In Vivo Monitoring of Fabp7 Expression in Transgenic Zebrafish

Abstract

In zebrafish, like in mammals, radial glial cells (RGCs) can act as neural progenitors during development and regeneration in adults. However, the heterogeneity of glia subpopulations entails the need for different specific markers of zebrafish glia. Currently, fluorescent protein expression mediated by a regulatory element from the glial fibrillary acidic protein (gfap) gene is used as a prominent glia reporter. We now expand this tool by demonstrating that a regulatory element from the mouse Fatty acid binding protein 7 (Fabp7) gene drives reliable expression in fabp7-expressing zebrafish glial cells. By using three different Fabp7 regulatory element-mediated fluorescent protein reporter strains, we reveal in double transgenic zebrafish that progenitor cells expressing fluorescent proteins driven by the Fabp7 regulatory element give rise to radial glia, oligodendrocyte progenitors, and some neuronal precursors. Furthermore, Bergmann glia represent the almost only glial population of the zebrafish cerebellum (besides a few oligodendrocytes), and the radial glia also remain in the mature cerebellum. Fabp7 regulatory element-mediated reporter protein expression in Bergmann glia progenitors suggests their origin from the ventral cerebellar proliferation zone, the ventricular zone, but not from the dorsally positioned upper rhombic lip. These new Fabp7 reporters will be valuable for functional studies during development and regeneration.

1. Introduction

Radial glial cells (RGCs) play essential roles as a source for neural progenitors and glia subpopulations, as a scaffold for cell migration, and as a reservoir during regeneration. RGCs are derived from neuroepithelial cells (NECs) [1] and have been shown in mice and zebrafish to express several evolutionarily conserved RGC marker genes including Glial fibrillary acidic protein (Gfap), S100 calcium binding protein B (S100b), the estrogen-synthesizing enzyme Aromatase B, and Brain Lipid Binding Protein (Blbp)—also named Fatty acid binding protein 7 (Fabp7) [2,3,4,5]. Differing from mammalians, the production of new neurons from radial glial cells is maintained in multiple areas of the central nervous system (CNS) of zebrafish until and during adulthood, under both physiological and regenerative conditions [3].

Fatty acid binding protein 7 is involved in lipid metabolism and membrane synthesis, strongly binding to omega-3 polyunsaturated fatty acids, such as docosahexaenoic acid (DHA) [2], is expressed in RGCs at very early stages, and is co-expressed with well-established radial glia markers such as gfap and aromatase B [2,3,4,5].

The expression of Fabp7 in teleosts (zebrafish) differs from other glial markers, such as the intermediate filament Glial fibrillary acidic protein (Gfap), since Gfap expression is activated only in RGCs, while Fabp7 is already expressed in the NECs, the RGC progenitors [6]. In some zebrafish studies, though, Gfap cells colocalizing with Nestin (another neuroepithelial cell marker) were detected, for example, in the embryonic spinal cord [7]. In fact, the transition of neuroepithelial cells to radial glial cells to mature glia is poorly defined based on marker gene expression in species different than mice. Therefore, further markers and in vivo tools to study radial glia are needed to better understand the developmental time course of glia in zebrafish.

Currently, the only regulatory elements for driving expression in radial glial cells in zebrafish are the enhancer of the glial fibrillary acidic protein (gfap), as the most commonly used glial specific enhancer, and aromatase-B, used only in few studies so far, which drive expression in fabp7 expressing RGCs but not in NECs [8]. gfap mRNA and Gfap protein expression show similar patterns, but it has also been reported that Gfap expression could be regulated at the level of translation, and cells producing gfap mRNA may not translate it, which could lead to misinterpretations about the glial identity of a cell [9,10].

The RGC marker Fabp7 is involved in brain development and adult neurogenesis [11], as well as in the maintenance of neural stem cell fate, acting as a downstream target of Notch signaling and Pax6 [12,13]. The Fabp7 protein is encoded by the gene brain-type fatty acid binding protein 7—Fabp7 [12]. Due to the genomic duplication that occurred in teleost fish [14], the fabp7 gene in zebrafish was duplicated to fabp7a and fabp7b. According to the levels of mRNA expression, fabp7a was determined to be the main isoform present in the adult zebrafish brain [2,11]. The specificity of the antibody against Fabp7 and its distribution in the whole brain of adult zebrafish has already been described in detail [2]. Thus, this RGC marker has been commonly used in multiple zebrafish studies about glio- and neurogenesis [4,7]. As informative as such immunohistochemical studies are, they are limited to fixed tissue and provide static information but do not allow for in vivo observations of fabp7-expressing cells and their descendants for assessing their behavior in longitudinal studies.

We have tested a Fabp7 regulatory element from a mouse, which mediates expression in radial glia, immature astrocytes, and Bergmann glia [13,15]. This regulatory element was combined with an adenoviral basal promoter (E1b) [16] in zebrafish. Three different transgenic reporter lines for detecting cells expressing fluorescent proteins under the control of the mouse Fabp7 regulatory element in zebrafish were created, and their prominent fluorescent reporter protein expression confined to the central nervous system was compared to endogenous Fabp7 and early gliogenic and neurogenic markers. These lines will be powerful genetic tools for dissection brain development, glia differentiation, neurogenesis, and regeneration.

2. Materials and Methods

Fish husbandry: All animals were raised and kept in the zebrafish facility in accordance with established guidelines adhering to the regulations of the local government [17,18]. AB wild-type or brass pigmentation mutant zebrafish were used for mating. No selection criteria were used to allocate zebrafish of both sexes to any experimental group. Embryos and larvae were maintained in zebrafish 30% Danieau rearing medium [100% Danieau: 58 mM NaCl, 0.7 mM KCl, 0.4 mM MgSO₄, 0.6 mM Ca(NO₃)₂, and 5 mM HEPES (N-2-hydroxyethylpiperazine-N-2-ethane sulfonic acid), pH7.2]. All imaging and analysis of larvae were performed at 2 and 5 days post-fertilization (dpf). Adults were maintained in fish tanks according to standard protocols at 28 °C, under light–dark conditions (simulating day-night cycle), and constant running water exchange. Stable transgenic reporter lines used in this work include Tg(Mmu.Fabp7-E1B:MScarlet-MYC)^bz23, Tg(Mmu.Fabp7:E1B:mClover)^bz25 and Tg(Mmu.Fabp7-E1B:FMA-mseCFP-2A-H2B-mseCFP)^bz24, approved by the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit LAVES (AZ 33.19-42502-04-17/2693), Tg(gfap:EGFP)^mi2001 [19], Tg(olig2:EGFP)^vu12 [20], Tg(nkx2.2:GFP) [21], Tg(Xla.Tubb:DsRed, previously NBT-dsRED) [22], Tg(zic4:Gal4TA4, UAS:mCherry)^hzm5, Tg(4xUAS:GFP)^hzm3 both by [23], Tg(ptf1a:EGFP) [24], and Tg(atoh1a:GAL4TA4)^hzm2 [25]. For simplicity reasons double transgenic Tg(atoh1a: GAL4TA4)^hzm2 × Tg(4xUAS:GFP)^hzm3 larvae were named Tg(atoh1a:GFP). All efforts were made to use only the minimum number of experimental animals necessary to obtain reliable scientific data.

Generation of transgenic lines: Previously a regulatory element from the mouse Fabp7 gene driving expression in radial glial cells has been identified and analyzed in detail [15,26]. This sequence of the mouse Fabp7 regulatory element followed by the E1b basal promoter is shown in Supplementary Figure S1. mClover3 [27] and mScarlet [28] were chosen as fluorescent reporters as they are among the brightest fluorescent proteins available in the green and red emission range, mseCFP is a derivative of ECFP [29], but it can be excited as well in the UV range (405 nm) offering co-expression analysis with a greater range of fluorescent reporter proteins. Zebrafish embryos were injected at the one-cell stage with plasmids containing a Tol2-transgene cassette and Tol2-encoding mRNA (25 ng/l each, volume: 1.5 nl). After raising fluorescent embryos to adulthood, germline transmission of the transgene was tested by individual crosses against wild-type fish. Transgenic founders were raised and used to establish stable transgenic strains [30]. Reporter fish with fluorescent protein expression mediated by the Fabp7 regulatory element are currently maintained in the F2 and F3 generations, respectively.

Immunohistochemistry: Larvae were processed for immunostaining as whole mount isolated brains. Brain isolation was performed after prefixing in 4% paraformaldehyde (PFA) for 20–30 min on ice. After washing with PBS, antigen retrieval with 10 mM citrate buffer was achieved by heating samples in a water bath until boiling. After cooling to room temperature, the brains were rinsed 5 times in PBS-T (PBS with 1% Triton X-100) for 5 min each, incubated in 100% precooled acetone for 15 min at −20 °C to improve tissue permeabilization, followed by additional 5 PBS-T washing steps. Nonspecific protein binding sites were blocked in 5% normal goat serum in PBS-DT-I (PBS, 1% bovine serum albumin (BSA), 1% DMSO, 1% Triton X-100) for 1 h at room temperature followed by primary antibody incubation in PBS-DT-II (PBS, 1% BSA, 1% DMSO, 0,3% Triton X-100) overnight at 4 °C and additional 3 h at room temperature. The primary antibodies are as follows: rabbit anti-FABP7 (IF 1:500, commercially named as anti-BLBP, Millipore ABN14, Burlington, MA, USA), rat anti-RFP—for detecting mScarlet fluorescent protein—(IF 1:1000, Chromotek 5F8-100, Planegg, Germany), and chicken anti-GFP (IF 1:1000, Aves Labs GFP-1010, Davis, CA, USA). Subsequently, brains were washed 5 times in PBS-T for 5 min each at room temperature with constant agitation and incubated with the secondary antibody diluted in PBS-DT-II overnight at 4 °C. Secondary antibodies: Alexa goat anti-rabbit 488 (IF 1:1000, Thermo Fisher A21206, Waltham, MA, USA), Alexa goat anti-rat 546 (IF 1:1000, Thermo Fisher A11081, Waltham, MA, USA), donkey anti-chicken (IF 1:1000, Jackson ImmunoResearch, 703-545-155, West Grove, PA, USA). Finally, brains were thoroughly rinsed in PBS-T and PBS before imaging.

Imaging and data analysis: For in vivo imaging, larvae were anesthetized with 0.015–0.02% Tricaine (Sigma Aldrich, St. Louis, MO, USA) dissolved in 1% low melting agarose/Danieau solution. For fixed and immunostained samples, the brains were embedded in 1% low melting agarose/PBS. Images of the larval zebrafish (at 2 and 5 dpf) were acquired using an SP8 laser scanning confocal microscope (Leica Microsystems, Wetzlar, Germany) using a 40× objective. Images were processed with the Leica LasX (LAS X 3.5.7.23225) and FIJI imaging (ImageJ-win64) and analysis software. For quantitative analysis, semi-automated counting of cells was performed with the cell counter plugin of the FIJI software, in which every cell was initially identified manually, followed by their automated detection within a z-stack of images, to avoid that cells were counted twice.

The design of the figures was carried out with CorelDraw software (Corel Corporation Ottawa, Ontario, Canada). Statistical analysis and graphics design were undertaken using Prism software (Graph Pad Software, Version 6, and 9.0.0 (121), San Diego, CA, USA).

3. Results

3.1. Generation and Characterization of Transgenic Reporter Lines Expressing Fluorescent Proteins under Control of a Mouse Fatty Acid Binding Protein 7 (Fabp7) Derived Regulatory Element

For the generation of transgenic lines, a previously identified 750 bp regulatory element derived from the mouse Fabp7 genomic region [26] was cloned in front of the adenoviral basal promoter E1b [16]. This was followed by the open reading frame either of the green fluorescent protein mClover [27], the red fluorescent protein mScarlet [28] fused to a myc tag sequence at its C-terminus, or a combination of membrane targeted monomeric super-enhanced CFP (mseCFP) [29] Fyn-mseCFP followed by a self-cleaving T2A-sequence [31] and a nuclear-localized Histone2B-fusion of mseCFP (H2BmseCFP). Transgenic fish strains were obtained by established Tol2-transposon protocols [30] (Figure 1A).

While the transgenic reporter lines Tg(Mmu.Fabp7-E1B:MScarlet-MYC)^bz23 [abbreviated name: Tg(Fabp7:mScarlet)] and Tg(Mmu.Fabp7-E1B:mClover)^bz25 [abbreviated name: Tg(Fabp7:mClover)] express the mScarlet and mClover fluorescent protein in the cytoplasm (Figure 1B), mseCFP expression in the transgenic line Tg(Mmu.Fabp7-E1B:FMA-mseCFP-2A-H2B-mseCFP)^bz24 [abbreviated name: Tg(Fabp7:mseCFP)] is localized to the cell membrane and the nucleus respectively (Figure 1C). In all three lines, expression becomes visible already in the brain at early embryonic stages around 14 hpf, which is similar to gfap enhancer-mediated expression [19], and fluorescent cells at 30 hpf are distributed only in the central nervous system [32], which is extensively labeled through the whole brain and spinal cord of young larvae, in the three transgenic reporter lines (Figure 1D–K, for a direct comparison of the expression patterns in the CNS among the three established lines please see Supplementary Figure S2) reminiscent to antisense mRNA in situ hybridization against fabp7a and fabp7b in zebrafish [11,33,34]. It is noteworthy that cells with Fabp7 enhancer-mediated fluorescent protein expression are enriched in the most medial areas (in the midsagittal plane along the ventricular zone), such as those located in the telencephalon, habenula, optic tectum, cerebellum, caudal rhombencephalon, and spinal cord (arrowheads in Figure 1H–K), and show a more dispersed cell density in parenchymal areas of the aforementioned brain structures (Figure 1H–K). The strongest expression combined with a higher cell density of Fabp7 enhancer-mediated fluorescent protein-expressing cells appears located in the most rostral area of the telencephalon (Figure 1H, H’), in the caudal area of the cerebellum, and in dorsal-paramedian areas of the caudal rhombencephalon (Figure 1H–K). In the cerebellum, Fabp7-regulatory element-mediated expression can be found in radially oriented cells with a unique branching of their membrane process visible in the molecular layer—the most dorsal cell layer of the cerebellar cortex (Figure 1H,H’’,I–K). Long Fabp7 enhancer-regulated fluorescent protein-expressing processes without branches are also observed in the lateral areas of the optic tectum and caudal rhombencephalon (Figure 1E–F,I–K).

3.2. Colocalization among Different Fluorescent Fabp7 Enhancer-Controlled Reporter Lines

The consistency of the expression pattern among the three different Fabp7 enhancer-regulated reporter transgenic lines was tested by analyzing the level of colabeling of mScarlet, mClover, and mseCFP fluorescent proteins in respective double transgenic offspring.

Larvae derived from Tg(Fabp7:mScarlet) × Tg(Fabp7:mClover) and Tg(Fabp7:mScarlet) × Tg(Fabp7: mseCFP) crosses of heterozygous carriers showed largely overlapping expression patterns throughout various regions of the CNS (Figure 1H-M and Figure 2A–F), as illustrated in the telencephalon (rostral part in Figure 2A, D) habenula (mostly located in the marginal zone, and a few in the parenchymal area, Figure 2B,D), cerebellum (mostly in the area of the ventricular zone, Figure 2C,E), and spinal cord (Figure 2F), with the majority of cells double positive for the respective fluorescent proteins in 5 dpf larvae (arrows in Figure 2).

Three CNS areas were selected to support this observation by quantification. This analysis showed that in Tg(Fabp7:mScarlet) × Tg(Fabp7:mseCFP) double transgenic larvae, about 90% of cells expressing mseCFP also expressed mScarlet fluorescent protein without a significant difference between the cerebellum (90.6%), caudal rhombencephalon (91.9%), and spinal cord (89.8%), (Figure 2G). Of note, in order to test the consistency in the spatial expression of both lines, the percentage of the total mScarlet expressing cells double positive in the spinal cord was also quantified (93.4%), which did not appear significantly different from the total amount of mseCFP⁺ cells. That it could not be detected in all cell expressions of both fluorescent proteins likely depends on the different strength of fluorescence, with mseCFP being clearly weaker in its emission than mScarlet. In addition, the half-life of maturation and turnover of both reporter proteins may differ.

3.3. Specificity of Fabp7 Enhancer-Mediated Reporter Line Expression Pattern

The specificity of the expression pattern of the transgenic line Tg(Fabp7:mScarlet) was tested by colocalization analysis of mScarlet with the endogenous Fabp7 protein using immunohistochemistry and commercially available antibodies (Figure 3A–E). The quantification of fluorescent cells (anti-mScarlet+/anti-Fabp7+) revealed that the vast majority of mScarlet-expressing cells were double positive co-expressing both proteins (telencephalon 88.6%, habenula 96%, optic tecum 78.5%, cerebellum 91.2%, and rhombencephalon 95%). Almost perfect colabeling was observed in the ventricular/subventricular zones, paramedian and lateral or parenchymal areas of all brain structures analyzed: telencephalon (rostro-medial zone, Figure 3A), habenula (marginal zone, Figure 3B), optic tectum (median area, Figure 3C), cerebellum and caudal rhombencephalon (paramedian area, Figure 3D). These findings support the specificity of the mouse Fabp7 regulatory element mediating fluorescent reporter protein expression in cells expressing Fabp7 endogenously in zebrafish. A few exceptions, though, were noticed, for example, in the lateral areas of the optic tectum, where the percentage of colabeled cells was significantly lower compared to other brain regions for currently unclear reasons (Figure 3E).

The observation of about 10% of the cells being positive only for mScarlet could be due to the long half-life of fluorescent proteins, likely displaying a slower turnover compared to the Fabp7 protein. Yet, the possibility of ectopic expression in a few cells and/or insufficient fluorescence intensity detection of the anti-Fabp7 antibody into the tissue cannot be completely ruled out. Nevertheless, comparative analysis of the expression pattern in the transgenic Fabp7 regulatory element-controlled reporter line with other specific cell type reporter lines might help to provide further confirmation of the specificity of the Fabp7 enhancer-mediated expression pattern.

Altogether, the colocalization of the anti-Fabp7 antibody staining and Fabp7:mScarlet, as well as the overlapping expression pattern of the Tg(Fabp7:mScarlet) and Tg(Fabp7:mClover) and Tg(Fabp7:mseCFP) lines, suggest that the three reporter lines are indeed specific for cells expressing Fabp7 endogenously and can be used interchangeably.

3.4. Comparison of Fabp7 Enhancer-Regulated Reporter Lines to Other Cell Type-Specific Transgenic Reporter Lines

Comparison of transgenic fluorescent protein reporter strains helps to further characterize the spatio-temporal activity of the Fabp7 enhancer-mediated fluorescent protein expression. We have therefore examined different areas of the brain and the spinal cord, with particular attention to the cerebellum, by comparing the expression of our Fabp7 regulatory element-controlled reporter to fluorescent protein expression patterns in other cell type-specific reporter strains.

3.4.1. Coexpression of Fabp7- and gfap-enhancer-Mediated Reporter Strains in Radial Glia

To further confirm the expression of the Fabp7 regulatory element in radial glial cells, we compared Fabp7 enhancer-mediated reporter protein expression to the expression of the green fluorescent protein in the Tg(gfap:GFP)^mi2001 strain, in which a regulatory element of the glial fibrillary acidic protein-encoding gene (gfap) was shown to activate reporter protein expression in radial glia [19]. In the majority of Fabp7:mScarlet red fluorescent cells, colabeling by the green fluorescent protein GFP was observed at 5 dpf in double transgenic [Tg(fabp7:mScarlet) x Tg(gfap:GFP)] offspring. Comparison of the percentage of Fabp7-gfap mediated reporter protein colabeling of cells in the cerebellum (ventricular and subventricular zone, arrows in Figure 3G) and other brain regions and spinal cord (arrows in Figure 3F–H and Figure 7C), revealed similar degrees of cell colabeling in the telencephalon 87%, the habenula 89.5%, the cerebellum 86.1%, and the spinal cord 82.9% (Figure 3I and Figure 7A). The small number of some mScarlet single positive cells, like a few scattered cells in lateral areas of the telencephalon, might be an indication for the minor heterogeneity of Fabp7 enhancer-mediated and gfap enhancer-controlled expression in the different radial glia and progenitor cell subpopulations.

3.4.2. Oligodendrocytes: A Subset of Fabp7-Expressing Cells Become Oligodendrocytes

In mice, gfap-expressing radial glia mostly differentiate into astrocytes [35]. While in zebrafish, glia with a typical astrocyte morphology, as observed in mammals, are not present, astrocyte-like cells have been identified in zebrafish [36]. Besides such astrocytic-like cells, glia progenitors also give rise to oligodendrocytes, of which the progenitors are known to express the transcription factors Olig2 and Nkx2.2a [21]. Double transgenic larvae from matings of the Tg(Fabp7:mScarlet) reporter line with carriers of the Tg(olig2:GFP) and Tg(nkx2.2:GFP) reporter strains revealed colabeling of cells with Fabp7 regulatory element-mediated fluorescent reporter protein expression with both oligodendrocyte reporters, mostly located in lateral/parenchymal areas (Figure 4A–K), in the telencephalon (arrow in paramedian zone, Figure 4D), optic tectum (arrows in paramedian and submarginal zones, Figure 4A,E), cerebellum (arrows in lateral areas, Figure 4B,G), and caudal hindbrain (arrows in lateral zone, Figure 4C). Quantification of olig2 enhancer regulated GFP expressing cells co-expressing Fabp7 regulatory element-mediated mScarlet expression revealed a high degree of GFP and mScarlet co-expression in the telencephalon 85.8%, the optic tectum 78.2%, and the spinal cord 72.4%, and about half of the population in the caudal rhombencephalon (Figure 3A–C,L and Figure 7A,C). Intriguingly, the cerebellum presented an exception with only 4.7% of olig2 expressing cells co-expressing the Fabp7 enhancer-regulated reporter (Figure 4B,L). This observation is due to the fact that not only oligodendrocytes are very low in numbers in the cerebellum, but also that olig2 expression in the cerebellum is largely confined to a neuronal population of eurydendroid cells as the main cerebellar efferent neurons [37].

With respect to the oligodendrocyte reporter Tg(nkx2.2:GFP), double transgenic Tg(Fabp7:mScarlet) x Tg(nkx2.2:GFP) larvae and the majority of GFP-expressing cells were also positive for Fabp7:mScarlet expression with 97.6% of colabeled cells in the telencephalon and 84.4.% colabeled cells in the cerebellum (although representing only about 8% of the Fabp7:mScarlet cells colabeled with GFP, as the majority of Fabp7-enhancer expressing cells develop into radial glia and only a minor portion into oligodendrocytes, as it is the case in the telencephalon as well, Figure 7A) because nkx2.2a expression is confined to the oligodendrocyte lineage but is not expressed in olig2-expressing eurydendroid neurons. This suggests that oligodendrocyte progenitors are derived from Fabp7 enhancer-controlled fluorescent protein-expressing glial cells (Figure 4D–K,M and Figure 7A,C). Of note, in all structures analyzed, the double positive cells were mainly found in lateral areas of the central nervous system, suggesting that these glial cells differentiating into oligodendrocytes are separated from the proliferative regions. In the spinal cord, only 21.8% of nkx2.2:GFP expressing cells co-expressed Fabp7:mScarlet with many solely nkx2.2:GFP positive cells being located in ventral positions. These cells may not give rise to oligodendrocytes but have been shown in mice and chicken embryos to be involved in ventral neural tube patterning being expressed in floorplate cells and later in somatic and visceral motoneurons [38,39,40].

3.4.3. Neuronal Cells: Some Fabp7 Regulatory Element-Mediated Fluorescent Protein Expressing Cells Are Neurogenic

The regulatory element derived from the Xenopus laevis neural-specific beta-tubulin gene mediates pan-neuronal expression of the fluorescent protein DsRed in the transgenic zebrafish strain Tg(Xla.Tubb:DsRed) [22]. The carriers of this strain were used to compare Fabp7 regulatory element-derived expression to neuronal expression. The analysis of double transgenic larvae of the genotype Tg(Fabp7:mseCFP) × Tg(Xla.Tubb:DsRed) revealed mostly complementary and non-overlapping expression patterns since the Fabp7 enhancer-mediated mseCFP-expressing cells were more concentrated in the ventricular areas, and Xla.Tubb+ cells were mainly located in the paramedian and parenchymal areas of the telencephalon (Figure 5A), the cerebellum, the caudal rhombencephalon (Figure 5B), and the spinal cord (Figure 5C) (Figure 5A–C and Figure 7C). This supports the fact that the Fabp7 regulatory element-induced expression is mostly confined to glial and not neuronal lineages.

However, at least a few Fabp7 and XIa.Tubb enhancer-mediated fluorescent proteindouble positive cells were detected (arrows in Figure 5), mostly located in the subventricular zone, as would be expected for recently differentiated neuronal cells. For instance, in the habenula (5.6% of Fabp7: mseCFP-cells were positive for DsRed), the caudal rhombencephalon (2.4%) and the spinal cord (2.3%) co-expression of both fluorescent proteins were displayed by a small subset of cells. This rate was lowest for the cerebellum (0.2%) and highest for the telencephalon (6.4%) (Figure 5D and Figure 7A). These findings of Fabp7-Xla.Tubb enhancer-mediated fluorescent protein double positive cells suggest that some Fabp7 regulatory element-controlled fluorescent protein-expressing cells differentiate further into neurons, which can be observed due to the high stability and slow turnover of the mseCFP fluorescent protein compared to the endogenous Fabp7 protein detected by immunohistochemistry.

Indeed, different from immunohistochemistry with the anti-Fabp7 antibody, the mseCFP reporter allowed for temporarily tracing the offspring of at least some Fabp7⁺ regulatory element-mediated fluorescent protein-expressing cells to develop into a neuronal fate, as at least a few differentiated neurons showed remaining mseCFP fluorescence. This finding could further explain the aforementioned small percentage of cells from the Fabp7 regulatory element-regulated transgenic reporter line that was apparently negative for the anti-Fabp7 antibody (Figure 3E).

3.4.4. Fabp7 Regulatory Element-Mediated Fluorescent Protein Expression in Neural Progenitor Cells of the Hindbrain

Next, focusing on the cerebellum, the expression of Fabp7 enhancer-regulated reporter expression in neural progenitors of the dorsal hindbrain was analyzed by using a transgenic zebrafish strain Et(zic4:Gal4TA4, UAS:mCherry)^hzm5 [abbreviated name Tg(zic4:mCherry)]. In this transgenic strain, the red fluorescent protein mCherry is expressed under the control of a regulatory element of the zic1/zic4 genomic locus in zebrafish [23]. These zinc finger-containing transcription factors are expressed early during central nervous system development and are involved in the regulation of cell proliferation in the dorsal neuroectoderm and formation of the hindbrain ventricle prior to the appearance of gliogenic and neurogenic lineages [41]. In double transgenic [Tg(Fabp7:mseCFP) x Tg(zic4::mCherry)] larvae, cells expressing both fluorescent reporter proteins (Figure 6A–F and Figure 7) were observed in all areas of the brain (as illustrated in the rostral most telencephalon, arrow in Figure 6A, and ventricular zone of the cerebellum, arrow in Figure 6D) as well as in the spinal cord (arrow in dorsal area, Figure 6F).

As zic4-expression is mostly confined to the dorsal part of the ventricular zone, dorsal CNS structures, such as the habenula and the cerebellum, showed the highest percentage of Fabp7 regulatory element-controlled mseCFP fluorescent reporter protein-expressing cells also expressing zic4 enhancer-mediated mCherry, with ratios of 94% and 54% respectively (Figure 6C,D and Figure 7A,C). Since zic4-expression is hardly expressed in ventral, tegmental areas, compartments derived from both alar and basal plates showed a significantly lower average percentage of Fabp7 regulatory element-mediated fluorescent protein-expressing cells co-expressing the zic4 reporter such as 16% in the telencephalon, 30% in the rhombencephalon, and 14% in the spinal cord (Figure 6A–G and Figure 7A,C). Nevertheless, these findings reveal that Fabp7 enhancer-mediated reporter protein expression is expressed very early during cerebellum development in neural progenitors and is maintained in glia but not in neuronal lineages.

3.4.5. Fabp7-Regulatory Element Regulated Fluorescent Protein Expression is expressed in Cerebellar Bergman Glia, which are derived from Cerebellar Ventricular Zone

Specifically focusing on the cerebellum, two dorsally located germinal zones can be distinguished, the dorsal most positioned rhombic lip and the ventrally adjacent ventricular zone, which can be distinguished molecularly by their mutually exclusive expression of the transcription factor-encoding genes atonal1a (atoh1a) and pancreas transcription factor 1 (ptf1a) respectively [42]. The ptf1a enhancer-mediated GFP expressing ventricular zone cells have been shown to give rise to important neuronal populations of the cerebellum, including Purkinje cells, eurydendroid cells, and inhibitory interneurons. Furthermore, Bergmann glia have been shown to originate from gfap and ptf1a enhancer-mediated fluorescent protein-labeled progenitor cells [43]. To further corroborate the origin of Bergmann glia, we crossed Tg(Fabp7:mScarlet) fish with Tg(ptf1a:GFP) carriers. Double transgenic larvae revealed a clear coexpression of the two fluorescent reporter proteins in about 20% of the ptf1a enhancer-mediated GFP expressing cells derived from the ventricular zone at 5 dpf (Figure 6H,I and Figure 7A,C).

This was consistent with our previous observations of Fabp7 immunohistochemistry staining in ptf1a enhancer-mediated fluorescent reporter protein-expressing cells, in which a fraction of about 10% of ptf1a:GFP green fluorescent cells in the zebrafish cerebellum was positive for Fabp7 expression [44]. Instead, only about 1% of atoh1-regulatory element-mediated fluorescent reporter protein-positive cells in Tg(Fabp7:mScarlet)/Tg(atoh1a:GFP) double transgenic larvae were double positive for both fluorescent reporters (meaning a total of only 1 to 6 double positive cells in the cerebellum of each larvae, being only of 0.9% of the total Fabp7 enhancer-controlled fluorescent protein-expressing cell population as well), which may represent maintained Fabp7 regulatory element-mediated reporter expression in very early neuroepithelial cells in which atoh1a and ptf1a enhancer-mediated fluorescent reporter protein expressing lineages have not been separated [45]. This Fabp7 regulatory element-driven fluorescent protein-expressing subpopulation of ptf1a regulatory element-controlled fluorescent protein-expressing ventricular zone-derived cells represented about half (53%) of the Fabp7 enhancer-mediated fluorescent reporter protein-expressing cells in the cerebellum (Figure 6H,I and Figure 7A,C) and may well represent progenitors of Bergmann glia as these double positive cells were mainly localized along the medial and caudal ventricular zone. Bergmann glia represent the almost only glia population of the cerebellum besides the few oligodendrocytes and are somewhat special as Bergmann glia differentiate into astrocytes of the cerebellum, but maintain characteristics of radial glia with radial fibers projecting to the pial surface in mammals as well as in zebrafish [5,46]. In the latter organism, they further contribute to the radial migration of granule cells during adult stages [5], which is considered a typical property of radial glia. Indeed, Fabp7 regulatory element-controlled fluorescent protein-expressing cells from the Tg(Fabp7:mScarlet) reporter line are still detected in the adult cerebellum, which, as typical Bergmann glia in adult zebrafish, present the soma located at the boundary between Purkinje cell layer and granular cell layer, and ramified branching in the molecular layer (Supplementary Figure S3). Therefore, Fabp7 regulatory element-mediated reporter protein expression is maintained in mature Bergmann glia and these cells marked by Fabp7 regulatory element-mediated fluorescent protein expression can be found throughout the zebrafish cerebellar cortex. Currently, we cannot rule out that a small subpopulation of Bergmann glia is derived from rhombic lip progenitor cells, but our findings further confirm that Bergmann glia represent a subpopulation of cells being produced by the ptf1a enhancer-mediated fluorescent protein-expressing cells of the ventricular zone of the zebrafish cerebellum [43].

4. Discussion

So far, the prominent enhancer for studying radial glial cells in zebrafish was the regulatory element of the intermediate filament Glial fibrillary acidic protein-encoding gene gfap. Due to the heterogeneity of glial cells, the characterization of a new enhancer of the Fatty acid binding protein Fabp7 from mice, as a reliable marker of proliferating and differentiating radial glial cells, provides a new tool for detecting a wider range of glial cells in zebrafish. The high degree of evolutionary conservation of the Fabp7 genomic sequence allows its use in zebrafish in combination with a teleostian basal promoter. Its small size of only 742 bp (compared to 7 kb of the gfap enhancer), provides a valuable tool for further functional studies, for example, to create multi-cistronic constructs for monitoring consequences of transgene expression in radial glia.

The colocalization of the Fabp7 regulatory element-mediated fluorescent reporter protein-expressing cells from the transgenic line Tg(Fabp7:mScarlet) with the immunolabeling of the anti-Fabp7 antibody supports the specificity of the new Fabp7 promoter in zebrafish. Additional proof of the Fabp7 promoter specificity for radial glial cells is evident by its coexpression with reporter fluorescent proteins mediated by the gfap promoter in a transgenic reporter strain (mi2001). Yet, it should be noted that a different transgenic GFP reporter line exists (zf167Tg) in which a BAC clone covering the genomic gfap locus of zebrafish has been used as a larger regulatory element to drive reporter protein expression, and both gfap reporter strains show slight differences in their reporter protein expression [47]. Nevertheless, the expression patterns of our new Fabp7 regulatory element-controlled reporter lines in larvae appear to match with the distribution of the Fabp7 protein previously described within the adult zebrafish brain, in particular for cells in the ventricular zone of different adult brain regions [2,4,5]. Furthermore, the overlapping expression of the fluorescent proteins among the new different Fabp7 regulatory element-controlled reporter lines observed in double transgenic larvae [Tg(Fabp7:mScarlet) vs. Tg(Fabp7:mClover) and Tg(Fabp7:mScarlet) vs. Tg(Fabp7:mseCFP)], argues that the Tg(Fabp7:mClover) and Tg(Fabp7:mseCFP) reporter lines are also specific for the Fabp7-expressing cells. The generation of these three Fabp7 enhancer-controlled reporter lines, expressing reporter fluorescent proteins with different excitation and emission properties, greatly expands the possibilities of combinations with other cell type-specific transgenic reporter lines for research in zebrafish.

Despite the high degrees of overlapping expression of endogenous Fabp7 proteins detected by immunohistochemistry and fluorescent reporter protein expression mediated by the Fabp7 regulatory element, the limitations of these reporter strains have to be mentioned as well. While the lack of perfect colabeling could be derived from the different half-lives of the zebrafish Fabp7 protein and the fluorescent reporter proteins, it is also possible that the Fabp7 regulatory element is not representing the entire enhancer of the endogenous Fabp7 homologs in zebrafish. If some enhancer or silencer elements are missing in the genomic fragment, small subpopulations of cells may ectopically express the fluorescent reporter or, conversely, may lack the expression of the reporter. Hence, the fluorescent protein expression in the Fabp7 regulatory element containing transgenic zebrafish reporter lines may not fully reflect the entire spatio-temporal expression of the endogenous Fabp7 homologs in zebrafish. For example, expression of fluorescent proteins in retinal Müller glia could not be found, and some weak expression in the retina may point to ectopic expression in retinal cells. Therefore, interpretations of findings of coexpression need to be cautiously made and should be confirmed independently.

To clarify the contribution of Fabp7 regulatory element-controlled fluorescent protein-expressing cells to different cell populations in the zebrafish brain, coexpression analysis in double transgenic larvae with various transgenic reporter strains was performed (Figure 7A–C). These revealed that oligodendrocytes marked by olig2 and nkx2.2 enhancer-mediated reporter protein expression are derived from Fabp7-regulatory element-controlled fluorescent protein-positive radial glia. Indeed, colocalization of Fabp7/olig2/nkx2.2 enhancer-mediated fluorescent reporter protein expression was observed in all brain areas and spinal cord. Of note, in previous studies, olig2 enhancer-mediated fluorescent protein-expressing cells were negative for the radial glial cell line, in which fluorescent reporter protein expression is mediated by a gfap gene-derived regulatory element [3,48]. Yet, we recently showed that in the cerebellum, isolated cells with olig2 enhancer-mediated fluorescent protein expression could be colabeled by immunohistochemistry against the Fabp7 protein [44]. This further suggests that oligodendrocyte precursors express Fabp7, and thus, Fabp7 enhancer-mediated fluorescent reporter protein expression allows the discrimination of different radial glia-derived subpopulations of glial cells.

Since Fabp7 expression is known to occur in proliferating neuroepithelial and radial glial cells, we also carried out a combined analysis with the neurogenic marker gene zic4. Indeed, a subset of Fabp7 enhancer-mediated fluorescent protein-expressing cells coexpressed zic4 enhancer-controlled fluorescent reporter protein expression, demonstrating that Fabp7 regulatory element-mediated reporter expression can be found in early neuroepithelial cells. The long half-life of the fluorescent reporter protein allowed us to temporally follow Fabp7 regulatory element-derived reporter protein expressing descendants. Such studies revealed that, unlike gfap-positive cells, despite some exceptions [7,12], Fabp7 expression is found in neurogenic progenitors that give rise to Fabp7/Xla.Tubb enhancer-mediated fluorescent protein reporter double positive neurons, which escape the detection by Fabp7 immunohistochemistry [2]. Intriguingly, during the processes of neuroregeneration, cells derived from gfap regulatory element-controlled reporter protein expression are also strongly involved in the production of neurons as well [12].

In the cerebellum, two proliferation zones can be distinguished and marked by the mutually exclusive expression of neurogenic marker genes atoh1a and ptf1a. Here, half of the Fabp7-regulatory element-mediated fluorescent protein-expressing cells coexpressed the ventricular zone, marking fluorescent protein expression controlled by the ptf1a regulatory element, and hardly any coexpression of fluorescent reporter proteins mediated by the Fabp7 and the atoh1a enhancers was found. Since Bergmann glia, the almost only glial cell population of the cerebellum, remain radially organized in the mouse cerebellum and also in the mature zebrafish cerebellum [5,46] and are known to express Fabp7 endogenously, these Bergmann glial cells are likely to be derived from the ptf1a regulatory element-mediated fluorescent protein-expressing ventricular zone and not from the atoh1a regulatory element-mediated fluorescent protein-expressing upper rhombic lip. Since in zebrafish, two additional atoh1 homologs exist with atoh1b and atoh1c [49], the origin of some radial glial cells from upper rhombic lip precursors cannot be completely ruled out. Nevertheless, the overlap in expression of Fabp7 and ptf1a enhancer-controlled fluorescent reporter protein expression argues against the existence of a third cerebellar proliferation zone—different from the ventricular zone and upper rhombic lip—and that cerebellar glia are derived largely from the more ventrally positioned ventricular zone, which is supported by recent Fabp7 antibody staining found in cerebellar ptf1a enhancer-mediated fluorescent protein-expressing cells [44]. Moreover, these findings further support previous studies in zebrafish, where gfap regulatory element-mediated fluorescent protein-expressing glia were also detected to coexpress the ptf1a enhancer-controlled fluorescent protein expression in the cerebellar ventricular zone [43]. Indeed, in mammalians, astrocytes derive mostly from ventricular neuroepithelia as well [50]. Our findings of different glial cell types and neuronal cells derived from Fabp7-regulatory element-mediated fluorescent reporter protein-expressing cells further confirm the previously described heterogeneity within radial glial cells [2].

Furthermore, the newly generated Fabp7 enhancer-controlled transgenic reporter lines provide helpful tools for developmental and regeneration studies by in vivo imaging in transparent zebrafish larvae. The stability of the fluorescent reporter proteins allows for detecting descendants of Fabp7 enhancer-mediated fluorescent protein-expressing cells until their differentiation and could be used for comparing neuroepithelial and radial glia cell contributions under conditions of neurological diseases and processes of neuroregeneration. For example, the effect of Purkinje cell degeneration in the genetic zebrafish models of human neurodegenerative diseases Spinocerebellar Ataxia Type 1 and Type 13 could be analyzed in double transgenic specimens [51,52]. In addition, zebrafish have recently been shown to recover from specific ablation of cerebellar Purkinje cells, yet the contribution of radial glia as potential progenitors of cerebellar Purkinje cell regeneration has remained unclear [44] due to the lack of suitable genetic tools presented here. Similarly, the response of radial glia to local wounding or large injuries in the central nervous system [1,53,54,55], for example, in increasing proliferation, extending membrane protrusions, phagocytosing cellular debris, or migrating towards injury sites could be studied in vivo using bioimaging approaches. Thus, a number of informative applications could be foreseen for the radial glia reporter strains established here.

Moreover, the evolutionary conservation of the regulation of the mouse genomic Fabp7 enhancer element in the teleost zebrafish further promises to hold true for other recently established teleostian model organisms such as Danionella cerebrum and Notobranchius furzeri [56,57]. Since these model organisms are accessible for genetic manipulations such as the generation of transgenic strains [58,59,60], genetically targeting radial glia in these models may allow us to further understand the contribution of radial glia to processes of circuitry formation, plasticity, and function in the teleost brain using adult in vivo imaging of Danionella cerebrum [61] or to address the contribution of radial glia to processes of central nervous system aging in Notobranchius furzeri [62]. Therefore, the Fabp7 regulatory element-mediated reporter strains and the characterization of the Fabp7 regulatory element represent powerful tools for in vivo research on glial cells in various fish model organisms.