The left–right (LR) axis in vertebrate embryos is specified by a hierarchy of gene products that interact in complex biochemical pathways. Many of these genes exhibit LR asymmetric expression patterns in the embryo before organ formation and are believed to specify morphogenetic differences between the left and right lateral plate mesoderm (LPM) that are ultimately manifested in the LR asymmetries of developing organs (reviewed in Capdevila et al., 2000; Schneider and Brueckner, 2000). The distal-most component of the LR pathway, the transcription factor Pitx2 (specifically the Pitx2a and Pitx2c isoforms; Schweickert et al., 2000; Yu et al., 2001), is thought to directly impart LR information during organogenesis. Similar to earlier genes in the LR specification pathway, the first asymmetric expression of Pitx2 is detected in the left but not right LPM. However, unlike the other LR genes, Pitx2 is also found in asymmetric patterns in the actual organs that become lateralized (Logan et al., 1998; Meno et al., 1998; Piedra et al., 1998; Ryan et al., 1998; St. Amand et al., 1998; Yoshioka et al., 1998; Campione et al., 1999; Essner et al., 2000; Schweickert et al., 2000). Unfortunately, it has been difficult to infer the normal morphogenetic role of Pitx2 from the complex phenotypes found in Pitx2-misexpressing or Pitx2-deficient animals, because the organ anomalies are not always immediately obvious as LR defects and seem to vary depending on the isoform and residual expression level of the Pitx2 allele involved (Gage et al., 1999; Kitamura et al., 1999; Lin et al., 1999; Rankin et al., 2000; Liu et al., 2001; Lowe et al., 2001). Thus, although LR gene expression is crucial for the normal development of asymmetric organs in vertebrates, its exact role in shaping or positioning individual organs is still unclear.
This ambiguity might be at least partially attributed to the fact that the underlying mechanisms of asymmetric organ morphogenesis have not been identified. There is some speculation that differential cell fate specification between the left and right sides of the embryo may be a mechanism of generating morphologic asymmetries (Capdevila et al., 2000; Patterson et al., 2000). For example, the anatomically left-sided spleen is reportedly derived exclusively from cells located on the left side in Xenopus (Patterson et al., 2000). In addition, differential cell proliferation in the left and right sides of individual organs is also a commonly invoked mechanism for creating LR asymmetric morphology (Kitamura et al., 1999; Capdevila et al., 2000). However, the slight differences in proliferation reported for the left and right sides of the heart tube are not likely to drive the substantial morphogenetic dynamics of cardiac looping, suggesting that such a mechanism may not play a primary role in asymmetric morphogenesis (Sissman, 1966; Manasek, 1981). Furthermore, unlike the case of the heart, in most developing organs, the locations of cells derived from the original left or right sides of the embryo have not been defined; thus, it has not been possible to assign differential proliferation rates or other morphogenetic roles to contralateral tissues within particular organs. Thus, a direct link between specific LR asymmetric gene expression patterns and specific morphogenetic events in developing structures has yet to be demonstrated.
Knowledge of the mechanisms that shape the LR asymmetry of individual organs is vital for a complete understanding of LR patterning and the etiology of situs deformities. Here, we investigate the development of LR asymmetry in the digestive system of the Xenopus laevis embryo. We have defined the organ-specific contributions and topologic distribution of left and right LPM tissues within the primitive gut tube during the complex looping process that positions the tadpole's digestive organs. Our analysis shows that contralateral tissues do not contribute to different digestive organs but do form opposing concave and convex topologies during gut looping. Interestingly, we find that key concave surfaces correlate with distinct early and late phase Pitx2 expression patterns, and we show that misexpression of Pitx2 mRNA induces ectopic concavities. Finally, we demonstrate that the left concave and right convex surfaces of the gut tube undergo differential elongation during looping, with no detectable difference in cell proliferation. Taken together, our results suggest that the LR signaling pathway may influence the asymmetric morphogenesis of the Xenopus digestive system by differentially regulating the rate of elongation of the left and right sides of the primitive gut tube.
Fates of Cells in the Lateral Plate Mesoderm Are Similar for the Left and Right Sides
Previous work has suggested that the left and right sides of the Xenopus embryo may contribute differentially to specific LR asymmetric structures, such as the spleen (Patterson et al., 2000). Thus, as a first analysis of the development of asymmetric morphology in the digestive system, we wanted to compare the contributions of the left and right LPM within the individual digestive organs. To facilitate this comparison, we performed a novel bilateral fate mapping of the embryo at stage 23, when LR asymmetric gene expression patterns have been defined in the LPM but organ morphogenesis has not yet commenced. We labeled groups of cells in both the left and right LPM with different fluorescent dyes (DiI [1,1-dioctadecyl-3,3,3′,3′-tetramethyl-indocarbocyanine perchlorate], PKH2GL), using a grid system of 20-square regions on each side of the embryo (Fig. 1A). Because the mesodermal layer is thin at these stages, our injections also occasionally labeled underlying endoderm cells. Although fate maps performed by using an earlier neurula (stage 14) indicated that mesoderm and endoderm cells shift with respect to each other during development (Chalmers and Slack, 2000), potentially complicating fate analyses, we find that the mesodermal and endodermal contributions to the gut organs largely overlap by stage 23–24 (data not shown), thus simplifying our fate assignments.
The final position of the labeled cells was evaluated in whole guts after evisceration at stage 45–46, when all of the digestive organs are morphologically distinguishable and the major LR asymmetries are clearly established. For example, Figure 1B shows the distribution of labeled cells from both left and right square A3. Both dyes are found in several foregut derivatives in the same animal, including the lungs, stomach, liver, and pancreas. Each lung shows exclusive orange or green staining, whereas the esophagus and stomach regions harbor both dyes, and the liver and pancreas both show orange and green in relatively equal amounts.
The results for all 20 squares are summarized in Figure 1C. Qualitatively, there are no major differences in the range of digestive organs (i.e., esophagus, stomach, duodenum, intestine, dorsal pancreas, ventral pancreas, liver, and colon) to which contralateral squares contribute. In addition, the frequency with which labeled cells from each square populate individual organs is generally similar for the left and right sides. In the few cases where the contributions of left and right squares to a particular organ appear to be dissimilar (e.g., frequency of contribution to pancreas by left and right squares A1), only a few cells of this organ were actually labeled in any one embryo. Thus, we believe these minor disparities may be attributed to slight variations in the number of cells that were initially labeled within each square. Nevertheless, these results suggest that unilaterally specifying the development of particular organs in the left or right LPM is unlikely to be a major mechanism of generating LR asymmetries in the Xenopus digestive system, as both sides contribute to each digestive organ and differentiate into the same range of cell types. This finding suggests that LR asymmetric gene expression may be differentially specifying the morphogenetic programs, rather than the specific cell types, of the left and right sides of the embryo during digestive organ development.
Defining the Topologic Distribution of Left and Right Side Cells During Gut Looping
According to Branford et al. (2000), the striking LR asymmetry of the Xenopus gut can be defined in terms of the sidedness of the proximal intestine (i.e., “coil origin”) and the direction of intestinal rotation (i.e., “coil direction;” see Fig. 2A). How these complex coiling asymmetries are actually determined during gut tube morphogenesis and what different morphogenetic roles the left and the right sides of the embryo may play in establishing these asymmetries are currently unknown. However, a careful examination of the discrete phases of gut morphogenesis suggests that two important symmetry-breaking looping events may be instrumental in orienting the final digestive organ anatomy.
Although the general anatomic changes during Xenopus gut development have been described previously (Nieuwkoop and Faber, 1969; Chalmers and Slack, 1998), here we specifically illustrate the topologic dynamics and anatomic consequences of these early looping events with a series of three-dimensional (3D) cartoons of the successive phases of asymmetric morphogenesis (Fig. 2). The first symmetry-breaking event begins subtly at stage 38–40, well before gut coiling is observable, but is most obvious by stage 41 when a concavity appears in the dorsoanterior left side of the gut tube. The right side of the tube does not experience this topologic deformation but instead becomes more convex. This asymmetric curvature involves the prospective esophagus, stomach, and duodenum (Smith et al., 2000). With further elongation of the foregut in the confined space, this small asymmetry initiates the rotation that forms the C-shaped gastroduodenal (GD) loop at stage 42, with the duodenum located on the right side and the stomach oriented transversely. Because this early looping event places the duodenum on the embryo's right side, it establishes the sidedness of the proximal intestine, and thus the gut's “coil origin” by stage 45.
The second looping event which further breaks the gut tube's bilateral symmetry begins at approximately stage 42, when a large convex bulge is apparent on the right side of the midgut. As the gut elongates, the right-side convexity can be seen to contrast with a left-side concavity (stage 43). Because this curvature causes the apex of the midgut (MG) loop to turn medially (stage 44), these events thus establish the counterclockwise rotation of the intestine during successive stages of gut elongation and orient the gut's “coil direction.”
To compare the morphogenetic behavior of the left and right sides of the embryo during these key looping events, we used our bilateral labeling technique to compile 3D models of the relative topologic distribution patterns of left and right LPM-derived cells in the developing gut tube at successive stages (Fig. 2B). In general, we observed little mixing between left- and right-side cells, except at the midline boundaries. During elongation, the majority of the gut tube undergoes a rightward torsion, causing the original left side to face more ventrally and the original right side to face more dorsally, almost transposing the LR and dorsoventral axes. Notably, as the gut loops take shape, cells from the left and right sides of the LPM take on opposing concave vs. convex topologies within the loops. For example, the left side contributes cells to the concave surface of the forming GD loop at stage 38–42, while right-side cells are found on the convex side. Similarly, right-side cells occupy the convexity of the midgut bulge at stage 42–43, while the concavity is composed primarily of cells from the dorsal left side. Overall, these results suggest that specific regions of the left and right LPM execute distinct morphogenetic programs during key phases of gut morphogenesis to create the opposing topologies necessary to form important loops in the developing gut tube.
Defining the Role of Pitx2 Expression in Gut Looping
To determine whether there is a direct link between early and/or late phase LR asymmetric gene expression patterns and the regions of the gut tube that undergo the important looping events, we first aligned Pitx2 expression domains in stage 23 embryos with the fate map grid shown in Figure 1A. Pitx2 can be found with varying frequency in left squares A1, A2, A3, A4, B1, B2, C1, and C2 (data not shown). According to our fate map (Fig. 1C) and topology diagrams (Fig. 2B), these squares would be predicted to contribute precisely to the concavity of the GD loop, the first symmetry-breaking curvature to form in the gut tube.
We confirmed this by directly mapping the fates of the prospective left Pitx2-expressing squares in our grid vs. the right non–Pitx2-expressing squares during gut morphogenesis (Fig. 3A). The left-side Pitx2-expressing cells indeed become localized directly to the concavity of the developing GD loop by stage 40–41, undergoing a dramatic inward curvature on the dorsal left side of the gut, while the non–Pitx2-expressing right-side cells do not undergo this deformation, and the tissues that they occupy correspond with a convex surface.
This result suggests that Pitx2 expression in the left LPM may determine the direction of concavity of the GD loop. To further correlate Pitx2 expression with this important topologic change, we treated gastrula stage embryos with heptanol, a chemical known to perturb LR gene expression patterns and induce heterotaxic gut phenotypes in Xenopus (Levin and Mercola, 1998). Pitx2 expression is normally maintained in LPM derivatives on the left side of the foregut until stage 40 when the forming concavity of the GD loop can just begin to be observed (Fig. 3B). In heterotaxic animals, we found the sidedness of Pitx2 expression at this stage to correlate exactly with the concavity of the GD loop, whether the loop was normal, reversed, or symmetrical (100%; n = 10; Fig. 3C). These data are consistent with the possibility that the formation of the GD loop is a direct consequence of early phase Pitx2 expression in the left LPM, inducing a concavity to form on the left side of the gut tube.
To determine whether the later, organ-specific phase of Pitx2 expression plays a similar role in the morphogenesis of the MG loop that orients the direction of intestinal rotation, we characterized late-phase expression in isolated gut tubes at stage 42, when the MG loop is forming. At this stage, Pitx2 mRNA can be detected symmetrically on both sides of the hindgut. In contrast, Pitx2 is expressed more intensely along the left side of the curving midgut (see Fig. 4A). Therefore, the late phase of Pitx2 expression also corresponds well with the forming concave surface of the MG loop. Thus, both early and late phase Pitx2 expression domains can be correlated with the concavities of key symmetry-breaking looping events during gut morphogenesis.
Misexpression of Pitx2 Can Induce Specific Topologic Changes in the Gut Tube
To verify that Pitx2 can actually induce concavities in the gut tube, and thus play a direct role in orienting the key loops that determine digestive anatomy, we tested whether ectopic Pitx2 expression was sufficient to create concavities in aberrant positions during gut morphogenesis. Because the concavities normally form on the left side of the gut tube, we microinjected synthetic Pitx2c mRNA on the right side of the early Xenopus embryo (Fig. 4). It is well established that such ectopic Pitx2 expression can cause heterotaxia in Xenopus and other vertebrates (reviewed in Capdevila et al., 2000); however, it is not known whether the cells ectopically expressing Pitx2 can be directly correlated with specific topologic deformations during heterotaxic organ development, such as areas of concavity.
To localize the ectopic Pitx2 expression and, thus, to minimize any effects on the main body axis, we targeted our injections to the blastomeres at the 8- to 16-cell stage that are fated to contribute to the right LPM and the right side of the gut tube. We coinjected the Pitx2 mRNA with fluorescein–dextran (FLDx) to be able to track the cells expressing ectopic Pitx2 in living embryos at multiple stages of gut morphogenesis. Because broad misexpression of Pitx2 can perturb gastrulation, only embryos with normal head development and complete dorsoanterior character were scored. Guts were isolated at sibling stage 42–44, analyzed for LR deformities and any areas of abnormal topology (e.g., reversed or ectopic concavities) were compared with the position of the ectopic Pitx2 expression domain.
Control embryos (Fig. 4A), injected with FLDx alone, rarely showed heterotaxic anomalies (11 of 142). In contrast, coinjection of FLDx and Pitx2 mRNA caused gut organ specific heterotaxias in 23.6% of the injected embryos (30 of 127). Isolated heart reversals were also observed in some embryos but were not scored (J.K.M. and N.N.Y., unpublished observations). Strikingly, in 80% (24 of 30) of the embryos with abnormal gut topologies, the cells that contained the injected Pitx2 mRNA, as indicated by the FLDx label, were located in the concavities of the abnormal loops. The presence of Pitx2 mRNA within and around the area of topologic deformation was also verified by RNA in situ hybridization.
In some injected embryos, bilateral concavities were evident in the GD region. The heterotaxic phenotypes associated with this deformation included a shortened foregut; a more anterior midline stomach and duodenum; a symmetrical, hypoplastic liver; and/or an annular pancreas. In these guts, the ectopic Pitx2 mRNA was localized to the ectopic right-side concavity of the GD region (Fig. 4B). In other cases, the GD loop was completely reversed, with the ectopic Pitx2 mRNA located in the concavity of the reversed loop. In these guts, the accessory organs were also reversed (Fig. 4C).
Finally, some Pitx2-injected embryos developed abnormal concavities on the right side of the midgut, transposing the apex of the MG loop to the animal's right, instead of left, to initiate an intestinal malrotation (Fig. 4D). Depending on the severity, this defect also impeded normal GD looping by shortening the entire right side of the gut tube as it formed a large concave curvature occupied by Pitx2-expressing cells. This deformation sometimes caused the accessory organs to be found in variable positions as well. In all cases, as the gut elongated further and began to undergo more complicated coiling and torsion, the ectopic Pitx2-expressing cells became extended longitudinally and distributed in variable topologic locations (e.g., hindgut in Fig. 4D and data not shown). Overall, these results directly correlate ectopic Pitx2 mRNA with the initial formation of ectopic concavities throughout the developing gut tube and, thus, link normal Pitx2 expression patterns with the key looping events that occur during the asymmetric morphogenesis of the Xenopus gut.
Left and Right Sides of the Developing Gut Tube Proliferate at the Same Rate but Elongate Differentially
The above results strongly suggest that Pitx2 expression patterns normally specify distinct morphogenetic behaviors in left-side cells that create concavities in the gut tube. To begin to identify the mechanisms underlying this topologic change, we used our LR topographic models to identify the location of left and right LPM cells in the gut tube at different stages of development (see Fig. 2), and then compared various gross morphogenetic properties between the two sides during looping. Because the midgut forms the largest asymmetric curvature in the gut tube and is most accessible for morphometric measurements, we confined our analyses to the MG loop.
The first obvious morphogenetic factor we analyzed was cell proliferation. To compare cell proliferation on the left and right sides of the midgut, we used an anti-phosphohistone H3 antibody on whole gut tubes that had been flattened during fixation to distinguish the faces of the midgut that were derived from the original left and right LPM of the embryo. Because it is not feasible to ascertain the boundary line or exact number of left vs. right-side cells in an individual gut tube, we could not calculate global left and right mitotic indices. Instead, we directly compared the relative number of phosphohistone H3-positive cells within defined square areas on either side of the midgut. Surprisingly, after comparing proliferation in this way at three different stages of MG looping, we detected no significant differences in this parameter (Fig. 5A; P < 0.01). Although it is possible that differential apoptosis might also drive looping, we were unable to detect any apoptosis in the gut tube by whole-mount terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling (TUNEL) staining at these stages (data not shown), consistent with previous unsuccessful attempts to detect apoptosis in the Xenopus gut (Patterson et al., 2000). These results suggest that neither differential cell proliferation nor cell death in the LPM-derived midgut mesoderm plays a key role in large scale gut morphogenesis.
Because classic studies in heart morphogenesis suggest that the left and right sides of the looping heart tube elongate at different rates, we next evaluated the morphogenetic influence of tissue elongation during gut looping (Stalsberg and DeHaan, 1969). To quantify tissue elongation, we compared the lengths of the left and right sides of the midgut over time, at consecutive stages during the formation of the MG loop (41–43; Fig. 5B). We defined the midgut region as the segment of the gut tube between the duodenum (at its obvious junction with the wider proximal intestine) and the center of the prominent hindgut fold that will become the future apex of the intestinal coil. These morphologic landmarks correlate with the previously characterized expression domains of Nkx-2.5 and Wnt5A, respectively, at the midgut boundaries in multiple vertebrates (Smith et al., 2000). The lengths of the left and right sides of the midgut began to differ significantly as looping progressed, with the right side becoming approximately twice as long as the left by stage 43, the peak of midgut curvature. These length differences are closely reflected in measurements of the full lengths of the left and right sides of the gut tube at the same stages, suggesting that the differential elongation we perceive during the formation of this key asymmetric curvature is a real and significant developmental event (see Fig. 5B).
Of interest, during looping, the absolute length of each side of the midgut increases, suggesting that curvature formation does not involve the shrinking of the left side, but rather a decreased rate of elongation of the left side compared with the right. For example, from stage 42–43 (approximately 7 hr), the right side elongates, on average, 80 μm at the rate of approximately 11 μm/hr (n = 10). In contrast, the left side length only increases by an average of 16 μm, a rate of 2 μm/hr. Thus, the rate of elongation of the right side is, on average, over five times that of the left, at a critical phase of MG looping morphogenesis. The difference in the lengths of the two sides begins to diminish after stage 44, when coiling begins (data not shown). Although it is not feasible to accurately measure proliferation and elongation in the forming GD loop, due to the spatiotemporal constraints of its earlier development, it is likely that similar mechanisms would generate LR asymmetries throughout the contiguous tube. Overall, these results suggest that the mechanism of generating key LR asymmetric curvatures in the Xenopus gut involves a spatiotemporally defined, proliferation-independent differential elongation of the left and right sides of the primitive gut tube.
Differential Fates Versus Differential Morphogenetic Programs
It has been suggested that one of the functions of LR asymmetric gene expression is to cause cells on the left vs. right sides of the embryo to assume disparate fates (Capdevila et al., 2000). For example, the mature spleen is thought to arise from the splenic precursors located on the left but not right side of the embryo (Patterson et al., 2000). However, we found that the left and right LPM contribute equivalently to the individual digestive organs. Although we cannot rule out the possibility that finer-scale fate differences exist that are beyond the resolution of our map, our results suggest that the unilateral specification of particular digestive organs on the left or right side of the embryo is unlikely to underlie large-scale asymmetric morphogenesis in the Xenopus gut.
Our data show that the cells of the left and right LPM contribute to distinct topologic regions within the key loops of the forming digestive system. These observations are consistent with the idea that differential left and right morphogenetic programs are specified by LR asymmetric gene expression. For example, we show that the domain of Pitx2 expression in the left LPM, one of the most important sites of LR gene expression, contributes directly to the concavity of one of the most evolutionarily conserved LR asymmetries in the vertebrate digestive system, the gastroduodenal (GD) loop (Kemp, 1951; Larsen, 2001). This correlation is true in both normal and heterotaxic contexts. Moreover, we show that ectopic Pitx2 mRNA can locally induce ectopic concavities throughout the Xenopus gut tube, suggesting that Pitx2 expression is capable of being a cause, rather than merely a consequence, of such pivotal topologic changes. In addition, the localized distortion of gut topology associated with ectopic right side Pitx2 expression provides an immediate morphogenetic explanation for the digestive heterotaxias commonly associated with Pitx2 misexpression in Xenopus and other vertebrate embryos (Logan et al., 1998; Campione et al., 1999). Overall, these observations provide a foundation for higher resolution analyses of the cellular and molecular changes that occur as a result of Pitx2 expression during asymmetric morphogenesis.
LR Asymmetric Gut Tube Elongation
Previously, it was not possible to compare the proliferation rates or other morphogenetic properties of contralateral tissues within the digestive organs, because the location of the original left- and right-side cells was not defined. In this study, topographical models of the distribution of left- and right-side cells at successive stages of gut development allowed us to directly compare left and right morphometric parameters during looping. Our results indicate that localized differential elongation of the two sides of the gut tube may drive symmetry-breaking curvature formation in the Xenopus digestive system. The tadpole gut tube more than triples its length during the stages of asymmetric morphogenesis, and this elongation must be coordinated with looping and coiling to package the increased length inside the shrinking peritoneal cavity (Chalmers and Slack, 2000 and N.N.-Y., unpublished observations). Thus it is not surprising that the morphogenetic forces associated with elongation might be co-opted as a mechanism for generating anatomic asymmetries in this system.
Because we show that proliferation is equivalent in the left and right midgut, what are the cellular mechanisms underlying the large-scale differences in elongation? In our fate mapping experiments, we observed that clusters of labeled cells arranged themselves in a line parallel with the long axis of the gut tube during morphogenesis (for example, see label in stomach in Fig. 1B). Therefore, LR differences in convergent extension movements may be involved in differential gut elongation in this system. In addition, radial intercalation of the endoderm cells occurs during the lumination of the Xenopus gut and has been proposed to drive general gut elongation by expanding surface area (Chalmers and Slack, 2000). Thus, the differential elongation we observe on either side of the midgut could be the result of heretofore unrecognized LR differences in radial intercalation of the endoderm during gut lumen formation and expansion. Consistent with this possibility, we have observed intriguing asymmetries in the thickness and architecture of the left and right endodermal epithelium during midgut looping (N. Nascone-Yoder, unpublished observations). Detailed characterization of left and right cell rearrangements during gut morphogenesis will yield further insight into the role of elongation in asymmetric organ development.
A link between gut morphogenesis and the LR signaling pathway is suggested by the correlation of Pitx2 expression with the concave, more slowly elongating, surfaces of the gut loops. Pitx2 has been shown to influence radial intercalation during epiboly as well as the convergent extension-mediated elongation of the main body axis in multiple vertebrate embryos (Logan et al., 1998; Essner et al., 2000; Faucourt et al., 2001). In addition, Wei and Adelstein (2002) found that ectopic Pitx2a expression in cultured cells can modulate Rho GTPase signaling, induce actin reorganization, and strengthen cell–cell adhesion. Thus, Pitx2 activity may orchestrate the formation of concavities in the gut tube by inhibiting tissue elongation by means of localized influences on cell rearrangements and/or cell shape changes. Indeed, severe overexpression of Pitx2 in Xenopus has been reported to induce extremely shortened gut tubes (Campione et al., 1999), supporting the idea that localized Pitx2 activity normally inhibits gut tube elongation in defined areas to initiate looping. Further characterization of the downstream effectors of Pitx2 function and the molecular mechanisms of gut tube elongation will verify the role of LR gene expression in asymmetric digestive organ morphogenesis.
LR Asymmetric Morphogenesis in Other Organs
It is interesting to note that the left and right sides of both the primitive gut and the primitive heart tube undergo differential elongation during looping morphogenesis, and that Pitx2 is expressed in the concavities of both structures (Gormley and Nascone-Yoder, 2003). However, the looping tube represents only one example of the many lateralized organ topologies that are regulated by asymmetric gene expression in vertebrate embryos. We suspect that localized Pitx2 expression may generate morphologic asymmetry in organs that bud off of the primitive gut tube (i.e., lungs, liver) by similar mechanisms as those that influence gut tube elongation, because the same cell properties are also likely to influence branching and lobation patterns. For a complete understanding of LR asymmetric morphogenesis, it will be important to further compare topologies and cell properties in the left and right domains of the gut and other asymmetric organs, and to explore the role of Pitx2 isoforms in modifying these parameters in multiple developmental contexts.
Embryo Culture and Dissections
Xenopus laevis embryos were obtained by in vitro fertilization by using standard procedures and reared at either 14°C or 23°C in 0.1× MMR (Sive et al., 1998). To evaluate gut situs and dissect gut tubes, tadpoles were anesthetized with 0.05% benzocaine (Sigma) in 0.1× MMR. Whole gut tubes, esophagus through colon, were dissected out for in situ hybridization and morphometric measurements (see below) by using watchmaker's forceps and tungsten needles as described (Chalmers and Slack, 1998).
Bilateral Fate Mapping
To normalize labelings, a grid was constructed for each side of the stage 23/24 embryo (Nieuwkoop and Faber, 1969) based on the position of prominent embryonic landmarks, including the cement gland, pharyngeal arches, blastopore, and somite ridge. Each embryo was microinjected with a bolus of DiI solution (approximately 1% in ethanol; Molecular Probes) in the desired square on the left side, flipped over to expose its right side, and injected in the contralateral square with a separate needle containing undiluted PKH2GL solution (Sigma) as described (Gormley and Nascone-Yoder, 2003). All injections were performed by using a dual tool holder micromanipulator and a pressure microinjector (World Precision Instruments). Embryos were allowed to heal for 30 min at room temperature before the location of each label on each side of the embryo was verified with a stereo fluorescent microscope (Leica). At least 20 embryos were bilaterally labeled in this way for each square.
The later position of labeled cells in the digestive system was determined in live embryos through the transparent ventral integument at stages 41–44 (Nieuwkoop and Faber, 1969). At least two embryos were killed at each of these stages to dissect out the gut tubes and ascertain the labeling patterns in three dimensions. Finally, guts were dissected out of the remaining embryos (n = 7 to n = 18) at stage 45/46 and scored for the presence of each dye in the mature digestive organs. Mesodermal vs. endodermal labeling patterns were distinguished by the depth of the fluorescent cells in bisected gut tubes. Digital photographs of embryos and guts were captured by using a digital camera system (Optronics). Because guts are twisted and coiled, dissected tissues were flat-mounted with a coverslip fragment to improve photography of the area of labeling. Brightfield and fluorescent images were superimposed by using Photoshop 4.0 (Adobe) software.
Stage 41, 42, and 43 gut tubes were isolated, straightened, and fixed in a flat-mount position in MEMFA (Sive et al., 1998) such that the left and right sides of the gut tubes occupied opposite faces of the flat-mounted tissue. At least 10 gut tubes were flattened in this way for each stage, and then processed for immunohistochemical staining as described (Sive et al., 1998), using an anti-phosphohistone H3 antibody (Upstate Biotechnology) at a 1:500 dilution. Both the left and right sides of each processed gut were photographed with a digital camera. The number of labeled cells within three fixed square areas on both the left and right sides of the midgut was counted by using Scion image (NIH image) software, and the mean number of labeled cells on both sides was averaged over 10 gut tubes for each stage. A two-sample Student's t-test was performed to determine whether the mean proliferation on the left and right sides was significantly different (the means were hypothesized to be equal).
Stage 41, 42, and 43 gut tubes were isolated, straightened, and flat-mounted under a coverslip fragment to center the dorsal midline so that the left and right sides of the gut tube could be viewed simultaneously (n = 10 for each stage). Scion (NIH Image) image software was used to measure the length of the left and right sides of the entire gastrointestinal tract, as well as the left and right sides of the midgut from the prospective duodenum through the prospective hindgut. Prospective organ boundaries were selected based on published gene expression studies (Chalmers and Slack, 2000; Smith et al., 2000). Absolute lengths were calculated by comparison to a metric ruler.
Embryos were treated with heptanol (Sigma) during stages 5–12 as described (Levin and Mercola, 1998) with a 1:50,000 dilution of 1-heptanol in 0.1× MMR at 23°C, and then reared in 0.1× MMR with 50 μg/ml gentamicin. Because heptanol treatment can perturb gastrulation and delay development, we were careful to evaluate only those embryos with complete head development, a normal dorsoanterior index (Kao and Elinson, 1988), and guts with clear LR topologic abnormalities, as opposed to gut phenotypes that could have resulted from delayed morphogenesis. Adhering to these criteria greatly reduced the pool of scoreable guts. Treated guts at stage 40 (n = 10) were harvested and processed for in situ hybridization as described below.
In Situ Hybridization
Dissected whole guts were fixed in MEMFA for 1 hr at room temperature, gradually dehydrated through a methanol series, and stored in absolute methanol at −20°C until use. The in situ hybridizations were carried out as previously described (Sive et al., 1998), with modifications by C.V. Wright (personal communication). A digoxigenin-labeled Pitx2 riboprobe was synthesized from a linearized pCS2 plasmid (gift from M. Blum; Campione et al., 1999) with T7 RNA polymerase using the SP6/T7 DIG RNA labeling kit (Roche).
Ectopic Expression of Pitx2 mRNA
Synthetic Xenopus laevis Pitx2c mRNA was produced by in vitro transcription from the pCS2 plasmid (gift from M. Blum, Campione et al., 1999) by using the mMESSAGE mMACHINE SP6 kit (Ambion). The mRNA was mixed 1:1 with a solution of fluorescein-conjugated dextran (25 mg/ml) as a lineage tracer. A pressure microinjector was used to deliver 1–3 nl (0.5–1.5 ng) of Pitx2c RNA to both animal and vegetal right dorsal blastomeres at the 8-cell stage or both animal and vegetal right dorsolateral blastomeres at the 16-cell stage. Only embryos with normal anterior–posterior character and minimal curvature of the body axis were scored for gut situs.
The 3D models of whole digestive systems were created by observing multiple dissected gut tubes of both living and preserved specimens at each stage of development, and digitally “sculpting” mesh replicas by using the Biospheres feature of Amorphium (Electric Image). After rendering, the program was then used to “paint” the rendered models with different colors to indicate the topologic position of original left- and right-side cells, and/or gene expression patterns, as ascertained by multiple observations of dye labeled or in situ hybridized gut tubes.
We thank G.W. Litman, M. Mercola, D.M. Prather, K. Symes, and J.A. Yoder for their critique and comments on the manuscript. In addition, we also thank Martin Blum (University of Karlsruhe) for his kind gift of the Pitx2 plasmids. N.N.-Y. is supported by the American Heart Association, Fl/PR affiliate (no. 9930262V), and the Suncoast Cardiovascular Research and Education Foundation, founded by Helen Harper Brown.