On rotation, torsion, lateralization, and handedness of the embryonic heart loop: New insights from a simulation model for the heart loop of chick embryos

Authors

  • Jörg Männer

    Corresponding author
    1. Department of Anatomy and Embryology, Center of Anatomy, Georg August University of Göttingen, Göttingen, Germany
    • Department of Anatomy and Embryology, University of Göttingen, Kreuzbergring 36, D-37075 Göttingen, Germany
    Search for more papers by this author
    • Fax: 49-551-397043


Abstract

The internal organs of vertebrates show specific anatomical left-right asymmetries. The embryonic heart is the first organ to develop such asymmetries during a process called dextro-looping. Thereby the initially straight heart tube curves toward its original ventral side and the resulting bend becomes displaced toward the right side of the embryo. Abnormal displacement of the heart loop toward the left is rare and is called levo-looping. Descriptive studies have shown that the lateralization of the heart loop is driven by rotation around its dorsal mesocardium. However, nothing was known on the modes of this process. To gain insight into this subject, different modes of rotation were tested in a simulation model for the looping chick embryo heart. The morphological phenotypes obtained in this model were compared with normal and mirror-imaged embryonic hearts. The following observations were made. One, rotation of the heart loop around its dorsal mesocardium has two consequences: first, lateral displacement of its bending portion either toward the right (D-loop) or left (L-loop) side of the embryo, and second, torsion of the cardiac bend into a helical structure that is wound either clockwise (right-handed helix) or counterclockwise (left-handed helix). The normal loop presents as a D-loop with left-handed helical winding and its mirror image presents as an L-loop with right-handed helical winding. This conflicts with the use to define D-loops as right- and L-loops as left-handed structures. Two, dextro-looping might be driven almost exclusively by rightward rotation of the arterial pole of the loop. It becomes complemented by leftward rotation of the venous pole during the subsequent phase of looping. An inverse mode of rotation might drive levo-looping. Three, the two different helical configurations of heart loops both can occur as right-sided, median, or left-sided positional variants. When viewed from the front, all right-sided variants appear as D-loops and all left-sided variants appear as L-loops at the end of dextro- or levo-looping. Their true asymmetric phenotypes become fully apparent only after simulation of the subsequent phase of looping. The terms D- and L-loop obviously do not fully define the chirality of heart loops. The chirality of their helical configuration should be defined, too. The implications of these data with respect to molecular and experimental data are discussed. Anat Rec Part A 278A:481–492, 2004. © 2004 Wiley-Liss, Inc.

The internal organs of vertebrates show species-specific morphological and topographical left-right (L-R) asymmetries. Understanding the embryonic background of the normal L-R asymmetries, the so-called situs solitus, is not only of interest for basic medical sciences, such as anatomy and embryology, but also for clinical disciplines since deviations from the normal pattern of L-R asymmetries (situs inversus, heterotaxia, visceral isomerism) are frequently associated with severe malformations of the organ systems affected (Aylsworth, 2001). This is particularly true for the cardiovascular system (Anderson et al., 1998; Icardo et al., 2002). During the past few years, fascinating progress has been made in elucidating the embryonic background of L-R patterning in vertebrates. Using molecular biology techniques, several evolutionary highly conserved molecules and molecular signaling cascades have been identified that are involved in the establishment of side-specific identities of the left and right body halves (Ryan and Izpisua Belmonte, 2000; Mercola and Levin, 2001). Three principal steps are distinguished in this process: the breaking of the initially bilateral symmetry of the embryo; propagation of the molecular signaling cascades to establish side-specific molecular identities of the two body halves; and translation of the molecular L-R pattern into asymmetric organ morphogenesis.

Although it has been found that Hensen's node develops morphological L-R asymmetry before propagation of the L-R signaling cascades (Dathe et al., 2002), it is generally agreed that the heart is the first organ to develop morphological L-R asymmetry during embryogenesis. The early morphogenesis of the embryonic heart has therefore received considerable attention from those developmental biologists researching on the determination of L-R asymmetries. The process leading to the morphological polarization of the embryonic heart along the L-R axis can be named, in a general and nonspecific way, cardiac lateralization. During cardiac lateralization, the embryonic heart is transformed from a straight tube-like structure, oriented along the midline of the embryo, into a c-shaped loop whose convexity can principally point either toward the right or toward the left side of the embryo. In all vertebrate embryos studied so far, the convexity of the c-shaped heart loop normally points toward the right side of the embryo. Normal lateralization of the embryonic heart loop of vertebrates has therefore been named rightward or dextro-looping, whereas the abnormal lateralization process leading to the mirror-imaged situation is called leftward or levo-looping. Numerous studies have shown that the laterality of the heart loop, and of other developing organs, is controlled by the molecular L-R signaling cascades mentioned above (Ryan and Izpisua Belmonte, 2000; Mercola and Levin, 2001; Icardo et al., 2002). The question, however, of how the molecular signals are translated into asymmetric organ morphogenesis remains unanswered.

The search for the correct answers to this question might profit from a good knowledge of the anatomy and biomechanics of the morphogenetic events leading to the morphological polarization of the developing organs along the L-R axis. Unfortunately, insight into these events is frequently confounded by incorrect interpretations of the morphological situation. Dextro-looping of the developing heart, for example, has frequently been misinterpreted as a bending of the straight heart tube toward the right side of the embryo. Accordingly, it was thought that the original right half of the heart tube forms the outer convex curvature of the c-shaped loop and the original left half forms the inner concave curvature (Patten, 1922; Stalsberg, 1970; Tamura et al., 1999). Detailed morphological analyses, however, have shown that the bending of the heart tube is in fact directed toward its original ventral side, which in turn becomes displaced to the right of the body by rotation of the heart tube around its dorsal mesocardium (Männer, 2000). The morphogenetic event underlying cardiac lateralization is therefore not the bending process but a rotation movement around a craniocaudal axis. The incorrect morphological interpretation of dextro-looping as rightward bending might have contributed to the situation that the biomechanical factors driving the lateralization of the heart loop have still not been identified (Männer, 2000).

The identification of rotation as the morphogenetic process underlying cardiac lateralization can therefore be regarded as an important step on the way to understanding asymmetric morphogenesis of the embryonic heart. At the present time, however, progress is hindered by the fact that the exact modes of normal rotation as well as the morphological consequences of normal and abnormal modes of rotation are poorly understood. Insight into these subjects might be achieved by experiments in which the process of rotation is altered in a controlled and predictable fashion. At the present time, however, lack of knowledge on the biophysical mechanisms driving cardiac rotation does not allow performance of such experiments in the embryo. Consequently, I decided to analyze the morphological consequences of different modes of rotation in an artificial model for the heart loop of chick embryos, which facilitated direct mechanical simulations of rotation.

Dextro-looping only represents the first step in a sequence of complex positional and morphological changes of the embryonic heart tube that brings its subdivisions and the vessel primordia approximately into their definitive topographical relationship to each other (Fig. 1). Therefore, the present model included not only the simulation of dextro- or levo-looping, but, furthermore, the simulation of the subsequent process of shortening of the distance between the fixed arterial and venous ends of the heart, which normally contributes to the transformation of the c-shaped heart loop into the s-shaped heart loop (Männer, 2000). This facilitates analysis of not only the consequences of different modes of rotation for the phenotype of a c-shaped tube, but also the consequences for the phenotype of the s-shaped tube, which is much closer to the final phenotype of the embryonic heart. The studies on this model disclosed some unexpected relations between the mode of rotation and the subsequent development of an asymmetric morphology of a bent elastic tube. The present findings might have important consequences for the correct morphogenetic interpretation of normal and abnormal heart loops.

Figure 1.

Schematic drawing showing the development of the chick embryo heart from the stage of the straight heart tube up to the stage of the four-chambered heart. Looping of the chick embryo heart can be subdivided into three subsequent phases, each of them characterized by typical morphogenetic events (Männer, 2000). One, dextro-looping, which is characterized mainly by the transformation of the straight heart tube into a c-shaped loop and by the lateral displacement of its bent portion toward the right side of the embryo. Two, the transformation of the c-shaped loop into an s-shaped loop. This phase is characterized mainly by the shortening of the distance between the arterial and venous pole of the heart loop and by the displacement of its ventricular bend from its original position cranial to the atria toward its final position caudal to the atria. Three, the final looping phase, which is characterized mainly by the shift of the proximal portion of the cardiac outflow tract from the right lateral position with respect to the atria toward its definitive position ventral to the right atrium and by the morphological appearance of the tubular anlage of the great arteries (truncus arteriosus).

MATERIALS AND METHODS

The goal of the present studies was to obtain information on the modes of rotation possibly underlying the process of lateralization of the looping embryonic heart. It was not intended to obtain information on the mechanisms driving the process of bending of the tubular heart. In contrast to the real situation in chick embryos, the model therefore had a bent portion already at the beginning of simulation of cardiac looping.

A flexible elastic rubber tube was used for construction of the model. The model tube was oriented along a craniocaudal axis and was deformed into an omega-shaped structure consisting of three portions (Fig. 2). First, a straight cranial portion, which was fixed dorsally to a laboratory tripod. This portion represented the straight cranial portion of the c-shaped heart loop of chick embryos. Second, a bent mid portion, which curved toward its ventral midline. This portion represented the ventricular loop of the c-shaped heart loop of chick embryos. Third, a straight caudal portion, which was fixed dorsally to a laboratory tripod. This portion represented the venous pole of the c-shaped heart loop of chick embryos.

Figure 2.

Photographs showing the simulation model on a frontal (A) and right lateral view (B) before the starting of the simulation of rotation.

Starting from this situation, two subsequent phases of the looping process were simulated in the present study. First, the phase of lateralization of the c-shaped heart loop, which normally leads to the displacement of its bending portion toward the right side of the body (dextro-looping). Second, the subsequent phase of transformation of the c- into the s-shaped heart loop. Lateralization of the c-shaped heart loop is driven by rotation around its dorsal mesocardium (Männer, 2000). For the simulations of this phase of the looping process, the cranial and/or caudal portions of the tube were therefore rotated around their longitudinal axis. The choice of the modes of rotation to be tested in the simulation model was based on the considerations given below. Transformation of the c- into the s-shaped loop is driven mainly by shortening of the distance between the fixed cranial and caudal poles of the heart (Männer, 2000). Simulation of this phase of the looping process was therefore achieved by shortening the distance between the fixed cranial and caudal portions of the tube subsequent to cardiac lateralization.

Modes of Rotation

Rotation of the looping embryonic heart around its dorsal mesocardium is said to cause not only lateralization but also torsion of the heart tube (De la Cruz, 1998). In the past, however, this torsion has not been characterized in morphological terms. The choice of the modes of rotation to be tested in the simulation model was therefore based on the fact that torsion of a vertically oriented tube can principally result from three different modes of rotation around its longitudinal axis: first, rotation of its superior end combined with fixation (blocking simultaneous rotation) of its inferior end; second, rotation of its inferior end combined with fixation of its superior end; and third, rotation of its superior end combined with opposite rotation of its inferior end. Since rotations can occur in two directions, a total number of six different modes were simulated in the present model. These modes were (1 + 2) 90° left- or rightward rotation of its cranial portion only; (3 + 4) 90° left- or rightward rotation of its caudal portion only; and (5 + 6) combinations of opposite rotations of its cranial and caudal portions.

Documentation and Analyses of Phenotypes

The positional and morphological changes of the looping model tubes were photographically documented at the end of each of the two simulated phases of cardiac looping. The pictures obtained were then compared with pictures showing normal or mirror-imaged heart loops of chick embryos from corresponding developmental stages (stages 12 and 16 according to Hamburger and Hamilton (1951) respectively (HH stages)). Pictures from normal heart loops were made with a scanning electron microscope as described previously (Männer, 2000). These photographs were also used for the production of pictures showing mirror-imaged heart loops.

RESULTS

1 + 2: Right- or Leftward Rotation of Cranial Portion Only

Ninety degree rightward rotation of the cranial portion of the heart tube only causes, first, lateral displacement of its bent portion toward the right of the midline and, second, the deformation of its bent portion into a helical structure wound counterclockwise (Fig. 3). At the stage of the fully lateralized c-shaped loop, it appears as a D-loop that resembles normal HH stage 12 chick embryo hearts (Fig. 4). Subsequent to shortening of the distance between its fixed cranial and caudal portions, it appears as an s-shaped loop of normal morphology whose apex points toward the right of the midline (Fig. 5).

Figure 3.

Frontal (A and C), cranial (B), and right ventrolateral (D) views on the simulation model in comparison to photographs of segments of a spirally twisted column wound counterclockwise (C′ and D′). Ninety degree rightward rotation of the cranial portion of the tube causes, first, lateral displacement of its bent portion toward the right of the midline and, second, the deformation of its bent portion into a helical structure wound counterclockwise. The direction of the winding of a helical structure is defined by the pathway, which follows the winding while going away from the observer (B).

Figure 4.

Frontal views on the simulation model (A) and on a normal HH stage 12 chick embryo heart (B). The c-shaped D-loop produced by 90° rightward rotation of the cranial portion resembles the normal c-shaped D-loop of HH stage 12 chick embryos. The caudal half of its bent portion shows a bulge (dotted line), whereas its cranial half shows a groove.

Figure 5.

Frontal (A) and right ventrolateral views (B) on the simulation model subsequent to the shortening of the distance between its fixed cranial and caudal ends. The morphology of the loop produced by 90° rightward rotation of the cranial portion resembles that of the normal heart loop of HH stage 16 chick embryos (C and D). Its position with respect to the L-R axis, however, deviates from the normal position of the embryonic heart in such a way that its apex points toward the right of the midline instead of the midline. Consequently, it is the right ventrolateral view on the simulation model (B) that corresponds to the frontal view on the embryonic heart (D), whereas the frontal view on the simulation model (A) corresponds to a left ventrolateral view on the embryonic heart (C). a, atriums; o, outflow tract; s, sinus venosus; v, ventricular bend.

Ninety degree leftward rotation of the cranial portion of the heart tube only causes, first, lateral displacement of its bent portion toward the left of the midline and, second, the deformation of its bent portion into a helical structure wound clockwise (Fig. 6). At the stage of the fully lateralized c-shaped loop, it appears as an L-loop that resembles the true mirror image of normal HH stage 12 chick embryo hearts (Fig. 7). Subsequent to shortening of the distance between its fixed cranial and caudal portions, it appears as an s-shaped loop of mirror-imaged morphology whose apex points toward the left of the midline (Fig. 8).

Figure 6.

Frontal (A and C), cranial (B), and left ventrolateral (D) views on the simulation model in comparison to photographs of segments of a spirally twisted column wound clockwise (C′ and D′). Ninety degree leftward rotation of the cranial portion of the tube causes, first, lateral displacement of its bent portion toward the left of the midline and, second, the deformation of its bent portion into a helical structure wound clockwise.

Figure 7.

Frontal views on the simulation model (A) and on a reversed HH stage 12 chick embryo heart (B). The c-shaped L-loop produced by 90° leftward rotation of the cranial portion (A) resembles the mirror image of the normal c-shaped D-loop of HH stage 12 chick embryos (B). The caudal half of its bent portion shows a bulge (dotted line), whereas its cranial half shows a groove.

Figure 8.

Frontal (A) and left ventrolateral view (B) on the simulation model subsequent to the shortening of the distance between its fixed cranial and caudal ends. The morphology of the loop produced by 90° leftward rotation of the cranial portion resembles that of the mirror image of the normal heart loop of HH stage 16 chick embryos (C and D). Its position with respect to the L-R axis, however, deviates from the position of the corresponding embryonic heart in such a way that its apex points toward the left of the midline instead of pointing toward the midline. Consequently, it is the left ventrolateral view on the simulation model (B) that corresponds to the frontal view on the embryonic heart (D), whereas the frontal view on the simulation model (A) corresponds to a right ventrolateral view on the embryonic heart (C).

3 + 4: Right- or Leftward Rotation of Caudal Portion Only

Ninety degree rightward rotation of the caudal portion of the tube only causes, first, lateral displacement of its bent portion toward the right of the midline and, second, the deformation of its bent portion into a helical structure wound clockwise (Fig. 9). At the stage of the fully lateralized c-shaped loop, it appears as a D-loop. The morphology of this D-loop, however, does not correspond to that of the normal HH stage 12 chick embryo heart (Fig. 10). Due to the helical configuration of the looping heart tube, the cranial half of the c-shaped bend of true D- and L-loops shows a groove and the caudal half of their c-shaped bend shows as a bulge when viewed ventrally (Figs. 4 and 7). In the false D-loop produced by 90° rightward rotation of the caudal portion of the tube only, the situation is reversed in that way that the cranial half of its c-shaped bend shows a bulge, whereas the caudal half shows a groove (Fig. 10A). Subsequent to shortening of the distance between its fixed cranial and caudal portions, this false D-loop appears as an s-shaped loop of mirror-imaged morphology whose apex points toward the right of the midline (Fig. 11).

Figure 9.

Frontal (A and C), cranial (B), and right ventrolateral (D) views on the simulation model in comparison to photographs of segments of a spirally twisted column wound clockwise (C′ and D′). Ninety degree rightward rotation of the caudal portion of the tube causes, first, lateral displacement of its bent portion toward the right of the midline and, second, the deformation of its bent portion into a helical structure wound clockwise.

Figure 10.

Frontal views on the simulation model (A) and on a normal HH stage 12 chick embryo heart (B). The c-shaped D-loop produced by 90° rightward rotation of the caudal portion (A) does not correspond to the normal c-shaped D-loop of HH stage 12 chick embryos (B). The cranial half of its bent portion shows a bulge (dotted line) whereas its caudal half shows a groove.

Figure 11.

Frontal (A) and right ventrolateral view (B) on the simulation model subsequent to the shortening of the distance between its fixed cranial and caudal portions. The morphology of the loop produced by 90° rightward rotation of the caudal end resembles that of the mirror image of the normal heart loop of HH stage 16 chick embryos (C and D). Its position with respect to the L-R axis, however, deviates from the position of the corresponding embryonic heart in such a way that its apex points toward the right of the midline instead of the midline. Consequently, it is the right ventrolateral view on the simulation model (B) that corresponds to the frontal view on the embryonic heart (D), whereas the frontal view on the simulation model (A) corresponds to a left ventrolateral view on the embryonic heart (C).

Ninety degree leftward rotation of only the caudal portion of the tube causes, first, lateral displacement of its bent portion toward the left of the midline and, second, the deformation of its bent portion into a helical structure wound counterclockwise (Fig. 12). At the stage of the fully lateralized c-shaped loop, it appears as an L-loop (Fig. 13). The morphology of this L-loop, however, differs from that of the true mirror image of the normal HH stage 12 chick embryo heart in the same way as in the false D-loop produced by 90° rightward rotation of the caudal portion of the tube only. The cranial half of the c-shaped bend of the false L-loop therefore shows a bulge, whereas the caudal half shows a groove (Fig. 13A). Subsequent to shortening of the distance between its fixed cranial and caudal portions, this false L-loop appears as an s-shaped loop of normal morphology whose apex points toward the left of the midline (Fig. 14).

Figure 12.

Frontal (A and C), cranial (B), and left ventrolateral (D) views on the simulation model in comparison to photographs of segments of a spirally twisted column wound counterclockwise (C′ and D′). Ninety degree leftward rotation of the caudal portion of the tube causes, first, lateral displacement of its bent portion toward the left of the midline and, second, the deformation of its bent portion into a helical structure wound counterclockwise.

Figure 13.

Frontal views on the simulation model (A) and on a reversed HH stage 12 chick embryo heart (B). The c-shaped L-loop produced by 90° leftward rotation of the caudal portion (A) does not correspond to the mirror image of the normal c-shaped L-loop of HH stage 12 chick embryos (B). The cranial half of its bent portion shows a bulge (dotted line), whereas its caudal half shows a groove.

Figure 14.

Frontal (A) and left ventrolateral views (B) on the simulation model subsequent to the shortening of the distance between its fixed cranial and caudal portions. The morphology of the loop produced by 90° leftward rotation of the caudal portion resembles that of the normal heart loop of HH stage 16 chick embryos (C and D). Its position with respect to the L-R axis, however, deviates from the normal position of the embryonic heart loop in such a way that its apex points toward the left of the midline instead of the midline. Consequently, it is the left ventrolateral view on the simulation model (B) that corresponds to the frontal view on the embryonic heart (D), whereas the frontal view on the simulation model (A) corresponds to a right ventrolateral view on the embryonic heart (C).

5 + 6: Combined But Opposite Rotations of Cranial and Caudal Portions

The combination of 90° rightward rotation of the cranial with 90° leftward rotation of the caudal portion of the loop causes deformation of its bent portion into a helical structure wound counterclockwise (Fig. 15). The cranial half of its bent portion is displaced toward the right and the caudal half of its bent portion is displaced toward the left of the midline. At the stage of the fully lateralized c-shaped loop, this loop therefore does not appear as a c- but as an s-shaped loop (Figs. 15A and 16). Subsequent to shortening of the distance between its fixed cranial and caudal portions, it appears as an s-shaped loop whose morphological and positional characteristics correspond to those of normal HH stage 16 chick embryo hearts (Fig. 17).

Figure 15.

Frontal (A) and cranial (B) views on the simulation model in comparison to a photograph of a spirally twisted column wound counterclockwise (C). The combination of 90° rightward rotation of the cranial and 90° leftward rotation of the caudal portion produced an s-shaped loop. This loop is a helical structure wound counterclockwise.

Figure 16.

Frontal view on the s-shaped loop produced by the combination of 90° leftward rotation of the cranial and 90° rightward rotation of the caudal portion.

Figure 17.

Frontal (A) and right lateral views (B) on the simulation model subsequent to the shortening of the distance between its fixed cranial and caudal portions. The morphological as well as the positional characteristics of the loop produced by the combination of 90° rightward rotation of the cranial and 90° leftward rotation of the caudal portion correspond to those of the normal heart loop of HH stage 16 chick embryos (C and D). C: Frontal view on the embryonic heart. D: Right lateral view on the embryonic heart.

The combination of 90° leftward rotation of the cranial with 90° rightward rotation of the caudal portion of the loop causes deformation of its bent portion into a helical structure wound clockwise (Fig. 18). The cranial half of its bent portion is displaced toward the left and the caudal half of its bent portion is displaced toward the right of the midline. At the stage of the fully lateralized c-shaped loop, it therefore appears as an inverted s-shaped loop (Figs. 18A and 19). Subsequent to shortening of the distance between its fixed cranial and caudal portions, it appears as an s-shaped loop whose morphological and positional characteristics correspond to those of the true mirror image of normal HH stage 16 chick embryo hearts (Fig. 20).

Figure 18.

Frontal (A) and cranial (B) views on the simulation model in comparison to a photograph of a spirally twisted column wound clockwise (C). The combination of 90° leftward rotation of the cranial and 90° rightward rotation of the caudal portion produced an inverted s-shaped loop. This loop is a helical structure wound clockwise.

Figure 19.

Frontal view on the inverted s-shaped loop produced by the combination of 90° leftward rotation of the cranial and 90° rightward rotation of the caudal portion.

Figure 20.

Frontal (A) and left lateral views (B) on the simulation model subsequent to the shortening of the distance between its fixed cranial and caudal portions. The morphological as well as the positional characteristics of the loop produced by the combination of 90° leftward rotation of the cranial and 90° rightward rotation of the caudal portion correspond to those of the mirror image of the normal heart loop of HH stage 16 chick embryos (C and D). C: Frontal view on the embryonic heart. D: Left lateral view on the embryonic heart.

Torsion of Normal Heart Loop of Chick Embryos

The similarities between the loops with counterclockwise winding, produced in the simulation model, and the normal heart loops of chick embryos suggest that the latter loops might also be helical structures wound counterclockwise. This idea was indeed confirmed by analyzing the morphology of chick embryo hearts on craniocaudal views (Fig. 21A and C). The mirror images of normal heart loops were found to be helical structures wound clockwise (Fig. 21B and D).

Figure 21.

Cranial views on the normal heart loops of HH stage 12 (A) and HH stage 16 (C) chick embryos and on their corresponding mirror images (B and D). The normal heart loop is a helical structure wound counterclockwise (A and D), whereas the mirror image of the normal heart loop is a helical structure wound clockwise (B and D).

DISCUSSION

Handedness of Embryonic Heart Loop

In the present study, a bent elastic tube was used as a model for the simulation of looping of the chick embryo heart in order to obtain information on the modes of rotation possibly underlying the early positional and morphological changes of the embryonic heart loop. Use of a simulation model for the looping heart paved the way to seeing some aspects of the early morphogenesis of the embryonic heart loop in a new light. The first aspect to be discussed concerns the classification of the asymmetric morphology of the embryonic heart loop. Analogous to the right and left hands of human beings, asymmetric structures are regarded as handed if they can principally exist in two morphological variants, each of them being a mirror image of the other one. The embryonic heart loop is regarded as a handed structure since its lateral displacement can principally occur either toward the right or the left side of the embryonic body. Thereby the handedness of the c-shaped heart loop is generally determined in a two-dimensional rather than a three-dimensional fashion, taking the direction of its convexity as the defining feature. Accordingly, a c-shaped loop is classified as a right-handed structure if its convexity points toward the right of the embryo body and as a left-handed structure if its convexity points toward the left of the embryo body. These heart loops are called D-loops and L-loops, respectively, and it is generally believed that these position-based designations sufficiently characterize the morphological aspects of cardiac handedness (Van Praagh, 1972). The present study focussed on the rotational movements leading to the lateralization and torsion of the c-shaped heart loop. Focussing on torsion of the tubular heart shifted the attention of the observer from the two-dimensional toward the three-dimensional aspects of cardiac development. It became clear that, as a consequence of torsion, the bent portion of the model acquired the configuration of a helix (Figs. 3, 6, 9, and 12). The presence of a helical configuration was then also confirmed for the normal heart loop of chick embryos (Figs. 4 and 21A and C). Helices are handed structures. They can either be wound clockwise or counterclockwise. Structures wound clockwise are mathematically defined as structures with a right-handed sense of twist and those wound counterclockwise are defined as structures with a left-handed sense of twist. These definitions are generally applied to biological and technical structures (Ludwig, 1932). The fact that the embryonic heart loop has a helical configuration has previously been noted by some researchers (Bremer, 1928b; Grant, 1964). Surprisingly, however, no attempts have previously been made to classify the morphology of the heart loop in terms of its helical handedness. If this is undertaken, one has to realize that the normal D-loop has a left-handed sense of twist (Fig. 21A and C), whereas its mirror-imaged L-loop has a right-handed sense of twist (Fig. 21B and D). For reasons of morphological correctness and to facilitate comparison of the looping heart with other helical structures, it seems reasonable to abandon the use of the classification of D-loops as right- and L-loops as left-handed structures.

Modes of Rotation Possibly Underlying Dextro- and Levo-Looping

Realizing that the normal heart loop has a left-handed sense of twist facilitates new morphogenetic interpretations of normal and abnormal heart loops. To follow these interpretations, the conditions leading to a left-handed torsion of a tube-like structure should first be envisaged in a straight and vertically oriented elastic tube. Such a tube can be deformed into a twisted tube with left-handed torsion principally by three modes of rotation around its longitudinal axis: first, rotation of its superior end toward the right side of the tube combined with fixation (preventing simultaneous rotation) of its inferior end; second, rightward rotation of its superior end combined with leftward rotation of its inferior end; and third, leftward rotation of its inferior end combined with fixation of its superior end. The present data demonstrate that left-handed torsion of a bent elastic tube forces its curvature to acquire the configuration of a left-handed helix. However, due to the fact that the axis of rotation is located outside of the helix, the first and third of the above-mentioned modes of rotation force the curvature not only to acquire the configuration of a left-handed helix but also cause a lateral displacement of the spirally twisted curvature. This displacement is toward the right of the axis in the case of rightward rotation of the cranial portion only (mode A) and toward the left in the case of leftward rotation of the caudal portion only (mode C). Therefore, with respect to the L-R axis, each of the three above-mentioned modes of rotation is found to produce a distinct positional variant of a loop with left-handed torsion, which can be named the right-sided (mode A), the median (mode B), and the left-sided (mode C) variant (Fig. 22). Corresponding positional variants also exist for loops with mirror-imaged (right-handed) torsion (Fig. 23).

Figure 22.

Frontal (AC) and cranial (A′C′) views on the simulation model showing the right-sided (A and A′), median (B and B′), and left-sided (C and C′) variants of loops wound counterclockwise (left-handed sense of twist) at the stage of the c-shaped loop. Note that, when viewed from the front, the right- and left-sided variants appear as D- and L-loops, respectively.

Figure 23.

Frontal (AC) and cranial (A′C′) views on the simulation model showing the left-sided (A and A′), median (B and B′), and right-sided (C and C′) variants of loops wound clockwise (right-handed sense of twist) at the stage of the c-shaped loop. Note that, when viewed from the front, the right- and left-sided variants appear as D- and L-loops, respectively.

Analysis of the positional variants of loops with left-handed torsion at successive stages of the simulation of cardiac looping and comparison of them with normal heart loops of chick embryos from corresponding stages make apparent that, at the stage of the fully lateralized c-shaped loop, it is the right-sided variant that corresponds to the normal c-shaped heart loop (Fig. 4), whereas at the subsequent stage of the s-shaped loop it is the median variant that corresponds to the normal s-shaped heart loop (Fig. 17). This suggests that, in the chick embryo, the normal torsion of the heart loop might be driven by a sequential mode of rotation. Thereby the normal rightward displacement of the c-shaped heart loop and the initial steps of its left-handed torsion might be driven almost exclusively by a rightward rotation of its cranial portion, whereas the subsequent shift of the spirally twisted curvature from the normal right-sided position of the c-shaped heart loop back to the normal median position of the s-shaped heart loop might be driven by an opposite (leftward) rotation of its caudal portion which, additionally, strengthens its torsion. A corresponding but mirror-imaged mode of rotation might lead to the development of heart loops with mirror-imaged morphology (Figs. 7 and 20).

If opposite rotations of the cranial and caudal portions of the embryonic heart loop might drive its normal torsion, one would expect to find some L-R asymmetries with opposite polarities at the cranial and caudal portions of the embryonic heart loop. Such asymmetries have indeed been reported in the embryological literature but it seems that they have not received sufficient recognition until now. Several authors, for example, have noted that the caudal portion of the looping embryonic heart tube normally behaves oppositely to its cranial portion in such a way that it undergoes a leftward instead of a rightward looping/displacement (Schulte, 1916; Bremer, 1928a; Steding and Seidl, 1980; Fransen and Lemanski, 1989; Tsuda et al., 1998; Conway et al., 2003). Stalsberg (1969) determined the cellular contributions of the right and left precardiac mesoderm to the normal c-shaped heart loop of chick embryos. He noted a predominance of cellular contribution from the right precardiac mesoderm at the cranial portion of the heart tube and from the left precardiac mesoderm at its caudal portion. Tsuda et al. (1998) noted an asymmetric expression of the extracellular matrix molecule flectin in the looping heart of mouse embryos with a right-sided predominance at the outflow tract and a left-sided dominance at the ventricular region. Unfortunately, however, they did not find a corresponding pattern of flectin expression in the looping heart of chick embryos (Tsuda et al., 1996).

All in all, it can be stated that there are good reasons to postulate that a combination of rightward rotation of the cranial portion of the looping heart tube with leftward rotation of its caudal portion might cause its normal displacements and torsion during the early phases of the looping process. Therefore, a search for molecules showing an opposite L-R expression pattern at the cranial and caudal portions of the looping heart tube might be a promising approach to find an answer to the question of how the recently identified molecular L-R signaling cascades are translated into the asymmetric morphogenesis of the heart. In this context, it should be emphasized that the mechanisms driving the normal rightward displacement of the c-shaped heart loop are expected to be located at its cranial portion, which has recently been found to be derived from a distinct part of the precardiac mesoderm, called the anterior or secondary heart field (Mjaatvedt et al., 2001; Kelly and Buckingham, 2002). Therefore, research on the development of the outflow portion of the early embryonic heart tube might provide important contributions to the understanding of cardiac looping. The observations that some genetically induced defects in the formation of the outflow portion of the embryonic heart tube are associated with lack of dextro-looping but presence of the normal leftward shift of the cardiac inflow portion (Lyons et al., 1995; Lin et al., 1997) support this idea.

Relation Between Lateralization and Morphological L-R Asymmetry of Heart Loops

The relation between the direction of lateral displacement of the looping heart tube and its morphological handedness deserves special attention since it might be of the outmost importance for the correct morphogenetic interpretation of the phenotype of heart loops produced by experimental interventions in normal L-R development. As described above, cardiac handedness has traditionally been defined in a position-based way in terms of D- and L-loops, assuming that a right- or leftward orientation of the outer convex curvature of the c-shaped heart loop is regularly linked to the presence of normal or inverse morphological asymmetry of the heart, respectively. The present data show that the morphological asymmetry of the early embryonic heart loop can be described in terms of helical handedness. Tests on the simulation model have furthermore shown that, depending on the mode of cardiac rotation, heart loops of the same helical configuration can principally occur as right-sided, median, and left-sided positional variants. The morphology of heart loops is usually examined in frontal views. If this is undertaken at the end of simulation of dextro- or levo-looping, all right-sided variants appear as D-loops and all left-sided variants appear as L-loops, irrespective of their helical configuration (Figs. 22A, 23A and 22C, 23C). This is due to the fact that, at this time point, the helical winding of the c-shaped loop has a relatively steep pitch so that, viewed from the front, the twisted curvatures of right- and left-sided variants do not primarily appear as helical structures but rather as simple bends whose convexity points toward the right or left body side, respectively (Figs. 22A, 23A and 22C, 23C). As a consequence of this fact, four different types of lateralized c-shaped loops could be distinguished in the present model at the end of dextro- or levo-looping: D-loops with the normal helical configuration, D-loops with mirror-imaged helical configuration, L-loops with the normal helical configuration, and L-loops with mirror-imaged helical configuration. This shows that, when examined during dextro- or levo-looping, the presence of a D- or L-loop phenotype must not regularly correlate with the presence of normal or mirror-imaged cardiac morphology. The traditional approach to define the L-R asymmetry of the embryonic heart simply in terms of D- and L-loop obviously does not properly define the morphological asymmetry of the loop but simply its positional L-R asymmetry (right- or left-sided variants). It appears that for a proper definition of cardiac L-R asymmetry, at least two aspects of the looping heart tube should be considered: first, its positional L-R asymmetry (D-/L-loop) and, second, its morphological asymmetry (left-/right-handed helix).

The suitability of this new approach to define cardiac L-R asymmetry becomes fully apparent if one takes into account not only the data obtained by the simulation of dextro- or levo-looping but additionally considers the data obtained by the simulation of the subsequent looping phase. During the latter phase, only the D-loop phenotype with normal helical configuration acquired the normal morphology of an s-shaped heart loop, whereas the other D-loop phenotype developed into an s-shaped loop of mirror-imaged morphology (Figs. 10 and 11). A corresponding but inverse behavior was found in the two L-loop phenotypes (Figs. 13 and 14). According to the traditional view, these observations have to be interpreted such that, subsequent to dextro- or levo-looping, only 50% of the loops behaved in the expected fashion and retained their originally acquired morphological L-R asymmetry. The remaining 50% of the loops, however, must have undergone a conversion of their morphological L-R asymmetry from normal (D-loop) to mirror-imaged or from mirror-imaged (L-loop) to normal and therefore behaved in a completely unexpected fashion. Considering the facts, however, that the morphological L-R asymmetry of the looping heart is not properly defined simply in terms of D- and L-loop but rather in terms of its helical handedness, the impression of conversions of cardiac handedness turns out to be simply the product of a wrong interpretation of the situation and therefore must be regarded as a purely fictitious phenomenon. For simplicity of description, the types of D- and L-loops undergoing this phenomenon may be called false D- and L-loops, whereas the D-loops with the normal helical configuration and the L-loops with inverse helical configuration may be called true D- and L-loops, respectively.

The question remains of whether these observations made in a simulation model for the looping heart tube are fully transferable to the developing hearts of embryos. If so, we should expect that at least some of the abnormal cardiac phenotypes observed in the present simulation model would previously have been observed in embryos subsequent to experimental interventions in normal L-R development. Indeed, this seems to be the case. In a recent experimental study on chick embryos, for example, Linask et al. (2003) observed the formation of median s-shaped heart loops instead of lateralized c-shaped heart loops during the first phase of the looping process. Experimentally induced L-loops resembling the false L-loops observed in the present simulation model have been presented by other authors (Itasaki et al., 1991). Moreover, even the phenomenon of conversion of c-shaped L-loops into normal s-shaped heart loops was previously observed in chick embryos subsequent to the experimental induction of L-loops (Steding and Seidl, 1981). In the latter experiment, L-loops were produced by an external push of the bending portion of the heart to the left side of the embryo using forceps. Although the creation of these L-loops is regularly followed by a reversal of the normal embryonic body rotation (G. Steding, personal communication), which is typically found in animal models showing inverse looping of the embryonic heart, one might speculate that the external translocation of the bending portion of the heart tube does not change its intrinsic sense of twist. The L-loops produced in this experiment might then simply represent left-sided variants of heart loops with the normal sense of twist. This idea is supported by the morphological characteristics of these L-loops. Viewed from the front, the cranial half of their c-shaped bends did not show a groove, as in true mirror images of normal D-loops, but showed a bulge (see Fig. 4 in Steding and Seidl, 1981). This is a typical morphological feature of the false D- and L-loops observed in the present simulation model (Figs. 10 and 13). The above-mentioned experimental data therefore suggest that the looping heart tube of the chick embryo seems to behave, in principle, in the same way as the present simulation model. Caution must therefore be taken when determining the handedness of chick embryo hearts only at the stage of the c-shaped loop. This approach obviously harbors a potential for misclassifications of their true morphological handedness. Such misclassifications might have fatal consequences for the correct morphogenetic interpretation of experimental findings. The present data, for example, suggest that false L-loops might experimentally be induced by interventions leading to a precocious activation of the normal leftward rotation of the caudal pole of the heart. If such loops were classified as hearts with mirror-imaged morphology (true L-loops), they might be misinterpreted as consequences of reversal of the normal development of L-R body asymmetries. To avoid such misinterpretations, it might be good to determine the handedness of the chick embryo heart not at the stage of the c-shaped loop (HH stages 11–13) but merely at the stage of the s-shaped loop (HH stage 16 and older). This, however, might not always be possible since most experiments are performed on embryos cultured in vitro. In vitro cultures of chick embryos do not always facilitate normal embryonic development until these stages. In such cases, correct determination of cardiac handedness might be achieved by a detailed analysis of the morphology of the c-shaped heart loop. When viewed from the front, the cranial half of the c-shaped bend of true D- and L-loops shows a groove and the caudal half of their c-shaped bend shows as a bulge (Figs. 4 and 7). In false D- and L-loops, the situation is reversed such that the cranial half of their c-shaped bends shows a bulge, whereas the caudal half shows a groove (Figs. 10 and 13).

Limitations of Model

Models cannot copy every aspect of reality and therefore have specific limits. The present simulation model was designed to obtain information on the modes of rotation possibly underlying the process of lateralization of the looping embryonic heart. It was not designed to study the process of ventral bending of the tubular heart. In contrast to the real situation in chick embryos, the model therefore had a bent portion already at the beginning of simulation of cardiac looping. This design practically alleviates the simulations of rotations but might give the impression that the process of bending is driven by biophysical mechanisms different from those driving rotation. This, however, must not necessarily be the case in reality, as has been shown by other simulation models (Manasek et al., 1984). It therefore should be noted here that, at the present time, it is not known whether these two aspects of the looping process rely on the same or on a different biophysical background. In any case, the present study was not conducted to test whether the bending and rotation of the heart loop might be driven by the same or different physical mechanisms, but only to obtain information on the mode of rotation of the looping heart.

A second limitation of the present model results from the fact that it does not take into account the segmental structure and the longitudinal growth of the looping embryonic heart. Using in vivo labeling techniques, De la Cruz (1998) has shown that the early heart tube of chick embryos consists of the primordia of the trabeculated regions of the future left and right ventricle only and that this tube becomes elongated due to the addition of new cardiac segments at its venous and arterial poles during cardiac looping. In the present model, the original length of the heart tube was not changed during the simulation of cardiac looping and the looping heart tube was not subdivided into its known anatomical segments. These facts do not limit the validity of the model if one wishes to gain information on the morphological behavior of a looping tubular heart as a whole. However, due to these facts, it is not possible to define exactly the cardiac segments or subsegments that might primarily be involved in the rotation processes at the arterial and venous pole. There are, however, some data from chick embryos that might identify these segments. At the arterial pole of the looping chick embryo heart, for example, it has been found that the most prominent rotation occurs at the transition zone between the bent and straight portions, while the distal portions of the heart tube show only little rotation (Männer, 2000). A similar situation exists at the venous pole. Here, it is the transition zone between the ventricular bend and the common atrium, the so-called atrioventricular canal segment, that shows the most prominent rotation while the common atrium and the sinus venosus show little or no rotation (Campione et al., 2001). In a bent rubber tube, it is difficult to simulate axial rotations only at the transition zones between its fixed and bent portions. In contrast to the real situation, the fixed portions of the heart loop were therefore rotated along their entire length in the present simulation model.

Acknowledgements

The author thanks Mrs. Kirsten Falk-Stietenroth and Mr. Hannes Sydow for technical and photographic assistance, Professor Gerd Steding for his critical comments on the manuscript, and Mrs. Cyrilla Maelicke for correcting the English manuscript.

Ancillary