Hypothalamic radial glia function as self-renewing neural progenitors in the absence of Wnt/ß-catenin signaling

The vertebrate hypothalamus contains persistent radial glia that have been proposed to function as neural progenitors. In zebrafish, a high level of postembryonic hypothalamic neurogenesis has been observed, but the role of radial glia in generating these new neurons is unclear. We have used inducible Cre-mediated lineage labeling to show that a population of hypothalamic radial glia undergoes self-renewal and generates multiple neuronal subtypes at larval stages. While Wnt/ß-catenin signaling has been demonstrated to promote the expansion of other stem and progenitor cell populations, we find that pathway activity inhibits this process in hypothalamic radial glia, and is not required for their self-renewal. In contrast, Wnt/ß-catenin signaling is required for the differentiation of a specific subset of radial glial neuronal progeny residing along the ventricular surface. We also show that partial genetic ablation of hypothalamic radial glia or their progeny causes a net increase in their proliferation, which is also independent of Wnt/ß-catenin signaling. Hypothalamic radial glia in the zebrafish larva thus exhibit several key characteristics of a neural stem cell population, and our data support the idea that Wnt pathway function may not be homogeneous in all stem or progenitor cells. radial glial both self-renewing and multipotent, proliferative response to partial ablation of themselves their neuronal progeny. use multiple perturbations of Wnt/β-catenin necessity glial self-renewal, expansion, neuronal differentiation. stem ectopic Wnt/ß-catenin activity neurogenic radial glia in the hypothalamus. data the most conserved role for Wnt pathway activity in neural progenitors is in promoting neurogenesis, and that other functions may differ between diverse stem and progenitor cell populations.


Introduction
The postembryonic zebrafish brain is highly proliferative and regenerative, characteristics that have been attributed to the presence of radial glia that persist throughout the central nervous system (CNS) and generate neurons (Kizil et al., 2012;Than-Trong and Bally-Cuif, 2015). We previously characterized a population of neural progenitors in the postembryonic zebrafish hypothalamus, which produces multiple neuronal subtypes through adulthood (Wang et al., 2012). A similar process also occurs in the mammalian hypothalamus, where adult neurogenesis contributes to reproductive and feeding behaviors (Kokoeva et al., 2005;Lee et al., 2012;Cheng, 2013). However the underlying progenitor cell populations supporting hypothalamic neurogenesis remain poorly characterized. While radial glia have been proposed to fulfill this role in both zebrafish and mouse (Lee et al., 2012;Wang et al., 2012;Haan et al., 2013;Robins et al., 2013), their capacity for self-renewal and differentiation have not been comprehensively tested.
In addition, the molecular pathways regulating radial glial self-renewal, expansion, and neurogenesis in the hypothalamus are poorly understood. Previous work from our laboratory showed that Wnt/β-catenin signaling is required for postembryonic hypothalamic neurogenesis (Wang et al., 2012), and other studies have also led to the hypothesis that pathway activity promotes radial glial differentiation (Lee et al., 2006;Wang et al., 2011;Choe and Pleasure, 2012;Wang et al., 2012;Varela-Nallar and Inestrosa, 2013). In contrast, Wnt/ß-catenin signaling has also been shown to promote the self-renewal and expansion of neural stem cells in the mammalian telencephalic subventricular zone and dentate gyrus (Qu et al., 2010). The specific function of Wnt/ß-catenin activity in hypothalamic radial glia is therefore unclear, leaving an Development • Advance article   open question as to whether a general role for the pathway exists for all neural stem and progenitor cell populations.
Here we take a genetic approach to identify the neural progenitor cell population in the larval zebrafish hypothalamus, and to characterize the response and regulation of hypothalamic radial glia during tissue growth and regeneration. Our data show that the radial glial population is both self-renewing and multipotent, and exhibits a proliferative response to partial ablation of themselves or their neuronal progeny. In addition, we use multiple perturbations of Wnt/β-catenin signaling to test the necessity and sufficiency of pathway activity for radial glial self-renewal, expansion, and neuronal differentiation. Consistent with studies of non-neural stem cells (Lowry et al., 2005;Blanpain and Fuchs, 2009;Farin et al., 2012), our data show that Wnt/βcatenin signaling is only necessary for the terminal differentiation of specific neuronal progeny. Furthermore, and as shown for radial glia in other brain regions (Wang et al., 2011), we find that ectopic Wnt/ß-catenin activity inhibits the expansion of neurogenic radial glia in the hypothalamus. Together, these data suggest that the most generally conserved role for Wnt pathway activity in neural progenitors is in promoting neurogenesis, and that other functions may differ between diverse stem and progenitor cell populations.

Radial glia are multipotent neural progenitors in the postembryonic hypothalamus
In an effort to identify molecular markers of radial glia in the zebrafish hypothalamic posterior recess (Fig. 1A), we found that at 5 days post-fertilization (dpf), Glutamine Synthetase ( Figure 2H].
These data indicate that postembryonic neurons in the zebrafish hypothalamus arise Development • Advance article from a radial glial population that can self-renew, expand, and generate multiple types of progeny.

Wnt/ß-catenin signaling is only required for differentiation of a specific subset of neuronal progeny
We next tested whether Wnt/β-catenin signaling is necessary for the self-renewal, expansion, or neuronal differentiation of radial glia. Using heat shock-mediated As an alternative method to inhibit Wnt signaling we examined the radial glial lineage in lef1 mutants. After 4-OHT-mediated conversion from 5-6 dpf and lineage analysis at 9 dpf, we found that loss of lef1 also did not significantly change the percentage of GS+ radial glia ( Figure 3E) within mCherry-labeled progeny. As we reported previously, lef1 is critically required to generate a subset of HuC/D+ neurons in the posterior recess (Wang et al., 2012). Our lineage analysis confirmed that these lef1dependent neurons arise from radial glia and showed that they specifically reside Development • Advance article within two cell diameters of the ventricle ( Figure 3F,G). However they comprise only a small portion of radial glial progeny, and the number of non-ventricular neurons was not decreased ( Figure 3F). Combined with the results of Dkk1 overexpression, these data lead us to conclude that Wnt/β-catenin signaling is not necessary for radial glial self-renewal or expansion. In addition, while Lef1-mediated Wnt activity is required for the differentiation of ventricular neurons, it is not required for the majority of neurogenesis in the hypothalamic posterior recess.

Partial genetic ablation of radial glia leads to increased proliferation of GS+ cells
To To determine if the proliferative response of radial glia to partial ablation requires Wnt/β-catenin signaling, we repeated our experiments in the presence of Dkk1 overexpression and in lef1 mutants. In both cases we observed a similar increase in BrdU incorporation within GS+ cells as in control animals ( Figure 4G,H). These data Development • Advance article indicate that just as in during normal growth, the regenerative expansion of hypothalamic radial glia is also Wnt/β-catenin independent.

Radial glia proliferate in response to genetic ablation of progeny
We next wanted to determine whether hypothalamic radial glia exhibit a proliferative response to the loss of a progeny cell type. Since we had observed that the lineage included dopaminergic neurons labeled by the th2 enhancer/promoter ( Figure 2F), and we found that the enhancer was not expressed in GS+ cells ( Figure 5A

Wnt activation blocks expansion of the radial glial population
Based on evidence that ectopic Wnt/ß-catenin signaling leads to a decrease in the  Figure 6A).

Development • Advance article
Continuous wnt8a expression over multiple days resulted in lethality, so to test the longer-term consequences of pathway activation we incubated animals in 4µM BIO, a pharmacological activator of Wnt/ß-catenin signaling (Sato et al., 2004;Shimizu et al., 2012;Lush and Piotrowski, 2014) from 6-9 dpf. This experiment also produced a small but significant decrease in the number of GS+ cells compared to controls ( Figure 6B,C), suggesting that Wnt signaling either inhibits radial glial expansion or causes the loss of GS expression. To specifically test these possibilities, we next performed lineage analysis in the presence of BIO from 6-9 dpf after recombination from 5-6 dpf. We found that the total number of mCherry+ cells in animals treated with BIO was significantly decreased compared to controls ( Figure 6D), coupled with a relative increase in the proportion of GS+ cells within the labeled population ( Figure   6E). The smaller number of progeny was not due to cell death, as neither wnt8a induction (

Hypothalamic radial glia exhibit multiple features of neural stem cells
Our results demonstrate that hypothalamic radial glia in zebrafish are self-renewing neural progenitors that can undergo a regenerative response, characteristics that are hallmarks of a stem cell population. Because we were not able to follow the lineage of single cells, we cannot determine whether individual radial glia are multipotent with Development • Advance article respect to neuronal fate. However the expansion that we observe in the lineage over a 5 week labeling period indicates that radial glia contribute significantly to the growth in size of the posterior recess, and our marker analysis shows that the population as a whole generates several neuronal subtypes.
While previous studies from our laboratory and others have observed the presence of proliferating neural progenitors in the adult zebrafish hypothalamus (Wang et al., 2012;Perez et al., 2013), the work described here focused on an earlier period of larval development. The behavior of radial glia during this period is therefore not strictly equivalent to other adult stem cell populations, which are typically quiescent or support tissue homeostasis rather than growth. Future studies testing the lineage and injury response of radial glia in the adult zebrafish posterior recess will provide more insight into whether they function as true neural stem cells.

Wnt/β-catenin signaling is not necessary for hypothalamic radial glial selfrenewal or expansion
Studies in the CNS (Piccin and Morshead, 2011) and other tissues (Nusse, 2008;Holland et al., 2013) have resulted in the hypothesis that Wnt/β-catenin signaling may function generally to promote stem and progenitor cell proliferation. In order to achieve the increase in population size that we observe in our lineage analysis, radial glia must undergo amplifying self-renewing divisions, and as has been shown in the telencephalon (Barbosa et al., 2015), regeneration may require these divisions at an even higher frequency. However our data indicate that ectopic Wnt activity in fact inhibits the expansion of hypothalamic radial glia population size, while other studies suggest that signals such as FGF (Kaslin et al., 2009;Robins et al., 2013), and Sonic Hedgehog (Dave et al., 2011;Shikata et al., 2011;Komada, 2012) likely promote this process.

Development • Advance article
We found that a specific subset of lef1-dependent neurons located near the hypothalamic ventricle arise from the radial glial lineage (Wang et al., 2012).
Combined with other studies in the retina (Agathocleous et al., 2009), cerebral cortex (Munji et al., 2011Zhang et al., 2014), hippocampus (Seib et al., 2013, and midbrain (Castelo-Branco et al., 2003), our data suggest that the most widely conserved role for Wnt/β-catenin signaling in the CNS may be to regulate the differentiation of specific subsets of committed neural progenitors.

Wnt/ß-catenin activity does not act identically in all neural stem and progenitor cells
Our experiments support the idea that diverse neural stem and progenitor cell populations likely exhibit different responses to Wnt/ß-catenin signaling. While Wnt ligands and reporters are expressed at high levels in the hypothalamic ventricular zone (Wang et al., 2012), radial glia largely fail to respond to these signals. This low activity state could be regulated by extracellular or intracellular pathway antagonists, or radial glia may simply fail to express the appropriate receptors to transduce Wnt signals. Regardless of the mechanism, it appears that this characteristic of hypothalamic radial glia is similar to other radial glial populations in the zebrafish retina and spinal cord (Goldman, 2014;Briona et al., 2015), but different from radial glia in the mammalian dentate gyrus (Qu et al., 2010). Other studies have similarly shown that neural progenitor populations vary dramatically in their interpretation of pathway activity (Poschl et al., 2013). Understanding these differences may help provide insight into the basis of radial glial, and neural stem/progenitor cell, heterogeneity.

Treatment of embryos and larvae
Cre-mediated recombination was performed by incubation in 5µM 4- conical tubes in a 39.5C water bath for 20 minutes. For experiments from 5-9 dpf, larvae were not fed.

Immunohistochemistry and in situ hybridization
Embryos were fixed in 4% paraformaldehyde with 5% sucrose overnight at 4C.
Brains were then dissected for immunohistochemistry and trunks were placed in PCR tubes for genotyping. Whole brains were washed in water, incubated in 2N HCl for 20 Brains were imaged on a Nikon A1 confocal microscope with a 60X oil objective.
The entire posterior recess was imaged using 3µm steps encompassing roughly 40µm total, cell counting was performed, and images were exported to Adobe Photoshop, Adobe Illustrator and ImageJ (NIH) for figure generation.
In situ hybridization was performed as described previously (Wang et al., 2012) using an antisense probe for sp5l generated from a PCR-amplified cDNA template. A T7 RNA polymerase initiation sequence was added to the 5' end of the reverse primer. Development • Advance article sp5l F: GTTTCCCAGCCACATGCAAC sp5l R:

Statistical analyses
Microsoft Excel was used to perform two-tailed equal variance T-tests, and p-values < 0.05 were interpreted as statistically significant.