previous analyses of the wea mutant phenotype demonstrated that its reduction of blood flow inhibits the characteristic cardiomyocyte cell shape changes that underlie formation of the ventricular outer curvature (Auman et al.,2007), but these studies did not investigate the impact of the wea mutation on morphogenesis beyond 2 dpf.

To extend our previous studies, we examined ventricular morphology at later stages in wea mutant embryos (Fig. 6). At 3 dpf, we could detect some signs of lumenal protrusion in the wea mutant ventricle (Fig. 6B, yellow arrowheads; Supp. Movie S12). However, these protrusions do not seem to be located in reproducible positions, indicating aberrant trabecular morphogenesis rather than merely a developmental delay. Moreover, even at 4 dpf, these protrusions do not seem to progress toward creating myocardial ridges, and trabeculae do not form (Fig. 6E,F; Supp. Movie S13). Therefore, normal blood flow is required to advance the process of trabeculation, although some initial and irregular thickening of the ventricular wall can occur in the presence of suboptimal blood flow.

thumbnail image

Figure 6. Reduced blood flow in wea mutant embryos inhibits the progression of trabeculation. A–F: Confocal reconstructions of the ventricular myocardium in wea mutant embryos expressing Tg(myl7:egfp). Using a stereomicroscope, we confirmed the presence of suboptimal blood flow through the ventricle in all wea mutant embryos chosen for analysis. All reconstructions and sections are as shown in Figure 1. Similarly staged wild-type (wt) embryos are shown in Figure 2. White asterisks indicate areas where the acquisition of fluorescent signal was blocked by overlying melanocytes. Yellow asterisks indicate examples of cells displaying epigenetic silencing of Tg(myl7:egfp). A: At 75 hours postfertilization (hpf), the wea mutant ventricle is dysmorphic and has a narrow AV canal. B,C: A few cells on the lumenal surface of the wea mutant ventricle appear to protrude into the chamber (B, yellow arrowheads) and the ventricular wall exhibits some areas of subtle thickening (C, white arrowheads). See also Supp. Movie S12. D–F: At 96 hpf, the wea mutant ventricle lacks ridges or trabeculae; however, some protruding cells are still present (E, yellow arrowheads). See also Supp. Movie S13.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Taken together, our data provide a unique perspective on the dynamic process of trabeculation. By taking advantage of the optical accessibility of the zebrafish heart, we can observe trabeculation in four dimensions and define morphologically discrete stages for the initiation, propagation, and network elaboration that form the ventricular trabeculae. Furthermore, we can identify essential regulators of trabeculation and pinpoint when they impact morphogenesis. Our studies show that trabeculation begins with a stereotyped pattern of protrusions that transform the initially smooth surface of the ventricular lumen. As development proceeds, a radial pattern of ridges forms, and thickened bundles of cardiomyocytes coalesce and partially detach from the ventricular wall. We find that Neuregulin signaling, presumably delivered from the endocardium to the myocardium, is required for the initial displacement of cardiomyocytes into the lumen. Additionally, blood flow plays an important role in promoting the progression of trabeculation.

Overall, the distinct phases of trabecular morphogenesis in zebrafish are comparable to the previously characterized phases of trabeculation in amniotes. The pattern of ridge formation in the zebrafish ventricle correlates with the previously observed circumferential trabecular ridges in chick and mouse (Sedmera et al.,2000). Similarly, in both zebrafish and amniotes, the ventricular lumen matures from a ridged morphology to a more complex trabecular network (Sedmera et al.,2000). The use of zebrafish allowed us to visualize the initiation of myocardial protrusion into the lumen, an aspect of trabeculation that has not been readily observed in previous studies of fixed tissue (Sedmera et al.,2000). In future studies, it will be particularly important to determine which specific cellular behaviors drive the initial myocardial protrusions and control the progression of ridge formation. Several cellular mechanisms could be responsible for these morphogenetic transitions, including changes in cardiomyocyte shape and size, directed cardiomyocyte migration, and/or oriented myocardial cell divisions. We note that a recently published study proposes that zebrafish trabeculation is driven by directional delamination of cardiomyocytes into the ventricular lumen (Liu et al.,2010). The feasibility of following the dynamics of trabeculation in live zebrafish embryos will greatly facilitate further insight into the cellular dynamics of this intricate morphological process.

Ultimately, it will be crucial to identify the specific signaling pathways that trigger the cell behaviors underlying trabeculation. The clo mutant phenotype suggests that signaling between the endocardium and the myocardium plays a fundamental role in driving trabeculation. This model is consistent with our data demonstrating a conserved requirement for Neuregulin/ErbB signaling during the trabeculation process. Furthermore, our study reveals that Neuregulin/ErbB signaling is crucial for the initiation of myocardial protrusion, rather than being necessary only for propagation of ridges or maintenance of trabeculae.

Because only select cardiomyocytes protrude into the ventricular lumen, Neuregulin signaling likely interfaces with other pathways to trigger cell behaviors in specific locations. One possibility is that Neuregulin signaling may be regulated in a spatially restricted manner that results in the selection of certain cardiomyocytes for protrusion. Alternatively, Neuregulin signaling may play a permissive role in allowing an independent, spatially restricted signal to initiate trabeculation in the appropriate positions. Perhaps this is accomplished through interplay between the Neuregulin pathway and other pathways that have been implicated in the regulation of trabeculation, such as the Semaphorin/Plexin (Toyofuku et al.,2004) or Bmp10 (Chen et al.,2004) signaling pathways.

Our studies indicate that normal blood flow is also required for the progression of trabeculation. However, the molecular basis for the influence of blood flow on myocardial cell behavior remains unknown. It may be that the biomechanical forces associated with hemodynamics can trigger relevant signaling pathways. The shear forces generated by circulating blood are known to have a significant impact on heart development (Hove et al.,2003; Vermot et al.,2009; Culver and Dickinson,2010), and the exposure of the endocardium to shear forces could alter its signaling to the myocardium. Notably, it is not yet known whether alteration of hemodynamics can affect Neuregulin signal transduction between the endocardium and myocardium (Lai et al.,2010). In addition to the impact of shear forces on the endocardium, the stretch forces experienced during normal ventricular loading could provoke intracellular signaling within the myocardium to drive trabeculation (Culver and Dickinson,2010). Whether through shear or stretch forces, it is intriguing to consider that hemodynamics could contribute to site selection for the initiation of trabeculation. Perhaps the reproducibility of the initial location of myocardial protrusion into the lumen reflects a peak site of biomechanical impact resulting from normal ventricular loading, whereas the irregular protrusions observed in wea mutants reflect their irregular and weak patterns of blood flow.

Over the long term, we anticipate that the zebrafish embryo will serve as an influential model organism for deciphering the etiologies of congenital defects in trabeculation. For example, our identification of the impact of blood flow on trabeculation suggests intriguing possible mechanisms responsible for hypertrabeculation or hypotrabeculation in patients. Perhaps slight alterations in embryonic blood flow, potentially in concert with fluctuations in Neuregulin signaling levels, could lead to initially subtle defects in trabeculation that magnify over time. Further exploitation of the zebrafish system to model trabeculation defects will likely allow us to unravel their mechanistic underpinnings.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Zebrafish

We used the following zebrafish mutations and transgenes: weak atriumm58 (wea; Berdougo et al.,2003; Auman et al.,2007), clochesk4 (clo; Schoenebeck et al.,2007), Tg(myl7:egfp)twu277 (Huang et al.,2003), Tg(myl7:dsredt4)sk74 (Garavito-Aguilar et al.,2010), and Tg(myl7:egfp-hshras)s883 (Chi et al.,2008). In some experiments, we prevented pigmentation by including 0.003% phenylthiourea (PTU) in the embryo medium before 24 hpf (Elsalini and Rohr,2003).

We chose to use multiple transgenic lines to ensure that the observed phenotypes were reproducible in different genetic backgrounds. Indeed, comparable cardiac phenotypes were found in all three settings. Embryos shown represent the most common morphology that was observed in at least 20 wild-type embryos, at least 15 clo mutant embryos, at least 15 AG1478-treated embryos, and at least 5 wea mutant embryos at each stage examined. In each stable transgenic line, we found considerable variability in the level of transgene expression among individual cardiomyocytes; this degree of variegation is compatible with previous observations of transgenes driven by the myl7 promoter and is presumably due to epigenetic silencing (e.g., Holtzman et al.,2007).

Imaging

Live embryos were anesthetized with tricaine to the point of cardioplegia before being mounted for imaging. Confocal images were obtained with a Zeiss LSM510 microscope using ×25 and ×40 objectives. Confocal planes were collected in the z-slice orientation (Fig. 1G). Fifty to 100 focal planes were collected at intervals of 1–2 μm.

Image Processing

We used Imaris software (Bitplane, St. Paul, MN) for analysis and reconstruction of images from confocal z-stacks. For each z-stack, three different types of image processing were used to obtain different perspectives on cardiac anatomy. All of the image series shown in Figures 1–6 include all three perspectives presented in the same sequence. The specific confocal z-stacks used for the reconstructions presented in Figures 1–6 are provided as Supp. Movies S1–S13.

Volume reconstruction.

The first perspective shown in each series is a volume reconstruction that depicts the orientation of the ventricle. We strived to orient the ventricle similarly in all cases, although some variations in embryo positioning were unavoidable. As schematized in Figure 1G, each embryo was oriented laterally and viewed from its left side. Thus, this perspective provides a view of the overall dimensions of the imaged chamber.

Surface reconstruction.

The second perspective shown in each series is a surface reconstruction of the lumenal surface of the ventricular outer curvature. Imaris can generate a geometric object from a volume image by using an operator-specified intensity threshold to distinguish between background voxels (volume elements) and object voxels; a triangulated surface is then rendered from the object voxels and pseudocolored gray. By compiling a selected range of the optical slices, this perspective provides a view of the inner contours of the ventricle.

Optical sections.

The third perspective shown in each series is a set of optical sections that represent individual two-dimensional slices in all directions using selected crosshairs as a reference point. Figure 1G provides a schematic example of selected crosshairs and the resulting slices, and Figures 1H, 1H′, 1H″, and 1H″′ diagram the optical sections as they are arranged in Figures 1C, 1C′, 1C″, 1C″′, and all other corresponding panels. The orientation of the name of each slice in Figures 1H, 1H′, 1H″, and 1H″′ is consistent with the orientation of the images displayed in Figures 1C, 1C′, 1C″, 1C″′, and all other corresponding panels. Each z-slice (sagittal section) is attained by slicing the ventricle sequentially from its left side to its right side (Fig. 1H). Each x-slice (transverse section) is attained by slicing the ventricle sequentially from anterior to posterior (Fig. 1H′). Each y-slice (coronal section) is attained by slicing the ventricle sequentially from ventral to dorsal (Fig. 1H″). The x-slices and y-slices are derived from reconstructions of the original data set and therefore are not as sharp as the z-slices. Figure 1H″′ depicts the region of interest (yellow box) on a maximum intensity projection of the data from which the optical sections are taken. In the maximum intensity rendering, the ventricle seems to be in front of the atrium due to the higher transgene expression levels in the ventricle. Together, the optical sections provide confirmation of specific anatomical features from multiple views.

AG1478 Treatment

Dechorionated embryos were incubated with 4 μM AG1478 (Calbiochem, San Diego, CA) in 2% DMSO (Lyons et al.,2005; Scherz et al.,2008). Control embryos were incubated in only 2% DMSO. AG1478 and/or DMSO were applied at 27 hpf. Qualitative assessment did not indicate significant impediment of blood flow through the ventricle by either treatment, in contrast to the reduced flow observed in wea mutant embryos (Berdougo et al.,2003; Auman et al.,2007).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We thank L. Pandolfo, K. McCrone, and E. Reynolds for expert zebrafish care, and T. Evans, D. Fitch, A. Joyner, H. Knaut, J. Torres-Vázquez, and members of the Yelon lab for constructive discussions.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
DVDY_22526_sm_Suppmovie1.mov1488KSupporting movie 1 Confocal sections of the wild-type ventricle depicted in Figure 1A–C.
DVDY_22526_sm_Suppmovie2.mov2408KSupporting movie 2 Confocal sections of the wild-type ventricle depicted in Figure 1D–F.
DVDY_22526_sm_Suppmovie3.mov5232KSupporting movie 3 Confocal sections of the wild-type ventricle depicted in Figure 2A–C.
DVDY_22526_sm_Suppmovie4.mov2046KSupporting movie 4 Confocal sections of the wild-type ventricle depicted in Figure 2D–F.
DVDY_22526_sm_Suppmovie5.mov4002KSupporting movie 5 Confocal sections of the wild-type ventricle depicted in Figure 3A–C.
DVDY_22526_sm_Suppmovie6.mov2545KSupporting movie 6 Confocal sections of the wild-type ventricle depicted in Figure 3D–F.
DVDY_22526_sm_Suppmovie7.mov3363KSupporting movie 7 Confocal sections of the clo mutant ventricle depicted in Figure 4A–C.
DVDY_22526_sm_Suppmovie8.mov3497KSupporting movie 8 Confocal sections of the dimethyl sulfoxide-treated ventricle depicted in Figure 5A–C.
DVDY_22526_sm_Suppmovie9.mov4036KSupporting movie 9 Confocal sections of the AG1478-treated ventricle depicted in Figure 5D–F.
DVDY_22526_sm_Suppmovie10.mov3518KSupporting movie 10 Confocal sections of the dimethyl sulfoxide-treated ventricle depicted in Figure 5G–I.
DVDY_22526_sm_Suppmovie11.mov3570KSupporting movie 11 Confocal sections of the AG1478-treated ventricle depicted in Figure 5J–L.
DVDY_22526_sm_Suppmovie12.mov3362KSupporting movie 12 Confocal sections of the wea mutant ventricle depicted in Figure 6A–C.
DVDY_22526_sm_Suppmovie13.mov2235KSupporting movie 13 Confocal sections of the wea mutant ventricle depicted in Figure 6D–F.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.