Subdivision and Specification of the Splanchnic Mesoderm
In vertebrates, the lateral plate mesoderm is split by the coelom into the somatic mesoderm, later forming the body wall and limb buds, and the splanchnic mesoderm, forming the circulatory system, the respiratory system, and the digestive system. Funayama et al. (1999) demonstrated that the early splanchnic mesoderm, although morphologically separated and distinct from the somatic mesoderm, can be converted into the somatic mesoderm. In the present study, the properties of the wing, flank, and leg levels of the splanchnic mesoderm changed to those of the somatic mesoderm, and these levels of the splanchnic mesoderm then formed structures in accordance with their original levels, i.e., transplants of the wing-level and leg-level splanchnic mesoderm expressed Tbx5 and Tbx4 and formed wing-type and leg-type limbs, respectively, whereas transplants from the flank level failed to initiate outgrowth and never formed any skeletal structures. These findings using morphological phenotype as the limb cartilage pattern not only confirmed the results of a study by Funayama et al. (1999) but also demonstrated that region specificity exists along the rostrocaudal axis in the splanchnic mesoderm for making additional limb structures. Because Funayama et al. (1999) showed that the flank-level (23-somite) splanchnic mesoderm can also be redirected to the somatic mesoderm, the regional specificity would not be due to different abilities for splanchnic-to-somatic conversion but to regional difference in abilities to make limb buds. Therefore, information on regional difference to form an appropriate structure (i.e., wing, flank, or leg) seems to be given in the lateral plate mesoderm before its splitting into two layers and to be retained in both layers. This idea is supported by the results of our previous study, showing that wing and leg levels of the lateral plate mesoderm acquire the ability to express Tbx5 and Tbx4 before stage 9, a stage at which splitting of the lateral plate has not yet occurred (Saito et al., 2002). The limb-forming ability in the splanchnic mesoderm seems to have stage-dependency, and later-stage splanchnic mesoderm tends to form no limb structure (Table 1), suggesting that the information for regional specification that was inherited in the splanchnic mesoderm is getting lost as development proceeds.
If the same information as that given to the somatic mesoderm is also inherited by the splanchnic mesoderm, some questions arise: is this information used for subdivision of the splanchnic mesoderm into three distinct domains and what are the structures that each domain contributes to? The results of the DiI-labeling experiment (Fig. 4) revealed that the wing-level splanchnic mesoderm becomes the small intestine rostral to the yolk stalk and that the leg-level splanchnic mesoderm contributes to the formation of the caudal small intestine, cecum, and large intestine caudal to the yolk stalk. The flank-level splanchnic mesoderm participates in the formation of the small intestine around the yolk stalk. It should be noted that the flank-level splanchnic mesoderm accounts for a much narrower space than does the limb-level splanchnic mesoderm, although at stages 12–14 the wing (15–20 somites), flank (21–26 somites), and leg (27–32 somites) levels occupy areas of similar sizes. Similarly, the flank-level somatic mesoderm makes a relatively thin layer, but the limb-level somatic mesoderm outgrows to form a limb. The different states of extension of the limb region and flank region in both layers of the lateral plate reflect the position-specific properties along the rostrocaudal axis. Some Hox members have regionally restricted expression boundaries in the splanchnic mesoderm, demarcating morphological landmarks of the midgut. Chick Hoxc9 is expressed in the posterior midgut, and the anterior boundary of the gene expression in the splanchnic mesoderm is located at the yolk stalk (reviewed in Roberts, 2000). Of interest, the anterior boundary of Hoxc9 expression in the somatic mesoderm is at the anterior flank, and at later stages Hoxc9 is expressed in the hindlimb bud but not in the forelimb bud (Nelson et al., 1996; Cohn et al., 1997), indicating that the anterior border of Hoxc9 expression in the two layers of the lateral plate coincides with the boundary between the wing and flank subdivisions.
Because descendants of the flank-level splanchnic mesoderm were slightly mixed with cells derived from the wing and leg levels of the splanchnic mesoderm (Fig. 4E), it is not clear whether the boundary of each domain of the splanchnic mesoderm is established at stages 12–14 or not. During limb-developing stages, the anterior and posterior regions of the flank mesoderm are known to express Tbx5 and Tbx4, respectively. Therefore, the reason the wing/flank or flank/leg boundary is not clear might be that the wing and flank or the leg and flank share the competence to express Tbx5 or Tbx4. In any case, the wing and leg levels of the splanchnic mesoderm never mix with each other, suggesting that these levels are established as distinct domains.
The pattern formation along the AP axis in the limb bud is regulated by the ZPA. Hornbruch and Wolpert (1991) reported that slight ZPA activity in the lateral plate mesoderm first appears at stage 10 in chick embryos. The ZPA activity is almost ubiquitous at first, but in the presumptive wing region, the activity becomes localized in the posterior region around stage 12. On the other hand, it is not clear when ZPA activity emerges in the presumptive leg bud, although it appears in the posterior regions of leg buds at least by stage 16. The expression of shh, whose transcript is probably the actual morphogen from the ZPA, first appears in the ZPA regions of both wing and leg buds at stage 17, when limb buds start swelling (Riddle et al., 1993). Because the division of the lateral plate into the splanchnic mesoderm and somatic mesoderm is completed by stage 12 at the wing level and by stage 13 at the leg level (Funayama et al., 1999), region-specific ZPA activity and shh expression should be detectable after the lateral plate has been split. In this study, when the donor grafts from the splanchnic mesoderm and the somatic mesoderm were implanted in the same direction as that of the host rostrocaudal axis, 71% (39 of 55) of the splanchnic samples showed no shh expression, whereas 35% (6 of 17) of the somatic samples did not show shh expression, suggesting that it is much more difficult for the splanchnic mesoderm to express shh than it is for the somatic mesoderm. Moreover, it is likely that the splanchnic mesoderm implanted under the ectoderm tends to express shh more randomly than the somatic mesoderm (Fig. 3; Table 2). These results suggest that the splanchnic mesoderm has little information on shh expression. The AP polarity may be regulated after the somatic and visceral layers of lateral plate have split. In addition, it is known that the flank region has a very potent inverted ZPA-inducing gradient, as evidenced from the 100% inverted polarity of limb buds induced by FGF beads (Cohn et al., 1995), it could be that the variable polarity in the grafts is due to the competition between endogenous polarity and local effect of the flank host on the implanted mesoderm grafts. Therefore, it is also possible that the splanchnic mesoderm is susceptible to the potent ZPA-inducing ability in the flank.
Development of the Trunk Lateral Plate
Figure 5 shows a schematic representation of the hypothetical process of development of the trunk lateral plate. The single layer of the lateral plate mesoderm first subdivides into three distinct domains according to the positional information along the rostrocaudal axis. At this time, these three domains acquire the ability to form the wing, flank, and leg regions, respectively, as each distinct property. Then splitting of the lateral plate mesoderm occurs, and both layers retain the domain-specific property and form appropriate structures according to the secondary information in each domain. The domains of the somatic mesoderm form the wing, flank, and leg, respectively. In contrast, those of the splanchnic mesoderm change their property to “visceral” and form distinct parts of the small intestine, cecum, and large intestine. Specification of the AP polarity within the limb bud as regionalization of the ZPA occurs only in the somatic mesoderm after the split, to form the skeletal pattern of limbs along the AP axis. Probably, other modifications within each domain are also accomplished in each layer, resulting in the formation of appropriate organs along the rostrocaudal axis.
Figure 5. A model of the process of trunk lateral plate development. During early developmental stages, the positional information along the rostrocaudal axis is given to the lateral plate mesoderm. On the basis of this information, the trunk lateral plate is subdivided into three domains (indicated in blue, gray, and pink), and then the lateral plate is split into two layers: the splanchnic mesoderm and somatic mesoderm. After splitting, each domain in each layer is modified to form distinct organs. For example, a modification along the anteroposterior axis (the zone of polarizing activity induction) occurs only in the limb region of the somatic mesoderm (indicated in green) after splitting. See text for details. LPM, lateral plate mesoderm; So-M, somatic mesoderm; Sp-M, splanchnic mesoderm.
Download figure to PowerPoint
In conclusion, we propose that the lateral plate mesoderm has acquired information for subdividing the trunk lateral plate into three domains before splitting of the lateral plate occurs, meaning that the positioning of the limb is a very early event of development of the lateral plate. Although little is known about the molecular mechanism underlying the subdivision, it is possible that the median tissues, including the notochord, neural tube, and somites, may send out the information for subdividing the trunk lateral plate, because our previous study (Saito et al., 2002) suggested that median tissues are involved in restricted expression of Tbx5 and Tbx4 genes in the wing and leg buds. Another clue must be the Hox code. The Hox code in the paraxial mesoderm is known to be correlated with limb position (Burke et al., 1995; Gaunt et al., 1999). Limb-type–specific expression of Hox genes exists also in the lateral plate mesoderm (Nelson et al., 1996; Cohn et al., 1997). Furthermore, it has been shown that the axial position of the limb bud is shifted in Hoxb5-deficient mice (Rancourt et al., 1995). These facts suggest that the subdivision of the lateral plate mesoderm into three domains may occur according to the Hox code.