Boundary cells regulate a switch in the expression of FGF3 in hindbrain rhombomeres
- 4.2k Downloads
During formation of the vertebrate central nervous system, the hindbrain is organized into segmental units, called rhombomeres (r). These cell-lineage restricted segments are separated by a subpopulation of cells known as boundary cells. Boundary cells display distinct molecular and cellular properties such as an elongated shape, enriched extracellular matrix components and a reduced proliferation rate compared to intra-rhombomeric cells. However, little is known regarding their functions and the mechanisms that regulate their formation.
Hindbrain boundary cells express several signaling molecules, such as FGF3, which at earlier developmental stages is transiently expressed in specific rhombomeres. We show that chick embryos that lack boundary cells due to overexpression of truncated EphA4 receptor in the hindbrain have continued segmental expression of FGF3 at stages when it is normally restricted to hindbrain boundaries. Furthermore, surgical ablation of the boundary between r3 and r4, or blocking of the contact of r4 with boundary cells, results in sustained FGF3 expression in this segment.
These findings suggest that boundary cells are required for the downregulation of segmental FGF3, presumably mediated by a soluble factor(s) that emanates from boundaries. We propose that this new function of boundary cells enables a switch in gene expression that may be required for stage-specific functions of FGF3 in the developing hindbrain.
KeywordsBoundary Cell Segmental Expression Barrier Insertion EphA4 Signaling Hindbrain Boundary
During early stages of nervous system development, the hindbrain is subdivided into several segments, called rhombomeres (r). Individual rhombomeres are polyclonal compartments, defined both by cell-lineage-restriction and by segmental expression of transcription factors, such as Krox20, Kreisler and members of the Hox gene family. This network of genes regulates the formation of specific rhombomeres and their identities along the anterior-posterior (A-P) axis [1, 2, 3, 4]. The morphological and molecular segmentation of the hindbrain is essential for the establishment of specific patterns of neuronal differentiation and axon outgrowth and for the formation of distinct streams of migratory neural crest cells, implicated in the subsequent generation of neuronal networks and craniofacial structures (reviewed in [1, 2, 3, 4, 5, 6, 7, 8, 9]).
Concurrent with the establishment of hindbrain rhombomeres, a specialized population of boundary cells forms at their interfaces. A series of studies have characterized these cells, demonstrating that boundary cells have an elongated shape, increased extracellular spaces containing matrix components and that they display a reduced proliferation rate and interkinetic nuclear migration compared to intra-rhombomeric cells [10, 11, 12]. Much less is known regarding the mechanisms that regulate their formation or what the functions of boundary cells are during hindbrain development.
The signaling system of Eph tyrosine kinase receptors and their membrane-bound ephrin ligands have been shown to be required for boundary cell formation in zebrafish and Xenopus embryos [13, 14]. Eph receptors and ephrins are largely expressed in alternate rhombomeres such that they interact at their interfaces, and this restricts cells from mixing across hindbrain segments, possibly by mediating cell repulsion [15, 16, 17]. In addition, EphA4 was shown to sharpen hindbrain boundaries by regulating cell affinity within rhombomeres . Importantly, knocking down Eph/ephrin proteins or inhibition of their activation also leads to a decrease or loss of the expression of several boundary cell markers, such as pax6 and sema3Gb, in the zebrafish hindbrain [17, 18]. These results indicate a requirement for this signaling system in boundary cell formation, although it is not known if this is due to a direct role in cell specification or secondary to the increased mixing between segments. Whether Eph-ephrin interactions are also required for hindbrain boundary formation in higher vertebrates is not known.
Several soluble signals were shown to be localized to boundary cells of different species. For instance,Wnt1 and Wnt3a are expressed in zebrafish hindbrain boundaries [19, 20, 21], while fibroblast growth factor 3 (FGF3) and FGF19 are confined to mouse and chick hindbrain boundary cells, from around E10/Hamburger Hamilton stage 16, respectively [22, 23, 24, 25, 26]. Some modulators or inhibitors of signaling systems, such as the TGFβ inhibitor follistatin and the Notch modulator radical fringe also accumulate at hindbrain boundaries of chick, mouse or zebrafish embryos [26, 27, 28, 29, 30]. Little is known regarding the function of these factors at rhombomere interfaces. Interestingly, before boundary cells are formed, several of these signals, such as FGF3 and follistatin, have segmental expression within specific rhombomeres [23, 24, 26, 27, 28, 29]. The significance of these dynamic spatio-temporal expression patterns as well as the regulatory mechanisms by which these signals are turned on and off in different hindbrain regions are not clear.
In this study we set out to determine whether signaling by EphA4 is required for boundary cell formation in the chick hindbrain. We found that boundary cell markers and the formation of sharp interfaces were lost upon overexpression of dominant negative EphA4. Unexpectedly, we found that the segmental expression of the boundary cell marker FGF3 persisted, raising the possibility that lack of boundary cells may underlie the failure of rhombomeric FGF3 to become downregulated. The effect of ablation of boundary cells or blockage of the contact between a rhombomere and one of its boundaries confirmed this possibility. These findings suggest a new role for hindbrain boundary cells in inducing downregulation of the segmental expression of FGF3 in rhombomeres.
Fertile Loman chick eggs were incubated at 38°C until embryos reached the desired somite-stage (ss). Before performing experimental procedures, eggs were windowed and embryos were visualised by injecting black ink below the blastodisc. Following manipulations, embryos were incubated to the required stage, fixed in 4% paraformaldehyde (PFA) and stored at -20°C for further analysis.
pCAGGS-IRES-GFP (a gift from J. Briscoe) and pCAGGS-truncated EphA4-IRES-GFP  constructs were diluted in 10 mM Tris pH 8.5 to a working concentration of 2 μg/μl. Vectors were injected into the hindbrain lumen of different staged embryos by using a pulled glass capillary. Following injection, electrodes were placed at the right and left sides of the embryo at hindbrain levels to obtain unilateral transfection. Electroporation was performed using a BTX 3000 electroporator with four 45 millisecond pulses of 12–16 volts and pulse intervals of 300 milliseconds .
Whole-mount in situhybridization and immunohistochemistry
Whole-mount in situ hybridization was performed as described , using chick probes for hoxb1, FGF3 (EST clone 812g6, MRC Geneservice, UK), pax6 (a gift from J. Briscoe), Krox20 (a gift from P. Charnay), follistatin (Connoly et al., 1995), and NSCL1 (EST clone 474F24, MRC Geneservice, UK). Probes were labeled with digoxigenin (DIG)-UTP and detected using alkaline phosphatase-coupled antibody (1:2000, Roche, Basel, Switzerland) followed by NBT/BCIP (Roche, Basel, Switzerland) staining. Whole-mount and paraffin-section immunohistochemistry was carried out alone or following some in situ hybridizations. Briefly, embryos were incubated in PBS with 0.1% Tween20, 5% goat serum for 2 hours (hrs) prior to incubation for 16 hrs with the following antibodies: rabbit anti-GFP (1:400, Molecular Probes, CA USA), rabbit anti-EphA4 (1:250) , mouse anti-chondroitin sulphate proteoglycan (CSPG, 1:50, Sigma, MO USA), as well as recombinant human ephrin A5-Fc (5 μg/ml, R&D systems, MN USA). Following PBS washes, the following secondary antibodies were added: anti-rabbit or anti-mouse Alexa 488 and anti-rabbit Alexa 594 (all 1:400, Molecular Probes, CA USA) to be visualized under epi-fluorescent microscope, or anti-rabbit and anti-mouse-HRP (1:250, Sigma, MO USA) visualized with AEC substrate system (Lab Vision Corporation, CA USA).
For ablation of boundary cells, a rectangular cut was made with a pulled glass needle around the r3/r4 boundary region of 12–14 ss embryos and the tissue removed by aspiration. A silicon piece was cut to the precise size and inserted into the gap to prevent boundary regeneration. For barrier insertion, a transverse slit was made just posterior to r3/4 boundary or anterior to r4/5 boundary with a similar needle and the barrier was inserted into the slit. Barriers included either aluminum foil or a PCF membrane of 3 μm pores (Millipore, MA USA), both cut into adequate sizes. 1–2 days later, embryos were fixed in 4% PFA and prepared for in situ hybridization.
Expression of truncated EphA4 receptor disrupts hindbrain segment borders
To determine whether interfering with EphA4 signaling disrupts segmental gene expression, we examined the expression of the transcription factor Krox20 and its direct transcriptional target EphA4 , both normally sharply restricted to r3 and r5. In control electroporations, there was an expected r3 and r5 expression of Krox20 mRNA (Fig. 1E n = 12) and of EphA4 protein, which was detected with an antibody against the intracellular domain that binds the endogenous EphA4 but not the truncated ectopic protein (Fig. 1H n = 12). In contrast, there was a disruption to the formation of sharp borders on the electroporated side of embryos transfected with dnEphA4, with Krox20 and EphA4 expressing cells extending from their normal r3/r5 expression domains in the electroporated side into the adjacent r2, r4 or r6 territories (Fig. 1F, G, I, J and Fig. 2D; n = 12 for Krox20, n = 12 for EphA4). Notably, the data shows variability between embryos in the severity of the effects and in which segment(s) ectopic cells with Krox20 or EphA4 expression are observed in. This variability may be due to differences in electroporation efficiencies, intrinsic variability in cell mixing, and whether or not ectopic cells have switched their identity at the time of analysis . Together, these results suggest that EphA4 signaling is required to restrict cell mixing between r3/r5 and adjacent segments in the chick hindbrain, consistent with previous reports in zebrafish and Xenopus embryos .
Disruption of boundary cells in embryos expressing truncated EphA4
We next analysed the effect of dnEphA4 on the distribution of other boundary cell markers, the paired box gene pax6  and the bHLH transcription factor neural stem cell leukaemia 1 (NCSL1) . Both markers are expressed in a specific DV-restricted pattern along the hindbrain as well as in higher levels at rhombomere interfaces. We found that there is normal expression of both genes in control embryos (Fig. 2I, J n = 8 for each) and the non-electroporated side of embryos transfected with dnEphA4 (Fig. 2K, L right hand side n = 9 for each). In contrast, the elevated expression levels of pax6 (Fig. 2K n = 9) and of NSCL1 (Fig. 2L n = 9) at rhombomere borders decrease following overexpression of dnEphA4, whereas their DV-restricted expression within hindbrain segments appears unaffected. These findings suggest that disruption of Eph-ephrin signaling does not lead to a general alteration of gene expression throughout the hindbrain but rather to a localized effect at rhombomere interfaces. To further confirm this point, we evaluated the effect of dnEphA4 on the expression pattern of hoxb1, which is normally localized to r4. We found that in contrast to the sharp borders of r4 in controls (Fig. 2M n = 6), expression of dnEphA4 leads to a mild disruption in the formation of sharp domains of hoxb1 expression (Fig. 2N n = 6). Notably, as for Krox20 and EphA4, there was a disorganisation consistent with abnormal mixing rather than segmental changes to hoxb1 expression, suggesting that expression of dnEphA4 does not alter segmental identity.
Finally, the gross morphology of the manipulated hindbrain was examined by frontal paraffin sections of dnEphA4-expressing embryos. While the shape of rhombomere bulges separated by repeated boundary constrictions is evident in the intact side of the hindbrain (Fig. 2O, P right hand-side), a loss of this typical hindbrain morphology occurs on the dnEphA4 electroporated side (Fig. 2O, P left hand-side n = 5). Taken together, these results suggest that disruption of EphA4 signaling in the chick hindbrain leads to a decrease or loss of hindbrain boundary cells, in agreement with previous studies on zebrafish [17, 18].
The boundary cell marker FGF3shows persistent segmental expression in embryos transfected with truncated EphA4 receptor
Our previous results show a loss in the expression of several boundary cell markers upon expression of a truncated form of EphA4 (Fig. 2). However, examination of FGF3 expression in 28–36 ss embryos following transfection of dnEphA4 at 8 ss revealed a surprising result:FGF3 was expressed in even-numbered rhombomeres (Fig. 3G, I, n = 12) at stages when this segmental expression would normally be downregulated (Fig. 3A–D). Indeed, on the contra-lateral side of the dnEphA4-expressing hindbrains (Fig. 3G, I right hand side) as well as in control electroporated embryos (Fig. 3F, H, n = 10), FGF3 expression was confined to boundary cells, as expected at this embryonic stage. The effect of dnEphA4 misexpression was non-cell autonomous since following dorsal electroporation, the transfected cells were not overlapping with the more ventral segmental FGF3 expression domains (Fig. 3I, compare GFP labeled cells in green with FGF3 expression in red). This result rules out the possibility that FGF3 becomes upregulated in cells expressing dnEphA4. In contrast to these findings, transfection of dnEphA4 in older embryos at 22 ss, when FGF3 is about to become downregulated from r4 (Fig. 3D), did not result in any rhombomeric expression of FGF3. Instead, expression occurred only in the boundary cells in both GFP control and dnEphA4 expressing embryos (Fig. 3J, K, n = 7 for each), further eliminating the possibility that ectopic expression of truncated Eph directly induces FGF3 upregulation. Taken together, these findings suggest that overexpression of dnEphA4 at stages when FGF3 is segmentally expressed leads to failure of the normal downregulation of this aspect of FGF3 expression.
Downregulation of rhombomeric FGF3requires boundary cells
The finding that disruption of boundaries by expression of dnEphA4 leads to sustained segmental expression of FGF3 (Fig. 3F–I) argues against the alternative explanation that the effects of barrier insertion are due to disruption of signaling between segments rather than the absence of a normal border in-between. Nevertheless, to test this we analyzed whether inserting a barrier anterior to the r3/4 boundary affects FGF3 expression. We found that this did not result in sustained FGF3 expression in r4 or in ectopic expression in r3 (Fig. 5G, n = 5). This result further confirms that FGF3 expression in r4 (Fig. 5D, E) is not due to a non-specific effect of the operation procedure. Moreover, this indicates that it is not signaling from r3 to r4 that is necessary to downregulate FGF3 in r4, but rather the r3/4 boundary itself is required.
To further analyze whether signaling from boundaries in involved in downregulation of segmental FGF3 expression, we inserted a permeable membrane of 3 μm-diameter pore size, which allows diffusion of proteins, between the r3/4 boundary and r4. In these embryos, FGF3 was downregulated from this segment and localized to the boundaries, as in the contra-lateral side of the operated embryo (Fig. 5H, n = 11). This control further excludes the possibility that the sustained FGF3 expression in r4 observed upon the insertion of the non-permeable barrier (Fig. 5D–F) is a non-specific effect of the surgical manipulation. The differing effects of inserting a non-permeable (Fig. 5D–F) or a porous barrier (Fig. 5H) suggests that a diffusible factor(s) from r4 boundaries are involved in downregulation of FGF3 from this segment.
Finally, we analyzed whether boundary cells also regulate the expression of follistatin, which is initially expressed in a segmental pattern in several rhombomeres including r4, and later becomes restricted to boundary cells and a DV-restricted stripe . We found that prevention of the contact between r4 and its anterior boundary at 12 ss embryos resulted in higher levels of follistain expression in r4 in 35 ss embryos (Fig. 5I n = 8). Due to the normal DV-restricted expression of follistatin, this result is less clear-cut than for FGF3, but nevertheless argues that boundaries regulate the expression of multiple genes within r4.
FGF3 has dynamic expression in the chick embryo hindbrain in which it is first restricted to specific rhombomeres and later becomes downregulated from these segments and upregulated at rhombomere boundaries. This study shows that embryos that lack hindbrain boundary cells, due to either surgical ablation or overexpression of a truncated EphA4 receptor, maintain segmental expression of FGF3 at stages when it is normally restricted to hindbrain boundaries. Moreover, abnormal maintenance of FGF3 expression in r4 occurs upon insertion of a non-permeable barrier between this rhombomere and either of its boundaries, suggesting that the downregularion of segmental FGF3 is induced due to the secretion of signals from boundary cells. These results suggest a novel function of boundary cells in promoting the downregulation of segmental expression of FGF3 in hindbrain segments.
Eph receptors are required for boundary formation in the chick hindbrain
The inhibition of cell mixing between adjacent rhombomeres is crucial to establish sharp domains of segmental gene expression in the hindbrain. Signaling by Eph receptors and their ephrin ligands have a key role in the sharpening of boundaries in the zebrafish hindbrain, by interactions across boundaries that restrict cell mixing between segments as well as by regulating cell affinity within rhombomeres [17, 18]. Our results in the chick embryo, in which expression of dnEphA4 leads to mixing of cells with r3/r5 identities into even numbered territories, suggest that Eph-ephrin signaling is required for hindbrain boundary sharpening in a higher vertebrate.
Analysis of multiple markers of hindbrain boundaries and of hindbrain morphology shows that formation of boundary cells is severely disrupted in the chick hindbrain following expression of truncated EphA4. These results are in agreement with previous reports in zebrafish where inhibition or knockdown of EphA4 function leads to loss of pax6 or sema3Gb expression in boundaries [17, 18]. It is possible that Eph-ephrin signaling at rhombomere interfaces specifies boundary cells directly, or by inhibiting cell mixing, provides a stable interface between rhombomeres that is essential for boundary cells to form.
Potential significance of switch in FGF3gene expression
Our finding that signaling from hindbrain boundaries promotes the downregulation of the segmental expression of FGF3 raises the question of the potential role of this interaction. The dynamic regulation of FGF3 in the hindbrain may be an example of the widespread phenomenon of the redeployment of the same signals at sequential stages of development. FGF3 and several other secreted factors, such as FGF19 and follistatin, initially have rhombomeric expression that then switches to boundary-restricted expression in the chick and mouse embryo hindbrain [22, 23, 24, 25, 27, 28, 29, 40]. The early segmental expression of FGF3 is required for the induction and patterning of the otic vesicle, which develops from a placode adjacent to the hindbrain [41, 42, 43]. In addition, rhombomeric-derived FGFs are required for hindbrain patterning, such as segmental expression of Krox20 and Kreisler/MafB in specific segments [44, 45, 46]. Recently, we have found that the segmental expression of FGF3 is enabled by follistatin that is expressed in the same segments and blocks BMP signaling that would otherwise inhibit the upregulation of FGF3 . Knockdown of follistatin leads to a lack of segmental FGF3 expression and disruption to hindbrain patterning .
The downregulation of segmental FGF3 expression could simply reflect that its initial roles in hindbrain segmentation and inner ear induction have been fulfilled, so continued expression is not required. However, our finding that boundary signals promote the downregulation of FGF3 suggests a more active requirement. An attractive possibility is that continued segmental expression of FGF3 would interfere with the roles of the subsequent boundary-restricted expression. Currently, it is not possible to test this since the role of FGF3 expression at hindbrain boundaries is not known. The expression of a number of signaling molecules at hindbrain boundaries in chick [22, 23, 24, 25, 26, 27, 28, 29, 40, 47] is suggestive of roles as signaling centers that pattern adjacent rhombomeres. Such a role could be analogous to the midbrain-hindbrain boundary that acts as an organizing centre to pattern cell fate and neural differentiation in the midbrain and anterior hindbrain through the secretion of FGF8 [37, 39, 48, 49, 50]. However, it is not clear whether hindbrain boundaries organize neuronal patterns within rhombomeres in the chick since, unlike the situation in zebrafish hindbrain , there is no overt organization of repetitive neurogenesis or of mature neurons and glia along the AP axis within rhombomeres [1, 52]. Furthermore, there is normal formation of neuronal nuclei following retinoic acid treatment, surgical ablation or genetic alterations that disrupt hindbrain boundaries in chick or mouse embryos [36, 53, 54, 55]. Nevertheless, it is possible that FGFs expressed at hindbrain boundaries have roles in other aspects of tissue organization, such as axon pathfinding or formation of nerve exit points [36, 52, 53].
What is the signal that downregulates FGF3?
An important question raised by our results is the identity of boundary signals that induce the downregulation of segmental FGF3 in the hindbrain. One possibility is that segmental – but not boundary – FGF3 expression is itself downregulated by FGF signaling from boundaries. However, on the contrary we found that addition of exogenous FGF3 increased the level of segmental FGF3 expression whereas blocking of FGF receptors inhibited FGF3 expression . These results suggest that an autoregulatory positive feedback loop regulates FGF3 expression in the hindbrain, arguing against the possibility that FGF3 from boundaries downregulates its own expression in segment bodies. FGF3 downregulation may be mediated by the loss of follistatin from segment bodies, since both genes overlap in the hindbrain and we have recently found that FGF3 requires follistatin in order to be expressed in rhombomeres . Our present data are consistent with the possibility that a boundary signal(s) downregulates segmental follistatin expression, in turn leading to downregulation of FGF3. Further investigations are required to elucidate which secreted boundary signals are inducing the downregulation of FGF3 and follistatin from rhombomeres.
In conclusion, we demonstrate that interference with EphA4 signaling in the chick hindbrain prevents the formation of sharp rhombomere interfaces. In addition, boundary cell markers are lost upon overexpression of dominant negative EphA4. However, the segmental expression of the boundary cell marker FGF3 persists. Similar sustained expression of FGF3 in rhombomeres occurs upon ablation of boundary cells or blockage of the contact between a rhombomere and its boundaries. Together, we suggest that hindbrain boundary cells are required for the switching-off of rhombomeric FGF3, presumably mediated by a soluble factor(s) emanating from the boundaries. These findings imply for a new role of boundary cells in the downregulation of genes expressed at hindbrain segments. This boundary-cell activity may be required for stage-specific function of segmental genes in the developing hindbrain.
We thank Dr. J. Briscoe and Dr. P. Charnay for plasmids and probes. We also thank M. Hazenfratz for summarizing parts of the data. This study was supported by THE ISRAEL SCIENCE FOUNDATION and by THE MEDICAL RESEARCH COUNCIL, UK. DSD was a recipient of an EMBO post-doctoral fellowship.
- 40.Powles N, Marshall H, Economou A, Chiang C, Murakami A, Dickson C, Krumlauf R, Maconochie M: Regulatory analysis of the mouse Fgf3 gene: control of embryonic expression patterns and dependence upon sonic hedgehog (Shh) signalling. Dev Dyn. 2004, 230: 44-56. 10.1002/dvdy.20028.CrossRefPubMedGoogle 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.