Dbx1b defines the dorsal habenular progenitor domain in the zebrafish epithalamus

Background The conserved habenular nuclei function as a relay system connecting the forebrain with the brain stem. They play crucial roles in various cognitive behaviors by modulating cholinergic, dopaminergic and serotonergic activities. Despite the renewed interest in this conserved forebrain region because of its importance in regulating aversion and reward behaviors, the formation of the habenular nuclei during embryogenesis is poorly understood due to their small size and deep location in the brain, as well as the lack of known markers for habenular progenitors. In zebrafish, the bilateral habenular nuclei are subdivided into dorsal and ventral compartments, are particularly large and found on the dorsal surface of the brain, which facilitates the study of their development. Results Here we examine the expression of a homeodomain transcription factor, dbx1b, and its potential to serve as an early molecular marker of dorsal habenular progenitors. Detailed spatiotemporal expression profiles demonstrate that the expression domain of dbx1b correlates with the presumptive habenular region, and dbx1b-expressing cells are proliferative along the ventricle. A lineage-tracing experiment using the Cre-lox system confirms that all or almost all dorsal habenular neurons are derived from dbx1b-expressing cells. In addition, mutant analysis and pharmacological treatments demonstrate that both initiation and maintenance of dbx1b expression requires precise regulation by fibroblast growth factor (FGF) signaling. Conclusions We provide clear evidence in support of dbx1b marking the progenitor populations that give rise to the dorsal habenulae. In addition, the expression of dbx1b in the dorsal diencephalon is tightly controlled by FGF signaling.


Background
The habenular nuclei (habenulae) develop in the dorsal diencephalon of vertebrates. These bilaterally paired nuclei receive inputs from the limbic system and basal ganglia and send outputs to dopaminergic and serotonergic centers. Despite their small size, these nuclei play crucial roles in regulating aversion and reward behaviors [1]. Moreover, the fact that the habenulae are a nexus for monoamine circuits highlights the importance of this brain region for studies of neuromodulation and multicircuit integration.
The habenular nuclei can be divided into two functionally distinct subnuclei alternately referred to as medial and lateral (mammals) or dorsal and ventral (zebrafish) subnuclei.
Recent study of the habenulae has largely focused on the ventral habenulae, but interest in the dorsal habenulae is increasing [2]. Alterations in dorsal habenular structure and function during development have been linked to dopamine receptor expression as well as impulsivity, attention, aversion and spatial memory endophenotypes in rodents [3,4]. In zebrafish, the dorsal habenulae regulate fear and aversion [5,6]. Together these studies implicate the dorsal habenulae as a possible mediator of attention deficit hyperactivity disorder (ADHD), depression and anxiety [2]. Interestingly, the dorsal habenulae are also sites of intense acetylcholine receptor and transporter expression and control nicotine intake [7][8][9]. This raises the possibility of targeting the habenulae as a therapeutic intervention to nicotine addiction. Beyond translational research, the zebrafish habenulae also serve as an excellent model to study the basic mechanisms underlying the development of left-right brain asymmetry. Unlike the mammalian habenulae, teleost dorsal habenulae are robustly asymmetric in anatomy, gene expression and functional connectivity [10].
Interest in how the dorsal habenulae integrate into cholinergic and monoaminergic circuitry has put pressure on researchers to understand dorsal habenular development. Indeed, Beretta et al. [11] have shown that dorsal and ventral habenular neurons arise from distinct progenitor populations. While progress has been made in describing habenular neurogenesis, differentiation and elaboration of processes, there are no known markers for habenular progenitors [12][13][14][15][16]. Therefore, finding marker genes that label dorsal habenular progenitors will be fundamental to studying how the diverse set of dorsal habenular neurons are generated and integrated into neural circuits underpinning aversive behavior as well as pathological addictive and depressive behaviors.
The dbx homeodomain transcription factors play a central role in regulating progenitor status in several regions in the developing central nervous system, including the spinal cord [17,18]. However, the upstream regulatory pathways that regulate dbx gene-family expression are not known. Here we report that in zebrafish, dbx1b is expressed in the dorsal diencephalon where it marks dorsal habenular progenitors, and further, that dorsal diencephalic expression of dbx1b is controlled by fibroblast growth factor (FGF) signaling.

Results and discussion
Dbx1b is expressed in the presumptive habenulae During our ongoing efforts to characterize transcription factors (TFs) that are expressed in the dorsal diencephalic region between 24 and 48 hours post-fertilization (hpf), we focused on a family of homeodomain-containing TFs encoded by the dbx genes because of their known roles in neural progenitors. There are three dbx genes in the zebrafish genome, and we carefully examined the expression pattern of the two dbx1 paralogs, dbx1a and dbx1b. We excluded dbx2 from our study because its expression has been detailed previously [19]. At 28 hpf, dbx1a and dbx1b showed similar yet distinct expression patterns (Additional file 1). As shown previously [20], dbx1a was expressed in sharply restricted domains in the diencephalon with prominent expression in the prethalamus and thalamus (Additional file 1: panel A). Expression of dbx1a and dbx1b was similar in the prethalamic region, but in the thalamic region dbx1b was expressed at a much lower level than dbx1a (Additional file 1; panel C). A more striking difference between the patterns of these two paralogs was the expression of dbx1b in the dorsal diencephalon, where dbx1a was completely absent (compare arrowheads in Additional file 1). Expression of dbx1b was excluded from the otx5-positive pineal complex, the other major structure of the dorsal diencephalon (Additional file 1, panel C). These data suggested that dbx1b could be an early molecular marker for the presumptive habenulae.
Next, we closely examined the expression of dbx1b at different developmental stages (Figure 1). At 22 hpf, expression of dbx1b was not yet present in the presumptive habenular region, although the prethalamic and midbrain regions showed strong expression ( Figure 1A,B). By 24 hpf, habenular dbx1b expression appeared ( Figure 1C,D) and was maintained through at least 96 hpf ( Figure 1E-J and data not shown). Moreover, dbx1b expression remained highest adjacent to the third ventricle of the brain, and was absent from regions distal to the ventricle by 48 hpf (Figure 1H and 1J; dorsal views). Because neuronal progenitors are often found in regions adjacent to the ventricle in the developing brain, these data led us to speculate that the dbx1b expression pattern includes the progenitors of habenular neurons.

Dbx1b labels dorsal habenular progenitors
To support the hypothesis that dbx1b-expressing cells represent habenular progenitors, we examined if these cells are proliferative. Indeed, as shown by phospho-histone H3 staining at 32 hpf, the dbx1b-positive cells close to the ventricular surface were proliferative, which was consistent with progenitor cell identity (Figure 2A-D). To see if dbx1b expression is restricted to progenitors we compared the expression of previously described precursor and neuronal habenular markers. Cxcr4b has been proposed as a marker of progenitors as well as post-mitotic habenular precursors [12]. At 36 hpf, a subset of dbx1b-expressing cells co-expressed cxcr4b. Specifically, cxcr4b expression was restricted to the dorsal half of the dbx1b expression domain leaving a ventral region of dbx1b-only expression along the ventricle ( Figure 2E-F). At 48 hpf, Elavl3 (also known as HuC) marks post-mitotic neurons. Expression of Elavl3 and dbx1b was more distinct than cxcr4b and dbx1b, with Elavl3 expressed dorsally and laterally while dbx1b was present medially and ventrally along the ventricle ( Figure 2G-H). Kctd12.1 and Kctd12.2 mark different populations of differentiated habenular neurons. At 72 hpf, neither showed any overlap with dbx1b expression ( Figure 2I-L). Similar results were observed with another habenular differentiation marker, pou4f1 (or brn3a) (data not shown). Together these results indicated that dbx1b expression is maintained in a proliferative population of habenular progenitor cells that reside along the ventricle. Dbx1b expression was absent in fully differentiated neurons, suggesting that dbx1b is downregulated in mature habenular neurons. Thus, we conclude that dbx1b can serve as an early marker to identify habenular progenitors.
To substantiate our conclusion that dbx1b is expressed in habenular progenitors, we performed a lineage-tracing  [22], almost all Elavl3-expressing habenular neurons were co-labeled with GFP ( Figure 3A-C). Indeed, the only domain expressing GFP but not Elavl3 was found along the ventricle, in an area that coincides with dbx1b transcription (Figure 3D-G). To determine if dbx1bpositive cells give rise to both dorsal and ventral habenular neurons, we compared the lineage-labeled domain with cadps2, a dorsal habenular marker [23]. Almost all lineagelabeled cells were cadps2-positive ( Figure 3H-J). The only dbx1b-derived cells (GFP-positive) outside the cadps2 domain were medially located and along the ventricle; none were found laterally in the region of the ventral habenulae. Moreover, we also compared the lineage-labeled domain with the expression domain of aoc1, a known ventral habenular marker [24], and found that these two domains abutted but clearly did not overlap ( Figure 3K-M). In conclusion, these results strongly suggested that most if not all post-mitotic dorsal habenular neurons are derived from progenitor cells that express dbx1b prior to their differentiation, confirming that dbx1b is a marker of early dorsal habenular progenitors.
FGF signaling is required for proper development of the dorsal habenulae FGF signaling has been shown to play critical roles during the development of the zebrafish dorsal diencephalon, particularly in pineal complex specification and parapineal migration [25,26]. It has been suggested that the development of both the left and right habenulae requires FGF signaling, as shown by the reduced expression of habenular differentiation markers (Kctd12.1, pou4f1) in FGF mutants [26]. However, how the loss of FGF signaling impacts habenular development remains unclear. We found that in fgf8a mutants, in which brain patterning appeared to be normal (Additional file 2), the expression of dbx1b was completely lost at 24 hpf ( Figure 4A-D) and this loss of expression persisted at later stages. The latter observation suggested that the loss of FGFdependent dbx1b is not a result of developmental delay. Therefore, FGF signaling is absolutely required for the initiation of dbx1b expression.
To explore the regulatory relationship between FGF signaling and dbx1b in more detail, we investigated if FGF signaling is continuously required to maintain dbx1bpositive habenular progenitors. We took advantage of the FGF receptor antagonist, SU5402, to block FGF signaling in a temporally controlled manner. Since dbx1b expression was first detectable at 24 hpf, the embryos were treated with SU5402 for 8 hours, from either 28 to 36 hpf or 48 to 56 hpf. The expression of dbx1b was analyzed at the end of the treatments as well as after 12 hours of recovery post-treatment. As shown in Figure 5, both treatment regimes abrogated dbx1b expression by the end of the treatment window, yet dbx1b expression began to return after 12 hours of recovery ( Figure 5B,D and 4F,H). This result suggests that FGF signaling is required for not only the initiation, but also the maintenance of dbx1b expression. Moreover, at least some neuronal progenitors in the dorsal diencephalon remain FGF-responsive and capable of reactivating dbx1b expression upon exposure to FGF signal, even when they were previously deprived of FGF signaling.

Conclusions
This report describes the expression of dbx1b, which we believe is the first reported marker of the neuronal progenitors that give rise to the dorsal habenulae. In addition, we found that FGF signaling controls the expression of dbx1b in the dorsal diencephalon. Together with other existing genetic tools, including various dbx1b BAC transgenic lines, our discovery of dbx1b as a habenular progenitor marker will not only allow for more detailed and nuanced investigation of dorsal habenular development, but also provide an exciting way forward to study proliferation, specification and circuit formation of the diverse neuronal populations in the habenular nuclei, and how these processes influence developmental and adult habenulae-mediated behaviors.

Zebrafish maintenance and strains
Zebrafish were raised at 28.5°C on a 14/10 hour light/ dark cycle and staged according to hours post-fertilization. The following fish lines were used: the wildtype strain AB* [27], TgBAC[dbx1b:Cre-mCherry] nns13a [21] and Tg[−10actb2:LOXP-mCherry-LOXP-nlsEGFP] pd31 [22]. All experiments were approved by the Vanderbilt University's Institutional Animal Care and Use Committee (IACUC) and Office of Animal Welfare, and performed according to national regulatory standards.
Double fluorescent in situ hybridization was performed with the following modifications to the above colorometric in situ protocol: after hybridization of DIG and fluorescein-labeled probes, anti-DIG antibody was applied (1:5,000, Roche) overnight at 4°C. The following day, embryos were washed 4 × 20 minutes in PBS with Triton (PBSTr) and 3 × 5 minutes in Fast Blue Buffer and developed in Fast Blue Substrate diluted in Fast Blue Buffer. After color development, embryos were washed 2 × 10 minutes in PBSTr. The alkaline phosphatase was acid inactivated by a 10-minute wash in 0.1 M glycine HCl, pH 2.0. After 2 × 10 minute PBSTr washes, embryos were incubated in anti-fluorescein antibody (1:1,000, Roche) overnight at 4°C. The following day, color was developed in Fast Red substrate as in Doll et al. [13]. cxcr4b [31] probe was generated with EcoRV and SP6 RNA polymerase.

Inhibitor treatments
For whole-mount in situ hybridizations and antibody labeling, embryos were incubated in their chorions in 12 μM (for complete receptor inhibition; [27]) of SU5402 (Tocris Bioscience, Bristol, UK) dissolved in 0.3% dimethyl sulfoxide (DMSO) in egg water supplemented with 0.003% N-phenylthiourea (PTU; Sigma-Aldrich) to prevent melanin formation. Control embryos were treated with 0.3% DMSO in parallel with their SU5402-treated siblings. Embryos were either fixed immediately following treatment or SU5402/DMSO was washed off with 5 × 5 minute egg water before being returned to egg water with PTU to develop to the desired stage for fixation.

Imaging
All samples were cleared in a glycerol series (50%, 100%). Colorometric in situ images were captured on a Leica DM6000 B compound microscope (Leica Microsystems, Buffalo Grove, IL, USA) under a 20× air objective in bright field conditions. Fluorescent images were captured on a PerkinElmer spinning disk confocal microscope (PerkinElmer, Waltham, MA, USA) or a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss Microscopy, Oberkochen, Germany) with a 40× oil-immersion objective and analyzed with Volocity software (Improvision/PerkinElmer, Waltham, MA, USA).

Additional files
Additional file 1: Dbx1b is expressed in the dorsal diencephalon. Lateral and dorsal views of a 28 hpf wildtype embryo. (A and B) In situ hybridization for dbx1a (blue) revealed several expression domains, including the olfactory bulb (Ob), prethalamus (pT), thalamus (Th), and midbrain (Mb) throughout the brain, but no expression in the dorsal diencephalon (arrow heads). (C and D) dbx1b transcript (blue) was expressed in a similar pattern but with greatly reduced expression in thalamus and robust expression in the dorsal diencephalon and olfactory bulb. otx5 (red) marks the pineal complex (P), a component of the dorsal diencephalon. Scale bars are 10 μm.