Gut development is characterized by rapid elongation during organogenesis, which causes temporary herniation as the growth rate surpasses that of the whole body and abdominal cavity (O'Rahilly and Müller,2001; Stainier,2005). However, the underlying mechanism of the gut elongation remains unknown. It has been suggested that during the human embryonic period, the duodenum (i.e., rostral midgut) is elongated by a “convergent extension (CE)” mechanism (Matsumoto et al.,2002; Schoenwolf et al.,2009). CE is a common developmental phenomenon, and consists of a narrowing and elongation of tissue driven by polarized cell movement, as in notochord elongation during gastrulation (for review, see Keller,2002). Noncanonical Wnt signaling has been implicated in polarized cell movement (for review, see Wallingford et al.,2002 and Veeman et al.,2003), and Wnt5a, which is a representative noncanonical Wnt member, was recently found to be a critical regulator of mouse gut elongation by means of the CE mechanism, as evidenced by the finding of shortened and widened small intestines in the absence of Wnt5a (Cervantes et al.,2009).
Ror2 is a member of the Ror family of receptor tyrosine kinases and acts as a receptor or coreceptor for Wnt5a, and mediates noncanonical Wnt signaling (Hikasa et al.,2002; Yoda et al.,2003; Oishi et al.,2003; Mikels and Nusse,2006; for review, see Minami et al.,2009). Ror2 has been demonstrated to play essential roles in developmental morphogenesis. Ror2 is expressed in the primitive streak, cephalic neural crest cells, prosencephalon, facial primordia, and developing limbs during early embryogenesis, as well as in the brain, heart, and lung at later stages of development (Oishi et al.,1999; Takeuchi et al.,2000; Matsuda et al.,2001), whereas Wnt5a is expressed in a graded manner in the developing limbs, tail, face, and genitals, the development of which requires extension from the body axis (Yamaguchi et al.,1999). Many of the abnormalities found in Ror2−/− and Wnt5a−/− embryos overlap during embryonic morphogenesis, including abnormalities of extension of the outgrowing structures along the proximal–distal axis, CE movement, and planar cell polarity (PCP), although the Ror2−/− phenotypes are generally milder than those of Wnt5a−/− mice. Ror2−/− and Wnt5a−/− mice exhibit short snouts with cleft palates, short limbs and tails, truncation of the trachea with distal lung abnormalities, and outgrowth defects in the genitals (Yamaguchi et al.,1999; Nomi et al.,2001; Oishi et al.,2003; Suzuki et al.,2003; Schwabe et al.,2004; Li et al.,2008; He et al.,2008). In addition, both Wnt5a−/− mice (Qian et al.,2007) and Ror2−/− mice (Yamamoto et al.,2008) exhibit misoriented stereocilia and cochlear phenotypes, indicating that their PCP and CE movements are dysregulated (for review, see Minami et al.,2009). Although extensive studies have documented the spatiotemporal expression patterns and the pleiotrophic functions of Ror2, including its relation with those of Wnt5a (reviewed in Minami et al.,2009), little is known about the expression pattern of Ror2 protein and its involvement in gut development. Here, we report that Ror2 exhibited a characteristic, region-specific, and stage-specific expression pattern along the rostral–caudal (RC) axis of the gut during gut development. Ror2 was most intensely expressed in the midgut endoderm during early gut organogenesis, and a shortened and widened midgut in Ror2−/− embryos was observed during the same period. Detailed analysis of midgut histogenesis suggested that Ror2 plays a role in the CE mechanism during midgut elongation.
Ror2 Exhibited a Region-Specific Expression Pattern Along the RC Axis of the Gut Endoderm and Mesenchyme During Development
We first analyzed Ror2 protein expression along the RC axis of the developing gut from embryonic day (E) 9.5 to E18.5 by immunohistochemistry. The expression of Ror2 showed a region-specific pattern in both the epithelium and mesenchyme during gut development, which is schematically shown in Figure 1B for E10.5 to E12.5. While Ror2 was detected in the endoderm of the midgut region at E9.5 (data not shown), its expression in the endoderm was the most intense from E10.5 to E12.5, and it began to decrease with time after E13.5. This strong expression of Ror2 in the midgut endoderm was the most pronounced in the rostral to middle region (Fig. 1A-3,B,G,H). A similarly strong level of Ror2 expression was also detected in the caudal stomach (the future pyloric antrum; Fig. 1B,C) and pancreatic buds (data not shown). Staining for Ror2 was moderately positive in the endoderm of the caudal midgut and entire hindgut during this period (Fig. 1B). Weak Ror2 signals were detected in the endoderm of the esophagus, the rostral region of stomach (future fundus), and the junctional region between the midgut and hindgut (Fig. 1A-1, 1A-4, B). As the gastric glands developed, Ror2 became confined to the base of the crypts (Fig. 1D–F). In the small intestinal epithelium, Ror2 became confined to the basal side of the epithelium after E13.5 (Fig. 1I for E14.5). As villus formation progressed from E15.5 to E16.5, Ror2 was weakly expressed in the intervillus region of the epithelium (Fig. 1J for E15.5). However, at E18.5, the Ror2-expressing region had shifted to the tip of the villi (Fig. 1K). In the colon, the epithelial expression of Ror2 was moderate from E10.5 to E12.5 (Fig. 1A-5, 1A-6, B) and decreased with villi formation (Fig. 1M, N).
In addition to the epithelium, the mesenchyme exhibited a characteristic Ror2 expression pattern. Ror2 was differentially expressed in the mesenchyme along the RC axis between E10.5 and E12.5, as was seen in the epithelium, but in a manner that differed from that in the epithelium (Fig. 1A,B). The mesenchyme of the junctional region between the midgut and hindgut, including the cecum, and that of the rostral hindgut, exhibited strong expression of Ror2 during this period (Fig. 1A-5, B). At E10.5 and E11.5, Ror2 was diffusely expressed in the mesenchyme of the gut (Fig. 1A). At E12.5, the two Ror2-expressing cell layers became distinct in the mesenchyme of the rostral midgut: the inner layer was for the most part overlapped with an α-smooth muscle actin (αSMA) -positive cell layer, and the outer layer was overlapped with a c-Kit-positive cell layer, both of which were confirmed by double-immunostaining of Ror2 together with αSMA and c-Kit, respectively (Fig. 2A,B). It is known that c-Kit-positive cells at E12.5 form a layer around the outer perimeter of the midgut, and these cells have been demonstrated to differentiate into longitudinal muscle cells and interstitial cells of the myenteric plexus in later periods of development (Torihashi et al.,1997). After E13.5, reduced Ror2 expression was observed in the outer peripheral layer, whereas stronger expression of Ror2 was seen in the inner circular muscle layer than in the outer longitudinal layer (Fig. 2A). The only exceptions were in the cecum and pyloric antrum, where Ror2 was expressed at comparable levels in both the inner and outer muscle layers (data not shown). The stronger Ror2 expression in the inner circular layer than in the outer longitudinal layer was maintained in the adult gut, although adult levels of expression were not as high as those seen during the embryonic period (data not shown).
A previous study (Matsuda et al.,2001) reported that, at E8.5, Ror2 mRNA was expressed in the cephalic neural crest, which migrates into the intestines to form the enteric nervous system (Young and Newgreen,2001). However, double fluorescence staining with antibodies against Ror2 and Ret, a marker of neural crest-derived cells, revealed that Ret was not expressed in the Ror2-expressing cells, but rather was expressed only in Ror2-negative cells, at E10.5 and E18.5 (Fig. 2C), suggesting that Ror2 does not play a positive role in the developing enteric nervous system on these days.
The expression of Wnt5a, which encodes a ligand for Ror2 (Oishi et al.,2003), also displays a region-specific pattern along the RC axis of the gut during development, as reported at E9.5, E10.5, E12.5, E14.5, and E16.5 (Yamaguchi et al.,1999; Lickert et al.,2001; Cervantes et al.,2009). In the present study, we analyzed the spatial expression of Wnt5a in the gut by in situ hybridization at E11.5, a time at which its expression in the gut has not been reported in the previous studies, and when Ror2 is differentially expressed in both the gut endoderm and mesenchyme. Wnt5a was detected in the mesenchyme of the rostral stomach (future fundus), caudal—but not rostral—midgut, and rostral hindgut (except for the cecum; Fig. 3); thus, this pattern differed from those of Ror2 expression in the endoderm and mesenchyme.
Ror2−/− Embryos Exhibited a Shortened Gut
To assess the functions of Ror2 in gastrointestinal development, we next analyzed gastrointestinal phenotypes in Ror2−/− embryos. The overall length of the intestines was reduced in Ror2−/− embryos as compared with that of controls at each stage (61.8 ± 7.9% at E10.5, n = 3 for each; 63.8 ± 13.8% at E11.5, n = 6 for the controls, n = 5 for Ror2−/−; 60.8 ± 10.8% at E12.5, n = 3 for each; 62.9 ± 7.5% at E13.5, n = 3 for each; 68.1 ± 6.4% at E15.5, n = 4 for the controls and n = 3 for Ror2−/−; and 68.1 ± 5.7% at E18.5, n = 3 for each; P < 0.05 at each date). At E18.5, in addition to the short gut, fundus of the stomach (Fig. 4A) and the cecum (Fig. 4B) were shortened in Ror2−/− embryos. We also observed that the gut of Ror2−/− embryos at E15.5 was malrotated and did not possess the herniated portion, whereas the control embryos showed convoluted intestines forming an umbilical herniation (Fig. 4C).
Middle Midgut of Ror2−/− Embryos Was Shortened and Widened at E11.5
As described above, Ror2 was most intensely expressed in the endoderm of the rostral to middle region of the midgut from E10.5 to E12.5, whereas its ligand, Wnt5a, was expressed in the mesenchyme of the caudal midgut during the same period. We, therefore, focused on the analysis of the E11.5 midgut of Ror2−/− embryos. The midgut length of Ror2−/− embryos was shortened to 62% of that of the control embryos (2.28 ± 0.42 mm in control embryos, n = 6; 1.42 ± 0.33 mm in Ror2−/− embryos, n = 5; P < 0.01), whereas the body weight (BW) and crown-rump length (CRL) showed no significant difference (BW 44.43 ± 6.74 mg for the controls, n = 21, vs. 41.26 ± 6.05 mg for Ror2−/−, n = 18, P = 0.13; CRL 6.92 ± 0.41 mm for the controls, n = 7, vs. 6.83 ± 0.28 mm for Ror2−/−, n = 7, P = 0.63). The cross-sections vertical to the RC axis revealed that the outer diameter of the epithelial tube in the middle part of the midgut was larger in the Ror2−/− embryos (Fig. 5B,C) than in the control embryos (Fig. 5A,C; 118.5 ± 9.1 μm for Ror2−/− vs. 91.5 ± 8.7 μm for controls, P < 0.01, Fig. 6A). The diameter of the epithelium rostral to the widened middle midgut was not significantly different between the Ror2−/− and control embryos (Fig. 6A). Intriguingly, at E11.0, the maximum diameter of the midgut epithelium in Ror2−/− embryos was not different from that in control embryos (113 ± 10 μm for Ror2−/− embryos, n = 3; 104 ± 8 μm for control embryos, n = 4, P = 0.2, Fig. 6A). As regards control gut development from E11.0 to E11.5, the midgut length increased 1.78-fold (P < 0.005, n = 4 for each stage), while the maximum diameter of the midgut tended to be reduced (87.7 ± 8.8%, n = 4 for each stage, P = 0.085; Fig. 6A). On the other hand, in Ror2−/− embryos, the maximum diameter did not change from E11.0 to E11.5 (113 ± 10 μm at E11.0 vs. 119 ± 9 μm at E11.5, P = 0.51), while the length increased only 1.39-fold (P < 0.05; Fig. 6A). These findings suggested that intestinal morphogenesis was disrupted in Ror2−/− embryos in terms of the proportion of elongation to diameter.
Defective CE Mechanism in the Absence of Ror2 Suggested by a Shortened and Widened Middle Midgut
Because the relative shortening and widening of the middle midgut in Ror2−/− embryos suggested a defective CE mechanism, we examined the histogenetic events in the midgut in closer detail. We initially counted all of the epithelial cells of the midgut using serial cross-sections along the RC axis. The total cell numbers of the midgut epithelium of Ror2−/− embryos at E11.0 and E11.5 were not significantly different from those of the control embryos (control vs. Ror2−/−; at E11.0, 5,484 ± 776 vs. 5,319 ± 332 cells, n = 4 for the controls and n = 3 for Ror2−/−, P = 0.75; at E11.5, 9,667 ± 1,485 vs. 8,705 ± 1,047 cells, n = 4 for each, P = 0.33). However, when we plotted the epithelial cell numbers in the cross-sections along the RC axis of the midgut, it was shown that, in the absence of Ror2, the epithelial cells accumulated in the middle part of the midgut (Fig. 6B), corresponding to the larger diameter of the same region than in control embryos (Fig. 6A).
We next examined cellular proliferation and apoptosis by 5-bromo-2′-deoxyuridine (BrdU) labeling and apoptotic cell counting, respectively. At E11.5, the overall proliferation and apoptosis in the midgut epithelium in Ror2−/− embryos were comparable to those in the control embryos (control vs. Ror2−/− embryos for percentage of BrdU-positive cells, 47.6 ± 5.6% vs. 42.0 ± 7.8%, n = 6 for each, P = 0.19; for the percentage of apoptotic cells, 0.48 ± 0.46% vs. 0.75 ± 0.46%, n = 4 for each, P = 0.45). The mean cell volume of the midgut epithelium did not significantly differ between the control and Ror2−/− embryos at E11.5 (628 ± 102 vs. 584 ± 57 μm3, n = 4 for each, P = 0.48). Thus, the short and dilated phenotype of the midgut in Ror2−/− embryos did not result from changes in cell proliferation, apoptosis, or cell size.
Another mechanism which could account for the shortened and widened shape of the epithelium in the middle midgut of Ror2−/− embryos would be the divergence of cell division axes (Matsuyama et al.,2009). We, therefore, measured the angle between the cell division axis and the RC axis of the E11.5 midgut in three-dimensional space. The distribution and the average of the cell division angles did not significantly differ between control and Ror2−/− embryos (F = 0.62, 56.1 ± 25.5 vs. 52.0 ± 27.3 degrees, n = 50 from 3 embryos for each, P = 0.44). Therefore, Ror2 does not appear to be involved in midgut elongation by means of regulation of the orientation of epithelial cell division.
Moreover, the patent vitelline duct was observed in the midgut of all of the Ror2−/− embryos (n = 10 at E11.5 and n = 3 at E11.0), while the lumen of the vitelline duct had already disappeared in all of the control embryos examined at E11.5 (n = 14; for E11.5, Fig. 5A,Biii,C). In half of the Ror2−/− embryos examined at E11.5, the cross-sections showed double lumens, i.e., the proper midgut and the vitelline duct that was bifurcated from the midgut. The outer diameter of the vitelline duct and the diameter of its lumen were wider than those of the proper midgut (Fig. 5Biii), which is similar to the findings in Wnt5a−/− embryos (Cervantes et al.,2009). In the region immediately caudal to the bifurcation of the patent vitelline duct, the diameter of the midgut epithelium became abruptly smaller in Ror2−/− embryos, whereas this narrowing was gradual in the control embryos (Fig. 5Biv,C).
Disrupted Cell Polarity in an Aberrant Cell Mass Within a Limited Region of the Midgut of Ror2−/− Embryos
One characteristic finding suggestive of the disrupted apicobasal polarity of the epithelium in the Ror2−/− gut was the aberrant epithelial cell clumps at E11.5, which were again similar to those reported in Wnt5a−/− embryos (Matsuyama et al.,2009). Interestingly, a single clump per embryo was observed in a region limited to that immediately rostral to the bifurcation of the vitelline duct (12 cases of 14 Ror2−/− embryos, vs. none of 14 control embryos; Figs. 5Bii,C, 7). In these cell clumps, the cells appeared to have lost their polarity and to be oriented in random directions. Phosphorylated aPKC, a marker for apicobasal polarity (Suzuki and Ohno,2006), was partially absent in the apical region of these clumps, suggesting disrupted apicobasal polarity of the cells (Fig. 7B; Supp. Fig. S1, which is available online). Moreover, there was no BrdU incorporation into the aberrant cell clumps (Fig. 7C), while BrdU was frequently observed in the other parts of the epithelium. On the other hand, apoptotic cells were frequently observed in the clumps (Fig. 7A), while they were rarely observed in other parts of the midgut epithelium of Ror2−/− embryos and in the midgut epithelium of control embryos. Interestingly, at E11.0, this type of clump of disoriented cells was not observed in the midgut of Ror2−/− embryos (n = 4; data not shown). Regarding the subcellular localization of Ror2 in the control midgut at E11.5, Ror2 was enriched in the basal membrane, moderately enriched in the lateral membrane, and sparsely present in the apical region (Fig. 2D,E; see Discussion).
To examine possible defects in the differentiation of gut components in Ror2−/− embryos, we examined the histological morphology. Hematoxylin and eosin (H&E) staining of sections showed a well-developed tissue architecture, including the villi and connective tissues, at E18.5, both in the control and Ror2−/− intestines (data not shown). Well-developed goblet cells in the small intestine stained with alcian blue were also observed both in the control and Ror2−/− embryos (data not shown). Anti-αSMA staining showed the two smooth muscle layers of both the control and Ror2−/− embryos (data not shown). Enteric neurons stained with Tuj1, an early neuronal marker, were observed in the small intestines, and there were no apparent differences in Tuj1 staining of control and Ror2−/− embryos at E15.5 (data not shown).
In human gut organogenesis, the total midgut elongates approximately three- to four-fold within a short period of time (from 1.21 mm at Carnegie stage 14 to 4.23 mm at Carnegie stage 17; approximate gestational age: 5 to 6 weeks after fertilization), whereas the crown-rump length increases only two-fold during the same period (O'Rahilly and Müller,1987), thus causing umbilical herniation at Carnegie stage 16 (O'Rahilly and Müller,2001), which indicates that this period is critical for midgut elongation. It has been shown that during this period, the duodenum undergoes a series of morphogenetic events while it elongates, i.e., a decrease in diameter of the epithelial tube, occlusion of the epithelial lumen, and recanalization leading to villi formation; disturbances in these events have been implicated in congenital anomalies such as duodenal atresia and duplicated duodenum (Matsumoto et al.,2002). That study showed that during the first two morphogenetic events, i.e., the decrease in diameter and luminal occlusion, there was neither a predominance of apoptosis nor an increase in cell numbers, suggesting that the epithelial cells converged toward the RC axis, resulting in gut elongation (Matsumoto et al,2002). Similarly, we found here that, during mouse gut morphogenesis from E11.0 to E11.5 (corresponding to the Carnegie stage 14 to 16 in humans; Theiler,1989), the epithelial tube diameter of the midgut had a tendency to narrow while it elongated, suggesting that the midgut elongates at least in part by means of the CE, as has been suggested in humans. These corresponding findings in humans and mice appear to verify that this dynamic morphogenesis, i.e., elongation of the midgut, occurs during this critical period by means of the CE mechanism. In the present study, Ror2 was strongly expressed in the epithelium of the rostral to middle midgut during this critical organogenetic period of the gut. Ror2−/− embryos exhibited a shortened and widened epithelial tube in the middle part of the midgut at E11.5. While the total epithelial cell numbers did not differ significantly from those of the control midgut, the epithelial cells were accumulated in the shortened middle midgut in Ror2−/− embryos. This morphogenetic abnormality suggests that, in the absence of Ror2, the epithelial cells were unable to rearrange themselves for elongation of the tube along the RC axis. Thus, the findings that both the strongest Ror2 expression and Ror2−/− phenotypes in the epithelium were observed in this period and region are consistent with the notion that Ror2 plays a role in the critical midgut elongation mechanism, at least in part by means of the mechanism of CE.
Although the abnormal phenotypes found in the gut of Ror2−/− embryos were not as severe as those of Wnt5a−/− embryos, many similarities between the two mutants were observed in the gut phenotype, including a shortened and widened midgut, dilated vitelline duct remnant, undisturbed differentiation of gut components (Cervantes et al.,2009), and aberrant cell clumps in the epithelium (Matsuyama et al.,2009). These similarities suggest that Ror2 indeed acts as a receptor (or coreceptor) for Wnt5a to mediate downstream signaling during gut development. However, we also observed several distinct differences between Ror2−/− and Wnt5a−/− embryos, which suggested that other downstream pathways for Wnt5a signaling and/or other ligands for Ror2 exist in gut development. BrdU-labeling assays failed to detect any changes in the overall cell proliferation rates in the middle midgut of Ror2−/− embryos. Therefore, Ror2 does not make an essential contribution to the regulation of proliferation in the epithelium during early midgut morphogenesis in mice. On the other hand, the epithelial proliferation is compromised in Wnt5a−/− embryos during gut development (Cervantes et al.,2009). Another difference between Ror2−/− and Wnt5a−/− embryos is the overall shortening of the gut in Ror2−/− embryos. In the absence of Wnt5a, the length of the small intestine is more severely affected than is the length of the other parts of the gastrointestinal tracts (Cervantes et al.,2009). Ror2 was expressed throughout the gut mesenchyme during all stages of embryonic development, while Wnt5a expression was localized to the mesenchyme in the junctional region between the midgut and hindgut. The difference in the affected regions may reflect the differences in the expression patterns and/or in the roles played in gut elongation by means of the mesenchyme (e.g., muscle layer formation), which warrant further investigation.
In Ror2−/− embryos, fundus of the stomach and the cecum were shortened at E18.5. These organs were regions exhibiting only low levels of Ror2 expression in the endoderm from E10.5 to E12.5. On the other hand, high levels of Wnt5a expression were seen in the mesenchyme of these regions. One possible mechanism for how these organs are affected in Ror2−/− embryos could be that high levels of expression of the ligand may compensate for the low levels of expression of the receptor; this mode of compensation may play a significant role in the morphogenesis of these structures. Another possible explanation could be that an intra-mesenchymal interaction between Wnt5a and Ror2 may be required for the development of the fundus and cecum. In fact, in the mesenchyme of the forestomach and the cecum, moderate to high levels of Ror2 expression were observed from E10.5 to E12.5.
It has recently been demonstrated that Ror2 mediates Wnt5a signaling to regulate mesenchymal cell migration along the posterior-to-anterior axis of the palate (He et al.,2008). In cultured mesenchymal cells, it has also been shown that Ror2 is required for Wnt5a-induced polarized cell migration, which it mediates by activating c-Jun N-terminal kinase (Nishita et al.,2006; Nomachi et al.,2008). In the present study, at E11.5, Wnt5a was expressed in the mesenchyme of the caudal midgut, whereas strong Ror2 expression was observed in the epithelium of the rostral to middle midgut. This reciprocal pattern of expression may create the morphogenic Wnt5a gradient for Ror2-expressing cells. There has been no evidence that Ror2-expressing epithelial cells move along the RC axis of the gut. Rather, epithelial cells are considered to move perpendicularly to the longitudinal axis during CE. It is, therefore, possible that with stronger Wnt5a signaling comes a greater extent of convergence of the tube in association with the gradient of Wnt5a along the RC axis.
In zebrafish, oriented cell division is a driving force for axis elongation during gastrulation (Gong et al.,2004). Oriented cell division along the RC axis may also contribute to organ lengthening in forestomach morphogenesis in mice at E12.5 (Matsuyama et al.,2009). However, our results revealed no differences in the distribution of the cell division angles between the control and Ror2−/− embryos, indicating that Ror2 does not play a role in controlling cell division orientation for midgut elongation.
In the embryos with Sfrps compound mutations (i.e., Sfrp1−/−Sfrp2−/−, and Sfrp1−/−Sfrp2−/−Sfrp5+/− embryos) as well as in Wnt5a−/− embryos, cell clumps with disrupted apicobasal polarity were observed, and it has been proposed that Wnt5a signaling, including Sfrp modulations, regulates the coordination of apicobasal polarity in the developing small intestines (Matsuyama et al.,2009). We also found polarity-disrupted cell clumps in the midgut epithelium of Ror2−/− embryos, suggesting that Ror2 may also be involved in the regulation of cell polarity. In the epithelial cells, Ror2 was enriched in the basal membrane, moderately enriched in the lateral membrane, and sparsely present in the apical region. This Ror2 subcellular localization along the apicobasal cell axis suggests its function in the establishment and/or maintenance of apicobasal cell polarity. While previous reports have demonstrated that Wnt5a interacts with Frizzled3 (Kawasaki et al.,2007), which is enriched in the apical region of the gut epithelium (Matsuyama et al.,2009), the present Ror2 distribution differed from that of Frizzled3, again suggesting multiple signaling pathways involving Wnt5a and/or Ror2.
It has been demonstrated that during early stomach development in humans, the epithelium is composed of pseudostratified columnar epithelial cells with their nuclei situated at different levels in the epithelium (Otani et al.,1993). In our present observations, during normal midgut development, mitotic cells were located on the luminal side of the epithelium, and DNA-synthesizing cells, detected as BrdU-positive cells, were on the basal side of the epithelium (for examples, see Fig. 7A,C), suggesting that gut epithelial cell nuclei may change their position synchronizing with the cell cycle, much like the interkinetic nuclear migration that occurs in the neuroepithelium during brain development (Jacobson,1991). The epithelial cells in the cell clumps that had lost their polarity in the Ror2−/− embryos showed no proliferation, and in fact showed increased apoptosis. These findings suggested that these cells had escaped from the normal cell kinetics and were in the process of elimination by apoptosis. Interestingly, the clumps were reproducibly observed in a region limited to that immediately rostral to the bifurcation of the dilated vitelline duct from the midgut, the diameter of which was abruptly reduced immediately caudal to the bifurcation in Ror2−/− embryos, in contrast to the more gradual narrowing observed in the control embryos. Thus, there was a reproducible and close spatial relationship between disrupted midgut morphogenesis and these aberrant cell clumps. This close relationship may indicate that these abnormal morphogenetic events are mechanistically linked to each other.
ICR (Jcl:ICR; CLEA Japan, Tokyo, Japan) mouse embryos were used for the Ror2 expression analysis. Noon of the day when a vaginal plug was observed was defined as E0.5. Mice lacking Ror2 expression were established as described previously (Takeuchi et al.,2000). Heterozygous mice were maintained on a C57Bl/6J background and crossed to generate embryos with the desired genotypes. Ror2−/− embryos were identified by appearance and PCR genotyping (Takeuchi et al.,2000). This study was approved by the Ethics Committee for animal experimentation of Shimane University, and the animals were handled in accordance with the institutional guidelines.
Histology, Immuno-histochemistry, and BrdU Labeling
After the dams were deeply anesthetized, the control and Ror2−/− embryos (n = 4 or more for each embryonic day) were collected and fixed in 10% formalin and 70% methanol overnight at 4°C. The gut was dissected out, processed into a paraffin block, and sectioned at 5 μm for H&E staining, alcian blue staining, and immunohistochemistry for αSMA, γ tubulin, and Tuj1. Immunohistochemistry for Ror2, Ret, c-Kit, E-cadherin, and laminin-1 was performed using cryosections. The gut was dissected out in normal saline, immersed in 20% sucrose, and embedded in OCT compound. After cryosectioning the samples at 10 μm, the sections were fixed in 2% paraformaldehyde for 30 min. For phosphorylated aPKC (phospho-PKCζ/λ), the embryo was fixed with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS), after which the gut was dissected out, processed into OCT blocks, and cryosectioned at 10 μm. The following antibodies were used for immunohistochemistry: αSMA (1A4, 1:50,000; Sigma, St. Louis, MO), c-Kit (1:50; R&D Systems, Minneapolis, MN), Ret (1:50; R&D Systems), Tuj1 (1:200; Covance, Princeton, NJ), E-cadherin (ECCD-2, 1:200; Invitrogen, Carlsbad, CA), laminin-1 (1:200; R&D Systems), phospho-PKCζ/λ (1:100; Cell Signaling, Beverly, MA), γ tubulin (GTU-88, 1:1,000; Sigma), and BrdU (3D4, 1:1000; BD Pharmingen, San Diego, CA). Rabbit polyclonal anti-mouse Ror2 antibody was prepared as previously described (Kani et al.,2004). BrdU (50 mg/kg) was injected intraperitoneally 2 hr before killing. For BrdU staining, the gut was fixed with 4% PFA and the paraffin block was sectioned at 5 μm. For antigen retrieval, 0.05% trypsin and 2N HCl were used. BrdU-positive and -negative cells were counted in 5 sections from the first flexure of the midgut, and in every 15 sections thereafter. Apoptotic cells or pycnotic nuclei were counted in 25 sections from the first flexure of the midgut, and in every 5 cross-sections.
Whole-mount In Situ Hybridization
The mouse Wnt5a probe plasmid, a gift from Dr. S. Takada (Takada et al.,1994), comprises a 360-bp fragment (positions 1260-1620 of GenBank Acc. No. NM_009524.2). To generate antisense and sense digoxigenin-labeled riboprobes for Wnt5a mRNA, plasmids were linearized with EcoRI and XbaI and subjected to in vitro transcription with SP6 and T7 RNA polymerases, respectively. The whole gut was fixed with 4% PFA in PBS at 4°C overnight. After pretreatment with proteinase K, it was hybridized overnight with a sense or an anti-sense RNA probe at 65°C in hybridization solution (50% formamide, 5× SSC, 1% SDS, and 50 μg/ml yeast tRNA). The gut was incubated with alkaline phosphatase-conjugated anti-digoxigenin Fab fragments (Roche, Mannheim, Germany) at 4°C overnight, followed by incubation with 5-bromo-4-chloro-3-indoyl phosphate (BCIP)/nitroblue tetrazolium (NBT) for visualization.
Epithelial Cell Counting and Measurements of the Gut Length, Mean Cell Volume, and Cell Division Axis
Total epithelial cells were counted in every 15 cross-sections of the midgut along the RC axis starting from the point just caudal to the bile duct to the end of the ileum. The total cell numbers were later corrected by Abercrombie's method: the corrected cell numbers = total cells counted × section thickness/(section thickness + nuclear length on the Z axis; Abercrombie,1946). For this correction, the average length of the 100 epithelial cell nuclei on the Z axis was measured using NIH ImageJ software (4.84 ± 0.74 μm in control embryos, 4.73 ± 0.76 μm in Ror2−/− embryos, P = 0.33). The gut length was measured after fixation. The convoluted gut was cut into segments such that the three-dimensional gut could be processed into segments in a two-dimensional plane, which rendered it possible to accurately measure the convoluted portions. Pictures of all segments were taken by a camera (Olympus E-330), and the length of each segment was measured by ImageJ software, after which the total length was calculated. Cross-sections of the midgut were prepared (25 μm thickness) and stained with an anti-γ tubulin antibody and hematoxylin for visualizing two spindle poles and chromosomes of the dividing cells. The coordinates of two spindle poles (i.e., the cell division axis) in the three-dimensional epithelium were provided by a microscope equipped with software for measuring the Z distance (Olympus BX51 with StereoInvestigator; MBF Bioscience, Williston, VT). The RC axis was provided by two coordinates: the centers of the areas in the upper and lower lumens of the gut section, as calculated by ImageJ software. The angle between the two vectors (the RC axis and the cell division axis) was calculated for the control and Ror2−/− embryos, respectively. The difference in distribution was evaluated by F-test. The mean cell volume in the midgut epithelium was calculated by dividing the total epithelial volume of the midgut (μm3) by the total epithelial cell numbers. The total volume of the midgut epithelium was yielded by Cavalieri's point counting technique (Slomianka and West,2005). Student's t-test was applied to evaluate the statistical significance of the results, which was set at the level of P < 0.05. Error bars in the figures indicate 1SD. The values are shown as the mean ± 1 SD.
We thank Dr. Shinji Takada for the Wnt5a probe, Ms. Yumiko Takeda for excellent tissue processing for the histological analyses, Ms. Mariko Ajiki for caring for the animals, Dr. Takeshi Urano and Dr. Shigefumi Yokota for advice on immunostaining, and Mr. Fumio Satow as well as all the members of the Department of Developmental Biology for their essential cooperation. We especially acknowledge the valuable advice of Dr. Masaaki Kurahashi and Dr. Akihiko Shimono on immunostaining during the revision. Y.M. was funded by the Ministry of Education, Culture, Sports, Science, and Technology, Japan.