Regulation of Hoxb4 induction after neurulation by somite signal and neural competence
© Amirthalingam et al; licensee BioMed Central Ltd. 2009
Received: 30 September 2008
Accepted: 25 February 2009
Published: 25 February 2009
While the body axis is largely patterned along the anterior-posterior (A-P) axis during gastrulation, the central nervous system (CNS) shows dynamic changes in the expression pattern of Hox genes during neurulation, suggesting that the CNS refines the A-P pattern continuously after neural tube formation. This study aims at clarifying the role of somites in up-regulating Hoxb4 expression to eventually establish its final pattern and how the neural tube develops a competence to respond to extrinsic signals.
We show that somites are required for the up-regulation of Hoxb4 in the neural tube at the level of somites 1 to 5, the anterior-most domain of expression. However, each somite immediately adjacent to the neural tube is not sufficient at each level; planar signaling is additionally required particularly at the anterior-most segments of the expression domain. We also show that the dorsal side of the neural tube has a greater susceptibility to expressing Hoxb4 than the ventral region, a feature associated with dorsalization of the neural tube by BMP signals. BMP4 is additionally able to up-regulate Hoxb4 ventrally, but the effect is restricted to the axial levels at which Hoxb4 is normally expressed, and only in the presence of retinoic acid (RA) or somites, suggesting a role for BMP in rendering the neural tube competent to express Hoxb4 in response to RA or somite signals.
In identifying the collaboration between somites and neural tube competence in the induction of Hoxb4, this study demonstrates interplay between A-P and dorsal-ventral (D-V) patterning systems, whereby a specific feature of D-V polarity may be a prerequisite for proper A-P patterning by Hox genes.
The anterior-posterior (A-P) identity of the body axis at the level of the hindbrain and the spinal cord is largely dependent upon the regulated expression of Hox gene clusters [1, 2]. At early embryogenesis, Hox genes are up-regulated sequentially in the epiblast and establish their ordered expression patterns along the A-P axis [3, 4]. They also play an instructive role in distributing cells in an ordered manner along the A-P axis during ingression of epiblast cells . As a consequence, Hox gene expression exhibits nested patterns in the paraxial mesoderm as well as in the neuroepithelium. One unique feature of conferring A-P identity by Hox genes is that these nested expression patterns display sharp anterior boundaries, creating codes of expression along the A-P axis [6, 7]. For example, expression of paralogue 4 Hox genes, such as Hoxb4, have an anterior-most limit at the rhombomere 6/7 boundary, while the anterior most limit of paralogue 5 genes lies at the rhombomere 7/8 boundary. Thus rhombomere 7 is defined as a Hox paralogue 4 positive and Hox paralogue 5 negative segment. As evidence of this code-dependent positional identity, null mutant mice of Hox genes exhibit the loss of a segmental identity only of the anterior-most domain of the gene expression [8–11]. Hence, regulation of Hox expression at the anterior-most domain is the most crucial step in the process of conferring A-P identity.
What is the possible factor responsible for the dynamic change in Hox gene expression in the neural tube? One strong candidate is the influence from flanking somites. It has been shown in chick embryos that transposition of regions of the neural tube along the A-P axis results in the reprogramming of Hox codes [15, 16]. Furthermore, somites have been shown to be able to up-regulate Hoxb4 when grafted ectopically in regions that do not normally express Hoxb4 [16, 17]. Similar results were obtained in zebrafish embryos, where grafting of non-axial mesoderm causes transformation of forebrain to a hindbrain character . These observations led to the idea that the neural tube undergoes continual assessment of its environmental signals in order to establish the correct pattern of Hox gene expression in the CNS after neural tube formation .
Retinoic acid (RA) is the most likely molecule responsible for the up-regulation of Hox genes by the somites. An enzyme retinaldehyde dehydrogenase-2 (RALDH2), which converts the inactive form retinaldehyde to RA, is expressed in the somites from early stages of development [19–21]. RA appears to be abundant in the neural tube as well as in the somites . Cyp1B1, another RA synthesizing enzyme, is also expressed in somites during early embryogenesis . In mice, at least, RALDH2 seems to be the main RA synthesizing enzyme in early embryogenesis at 7.5 dpc and 8.5 dpc, since the RA-responsive transgene RARE-hsp-LacZ  does not show expression in RALDH2-/- embryos at these stages except in the eye . RA deficiency caused either by genetic deletion of RALDH2 in mice [25, 26] or by placing quail hens on a RA-deficient diet  results in defects in axial development and patterning. The defects do not span the whole Hox-territory; rather, the defect is restricted to the posterior hindbrain (rhombomeres 4–8 including the level of somites 1–5), demonstrating that this region requires correct RA levels [28, 29].
To account for the complex organization of Hox genes by signaling mechanisms, a number of models have been proposed. For example, individual Hox genes have specific retinoic acid response elements (RAREs) with different sensitivities to RA, thereby allowing each Hox gene to be controlled differently depending on the concentration of RA . Another example of differential regulation of Hox genes is by FGF signaling, where some Hox genes have the binding sequence for a downstream transcription factor, Cdx, in their enhancers [31–33]. In addition to the above, it is likely that there are more mechanisms that are responsible for establishing the correct A-P pattern, such as the duration of exposure to signals, the degree of dependence on signals, the involvement of planar signaling, and the competence of the neural tube to respond to signals. This study aims at clarifying two issues. First, to what extent does endogenous Hox expression depend upon the somite signal? Second, what determines the competence of the neural tube to respond to the somite signal? In order to address these questions, Hoxb4 has been chosen as a model because the dynamic changes in its expression pattern occur at the stages when the tissues involved are accessible for refined dissections. The chick explant culture system was employed to identify tissue interactions that are responsible for the up-regulation of Hoxb4 expression in the neural tube.
The dynamics of Hoxb4expression in the developing chick neural tube
We first investigated in detail the changes in the Hoxb4 expression pattern that occur after neurulation. Hoxb4 is first detectable at the full streak stage (stage 4 of Hamburger and Hamilton, HH 4)  and the anterior-most boundary resides at the level of the future 6th somite both in the mesoderm and the neural tube until the 5 somite stage (HH 8+)  (Fig. 1A). This level is about 5 somite segments more posterior than the final anterior-most boundary of neural tube expression, the rhombomere 6/7 boundary, which corresponds to the anterior edge of the 1st somite. During somite stages 6 to 10 (HH 9- to 10), which is approximately a difference of 6 hours, the expression pattern rapidly changes exclusively in the neural tube; the domain here extends anteriorly while mesodermal expression remains at the same level (Fig. 1B–E). Expression extends until it finally establishes its anterior most limit at the future rhombomere 6/7 boundary. This does not involve cell movement as cells in the neural tube maintain their relative positions to the flanking somite at these stages . During somite stages 10–20 (HH 10–13), the expression becomes stronger while the domain remains unchanged (data not shown). By the 22 somite stage (HH 14), the rhombomere 6/7 boundary is formed and the Hoxb4 domain is clearly defined (Fig. 1F).
Signals from the mesoderm are necessary for the initial up-regulation of Hoxb4
At a given A-P level, signals from the somites are not sufficient to up-regulate Hoxb4expression
The above experiments suggest that the ability of somites to induce Hox genes in the adjacent neural tube is not simply determined by the production of RA. Other factors to be considered include the identity of the Hox gene, the relative timing of up-regulation of Hox genes and RA production, and the A-P position within the Hox gene expression domain. In order to address these, we used Hoxb3 as another Hox marker. Hoxb3 has an anterior-most expression domain at the rhombomere 5/6 boundary and establishes its final expression pattern at the 9 somite stage, slightly earlier than Hoxb4 . We first confirmed using the explant culture method that up-regulation of Hoxb3 requires flanking somites at the 5 somite stage, in a similar manner to Hoxb4 (Fig. 3S, T). By dissecting somite level 1 neural tube along with flanking somites, it was observed that somite 1 is sufficient to up-regulate Hoxb3 in the adjacent neural tube (Fig. 3U). Hence, the 1st somite is not incapable of inducing Hox genes; the failure to up-regulate Hoxb4 in somite level 1 neural tube is not due to the specific feature of the 1st somite; rather, different signal strengths are required for the up-regulation of different Hox genes. Nevertheless, the result of dissecting neural tube and flanking somites at each of the somite levels 1 to 5 (Fig. 3A–L) is in clear contrast to the result of explanting the neural tube and somites as a whole (Fig. 2G–I), where the neural tube displays homogenous expression along its axial length. These data suggest that vertical signals alone are not sufficient and planar communication or signaling is required.
Involvement of planar signaling
Barrier-placing experiments were further conducted using whole embryos at the 2 somite stage, by placing a barrier posterior to the 2nd somite (future somite 2/3 boundary). The barrier was placed across the neural tube as well as all other tissues including somites. Embryos were incubated ex ovo on albumen-agar plates , which helped stable positioning of the barrier. After 24–26 hours of incubation, Hoxb4 induction on the anterior side of the barrier was found to be clearly blocked, either completely (n = 4/12) or to significantly low levels compared to the posterior side of the barrier (n = 7/12) (Fig. 4B, Additional file 1). This result is consistent with the above experiment of explant culture suggesting planar signaling; however, it is not in agreement with the study by Gaunt and Strachan (1994) where the Hoxd4 expression domain was shown to 'spread forward' regardless of physical barrier after 24 hours of culture. The reason for this discrepancy is not clear. One possible explanation is that ex ovo culture may cause delays in chick development and Hoxd4/b4 might be up-regulated after a longer incubation, although embryos incubated ex ovo in the same condition without barriers showed a clear up-regulation of Hoxb4 as normal (Additional file 1). However, there was one case showing strong Hoxb4 induction anterior to the barrier (n = 1/12, Additional file 1), suggesting that the results may vary among the cases. Nonetheless, both our explant and in vivo studies suggest the requirement of tissue continuity in up-regulating Hoxb4 in the neural tube, suggesting involvement of planar signaling. The planar signaling may be required for the initial extension in expression, and/or the maintenance of the expression initiated by signals from flanking somites.
Given the importance of planar signaling in the neural tube, it was next asked whether anterior somites are required for the planar signaling or not. Somites 1 and 2, although unable to up-regulate Hoxb4 in the adjacent neural tube in a vertical manner (Fig. 3), are still required for Hoxb4 expression, since removal of somites 1 and 2 from the explant of somite level 1–5 abolished up-regulation of Hoxb4 in the anterior neural tube (Fig. 4C, compare with Fig. 2I). This could be because somites are either required for instructive vertical signals or for sending permissive cues that allow the anterior neural tube to respond to planar signals. We have noticed that the neural tube and somites from level 4 or 5 cause a stronger induction in the level 1 neural tube compared to those from more anterior levels such as levels 1–3 (Additional file 2). We therefore designed experiments where a neural explant dissected at somite levels 1–2 (without flanking somites) was combined with a level 5 neural tube with flanking somites (Fig. 4D). Following a 24 hour culture, significant up-regulation of Hoxb4 was seen in the anterior neural tube in an area approximately one somite diameter in length (Fig. 4D). Hence, the neural tube is able to up-regulate Hoxb4 in the absence of flanking somites when combined with a more posterior neural tube with somites. In a similar experiment where the neural explant of somite level 1–2 was combined with a level 5 neural tube, where all levels included somites, it was noted that up-regulation of Hoxb4 was fully extended to the anterior-most end of the explant (Fig. 4E), despite that these anterior somites are not capable of sufficiently up-regulating Hoxb4 in the adjacent neural tube (Fig. 3H, I). These results demonstrate that somites 1 and 2 are required for up-regulation of Hoxb4 in the neural tube, at least in part to assist in the response of the neural tube to planar signaling, suggesting a synergistic effect between planar and vertical signals.
Dorso-ventral difference of Hoxgene expression
There is additional evidence indicating that the dorsal side of the neural tube precedes the ventral side in Hoxb4 expression: When the neural tube is translocated posteriorly, the graft up-regulates Hoxb4 in response to the new posterior environment. In this situation, the dorsal side of the graft expresses Hoxb4 earlier than the ventral side . In another situation, where somites are grafted anteriorly into the pre-otic region causing ectopic induction of Hox genes, Hoxb4 up-regulation can be seen predominantly in the dorsal edge of the neural tube, which is particularly the case when anterior somites with weak inductive abilities are used . Furthermore, during the normal course of development (HH 18–20), there is a transient up-regulation of Hoxb4 in the dorsal rim of rhombomere 6 while the anterior-most limit of the main expression domain is at the rhombomere 6/7 boundary . These observations collectively imply that the dorsal side of the neural tube might have a greater susceptibility to expressing Hox genes than the ventral region.
BMP signals are involved in up-regulating Hoxb4in the neural tube
An obvious feature of the dorsal neural tube is its contact with the surface ectoderm and subsequent expression of BMP ligands and other dorsal neural tube markers such as Msx1/2 and Pax7 . It has been shown that a neural explant induces dorsal markers Msx1/2 in response to contact with the surface ectoderm . In fact, transverse sections of an embryo at the 12 somite stage revealed that up-regulation of Hoxb4 appears to occur most strongly in the region where the neural tube is in contact with the surface ectoderm (Fig. 5E–G). Hence, an experiment was conducted to determine whether Hoxb4 expression is up-regulated in response to the dorsalizing signal from the ectoderm in vivo. The dorsal neural tube (approximately 10–15% of the neural tube) was ablated at the level of prospective somites 1–5 on one side together with the surface ectoderm immediately overlying the neural tube (Fig. 6G). In some cases, a foil barrier was placed at the edge of the surface ectoderm to prevent it from regenerating and fusing to the dorsal neural tube (Fig. 6J). The ablation was conducted on 3 to 4 somite stage embryos, after which they were incubated until the 10–12 somite stage. Without a foil barrier, the surface ectoderm rapidly regenerated within 6 hours following ablation and formed a continuous epithelial layer with the ablated end of the neural tube. In these cases, up-regulation of Msx1 was seen at the dorsal tip of the ablated side of the neural tube (Fig. 6H), as was Hoxb4 (Fig. 6I). In contrast, in cases where regeneration of the surface ectoderm was inhibited by the barrier, neither Msx1 (Fig. 6K) nor Hoxb4 (Fig. 6L) were up-regulated. This suggests that up-regulation of Hoxb4 in the dorsal neural tube is associated with contact with the surface ectoderm and subsequent dorsalization of the neural tube.
BMP4 requires RA signalling to up-regulate Hoxb4
Since RA is able to up-regulate the expression of Hoxb4 and is known as the likely candidate for the somite-derived signals, the possible mechanism for up-regulating Hoxb4 expression by BMP4 was further examined to determine if BMP4 is capable of up-regulating Hoxb4 independently of RA or whether BMP4 requires RA in order to up-regulate Hoxb4. With the aim of blocking RA signaling, embryos at 4–5 somite stages were cultured ex ovo in the presence of a RA receptor antagonist, BMS493 . BMS493-treated embryos failed to up-regulate Hoxb4 in the neural tube at the level of somites 1–6, when observed at both the 10–12 and 22 somite stages (Fig. 7A, E and 7D, H, respectively). This was consistent with the result seen in vitamin A deficient quail and RALDH2-/- mouse embryos [25–28, 41, 42]. In the presence of BMS493, additional BMP4 protein did not up-regulate Hoxb4 expression at the level of somites 1 to 6 (Fig. 7F, G). Therefore it is suggested that, during the course of Hoxb4 up-regulation at the level of somites 1 to 6, RA is absolutely required and BMP4 cannot compensate for its absence. This was supported by explant culture experiments, where the neural tube at the level of somites 1 to 5 was dissected without flanking somites and cultured in the presence of RA or BMP4 (Fig. 7I–K). While exogenous RA sufficiently up-regulated Hoxb4 expression in the neural tube explant (Fig. 7J), BMP4 was not able to do so (Fig. 7K), suggesting that BMP4 cannot exert its function to up-regulate Hoxb4 in the absence of somites. Hence, while RA plays an instructive role in the up-regulation of Hoxb4, BMP4's role is likely to be permissive rather than instructive.
We further investigated whether RA is able to promote ventral expansion of Hoxb4 expression in a similar manner to BMP4 in vivo. While BMP4 showed clear up-regulation in the ventral side of the neural tube (Fig. 7L, N), exogenous RA only enhanced the dorsally dominant Hoxb4 expression and did not show up-regulation in the ventral neural tube as significantly as BMP4 did (Fig. 7M). The result that exogenous RA cannot up-regulate Hoxb4 in the ventral neural tube while BMP4 can, underscores a distinct role of BMP signaling in the in vivo context. Collectively, these results suggest a two-phase model in establishing Hoxb4 expression in the axial level of somite 1–6. First Hoxb4 is up-regulated at the dorsal neural tube by signals from the surface ectoderm, likely mediated by BMP or TGFβ signaling (Fig. 6G–L). RA is required for the dorsal patterning process (Wilson et al., 2004), and hence this initial phase likely employs both signals. Second, Hoxb4 expression spreads more ventrally, which can be promoted by exogenous BMP4 but not by RA. However, this process does not occur in the absence of RA, at least at the somite level 1 to 6. It is not clear in the experiment of Figure 7 using BMS493 in vivo, whether the requirement of RA is only in the initial step at the dorsal side, or also in the up-regulation at the ventral side. However, the result that after removal of dorsal neural tube, the remaining ventral neural tube shows Hoxb4 expression in a comparable manner to the control side (Fig. 6L) suggests that RA/somite up-regulates Hoxb4 at the ventral side independently of the preceding dorsal expression. Given the direct role of RA on the Hoxb4 enhancer , the data suggest distinct functions of RA and BMP signals for up-regulating Hoxb4 in the ventral neural tube, where RA provided by somites functions as an essential signal, while BMP4 functions as a factor facilitating the neural tube to respond to the RA/somite signal.
It was noted that, in the neural tube at the level posterior to the 7th somite, BMP4 is able to up-regulate Hoxb4 even in the presence of BMS493 (Fig. 7E–G, arrows). This axial level does not require RA signaling for Hoxb4 expression [25–28, 41, 42]. Hence there remains a possibility that BMP signals may be able to up-regulate Hoxb4 independent of RA.
Dorsalization of the neural tube precedes up-regulation of Hoxb4
Many previous studies have focused on vertical signals during the process of neural A-P patterning. While RA is likely to be the main signal derived from somites, it has been difficult to explain the neural A-P patterning process solely by RA/somite signals. The aim of this study is to shed light on new aspects other than the factor(s) derived from somites, that is, ectoderm-derived BMP/TGFβ signals and the subsequent acquired competence of the neural tube to respond to the somite/RA signal, together with planar signaling. This work also highlights a mechanism where RA and BMP4 act in a concerted manner to initiate neural Hoxb4 expression.
Vertical and planar signaling
The classical idea of vertical and planar signaling has been proposed in studies of neural induction and patterning in amphibian embryos [44–46]. With regard to neural A-P patterning, Nieuwkoop proposed that extrinsic 'caudal influences' originate from the mesoderm  whereas others showed that planar signals alone can induce neural A-P pattern based on experiments with exogastrula embryos and Keller's explants . Hence the role of vertical and planar signals in amphibian gastrulae remains debatable.
This work has focused on the process of neural patterning long after neural induction, revealed by up-regulation of Hoxb4. It has been demonstrated that vertical signals from somites are required but not solely responsible for the establishment of the Hoxb4 pattern (Fig. 3A–L). Additional signals emanating from more posterior tissues (neural tissue and/or somites) work in conjunction with the vertical signals to up-regulate Hoxb4 expression in anterior regions. These signals travel within the plane of the neural tube (planar signals), however, the source and identity of the planar signal remains to be determined. Based on the requirement of flanking somites for planar signaling (Fig. 4D, E), we suggest that somites provide not only instructive signals for Hoxb4 up-regulation but also permissive signals that assist in planar signaling.
Other studies have also suggested the existence of planar signals. This includes experiments where rhombomeres were transposed and incorporated at different A-P levels of the neuroepithelium. Induction of Hox genes was only observed in the grafted fragment of tissue providing the graft was perfectly incorporated into the host's neuroepithelium . This suggested that the inducing signals are being transduced along the plane of the neural tube. In contrast, studies by others have discounted the possibility of planar signals. Analysis of the anterior extension of the Hoxd4 expression domain demonstrated that tissue continuity was not required in order for Hoxd4 expression to be established . Implantation of a glass barrier in the neural tube of a 2 somite stage embryo posterior to the 2nd somites did not prevent the extension of Hoxd4 expression, thus implying that planar signals are not necessary after the stage at which the glass barrier was placed. This experiment was reassessed in the present study using the ex ovo culture system with Hoxb4 as a marker (Fig. 4B and Additional file 1). The result showed a variable yet significant block of Hoxb4 induction at the anterior side of the barrier, suggesting that tissue continuity is indeed required during the normal course of Hoxb4 up-regulation.
The actual mode of action of the planar signal remains elusive. It is possible that one Hoxb4-expressing cell activates Hoxb4 expression in the cell(s) adjacent to it. This idea comes from an observation in Xenopus embryos, where injection of Hoxb4 mRNA into one blastomere causes induction of endogenous Hoxb4 expression outside of the injected lineage , suggesting that Hoxb4 expression is able to induce its expression in adjacent cells in a cell non-autonomous manner.
The mode of action of the somite signal is influenced by the competence of the neural tube
This study, as well as previous work by others, has shown that in order to establish Hoxb4 expression in the neural tube, RA from the adjacent somites is required [12, 30]. However, as seen from transverse sections of normal embryos (Fig. 5E–G), the somite, although adjacent to the neural tube, is not necessarily close to the dorsal side of the neural tube, which is where Hoxb4 is strongly expressed initially. This raises the question of how the neural tube exploits the RA/somite signal to initiate Hoxb4 expression. Somite rotation experiments showed no evidence of D-V difference in the strength of inducing ability in somites. This is in agreement with the expression pattern of RALDH2, which shows homogeneous expression along the D-V axis of the somites . Our data suggest that RA is provided evenly at the dorsal and ventral sides of the neural tube, and it is the action of BMPs or other members of the TGFβ super family that may sensitize the dorsal neural tube to the RA signal, causing stronger Hoxb4 expression dorsally.
The effect of D-V differences in establishing the regional specificity along the A-P axis
This work has suggested that BMP signaling is involved in rendering the neural tube competent to express Hoxb4 in response to RA or somite signals. Since BMP signaling is a specific feature of the dorsalization of the neural tube, this provides evidence that establishment of the Hoxb4 expression pattern, and hence establishment of A-P positional identity, is under the influence of D-V specific cellular characters, demonstrating interplay between the patterning of these two axes.
The phenomenon of A-P positional markers being initially up-regulated at the dorsal side of the neural tube is common in many Hox genes in the spinal cord and in Krox20 in rhombomere 5 [12, 50]. However, it should be noted that the D-V difference might be only to influence the initial up-regulation and not to affect the expression domains of each Hox gene along the A-P axis. Genetically modified animals with affected D-V patterning in the neural tube, such as zebrafish embryos with compromised BMP signaling  and mouse embryos with deficiency in shh signals , exhibit correct A-P patterns. Therefore the dorsal specific features are, at most, to facilitate the establishment of the A-P patterns and not to give a clue for the correct A-P patterns. In fact, over-expression of BMP4 causes a ventral expansion of Hoxb4 only at the specific A-P level where Hoxb4 is normally expressed, and never anteriorly beyond the normal rhombomere 6/7 boundary (Figs. 7B, C; 8A, B). Thus the mechanism to prevent the anterior extension of expression domains is yet to be clarified.
The effect of RA on Hoxb4up-regulation
RA signaling is important not only in rhombomere patterning along the A-P axis but also in specifying the dorsal neural tube. Quail embryos deficient in RA exhibit a great loss of dorsal neural tube-specific markers such as BMP4/7, Msx2 and Pax3/6/7 . However, RA does not appear to be sufficient for dorsalization of the neural tube. It is rather BMP4 that is responsible for dorsalization [38, 53]. Hence it appears that RA is required to exert BMP4's dorsalizing activity. Another example of RA functioning in such a supporting manner is observed in the ventral neural tube during motor neuron differentiation. RA is required for the shh signal to induce olig2 in the ventral spinal cord [54, 55]. Again, RA is not sufficient to induce olig2 in the absence of shh signaling; it is shh which induces olig2. Hence, in these contexts, RA acts as a factor to render the neural tube competent to other extrinsic signals such as shh and BMP4. In the case of Hoxb4, it has been shown that RA works directly as an inducer of Hoxb4 expression through an RARE in its enhancer . In addition to this, the present study proposes that RA might also act to maintain the dorsal-specific domain that serves as an 'initial up-regulation area' for Hoxb4, which is likely induced by ectoderm-derived BMP4/TGFβ signals. Furthermore, our data showed distinct functions of RA and BMP signals: excess RA expands the Hoxb4 expressing domain anteriorly  but not ventrally (Fig. 7M), while excess BMP4 facilitates its ventral expansion. However, BMP4 cannot accomplish ventral expansion of Hoxb4 in the absence of RA (Fig. 7F, G). Hence RA and BMP4 are mutually required for Hoxb4 expression both in the dorsal domain as well as in the more ventral side of the neural tube.
The mechanism of up-regulating Hoxb4 after neural tube formation was investigated. While vertical signals from somites are necessary for up-regulating Hoxb4 expression in the adjacent neural tube, these signals are not always instructive in nature; especially in the anterior-most region of the Hoxb4 expressing domain, the flanking somites do not sufficiently up-regulate its expression in the adjacent neural tube, yet they are necessary to provide permissive cues that allow the neural tube to respond to planar signals. Hoxb4 is initially up-regulated at the dorsal neural tube, with this up-regulation correlating with the dorsalized character of the neural tube. Moreover, somite/RA-dependent up-regulation of Hoxb4 is promoted by BMP signals. These data suggest that establishment of the Hoxb4 expression pattern, and hence establishment of A-P positional identity, is under the influence of D-V specific cellular characters, demonstrating interplay between the patterning of these two axes.
Chick explant culture
Fertilised chick eggs were incubated at 38.5°C until they reached the required somite stages. All dissections were performed in L-15 medium in a petri dish. The chick explant culture system utilises the two-drop collagen method: Collagen was extracted from rat tail by dissolving 1 g of tendon in 100 ml of 0.2% acetic acid, followed by dialysis against 0.1× DMEM (pH 4). The collagen solution was used at a concentration of 2 parts stock collagen to one part 0.1× DMEM. 270 μl of this working solution was then mixed with 30 μl of 10× DMEM. 5–10 μl of 1 M sodium bicarbonate was then added until the solution turned a pale orange colour. 20 μl drops of collagen were placed in each well of a 4-well plate (where each well measures 16 mm in diameter) and left to set for 30 minutes at 37°C with 5% CO2. Explants were placed on top of the first drop. A small groove was cut into the surface of the first collagen drop and the explant was placed carefully inside. In cases where somites should be removed from the neural tube, Dispase I (Roche) was used at 5 units/ml for no more than 3 minutes. Once somites and other surrounding tissues were removed, the neural tube was placed in a fresh drop of L15 thus removing any residual Dispase, this was followed by embedding the neural tube in the collagen drop. A second drop of collagen solution was added (approx. 15 μl) on top of the explant. The collagen was allowed to set as described above after which 400 μl of F-12 solution was added containing penicillin/streptomycin and 0.1% Mito serum extender (BD Biosciences). Explants were cultured for 24 hours at 37°C with 5% CO2, after which the explant was fixed with 4% paraformaldehyde in PBS for 15 minutes, gradually dehydrated to 100% methanol and stored at -20°C. In situ hybridization was performed in the same way as whole embryos as described previously .
For the purpose of placing a barrier across the axis in vivo, chick embryos at 2 or 10 somite stage were transferred to albumen-agar plates  and a piece of aluminium foil (approximately 300 mm × 900 mm) was placed at the level posterior to the 2nd somite, separating all three germ layers into anterior and posterior sides. The embryos were incubated for 24–26 hours and processed for in situ hybridization.
Ablation experiments were performed in ovo using 3–4 somite stage embryos. Embryos were fixed at 10–12 somite stages and processed for in situ hybridization.
Treatment of embryos with BMS493 and BMP4
A RA-receptor antagonist BMS493 was prepared following the same synthetic sequence described for related analogues . Addition of the lithium derivative of 1-ethynylbenzene to 7-bromo-4,4-dimethyl-3,4-dihydronaphthalen-1-(2H)-one  to give the propargyl alcohol (82%), followed by dehydration with p-TsOH (83% yield), formylation (n-BuLi, then DMF, 76%) and Horner-Wadsworth-Emmons condensation with the anion of methyl 4-(diethylphosphonyl-methyl)-benzoate (n-BuLi, DMPU) afforded the entire arotinoid skeleton (89% yield). Saponification of the ester group afforded BMS493 (89% yield).
For treatment of embryos with BMS493, BMP4 or RA, embryos were dissected at 2 to 5 somite stages in L15 and cultured ex ovo  in optiMEM (GibcoBRL) with relevant compounds and/or proteins for 8–24 hours until they reached the 10–12 or 22 somite stages. BMS493 was used at the concentration of 4 μM. As the stock of BMS493 was dissolved in a 1:1 mix of DMSO and ethanol, control embryos were also cultured with the equivalent amount of DMSO and ethanol (0.004% each). Human recombinant BMP4 (rBMP4, R&D Systems) was used at a concentration of 60 ng/ml. BMP4 conditioned medium was obtained by transfecting HEK293 cells with a plasmid encoding BMP2/4 fusion for improved secretion of processed BMP4 . Cells were incubated for two days after transfection with DMEM and 10% foetal calf serum before the conditioned medium was collected. Retinoic acid (all-trans, SIGMA) was used at the concentration of 500 nM.
Flat mounting was performed after in situ hybridisation staining by dissecting the neural tube in 80% glycerol and opening it from the dorsal side. Vibratome sectioning (Fig. 5E–G, Fig. 6H, I, K, L) was performed by embedding embryos in 3–4% agarose and cutting at a thickness of 40 μm. Cryosectioning (Fig. 6C, F) was performed by embedding embryos in OCT compound (BDH) and cutting at a thickness of 13 μm.
We are grateful to Profs C. D. Stern and N. Papalopulu for valuable suggestions and discussion. We thank G.M. Ross, D.W. Ballard and S. Guidato for suggestions, discussion and performing preliminary experiments. This work was supported by the Medical Research Council, UK.
- Lumsden A, Krumlauf R: Patterning the vertebrate neuraxis. Science. 1996, 274: 1109-15. 10.1126/science.274.5290.1109.View ArticlePubMedGoogle Scholar
- Krumlauf R: Hox genes in vertebrate development. Cell. 1994, 78: 191-201. 10.1016/0092-8674(94)90290-9.View ArticlePubMedGoogle Scholar
- Gaunt SJ, Strachan L: Temporal colinearity in expression of anterior Hox genes in developing chick embryos. Dev Dyn. 1996, 207: 270-80. 10.1002/(SICI)1097-0177(199611)207:3<270::AID-AJA4>3.0.CO;2-E.View ArticlePubMedGoogle Scholar
- Wacker SA, Jansen HJ, McNulty CL, Houtzager E, Durston AJ: Timed interactions between the Hox expressing non-organiser mesoderm and the Spemann organiser generate positional information during vertebrate gastrulation. Dev Biol. 2004, 268: 207-19. 10.1016/j.ydbio.2003.12.022.View ArticlePubMedGoogle Scholar
- Iimura T, Pourquie O: Collinear activation of Hoxb genes during gastrulation is linked to mesoderm cell ingression. Nature. 2006, 442: 568-71. 10.1038/nature04838.View ArticlePubMedGoogle Scholar
- Keynes R, Krumlauf R: Hox genes and regionalization of the nervous system. Annu Rev Neurosci. 1994, 17: 109-32. 10.1146/annurev.ne.17.030194.000545.View ArticlePubMedGoogle Scholar
- Hunt P, Krumlauf R: Hox codes and positional specification in vertebrate embryonic axes. Annu Rev Cell Biol. 1992, 8: 227-56. 10.1146/annurev.cb.08.110192.001303.View ArticlePubMedGoogle Scholar
- Krumlauf R, Marshall H, Studer M, Nonchev S, Sham MH, Lumsden A: Hox homeobox genes and regionalisation of the nervous system. J Neurobiol. 1993, 24: 1328-40. 10.1002/neu.480241006.View ArticlePubMedGoogle Scholar
- McIntyre DC, Rakshit S, Yallowitz AR, Loken L, Jeannotte L, Capecchi MR, Wellik DM: Hox patterning of the vertebrate rib cage. Development. 2007, 134: 2981-9. 10.1242/dev.007567.View ArticlePubMedGoogle Scholar
- Pasqualetti M, Rijli FM: Homeobox gene mutations and brain-stem developmental disorders: learning from knockout mice. Curr Opin Neurol. 2001, 14: 177-84. 10.1097/00019052-200104000-00007.View ArticlePubMedGoogle Scholar
- Crawford M: Transformations in null mutants of Hox genes: do they represent intercalary regenerates?. Bioessays. 1995, 17: 1065-73. 10.1002/bies.950171211.View ArticlePubMedGoogle Scholar
- Bel-Vialar S, Itasaki N, Krumlauf R: Initiating Hox gene expression: in the early chick neural tube differential sensitivity to FGF and RA signaling subdivides the HoxB genes in two distinct groups. Development. 2002, 129: 5103-15.PubMedGoogle Scholar
- Oosterveen T, Niederreither K, Dolle P, Chambon P, Meijlink F, Deschamps J: Retinoids regulate the anterior expression boundaries of 5' Hoxb genes in posterior hindbrain. Embo J. 2003, 22: 262-9. 10.1093/emboj/cdg029.PubMed CentralView ArticlePubMedGoogle Scholar
- Oosterveen T, Meijlink F, Deschamps J: Expression of retinaldehyde dehydrogenase II and sequential activation of 5' Hoxb genes in the mouse caudal hindbrain. Gene Expr Patterns. 2004, 4: 243-7. 10.1016/j.modgep.2003.11.007.View ArticlePubMedGoogle Scholar
- Grapin-Botton A, Bonnin MA, McNaughton LA, Krumlauf R, Le Douarin NM: Plasticity of transposed rhombomeres: Hox gene induction is correlated with phenotypic modifications. Development. 1995, 121: 2707-21.PubMedGoogle Scholar
- Itasaki N, Sharpe J, Morrison A, Krumlauf R: Reprogramming Hox expression in the vertebrate hindbrain: influence of paraxial mesoderm and rhombomere transposition. Neuron. 1996, 16: 487-500. 10.1016/S0896-6273(00)80069-0.View ArticlePubMedGoogle Scholar
- Grapin-Botton A, Bonnin MA, Le Douarin NM: Hox gene induction in the neural tube depends on three parameters: competence, signal supply and paralogue group. Development. 1997, 124: 849-59.PubMedGoogle Scholar
- Woo K, Fraser SE: Specification of the zebrafish nervous system by nonaxial signals. Science. 1997, 277: 254-7. 10.1126/science.277.5323.254.View ArticlePubMedGoogle Scholar
- Niederreither K, McCaffery P, Drager UC, Chambon P, Dolle P: Restricted expression and retinoic acid-induced downregulation of the retinaldehyde dehydrogenase type 2 (RALDH-2) gene during mouse development. Mech Dev. 1997, 62: 67-78. 10.1016/S0925-4773(96)00653-3.View ArticlePubMedGoogle Scholar
- Swindell EC, Thaller C, Sockanathan S, Petkovich M, Jessell TM, Eichele G: Complementary domains of retinoic acid production and degradation in the early chick embryo. Dev Biol. 1999, 216: 282-96. 10.1006/dbio.1999.9487.View ArticlePubMedGoogle Scholar
- Berggren K, McCaffery P, Drager U, Forehand CJ: Differential distribution of retinoic acid synthesis in the chicken embryo as determined by immunolocalization of the retinoic acid synthetic enzyme, RALDH-2. Dev Biol. 1999, 210: 288-304. 10.1006/dbio.1999.9286.View ArticlePubMedGoogle Scholar
- Maden M, Sonneveld E, Saag van der PT, Gale E: The distribution of endogenous retinoic acid in the chick embryo: implications for developmental mechanisms. Development. 1998, 125: 4133-44.PubMedGoogle Scholar
- Chambers D, Wilson L, Maden M, Lumsden A: RALDH-independent generation of retinoic acid during vertebrate embryogenesis by CYP1B1. Development. 2007, 134: 1369-83. 10.1242/dev.02815.View ArticlePubMedGoogle Scholar
- Rossant J, Zirngibl R, Cado D, Shago M, Giguere V: Expression of a retinoic acid response element-hsplacZ transgene defines specific domains of transcriptional activity during mouse embryogenesis. Genes Dev. 1991, 5: 1333-44. 10.1101/gad.5.8.1333.View ArticlePubMedGoogle Scholar
- Niederreither K, Subbarayan V, Dolle P, Chambon P: Embryonic retinoic acid synthesis is essential for early mouse post-implantation development. Nat Genet. 1999, 21: 444-8. 10.1038/7788.View ArticlePubMedGoogle Scholar
- Niederreither K, Vermot J, Schuhbaur B, Chambon P, Dolle P: Retinoic acid synthesis and hindbrain patterning in the mouse embryo. Development. 2000, 127: 75-85.PubMedGoogle Scholar
- Maden M, Gale E, Kostetskii I, Zile M: Vitamin A-deficient quail embryos have half a hindbrain and other neural defects. Curr Biol. 1996, 6: 417-26. 10.1016/S0960-9822(02)00509-2.View ArticlePubMedGoogle Scholar
- Maden M: Heads or tails? Retinoic acid will decide. Bioessays. 1999, 21: 809-12. 10.1002/(SICI)1521-1878(199910)21:10<809::AID-BIES2>3.0.CO;2-0.View ArticlePubMedGoogle Scholar
- Gavalas A, Krumlauf R: Retinoid signalling and hindbrain patterning. Curr Opin Genet Dev. 2000, 10: 380-6. 10.1016/S0959-437X(00)00100-3.View ArticlePubMedGoogle Scholar
- Gould A, Itasaki N, Krumlauf R: Initiation of rhombomeric Hoxb4 expression requires induction by somites and a retinoid pathway. Neuron. 1998, 21: 39-51. 10.1016/S0896-6273(00)80513-9.View ArticlePubMedGoogle Scholar
- Subramanian V, Meyer BI, Gruss P: Disruption of the murine homeobox gene Cdx1 affects axial skeletal identities by altering the mesodermal expression domains of Hox genes. Cell. 1995, 83: 641-53. 10.1016/0092-8674(95)90104-3.View ArticlePubMedGoogle Scholar
- Isaacs HV, Pownall ME, Slack JM: Regulation of Hox gene expression and posterior development by the Xenopus caudal homologue Xcad3. Embo J. 1998, 17: 3413-27. 10.1093/emboj/17.12.3413.PubMed CentralView ArticlePubMedGoogle Scholar
- Shimizu T, Bae YK, Hibi M: Cdx-Hox code controls competence for responding to Fgfs and retinoic acid in zebrafish neural tissue. Development. 2006, 133: 4709-19. 10.1242/dev.02660.View ArticlePubMedGoogle Scholar
- Hamburger V, Hamilton HL: A series of normal stages in the development of the chick embryo. Journal of Morphology. 1951, 88: 49-92. 10.1002/jmor.1050880104.View ArticlePubMedGoogle Scholar
- Gaunt SJ, Strachan L: Forward spreading in the establishment of a vertebrate Hox expression boundary: the expression domain separates into anterior and posterior zones, and the spread occurs across implanted glass barriers. Dev Dyn. 1994, 199: 229-40.View ArticlePubMedGoogle Scholar
- Chapman SC, Collignon J, Schoenwolf GC, Lumsden A: Improved method for chick whole-embryo culture using a filter paper carrier. Dev Dyn. 2001, 220: 284-9. 10.1002/1097-0177(20010301)220:3<284::AID-DVDY1102>3.0.CO;2-5.View ArticlePubMedGoogle Scholar
- Timmer JR, Wang C, Niswander L: BMP signaling patterns the dorsal and intermediate neural tube via regulation of homeobox and helix-loop-helix transcription factors. Development. 2002, 129: 2459-72.PubMedGoogle Scholar
- Liem KF, Tremml G, Roelink H, Jessell TM: Dorsal differentiation of neural plate cells induced by BMP-mediated signals from epidermal ectoderm. Cell. 1995, 82: 969-79. 10.1016/0092-8674(95)90276-7.View ArticlePubMedGoogle Scholar
- Connolly D, McNaughton LA, Krumlauf R, Cooke J: Improved in vitro development of the chick embryo using roller-tube culture. Trends Genet. 1995, 11: 259-60. 10.1016/S0168-9525(00)89070-8.View ArticlePubMedGoogle Scholar
- Germain P, Iyer J, Zechel C, Gronemeyer H: Co-regulator recruitment and the mechanism of retinoic acid receptor synergy. Nature. 2002, 415: 187-92. 10.1038/415187a.View ArticlePubMedGoogle Scholar
- Wilson L, Gale E, Chambers D, Maden M: Retinoic acid and the control of dorsoventral patterning in the avian spinal cord. Dev Biol. 2004, 269: 433-46. 10.1016/j.ydbio.2004.01.034.View ArticlePubMedGoogle Scholar
- Wilson L, Maden M: The mechanisms of dorsoventral patterning in the vertebrate neural tube. Dev Biol. 2005, 282: 1-13. 10.1016/j.ydbio.2005.02.027.View ArticlePubMedGoogle Scholar
- Imamura T, Takase M, Nishihara A, Oeda E, Hanai J, Kawabata M, Miyazono K: Smad6 inhibits signalling by the TGF-beta superfamily. Nature. 1997, 389: 622-6. 10.1038/39355.View ArticlePubMedGoogle Scholar
- Doniach T: Planar and vertical induction of anteroposterior pattern during the development of the amphibian central nervous system. J Neurobiol. 1993, 24: 1256-75. 10.1002/neu.480241003.View ArticlePubMedGoogle Scholar
- Ruiz i Altaba A: Planar and vertical signals in the induction and patterning of the Xenopus nervous system. Development. 1992, 116: 67-80.PubMedGoogle Scholar
- Grunz H, Schuren C, Richter K: The role of vertical and planar signals during the early steps of neural induction. Int J Dev Biol. 1995, 39: 539-43.PubMedGoogle Scholar
- Nieuwkoop PD, Others: Activation and organization of the central nervous system in amphibians. J Exp Zool. 1952, 120: 1-108. 10.1002/jez.1401200102.View ArticleGoogle Scholar
- Doniach T, Phillips CR, Gerhart JC: Planar induction of anteroposterior pattern in the developing central nervous system of Xenopus laevis. Science. 1992, 257: 542-5. 10.1126/science.1636091.View ArticlePubMedGoogle Scholar
- Hooiveld MH, Morgan R, in der Rieden P, Houtzager E, Pannese M, Damen K, Boncinelli E, Durston AJ: Novel interactions between vertebrate Hox genes. Int J Dev Biol. 1999, 43: 665-74.PubMedGoogle Scholar
- Irving C, Nieto MA, DasGupta R, Charnay P, Wilkinson DG: Progressive spatial restriction of Sek-1 and Krox-20 gene expression during hindbrain segmentation. Dev Biol. 1996, 173: 26-38. 10.1006/dbio.1996.0004.View ArticlePubMedGoogle Scholar
- Barth KA, Kishimoto Y, Rohr KB, Seydler C, Schulte-Merker S, Wilson SW: Bmp activity establishes a gradient of positional information throughout the entire neural plate. Development. 1999, 126: 4977-87.PubMedGoogle Scholar
- Lebel M, Mo R, Shimamura K, Hui CC: Gli2 and Gli3 play distinct roles in the dorsoventral patterning of the mouse hindbrain. Dev Biol. 2006Google Scholar
- Liem KF, Tremml G, Jessell TM: A role for the roof plate and its resident TGFbeta-related proteins in neuronal patterning in the dorsal spinal cord. Cell. 1997, 91: 127-38. 10.1016/S0092-8674(01)80015-5.View ArticlePubMedGoogle Scholar
- Novitch BG, Wichterle H, Jessell TM, Sockanathan S: A requirement for retinoic acid-mediated transcriptional activation in ventral neural patterning and motor neuron specification. Neuron. 2003, 40: 81-95. 10.1016/j.neuron.2003.08.006.View ArticlePubMedGoogle Scholar
- Diez del Corral R, Olivera-Martinez I, Goriely A, Gale E, Maden M, Storey K: Opposing FGF and retinoid pathways control ventral neural pattern, neuronal differentiation, and segmentation during body axis extension. Neuron. 2003, 40: 65-79. 10.1016/S0896-6273(03)00565-8.View ArticlePubMedGoogle Scholar
- Vuligonda V, Standeven AM, Escobar M, Chandraratna RA: A new class of potent RAR antagonists: dihydroanthracenyl, benzochromenyl and benzothiochromenyl retinoids. Bioorg Med Chem Lett. 1999, 9: 743-8. 10.1016/S0960-894X(99)00077-3.View ArticlePubMedGoogle Scholar
- Liem KF, Jessell TM, Briscoe J: Regulation of the neural patterning activity of sonic hedgehog by secreted BMP inhibitors expressed by notochord and somites. Development. 2000, 127: 4855-66.PubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.