Time-lapse analysis reveals local asymmetrical changes in C-looping heart tube

Authors

  • Hinako Kidokoro,

    1. Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Aobayama, Aoba-ku, Sendai, Japan
    Search for more papers by this author
  • Masataka Okabe,

    1. Department of Anatomy, The Jikei University School of Medicine, Minato-ku, Tokyo, Japan
    Search for more papers by this author
  • Koji Tamura

    Corresponding author
    1. Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Aobayama, Aoba-ku, Sendai, Japan
    • Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Aobayama Aoba-ku, Sendai 980-8578, Japan
    Search for more papers by this author

Abstract

Heart development has long served as a model system of left–right asymmetrical morphogenesis, and many key laterality genes have been shown to be involved in the process of asymmetrical heart looping. We established a time-lapse imaging system to observe the process of C-looping during chick heart development, and our observations showed that the C-looping is a very complicated process that involves several local changes in shape: the process can be divided into dextral rotation of the rostral and caudal segments with ventral bending in the rostral part and horizontal anti-clockwise rotation with enlargement of the left part in the caudal segment. Further experimental manipulations revealed characteristics of these morphological changes and regional interactions for the events, and we propose that asymmetrical enlargement of the caudal part is one of the targets of the laterality genes in the C-looping process. Developmental Dynamics 237:3545–3556, 2008. © 2008 Wiley-Liss, Inc.

INTRODUCTION

In contrast to symmetrical morphology of outer appearance, the morphology and location of each internal organ in vertebrates are mostly asymmetrical. The molecular pathway that controls the asymmetrical morphogenesis of internal organs has become clear (Schlueter and Brand,2007). Asymmetrical morphogenesis along the left–right (L–R) axis progresses through three sequential steps (Mercola and Levin,2001; Hamada et al.,2002). (1) The first step is initial breaking of bilateral symmetry. This begins at a local small part of the embryo and takes place at or around the node. (2) The second step is transmission and propagation of local polarity. The local polarity established at the node is transmitted to asymmetric gene cascades and propagated throughout the lateral plate mesoderm (LPM). In the left-specific pathway, Nodal, a TGF-superfamily member, is transiently expressed in the left side of the LPM and induces a bicoid homeobox transcription factor, Pitx2c (Logan et al.,1998; Meno et al.,1998; Piedra et al.,1998; Ryan et al.,1998; Yoshioka et al.,1998). While the expression of other upstream genes disappears before the organogenesis, Pitx2c continues to be expressed on the left side of each organ throughout the next step. (3) The third step is asymmetrical organogenesis. Asymmetrical morphogenesis in each organ is carried out based on the L–R information. Pitx2c is thought to play a key role in this process because both the gain- and loss-of-functions of Pitx2 result in the failure of normal L–R morphogenesis in the heart and other asymmetrical organs (Logan et al.,1998; Ryan et al.,1998; Campione et al.,1999; Kitamura et al.,1999; Lin et al.,1999).

In the vertebrate embryo, the developing heart is the first organ that exhibits morphological L–R asymmetry. In early cardiac development, a pair of heart primordia originating from the two sides of the embryo body fuses to form a single straight tube at the center of the body. In the chick embryo, the fused straight heart tube can be first observed at Hamburger and Hamilton (HH) stage HH 8+/9− (6–7 somite stage, Hamburger and Hamilton,1992; Manner,2000). Although the early heart tube is relatively short at this stage, rapid elongation along the craniocaudal axis occurs in tandem with the progression of caudal fusion. Until stage HH 9−/9+ (7–9 somites), the heart tube is almost bilaterally symmetrical, and then it begins to deform at HH 9+/10− (9–10 somites), the straight morphology transforming into a C-shaped tube (Manner,2000; Martinsen,2005). In the C-looping process, the heart tube twists to turn the original ventral midline to the right side of the embryo body (dextral rotation or rightward rotation) and bends toward its ventral side (Manner,2000). Thus, the original ventral side of the tube becomes the outer convex curvature of the C-shaped heart rudiment. The C-looping process has long been investigated as a fascinating asymmetrical phenomenon, and most studies have focused on the dextral rotation mechanism. Itasaki et al. (1991) suggested that actin bundles on the right caudal part of a heart tube generate tension and cause dextral rotation (Itasaki et al.,1991). In another study, Voronov and his colleagues examined biomechanical forces in cardiac rotation. The results suggested that the rotation is driven by external forces exerted by the splanchnopleure (SPL) and the omphalomesentric veins (OVs; Voronov et al.,2004; Nerurkar et al.,2006), and unbalanced forces in the OVs are responsible for the looping directionality. Based on data obtained from inhibition experiments of MMP-2, Linask et al. (2005) suggested that asymmetric cell proliferation in the dorsal mesocardium, through which the heart tube is suspended in the embryo body, can drive dextral looping (Linask et al.,2005). However the dextral-rotation mechanism remains elusive and little is known about the function of laterality genes in the C-looping process.

Although much attention has been given to laterality in the cardiac C-looping process, there have been few reports in which morphological differences between the left and right sides of the heart rudiment other than the looping direction are described. However, due to the multiplicity of this process that characterizes heart development, the complex morphological changes that occur between sided gene expression and the final cardiac rotation phase should be elucidated to understand the cellular mechanisms of asymmetrical morphogenesis. Moreover, because the delicate nature of the heart primordium can easily cause irregularities in the looping direction as a result of experimental manipulation (Lepori,1967; Voronov and Taber,2002), the use of only the directionality of looping for judging heart laterality in the C-looping stage is sometimes insufficient and inaccurate. Thus, more morphological descriptions regarding asymmetry other than the looping direction would be useful for investigation of the function of laterality genes. To gain insights into the downstream mechanisms of the laterality genes and to find novel indicators of L–R asymmetry in the developing heart tube, this study was carried out with the aim of elucidating the side-specific morphological character provided by the laterality genes in the C-looping heart rudiment.

RESULTS

Time-Lapse Observation of the C-Looping Process of the Chick Heart Tube

L–R asymmetrical morphology of the early heart tube is described in some reports on the early process of development of the chick heart tube (Stalsberg,1969a; Stalsberg and DeHaan,1969; Manner,2000), but the information is fragmented. We began this study with observation of the C-looping process of the normal chick heart using a time-lapse microscopy, focusing on the detailed morphological change and time course of this change. Chick embryos were taken out from eggs using a paper culture technique (Chapman et al.,2001), and a series of time-lapse images were taken every 15 min (see the Experimental Procedures section). In our time-lapse observation, embryos often showed a slight delay or defect in rotation (data not shown). This may have been caused by surface tension because it has been reported that embryos placed on a medium without a layer of fluid are exposed to surface tension which can influence rotation and morphology of hearts (Voronov and Taber,2002). As found in the previous work, in our time-lapse observation some embryos showed a delay (n = 3/10) or defect (n = 2/10, data not shown) in rotation. These embryos showed abnormal morphology and movement of the heart that appeared to be compressed and to be dragged by SPL. However, it was impossible to culture embryos under the recommended condition (under a layer of fluid in 95% O2 and 5% CO2; Voronov and Taber,2002) to eliminate surface tension in our systems. We, therefore, put a few drops of medium on the embryos instead a layer of fluid, and we repeated filming the C-looping process several times and showed one that appeared to be normal.

At the eight-somite stage (HH 9+ stage, time point 0:00), the morphology of the heart rudiment appeared to be almost symmetrical, and the midline of the heart tube was directed to the ventral side of the embryo body (Fig. 1A). As described previously, the primitive straight heart tube bent ventrally and turned to direct the ventral midline toward the right (Manner,2000) in the subsequent steps. It took approximately 10 hr for the formation of a completely rotated heart tube with a C-looped shape in our culture system (Fig. 1A–E, see also Supplementary Video S1, which can be viewed online). To observe the morphological change more clearly in this process, four rostrocaudal points at the ventral midline of the heart rudiment were labeled by the fluorescent dye DiI (1,1-dioctadecyl-3,3,39,39-tetramethylindocarbocyanineperchlorate; Fig. 1F–M, see also Supplementary Video S2). Each labeled point in the ventral midline turned rightward, almost evenly along the rostrocaudal axis (Fig. 1F–M).

Figure 1.

A series of time-lapse images of the normal C-looping heart tube. All of the pictures are ventral views, and the left side of each picture is the right side of the embryo. AE: A series of brightfield time-lapse observations: (A) 0h00min (8-somite stage), (B) 2h30min (9-somite stage), (C) 5h00min (10- to 11-somite stage), (D) 7h30min (12-somite stage), (E) 10h00min (13-somite stage). See also Supplementary Video S1. FM: Four points at the ventral midline of a symmetrical straight heart tube of a 9-somite stage embryo were labeled by DiI (1,1-dioctadecyl-3,3,39,39-tetramethyl-indocarbocyanineperchlorate; F, 0h00min). See also Supplementary Video S2. G: 1h00min. Note that the frontal edge of the caudal rudiment in the left side is located at a slightly more anterior level than that in the right side (arrows). H: 2h15min. I: 3h30min. A mound-like bulge can be seen in the left side of the caudal part (arrow). J: 4h45min. K: 7h15min. L: 9h30min. M: 12h00min. The angle between the rostral and caudal parts in the left side gradually becomes steeper than that in the right side, and the midline of the caudal part (indicated by a green line) becomes slanted to the right. DiI-labeled points were completely directed to the right side of the body within approximately 9 hr (L). N: Schematic representation of the C-looping process of the heart tube. Left rostral (Lr), left caudal (Lc) and right caudal (Rc) parts of the heart tube are depicted by red dots, red stripes and blue stripes, respectively. L, left; R, right. Scale bar = 200 μm.

In this process, an asymmetrical symptom could already be seen at the time point 0:00. The heart tube has a handstand Y-shape with bilateral inflow tracts at the caudal end, and there is a pair of lateral furrows at the hinge of bifurcation (interventricular groove, indicated by arrowheads in Fig. 1F). In the following, the part rostral to the furrows will be called rostral segment and the part caudal to the furrows will be called caudal segment for descriptive purposes. At the nine-somite stage (HH 10− stage, time point 0:00), the caudal rudiment in the left part was located at a slightly more anterior level than that in the right part (Fig. 1F). One hour later, this rostrocaudal lag between the left side and right side became more prominent (arrows in Fig. 1G). In subsequent steps when the heart tube deformed to bend ventrally and rotated rightward while continuing to fuse at the caudal end, the left part of the two-forked caudal rudiment kept forming a mound-like shape at the midline (indicated by an arrow in Fig. 1I) and elongated in the rostral-right direction. Simultaneously, the right corresponding part shifted caudal-left, giving rise to a partial anti-clockwise rotation of the caudal part at the horizontal plane (see arrows in Fig. 1K). The midline of the caudal part (indicated by a green line in Fig. 1G and M) was tilted to the right. In fact, the caudal segment was tilting concomitantly with the rightward movement of DiI labels, indicating that counter-clockwise movement was accompanied by rightward rotation of the entire heart tube. The caudal segment, therefore, showed horizontal rotation with rightward vertical torsion, and the left rudiment overrode on the right rudiment, while the right tube slipped under the fused heart tube. DiI-labeled points were completely directed to the right side of the body at 9 hr, indicating that dextral rotation was achieved in this time frame (Fig. 1F–M).

Difference Between Sizes of the Right Part and Left Part of the Caudal Heart Rudiment

Because the above time-lapse analysis showed an early sign of asymmetric morphology in the cardiac caudal part of the heart tube, we carefully compared the morphology of the left caudal part with that of the right caudal part at some stages. At the 10-somite stage (HH 10 stage), when the heart rudiment is a bilateral single tube, the left part of the caudal rudiment is slightly larger than the right part (Fig. 2A, asterisk). At the 12-somite stage (HH 11− stage), when the dextral rotation is completed, a small bulge was observed in the left caudal region (Fig. 2B, arrow), but such a bulge structure could not be found in the right caudal region (Fig. 2C, arrowhead). To further clarify the morphological difference in left and right caudal parts, the heart tube at the 13-somite stage (HH 11 stage) was excised from the embryo body and observed in vitro. Observation of the excised heart rudiment revealed that the caudal part in the left side is noticeably larger (Fig. 2D. arrow) than that in the right side (Fig. 2D, arrowhead), indicating that the above in vivo observation had no bias due to the fixed pose in terms of morphology. We further investigated the sizes of the left and right parts of the caudal region using a gene marker, atrial-specific myosin heavy chain (AMHC1), which is exclusively expressed in the caudal region of the heart tube at the C-looping stage (Yutzey et al.,1994). We used AMHC1 merely as a marker for the caudal region, not suggesting a function of AMHC1 in morphogenesis. At the eight- to nine-somite stage (HH 9+/10− stage), AMHC1 was almost evenly expressed in the two sides of the caudal rudiment (Fig. 2E). The asymmetric expression pattern of AMHC1 was first observed at around the 9- to 10-somite stage (HH 10−/10 stage), and the expression domain of AMHC1 was slightly larger in the left side than in the right side of the caudal part (data not shown). The asymmetric expression pattern of AMHC1 was more prominent at the 11- to 14-somite stage (HH 10+/HH 11− stage, Fig. 2F–G). These results supported the above morphological observations, showing that the left caudal part of the heart tube was larger than the right caudal part in the C-looping stage embryo.

Figure 2.

A: A straight heart tube at the 10-somite stage before C-looping. The ventral view appears asymmetrical, with the left part of the caudal rudiment being slightly larger (asterisk) than the right part. Note that the rostral portion also shows asymmetry in size (red brackets). B,C: A C-looped heart tube in the normal chick embryo (12-somite stage). The heart tube has a bulge in the left caudal region (B, arrow), but such a bulge structure cannot be observed in the right posterior region (C, arrowhead. In C, the C-looped tube in B was turned over by a fine needle.). D: A heart tube excised from the body at the 13-somite stage. The left part in the caudal rudiment (arrow) is clearly larger than the right part (arrowhead). EG: Whole-mount in situ hybridization of AMHC1 for C-looping stage embryos. E: Eight- to nine-somite stage embryo. AMHC1 is expressed almost evenly in the two sides of the caudal heart rudiment. F: 11-somite stage. G: 14-somite stage. Note the asymmetric expression of AMHC1. L, left; R, right. Scale bars = 200 μm.

However, it was still possible that the size difference in the caudal part is merely due to a secondary consequence of looping, caused by an effect of tension. To address this issue, we prevented fusion of the left and right heart rudiments at the midline by ablation of the central endoderm beneath the heart rudiment (DeHaan,1959, and see the Experimental Procedures section). After this operation, each heart rudiment, which separately developed at each side of the embryo, still showed a size difference in the caudal part (n = 46, Fig. 3). Although it was not clear whether the wrinkles seen in the cardia bifida correspond to the border of the rostrocaudal compartment, the most-posterior part of the heart divided by the wrinkles was almost always larger in the left rudiment than in the right rudiment (Fig. 3A, compare red brackets). We next examined AMHC1 expression in the cardia bifida. The expression domain of AMHC1 was slightly larger in the left side than in the right side at 10-somite stage (HH 10 stage, Fig. 3B) and the asymmetric expression pattern of AMHC1 was more prominent at 13-somite stage (HH 11 stage, Fig. 3C) after this operation. These findings suggest that the size difference in the left and right sides of the caudal part is not due to a secondary consequence but to a genetically programmed morphological change.

Figure 3.

Double-hearted embryos (cardia bifida) in which fusion of the left and right heart primordia had been prevented by ablation of the central endoderm beneath the heart rudiment. A: After this operation, separately developed heart tubes showed a difference in size of the caudal parts (indicated by red brackets). B,C: Expression pattern of AMHC1 after this operation (10-somite stage embryo [B], and 13-somite stage embryo [C]). Arrowheads in B and C indicate the rostral and caudal ends of the heart region. Asymmetrical expression domain of AMHC1 supported the morphological observation. L, left; R, right. Scale bars = 200 μm.

Role of Nodal-Pitx2 Pathway in Asymmetrical Morphogenesis in the Caudal Part

Shh is known to be expressed only at the left side of the chick Hensen's node from stage HH 4+ and to induce the left-specific Nodal-Pitx2 pathway on the left LPM (Levin et al.,1995). To determine whether the Nodal-Pitx2 pathway is responsible for the increase in size of the left caudal part of the heart tube, we implanted a bead soaked in Shh-containing solution on the right side of Hensen's node in HH 4–5 stage embryos (Fig. 4) and examined the effect on the morphology of the heart rudiment at the 13–14 somite stage (HH 11/11+ stage), when the C-looping process is completed. This manipulation of the embryos resulted in bilateral expression of Pitx2 in the heart tube and the LPM (Fig. 4B, arrowheads). Shh treatment of the right side randomized the looping direction. The ventral midline was directed in the rightward direction in 64% (39/61) of the hearts, reversed rotation occurred in 33% (20/61) of the hearts, and in some cases hearts that had hardly turned remained in the midline (3%, Fig. 4D, n = 2/61). In the rightward looped hearts, 38% (n = 15/39) showed normal morphology and looping, while the right caudal part of the remainder (62%, n = 24/39) was expanded to the same extent as that of the corresponding left part (Fig. 4F, arrow). Enlargement of the right caudal part was also observed in the reversed heart tubes (Fig. 4G, arrow; 19/20 = 95%) as well as in unrotated heart rudiments (2/2 = 100%). Control embryos never showed expansion of the right caudal part (n = 29/29) even when they were reversely looped (n = 6/6). It should be noted that the bulge in the left side was still observable (Fig. 4F,G, arrowheads) in the Shh-treated embryos. Furthermore, in these embryos, the AMHC1 expression domain was broad also in the right side of the heart tube compared with that in the normal embryo (Fig. 4H–J), corresponding to the morphological observations. Transverse sections clearly showed that the heart tube of a Shh-treated specimen, which had a bilateral morphology (Fig. 4D), had an expanded right caudal part (compare Fig. 4D with Fig. 4C). These results strongly suggest that the Nodal-Pitx2 pathway is involved in the increase in size of the left caudal part during the cardiac C-looping process.

Figure 4.

Shh protein-soaked bead implantation in the early-stage chick embryo. A,B: Pitx2 expression in embryos with a phosphate buffered saline (PBS) -soaked bead (control; A) and a Shh-soaked bead (B). Pitx2 is bilaterally expressed in the Shh-applied embryo (B, arrowheads), whereas control embryo shows left-sided expression of Pitx2 (A, arrowheads). C,D: Hematoxylin and eosin (H&E) -stained paraffin sections of a normal C-looped heart tube (C, 11- to 12-somite stage) and a double left heart tube (D, 12-somite stage). The Shh bead-implanted embryo shows expansion of the right caudal part of the heart tube (arrow). EG: Ventral views of a control embryo with a PBS-bead (E, 13-somite stage) and embryos with a Shh-bead (F, rightward looped heart (13-somite stage); G, reversed tube (13-somite stage)). To clearly observe the morphology of hearts, SPL over the heart rudiments were removed in these embryos. In Shh-treated embryos, a bulge structure was observed in both sides of the caudal rudiment (arrows and arrowheads in F and G). HJ: Expression of AMHC1 in a normal embryo (H, 12- to 13-somite stage) and Shh bead-implanted embryos (I, 10- to 11-somite stage; J, 12-somite stage). AMHC1-expressing domain in the right side of the caudal part in I and J is as large as that in the left side. L, left; R, right. Scale bars = 200 μm.

Dependency of the Rostral and Caudal Parts for their Asymmetrical Morphogenesis

The dextral rotation of the heart tube has been regarded as a landmark of the cardiac laterality in many studies. Our time-lapse observations, the results of which are shown in Figure 1, demonstrated that the main part for dextral rotation is the rostral part of the heart tube and that the caudal part of the tube gives rise to an asymmetrical morphogenesis different from that in the case of the rostral part, size change and partial anti-clockwise rotation. These findings in the present study gave rise to the question of whether the morphological asymmetry of the caudal part is related to the dextral rotation. To assess the role of the caudal part in the dextral rotation, we performed an ablation experiment in which we divided the heart tube into rostral and caudal segments at the level of the interventricular groove and removed the caudal part at around the 9-somite stage (HH 10− stage, Fig. 5A). Time-lapse analysis showed that the residual rostral part of the heart tube twisted and completely turned its ventral midline to the right side within 9 hr (Fig. 5B–G, see also Supplementary Video S3, n = 8/14), although in a few cases the anterior remaining part showed a delay in rotating (n = 3/14) or failed to make a complete rotation (n = 3/14, not shown). It is likely that the rostral part of the heart tube after the 9-somite stage (HH 10− stage) can achieve dextral rotation without the caudal part.

Figure 5.

A: Ablation experiments for examining the role of the caudal part in asymmetrical morphogenesis of the rostral portion, in which the heart tube was divided into two portions at the interventricular grooves (IG, arrowheads in A) and the caudal part was ablated at the nine-somite stage. BG: Time-lapse series of the caudal part-ablated heart tube. See also Supplementary Video S3. (B) 0h00min. Arrowheads indicate the excision level. C: 2h15min. D: 4h45min. E: 7h15min. F: 9h45min. G: 12h15min. DiI (1,1-dioctadecyl-3,3,39,39-tetramethyl-indocarbocyanineperchlorate) labeling revealed that the residual anterior part of the heart tube completely turns its ventral midline toward the right side within 7.5h. Scale bar = 200 μm.

Next, we examined the contribution of the rostral part to asymmetrical morphogenesis in the caudal part of the heart tube. The rostral part of the heart tube was ablated at the nine-somite stage (HH 10− stage), and morphological change of the remaining caudal part was analyzed by time-lapse imaging (Fig. 6, see also Supplementary Video S4). The left part of the two-forked caudal segment in the operated heart primordium elongated in the rostral-right direction, and the right part moved caudal-leftward (Fig. 6A–F). The DiI-labeled cell populations shifted toward the right side, and the ventral midline was tilted rightward (Fig. 6A–F, indicated by a green line), indicating that the operated heart tube showed morphological changes. However, the DiI-labeled cell population hardly turned to the right, and the angle of the tilt was much smaller than that in the normal process (n = 8/8; Fig. 6F, compare to Fig. 1M). These results suggest that the rostral segment is needed for the caudal part to exert dextral rotation as well as anticlockwise movement. It should be noted that the size of the remaining left side was larger than that of the right side (Fig. 6F), and the difference in size is consistent with the results of fusion inhibition experiments (Fig. 3). These results suggest that the difference in size of the caudal part, which is probably not due to a secondary consequence raised by a mechanical effect of the C-looping, is produced by an intrinsic cellular property, although other asymmetrical events in the caudal part appear to be mediated by the influence of the rostral part.

Figure 6.

The rostral segment was ablated (A, arrowheads) to examine its role in morphogenesis of the caudal segment (see Fig. 5A). Time-lapse series of the rostral part-ablated heart tube. See also Supplementary Video S4. A: 0h00min. B: 2h15min. C: 4h45min. D: 7h15min. E: 9h45min. F: 12h15min. The size of the left side in the remaining caudal part (arrowhead) is still larger than that of the right side, while a series of DiI (1,1-dioctadecyl-3,3,39,39-tetramethyl-indocarbocyanineperchlorate) -labeled cell population shifting toward the right side show incomplete displacement (compare Fig. 6 with Fig. 1F). Scale bar = 200μm.

DISCUSSION

Time-Lapse Observation of C-looping

Remarkable progress has been made over the past decade in elucidation of the molecular basis for establishment of the L–R axis in vertebrates, and several studies have shown that gain- and loss-of-functions of key genes give rise to abnormal morphology of asymmetric organs, including isomerism and situs inversus (Levin et al.,1995; Supp et al.,1997; Nonaka et al.,1998; Ryan et al.,1998; Kitamura et al.,1999; Marszalek et al.,1999; Watanabe et al.,2003). However, it is difficult to trace the intermediate process connecting gene expression/function and phenotypes, and downstream mechanisms to embody the gene function in the asymmetrical morphogenesis indeed remain largely unresolved. The dextral rotation of the heart tube has been regarded as a landmark of the asymmetrical morphogenesis, and the looping direction has been used as almost the only indicator of laterality in many studies on the C-looping process, but the delicate nature of the heart primordium, which makes the looping direction susceptible to mechanical influence, has made it difficult to solve this problem. More detailed morphological features and changes including features that are not susceptible to mechanical influence should, therefore, be recorded. Continuous observation of morphological change to obtain an overall picture of the developmental process would be helpful for identifying key actions responsible for morphogenesis, and we used this technique. In this study, to morphologically characterize each side of the heart tube and to define morphological features provided by the left-specific pathway in the C-looping stage, we successfully established a system for comprehensive observation of the C-looping heart rudiment with a combination of fluorescent and brightfield time-lapse imaging.

According to our time-lapse analysis of our culture system, the rotation started at around the 9-somite stage (HH 10− stage) and took approximately 9 hr, being completed by the 13-somite stage (HH 11 stage). At the end of the rotation, the points labeled at the ventral midline were observable at the right margin of the rotated tube (Fig. 1M). In the middle of this process, the rotating tube started beating at approximately 5 hr. It has been suggested that initiation of heart contraction and blood flow might be involved in the cardiac looping process (Hove et al.,2003; Linask and Vanauker,2007), and our results showing that the heart tube starts looping before the onset of heart beating suggest that the looping is initiated independently of the beating. Time-lapse analysis gave a whole picture of the complicated morphological change in the early heart rudiment and further revealed that this process includes multiple local changes in shape that are temporally coordinated to properly accomplish dynamic morphogenesis. The main element for the dextral rotation is the rostral part of the heart tube, but the caudal part also shows complex morphological changes. A morphological symptom of the left-right asymmetry in the bilateral straight heart tube was first seen at the nine-somite stage (HH 10− stage) before the onset of looping. The frontal edge of the left side in the bifurcated caudal rudiment was located at a more anterior level than was that of the right side (Fig. 1F), and the angle between the rostral segment and caudal segment (angle at the L-shaped hinge region) was steeper in the left side than in the right side. A similar observation was reported by Manner (2000), who pointed out in his review that a flattening of the right lateral furrow and a deepening of its left counterpart are the first signs of morphological left-right asymmetry. This asymmetrical shape of the caudal segment became prominent as looping proceeded. In the subsequent process, the left side of the caudal part moved in the rostral-right direction as if the left caudal part had pushed the rostral part toward the right side. The right side of the bifurcated caudal segment shifted in the caudal-left direction to fuse with its left counterpart and form a compact shape, and hence the overall caudal segment showed anti-clockwise movement at the horizontal plane. In this process, the left side of the caudal segment became larger than the corresponding region in the right side, supported by results from comparison of the AMHC1-expressing area (Figs. 2, 3). As described above, the rostral segment rotated in the rightward direction, and the caudal segment also showed partial rightward rotation (see Fig. 1F–M,N, Supplementary Video S2). The turning in the right direction appeared to occur in a manner involving corporation of the rostral and caudal segments. Meanwhile, the rostral segment of the heart tube bends ventrally, concomitantly with the elongation along the rostrocaudal axis of the embryo. Therefore, the C-looping process of the heart tube can be divided into at least two events: rightward vertical rotation of the rostral and caudal segments with ventral bending in the rostral part and horizontal anti-clockwise rotation with enlargement of the left part in the caudal segment.

Morphological asymmetry was also observed in the rostral segment of the heart rudiment in a prelooping phase: the size of the rostral portion in the right side was slightly but clearly larger than that in the left side (Fig. 2A, brackets). On the other hand, the left part in the caudal segment is larger than the right part (Fig. 2A, asterisk), and the total proportion of rostral and caudal parts is, therefore, different in the left and right sides before the heart looping and, moreover, this difference may be related to the later asymmetric morphological change in the tube. This observation is consistent with the results of a previous study showing that approximately 55% of epimyocardial cells in the cephalic part of the C-looped heart tube are derived from the right side and that only 40% of the cells in the caudal part are derived from the right side (Stalsberg,1969a).

Rostrocaudal Interactions and Mechanical Influence for C-looping

Various mechanisms of dextral rotation have been proposed, although which one is most appropriate for the machinery remains unclear. Itasaki et al. (1991) proposed that actin bundles in the caudal part of the heart tube (which seems to correspond to the posterior part of the rostral segment in this study) play an important role in driving the dextral rotation. On the other hand, Linask et al. (2005) suggested that asymmetrical cell proliferation in the dorsal mesocardial folds drives the looping direction, while a recent study suggested that the dextral rotation is driven mainly by the SPL and the left omphalomesentric vein (Voronov et al.,2004; Nerurkar et al.,2006).

In this study, we carried out some ablation experiments and showed that the rostral part of the heart tube could almost normally achieve dextral rotation without involvement of the caudal segment (Fig. 5). It should be noted that the results of these experiments performed in the absence of the SPL do not rule out the possibility of involvement of the SPL in dextral rotation as it is possible that surface tension plays an alternative role of the SPL as mentioned later. Thus, our results suggest that dextral rotation of the rostral part is driven mainly by intrinsic morphological change or force of the rostral part and/or a new segment added from the anterior heart field (Kelly and Buckingham,2002) in the presence of mechanical load of the SPL independently of the caudal part, although the possibility of participation of external force from the caudal segment cannot be excluded.

On the other hand, ablation of the rostral segment resulted in failure of dextral rotation as well as anti-clockwise movement of the residual caudal segment (Fig. 6), suggesting that these dynamic movements of the caudal segment depend on the rostral segment. One possibility is that the rostral segment positively pulls the caudal segment to lead to dextral rotation in the normal C-looping process. This hypothesis leads to the following interpretation of the anti-clockwise motion that we observed in the caudal segment: Pulled by the rostral segment, the midline of the caudal portion is tilted to the right, and subsequent fusion of the bifurcated caudal rudiment occurs along the titled midline. As a result, the right part of the two-forked caudal rudiment shifts to the caudal-left to fuse with its counterpart, while the left caudal rudiment, accompanied by dextral movement of the rostral portion, moves and elongates in the right direction with increase the size; thus, overall the caudal segment shows anticlockwise motion. Importantly, the size difference between the two sides of the caudal segment was still observed when the rostral segment was excised, whereas its ablation caused a defect of the above dynamic movements. This finding is consistent with our observations in double-hearted (cardia bifida) embryos, suggesting that the size difference in the caudal segment, which is not susceptible to mechanical influence, arise not from the secondary effects of the looping but from genetically programmed events. Although we cannot exclude the possibility that the rostral segment merely plays a supportive role in the anti-clockwise movement of the caudal portion, it seems that appropriate local changes in morphology plus the external force from the rostral segment allow the caudal segment to correctly achieve the C-looping process. Thus, the C-looping process is thought to be accomplished by coordinating the independent asymmetrical morphogenesis of the two parts.

We incised the SPL over the rostral part of the heart rudiment to access the hearts for DiI injection. Previous studies suggested that dissection of the SPL leads to a delay in C-looping because the SPL is an important extrinsic factor for C-looping (Nerurkar et al.,2006). According to the report, in the absence of surface tension and the SPL, torsion of the heart is temporarily suppressed, and several hours later the heart starts turning by adaptive response to the change in mechanical environment, resulting in a delay in looping. In our time-lapse analysis, such a delay was not observed in 60% of the embryos (n = 3/5). These hearts started to turn immediately after incubation (see Supplementary Video S2) and completed rotation in almost the same time frame as that of the embryo with the SPL intact (Supplementary Video S1). Thus, the hearts in our study did not seem to rotate by an adaptation mechanism. This difference between our results and results of a previous study may due to the different culture conditions described above. The surface tension that was thought to be present in our culture system might compensate for the absence of the SPL. On the other hand, 40% (n = 2/5) of the embryos showed a slight delay in looping in our culture system. In these embryos, the remaining SPL appeared to contract and to pull the hearts in the left-caudal direction as reported previously (Itasaki et al.,1991, Voronov and Taber,2002, See Supplementary Video S5). Thus, we speculate that SPL contraction might lead to the delay, although the precise reason for the delay is not clear.

Involvement of Left-Specific Pathway in C-Looping Morphogenesis

Approximately 10 years has passed since the involvement of the Nodal-Pitx2 pathway in the asymmetrical morphogenesis of heart and other internal organs was first suggested (Levin et al.,1995; Lowe et al.,1996), and several molecules upstream of Pitx2 have been reported in chick and mouse embryos (see a review by Schlueter and Brand,2007, and references therein). However little information is available about the morphological features that are provided by the left-specific pathway. Regarding heart development, some studies have shown that Pitx2 is essential for cardiac outflow tract development and for patterning of atrioventricular valves and cushions (Kioussi et al.,2002; Liu et al.,2002). However, these studies focused on the later phase of cardiogenesis after the completion of C-looping, and, to our knowledge, looping direction was the only clue for the function of laterality genes in the heart tube and few studies have shown what kinds of morphological changes the left-specific pathway controls in the C-looping stage. In this study, we observed a prominent difference between sizes of the two sides of the caudal rudiment in the C-looping process (Fig. 2). This size difference was not impaired by the inhibition of loop formation (Fig. 3), suggesting that genetically programmed events control the caudal size of the heart rudiment in the C-looping stage. We performed ectopic induction of the left-specific pathway on the right side of the chick embryo, in which Pitx2 was ectopically induced in the right side of the embryo. In the heart rudiment of those embryos, the right part of the caudal segment was enlarged, while the size of the left part was never reduced (Fig. 4). These results suggest a specific role of the laterality genes, by which the left caudal part of the heart rudiment is expanded in the C-looping process. We propose that the size of tissue is one of the targets of left-specific pathway functions. There are several possible mechanisms by which the laterality genes provide morphological asymmetry (difference in size) to the caudal part. It has been shown that Pitx2 is involved in promoting cell proliferation in several tissue and cell types (Kioussi et al.,2002; Martinez-Fernandez et al.,2006). Thus, Pitx2 might promote cell proliferation also in the early heart tube for giving rise to an increase in size of the caudal segment. However, a detailed analysis of the regional mitotic activity of the precardiac area have shown that there is no difference between proliferation rates in the left and right sides in the HH 4∼14 stage chick heart tube at any level along the rostrocaudal axis (Stalsberg,1969b). It seems that we need to consider another possibility to explain the cellular mechanism for the size asymmetry of the caudal segment.

There are several other explanations for the larger size of the left caudal part than that of the right one.

Cell size and shape.

An increase in tissue volume caused by a change in cell size and/or cell shape may be responsible. Soufan et al. (2006) reported that in the chick cardiogenesis, only two-thirds of the sixfold increase in volume that the myocardium undergoes from the HH 10 to HH 12 stage is accounted for by the number of cardiomyocytes and the remainder is accounted for by regional increase in cell size (Soufan et al.,2006). Another study showed that ectopic expression of Pitx2a in HeLa cells induced actin-myosin reorganization and resulted in dramatic change in cell morphology (Wei and Adelstein,2002). Although the Pitx2a isoform is different from Pitx2c, which controls asymmetrical morphogenesis in the heart and other internal organs, ectopic introduction of either Pitx2a or Pitx2c equally randomizes the heart looping direction (Yu et al.,2001).Thus, it is conceivable that Pitx2 in the caudal segment of the heart rudiment alters cell shape into a flattened morphology, leading to an increase in cell apical surface area and expansion of the left caudal part. However, more experiments are needed to confirm that Pitx2c can alter the shape of cardiomyocytes in vitro, and of course the possibility should also be examined by in vivo analysis.

Asymmetry in the cell density.

A previous study showed that the myocardium, which initially has large intercellular spaces between cells, gradually becomes more tightly packed with reduction in intercellular spaces in early heart development (Manasek,1968). The difference between the states of cell density in the two sides of the heart rudiment might be responsible for the difference in size or morphology of tissue. This assumption may be considered in conjunction with the above possibility of alteration in cell shape.

Difference in the number of the myocardial layers.

Assuming that a multilayered sheet is made by the same number of cells, the surface area would be inversely proportional to the number of layers. Thus, it is conceivable that the surface area increases in proportion to reduction in the number of layers caused by cell rearrangement. However, some studies have indicated that the myocardial cells are already firmly attached to each other at the onset of C-looping and are not free to migrate (Manasek,1968). It, therefore, might be more accurate to assume that the difference in the number of layers arises before the heart starts to loop.

Decrease of the programmed cell death.

As far as we know, there is no report in which detailed distribution of apoptotic cells during early heart development is described.

Asymmetrical incorporation of precardiac materials into the heart rudiment.

A fate mapping study at the head process stage showed that the subdivisions of the precardiac area that contribute to the rostral region of the stage 12 heart are larger on the right side and that the subdivisions that contribute to the caudal regions of the heart are larger on the left side (Stalsberg and DeHaan,1969). If this estimate is correct, the size difference in the caudal segment of the heart would exist from the very beginning of fusion of the heart-forming mesoderm derived from both sides. This prediction seems to be consistent with our observations that the left side of the caudal segment that has not yet fused was already larger than the right side from the very early stage (Fig. 2A) and that the positions of the two interventricular grooves were not at the same level along the anterior–posterior axis before the heart starts to loop (Figs. 1F, 2A). The primary difference in cell number is expected to expand exponentially with cell division, and it might be able to generate a regional increase of tissue size. It would be worthy to verify this possibility by mathematical analysis.

The above possibilities can be adapted to a cause of the dextral rotation in the rostral part of the heart rudiment, and it is of course possible that multiple factors are involved in asymmetrical morphological changes in the heart rudiment. Discussion of the cellular mechanisms of asymmetrical morphogenesis in the heart development is still in the speculative stage, but our results presented here, which dissected the complex process into several elements of morphological changes, will help further cellular investigations on this issue.

EXPERIMENTAL PROCEDURES

Embryos and Culture Preparation

Fertilized White Leghorn chicken eggs (Iwaya farm, Sendai, Japan) were incubated at 38° C until embryos reached the appropriate stage in accordance with Hamburger and Hamilton (1951). For each manipulation, the embryos were explanted using the modification of New's culture method (EC culture, Chapman et al.,2001). Embryos attached to a vitelline membrane were dissected out from eggs using filter paper rings (Whatman no.1) and put ventral-side-up on agar-albumen culture plates. The embryos were washed with Tyrode's solution to remove the yolk. After each manipulation, the embryos were incubated at 37°C in 5% CO2.

Fluorescent Labeling and Time-Lapse Imaging

Embryos of the 8–9 somite stage (HH 9+/10- stage) were prepared by the above-described whole embryo culture method. DiI (1,1-dioctadecyl-3,3,39,39-tetramethyl-indocarbocyanineperchlorate; Molecular Probes, Inc.) stock solution (0.5% in 100% ethanol) was freshly diluted 1:10 in 0.3 M sucrose containing 0.005% Nile blue sulfate for each experiment. To access the heart rudiment, the splanchnopleura over the rostral part was carefully cut using fine-sharpened tungsten needles, and small subpopulations of cells at the ventral midline of the heart rudiment were labeled by focal injection of DiI solution by means of pressure of expiration and a pulled micropipette. DiI-labeled embryos were visualized using a stereomicroscope (Leica), and a fluorescence imaging system (Leica FW4000). Images were taken every 15 min, and approximately 5–6 Z stacks were recorded for each time point. The most focused images were selected for conversion into movie format (using Leica FW4000 software).

For brightfield time-lapse imaging, six- to eight-somite stage(HH 9−/9+ stage) embryos were prepared, and pictures were taken using a stereomicroscope (Olympus SZX12) with a transmitted light illuminator. Images were taken every 15 min and converted into movie format (by means of NIH ImageJ software). In both imaging systems, a transparent warm plate (Kitazato MP10DM) was placed on the culture plate to maintain the environment at 38 ± 1° C for filming.

Cardia Bifida

To prevent fusion of the cardiac primordia originating from the two sides, microsurgery (DeHaan,1959) was applied to the head-fold embryo (1- to 4-somite stage: HH 7/8− stage), in which the foregut is a shallow crescent and the cardiac primordia have not yet fused in the bulboventricular region. After EC culture preparation, a fine-tungsten needle was inserted under the anterior intestinal portal, and the floor of the foregut was cut at the midline. These operations produced double-hearted embryos (cardia bifida) without interfering with the normal development of other structures of the embryo as described by DeHaan (1959).

Bead Implantation

Affigel blue beads (150–300 μm in diameter ([50–100 mesh], Bio-Rad Laboratories, Ltd.) were soaked in 1 mg/ml Shh (Shh N-terminal peptide of human recombinant, R&D Systems) in phosphate buffered saline (PBS) for more than 2 hr at 4°C. A bead (200–300 μm in diameter) was implanted on the right side of Hensen's node of stage 4–5 embryos explanted in EC culture. As controls, beads soaked in PBS were used. The bead-implanted embryos were incubated for a further 12–18 hr until embryos reached approximately the 13-somite stage (HH 11 stage).

In Situ Hybridization and Histology

For H&E staining, chicken embryos were fixed for 1 hr in freshly prepared Bouin's fixative at room temperature and embedded in paraplast (Sigma) after dehydration with a graded ethanol series and benzene. Sections were prepared at 7 μm, stained with Mayer's hematoxylin (Merck) and eosin (Chroma Gesellschaft Schmid & Co.) solution, and mounted with Entellan (O. Kindler,Germany).

For whole-mount in situ hybridization, embryos were fixed for 2 hr to overnight in freshly prepared 4% paraformaldehyde in PBT (PBS+0.1% Tween20) at 4°C. Whole-mount in situ hybridization was carried out essentially as previously described (Yonei et al.,1995), using RNA probes for chick Pitx2 (Ryan et al.,1998) and AMHC1 (Yutzey et al.,1994). Plasmids for AMHC1 and Pitx2 probes were kindly provided by Drs. K.E. Yutzey and Y. Kawakami.

Acknowledgements

We thank Dr. Sayuri Yonei-Tamura (Tohoku University) for helpful discussions and picture drawing. We also thank Dr. Y. Kawakami (The Salk institute, USA) and Dr. K.E. Yutzey (Cincinnati Children's Medical Center, USA) for kind gifts of the plasmids.

Ancillary