Adult neurogenesis and brain regeneration in zebrafish


  • Caghan Kizil,

    1. DFG-Center for Regenerative Therapies Dresden, Cluster of Excellence (CRTD), and Biotechnology Center (Biotec), Technische Universität Dresden, Tatzberg 47/49, 01307, Dresden, Germany
    Search for more papers by this author
  • Jan Kaslin,

    1. DFG-Center for Regenerative Therapies Dresden, Cluster of Excellence (CRTD), and Biotechnology Center (Biotec), Technische Universität Dresden, Tatzberg 47/49, 01307, Dresden, Germany
    Search for more papers by this author
  • Volker Kroehne,

    1. DFG-Center for Regenerative Therapies Dresden, Cluster of Excellence (CRTD), and Biotechnology Center (Biotec), Technische Universität Dresden, Tatzberg 47/49, 01307, Dresden, Germany
    Search for more papers by this author
  • Michael Brand

    Corresponding author
    1. DFG-Center for Regenerative Therapies Dresden, Cluster of Excellence (CRTD), and Biotechnology Center (Biotec), Technische Universität Dresden, Tatzberg 47/49, 01307, Dresden, Germany
    • DFG-Center for Regenerative Therapies Dresden, Cluster of Excellence (CRTD), and Biotechnology Center (Biotec), Technische Universität Dresden, Tatzberg 47/49, 01307, Dresden, Germany
    Search for more papers by this author


Adult neurogenesis is a widespread trait of vertebrates; however, the degree of this ability and the underlying activity of the adult neural stem cells differ vastly among species. In contrast to mammals that have limited neurogenesis in their adult brains, zebrafish can constitutively produce new neurons along the whole rostrocaudal brain axis throughout its life. This feature of adult zebrafish brain relies on the presence of stem/progenitor cells that continuously proliferate, and the permissive environment of zebrafish brain for neurogenesis. Zebrafish has also an extensive regenerative capacity, which manifests itself in responding to central nervous system injuries by producing new neurons to replenish the lost ones. This ability makes zebrafish a useful model organism for understanding the stem cell activity in the brain, and the molecular programs required for central nervous system regeneration. In this review, we will discuss the current knowledge on the stem cell niches, the characteristics of the stem/progenitor cells, how they are regulated and their involvement in the regeneration response of the adult zebrafish brain. We will also emphasize the open questions that may help guide the future research. © 2011 Wiley Periodicals, Inc. Develop Neurobiol 72: 429–461, 2012


Regeneration is one of the most intriguing and fascinating biological phenomena but what the molecular and cellular basis of it is still not yet fully elucidated. The ability to regenerate lost or damaged tissues is seen to varying degrees in many organisms. For instance, unlike the vertebrates with high regenerative capacity, such as the urodele amphibians and the teleost fish, we mammals have quite restricted ability to regenerate our lost cells, tissues, and organs in adult stages (Stocum,2006; Poss,2010). Therefore, understanding the molecular nature of regeneration in organisms that can invoke special biological programs to replenish lost structures is not only interesting for shedding light onto a longstanding question of regeneration mechanisms, but can also be instrumental for devising therapeutic applications for humans.

Zebrafish is an excellent vertebrate model system for many disciplines such as molecular and cellular biology, developmental biology, and genetics (Nüsslein-Volhard and Dahm, 2002). Beside these, zebrafish is also valuable for analyzing the mechanisms how to recuperate tissue loss in adult stages, due to its pronounced regenerative capacity in several organs and tissues including the muscle, heart, pancreas, liver, skin, pigment cells, fins, barbels, and the central nervous system (Poss et al.,2003; Raya et al.,2004; Keating,2004; Nakatani et al.,2007; Hata et al.,2007; Stoick-Cooper et al.,2007; White and Zon,2008; Becker and Becker,2008; Huang and Zon,2008; Brignull et al.,2009; O' Reilly-Pol and Jonhson,2009; Antos and Tanaka,2010; Antos and Brand, 2010; Curado and Stainier,2010; LeClair and Topczewski,2010; Tal et al.,2010). Although several organisms have been used for regeneration studies (Sanchez-Alvarado and Tsonis,2006; Poss,2010), zebrafish offers a unique combination of advantages as a vertebrate regeneration model because of its easy maintenance, sequenced genome, established transgenesis techniques, cell lineage-tracing technologies, feasible forward-reverse genetic approaches and emergent opportunities for real-time imaging and micromanipulation (Antos and Brand, 2010; Poss,2010).

Different tissues in zebrafish have diverse modes of regeneration. For instance, heart regeneration involves dedifferentiation of cardiomyocytes after amputation of the ventricular apex and those cells proliferate and are used for replenishing the lost regions of the heart together with other cells from, for instance, the epicardium (Poss et al.,2007; Jopling et al.,2010; Kikuchi et al.,2010). Conversely, liver regenerates by a compensatory growth mechanism that utilizes proliferation of existing hepatocytes in the uninjured lobe (Stoick-Cooper et al.,2007; Kan et al.,2009; Curado and Stainier,2010). An alternative and additional mechanism for regeneration is to employ stem cells that normally reside in the tissue. The stem cells in the tissue are activated (i.e., losing their quiescence and/or increase in the rate of proliferation) following the insult and can give rise to multiple cell types depending on their potency. This mode of tissue repair is seen in the spinal cord or the retina of the central nervous system (Becker and Becker,2008; Kaslin et al.,2008; Reimer et al.,2008; Fleisch et al.,2011). No matter which mode is employed, three important characteristics are common: (1) Progenitors that are competent to differentiate into the lost cell types must be available in the tissue. These progenitors may be resident in the tissue before the regeneration response, or they could be generated by dedifferentiation or transdifferentiation of existing cell types. (2) Sufficient proliferation of stem cells or precursors that would supply the reservoir of newborn cells is essential. Therefore, the cues that instruct a proliferation response in the injured tissue are important components of a regeneration response. (3) Proper cues must act spatially and temporally in concert for newly produced cells to differentiate into lost cell types and to restore the tissue architecture analogous to the original forms of the precedent structures. These features indicate that for regeneration to occur, progenitor or stem cells and their progeny must be tightly controlled in a spatial and temporal manner. Hence, one interesting and attractive research field is to identify the molecular mechanisms that choreograph the regeneration response of the progenitors in regenerating organisms.

Tissues must initiate recuperation mechanisms to restore their internal homeostasis following physiological cellular turnover or an acute injury. The central nervous system is one of the most important tissues that is critical to survival of the animal upon injury owing to its pivotal role in many vital processes; for instance, control of movement and complex behaviors and the regulation of the endocrine system. Therefore, restoration of the nervous system function by regeneration is an intriguing phenomenon. However, the current level of cellular and molecular knowledge is insufficient regarding five major questions: (1) whether or not the endogenous progenitor cells can be stimulated to regenerate lost parts of the neuronal circuitry in the mature CNS, (2) what is the nature and identity of the progenitor cells that are neurogenic after the injury insult, (3) what are the molecular players involved in activating progenitor cells for proliferation and neurogenesis, (4) what are the patterning cues that instruct the newly formed cells to differentiate into correct types of cells, and (5) whether the newly formed neurons are maintained and integrate into the existing circuitry functionally and if yes how. Adult zebrafish brain serves as an excellent system for exploring these questions.

In this review, we will focus on the current knowledge and the potentials of brain regeneration capacity in the adult zebrafish emphasizing how stem/progenitor cells are involved in this process, and how we can determine the underlying molecular regulatory mechanisms. Substantial knowledge on regeneration of two parts of the central nervous system—retina and the spinal cord—has been gained through detailed studies using different injury and assay methods in zebrafish and other organisms; and these are discussed in several recent reviews (Becker and Becker,2008; Blesch and Tuszynski,2009; Filoni,2009; Mladinic et al.,2009; Tanaka and Ferretti,2009; Rossi and Keirstead;2009; Fleisch et al.,2011). Here, we will focus mainly on the stem/progenitor cell activity during adult neurogenesis and regeneration, which is emerging as a new and relatively unexplored research field in zebrafish.


Adult Neurogenesis and Regeneration Capacity

All vertebrates studied so far were shown to have progenitor zones in their adult brains with varying degrees of abundance and neurogenic capacity (Doetsch and Scharff,2001; Chapouton et al.,2007; Kaslin et al.,2008). However, the reasons why different organisms have different levels of adult neurogenesis (see Fig. 1) are intriguing yet still unclear. Several hypotheses can partially explain this phylogenetic variation: (1) Progenitor cells that proliferate in a constitutive manner have to be present in the neurogenic regions. Therefore, it is possible that during the course of evolution the availability of the proliferation zones might have been reduced for some organisms, such as mammals (Rakic,2002; Lindsley and Tropepe,2006; Ortega-Perez et al.,2007; Kaslin et al.,2008; Tanaka and Ferretti,2009). It is also interesting that the extent of adult neurogenesis seems to follow a similar trend with the capacity for adult growth and regeneration of different tissues in organisms. For instance, zebrafish can continue to grow as adults (Golsmith et al., 2006; Tsai et al.,2007) and in its adult brain have several progenitor zones, which proliferate continuously during homeostasis andcontribute to life-long growth of the brain (Grandel et al.,2006). In contrast, mammalian brain has limited post-embryonic neurogenesis (Kirkwood and Holliday,1979; Jacobson,1991; Metcalde and Monaghan,2003); for instance, they have limited amount of proliferation in the adult brain that is confined to two subregions of the telencephalon. Therefore, the reason why some vertebrates have less adult neurogenesis compared to others might stem from the same reason why they have limited growth capacity and quiescence. Yet, what that reason is vague. (2) Through phylogeny, due to the increased risk of tumorigenesis the evolutionary pressure might have favored for tight control over cell proliferation at the expense of regenerative potential. Since both regeneration and adult constitutive neurogenesis require substantial levels of progenitor cell proliferation, they might produce a burden on the immune system and constitutive cell production may lead to cancerous growth (Pearson and Sanchez-Alvarado, 2008). By doing so, the organisms that have less regenerative potential might be exerting an unfavorable environment on cell proliferation and neurogenesis by keeping some regions of the brain quiescent and/or non-neurogenic. Such a mechanism is supported by experiments in which cell proliferation and neurogenesis could be induced in vivo and ex vivo in brain regions that are otherwise non-neurogenic (Rasika et al.,1999; Benraiss et al.,2001; Pencea et al.,2001; Chen et al.,2007; Jiao and Chen,2008; Berg et al.,2010). (3) The cell-intrinsic mechanisms regulating the cellular plasticity might also have an impact on the capacity for adult neurogenesis. Several studies have shown that the intrinsic mechanisms determining the restricted potential of progenitors to form restricted lineage of neurons is an important character of stem cells in the nervous system (Barbe and Levitt,1991; Frantz and McConnell,1996; Suhonen et al.,1996; Desai and McConnell,2000; Shihabuddin et al.,2000; Seidenfaden et al., 2000). However, when stem cells from one neurogenic region is transplanted into other neurogenic regions, transplanted cells can acquire new fates and give rise to the type of neurons in the recipient region (Lim et al.,1997; Clarke et al.,2000; Temple,2001; Peng et al.,2002). This suggests that the intrinsic mechanisms can be overridden under certain circumstances, which are likely to depend on the combination of cell-intrinsic and extrinsic factors (Temple,2001; Merkle et al.,2007; Alvarez-Buylla et al.,2008; Costa et al.,2010; Ma et al.,2010). (4) Another hypothesis emphasizes more on the permissive environment after the injury. Instead of regenerating the lost structures, mammals often form scar tissues at the wound site after injury (heart infarcts, skin wounds, stab lesions, etc.) (Dinsmore, 2001). This is also true for the brain, where in the process of reactive astrogliosis parenchymal astrocytes are activated and recruited to the injury. These cells lay extensive extracellular matrix to the wound site, which forms the gliotic scar that is impermeable to axon regeneration and neuronal cell infiltration (Bovolenta et al.,1992; Fawcett and Asher,1999; Silver and Miller,2004; Rolls et al.,2009). It is remarkable that the astrocytes, which form the scar tissue are very similar to the ones that constitute the neural stem cells or are found in the brain parenchyma (Liberto et al.,2004; Faijerson et al.,2006; Buffo et al.,2008; Robel et al.,2011). Indeed, reactive astrocytes can also generate neurons under certain conditions (Wilhelmsson et al.,2004; Mori et al.,2005; Berninger et al.,2007; Buffo et al.,2008; Robel et al.,2009; L'Episcopo et al.,2011; Heinrich et al.,2010; Blum et al.,2010). However, zebrafish does not form any obvious scar tissue after injury in the CNS, heart, skin, and fins (Becker and Becker,2002; Poss et al.,2003; Antos and Brand, 2010). Therefore, permissive environment might be a factor endowing zebrafish and other nonmammalian vertebrates a higher regeneration potential than that of mammals.

Figure 1.

Neurogenic regions of the zebrafish brain in comparison to mammals. In the CNS of adult mammals, stem cell niches are restricted to the subventricular zone (SVZ) and the dentate gyrus (DG) in the telencephalon. In zebrafish stem cell niches are more abundant and distributed along the whole rostro-caudal brain axis.

Although the reasons why adult brains of different vertebrates have different potentials of neurogenesis and regeneration are not clear, highly complex, tightly regulated, and still largely enigmatic nature of stem/progenitor cells during brain regeneration in zebrafish thus raises two fundamental questions. (1) Which cells in the tissues act as stem/progenitor cells? (2) What are the molecular programs governing the stem/progenitor cell activity? In next coming sections, we will elaborate on these questions in light of the current literature.

Stem/Progenitor Cell Niches in the Vertebrate Brain

It was for a long time assumed that little or no neurogenesis takes place in the adult vertebrate brain. However, this has been disproven and all studied vertebrates so far display some degree of adult neurogenesis. Ability to form new neurons in adult brain relies on the presence of constitutively active progenitor zones (Lindsey and Tropepe, 2007; Ihrie and Alvarez-Buylla,2008; Conover and Notti,2008; Kaslin et al.,2008; Kempermann,2010).

In general, nonmammalian vertebrates display more abundant adult neurogenesis than mammals (see Fig. 1). In adult mammals and birds, neurogenic stem cell niches are restricted to one brain region, the telencephalon (Goldman and Nottebohm,1983; Doetsch and Scharf,2001; Alvarez-Buylla et al.,2002; VelleMa et al.,2010). Similarly, in reptiles and amphibians, most of the progenitor activity takes place in the telencephalon, although some activity is detected in the diencephalon and mesencephalon (Font et al.,2001; Raucci et al.,2006; Berg et al.,2010). Teleost fishes have numerous proliferation zones and widespread capacity to produce new neurons (Rahmann,1968; Rahmann and Korfsmeier,1968; Birse et al.,1980; Raymond and Easter,1983; Alonso et al.,1989; Zupanc and Zupanc,1992; Zupanc et al.,1996; Marcus et al.,1999; Ekström et al., 2001; Zupanc et al.,2005; Grandel et al.,2006; Kaslin et al.,2008). In fact teleost fish exhibit the most pronounced and widespread adult neurogenesis of any vertebrate studied so far (summarized in Kaslin et al.,2008).

It is thought that the cells within the progenitor zones of the brain form specialized micro-environments, neurogenic niches that harbor and regulate neural stem cells (Garcia-Verdugo et al.,2002; Alvarez-Buylla and Lim,2004). In the mammalin telencephalon neural stem cells are found in two distinct neurogenic niches, the subventricular zone (SVZ, also known as the subependymal zone) of the forebrain lateral ventricle and the subgranular zone of the dentate gyrus in the hippocampus (SGZ) (Altman and Das,1965; Altman,1969; Luskin,1993; Cameron et al.,1993; Lois and Alvarez-Buylla,1994; Betarbet et al.,1996; Eriksson et al.,1998; Seri et al.,2001; Carleton et al.,2003; Kempermann et al.,2003; Seri et al.,2004). In rodents, the SVZ niche consists of an extensive domain of heterogeneous neural stem cells that give rise to diverse cell types (Doetsch et al.,1997; Merkle et al.,2007). The SVZ contains relatively quiescent astrocyte-like neural stem cells called B cells. Although the B cells are subependymally located, they maintain contact to the ventricle through an apical process. Clustered processes from B cells together with ependymal cells (E cells) form a peculiar pinwheel-like structure at the ventricular surface (Mirzadeh et al.,2008). The end feet of the basal processes of B cells frequently interact with the brain vasculature (Mirzadeh et al.,2008; Shen et al., 2008). The B cells give rise to actively cycling intermediate progenitors (C cells or transit amplifying cells) that in turn gives rise to immature neuroblasts (A cells). In addition the SVZ also gives rise to oligodendrocyte progenitors (Lois and Alvarez-Buylla,1994; Kriegstein and Alvarez-Buylla,2009). The neuroblasts are tangentially migrating in chains into the olfatory bulb where they replace different interneuron types (Carleton et al.,2003; Imayoshi et al.,2008). This distinct migratory pattern is also known as the rostral migratory stream (RMS). Similar to the SVZ niche, the neural stem cells in the SGZ (Type 1 cells) display astrocyte-like properties. However, the progenitors in the SGZ undergo a lower rate of amplification compared to SVZ. Two subtypes of neural stem cells, radial and horizontal have been identified in the SGZ (Suh et al.,2007). The hierarchy between these two cells types remains enigmatic. Interestingly, these two subtypes respond differently to physiological stimuli and aging (Suh et al.,2007; Ligert et al., 2010). However, both type 1 cells are able to give rise to intermediate neural progenitors (Type 2 cells) that migrate and integrate into the adjacent granular cell layer (Cameron et al.,1993; Seri et al.,2004; Kempermann et al.,2004).

Additionally, following an injury, these stem cell niches enhance the rate of cell proliferation and neurogenic capacity (Arvidsson et al.,2002; Vela et al.,2002; Bottai et al.,2003; Urrea et al.,2007; Kernie and Parent,2010; Aguirre et al.,2010). Nevertheless, the vast majority of the neurons formed from these zones appear not to survive long enough to functionally integrate into the existing circuitry (Arvidsson et al.,2002; Webber et al.,2007, 2008; Di Giovanni,2009; Vandenbosch et al.,2009; Huebner and Strittmatter,2009; Kernie and Parent,2010).

The composition of neurogenic niches outside the mammalian lineage is poorly understood. In the few studied nonmammalian vertebrates, progenitors are found along the brain ventricles, but very little is known about the cellular and molecular identity (Alvarez-Buylla et al.,1998; Doetsch and Scharf,2001; Garcia-Verdugo et al.,2002). In zebrafish, there are 16 different progenitor niches that are distributed along the entire rostro-caudal brain axis (Grandel et al.,2006). These progenitor niches contain label-retaining cells that are thought to give rise to many different subtypes of neurons (Zupanc et al.,2005; Adolf et al., 2006; Grandel et al.,2006). We will discuss some of the context specific data for the stem cell niches and neural stem cells in the zebrafish brain in the following sections.

Characteristics of Neural Stem Cells in the Vertebrate Brain

Recent studies have shown that cells with glial characterstics act as neural stem cells during development and in the adult brain of mammals (Alvarez-Buylla,2001; Kriegstein and Alvarez-Buylla,2009). During brain development radial glia serve as the major neural stem cell type, while astrocyte-like cells function as neural stem cells in the adult mammalian brain. It has been shown that astrocyte-like neural stem cells in the adult brain are derived from embryonic radial glia (Miyata et al.,2004; MerkLe et al.,2004; Noctor et al.,2004). Furthermore, they share epithelial properties with radial glia, such as apical contact with the ventricle and basal endings on blood vessels (Mirzadeh et al.,2008; Shen et al., 2008). However, towards the end of development in mammals, most radial glia loose their ventricular attachment and migrate into the parenchyma where they regress their radial process and attain the multipolar “star-shaped” morphology that is characteristic for astrocytes (Noctor et al.,2008; Kriegstein and Alvarez-Buylla,2009). Astrocytes in the parenchyma are quiescent in the intact adult brain. Additionally, another progenitor type—intermediate progenitor cells (transit amplifying cells)—can be found in the brain both during the development and in the adult. The intermediate progenitors are lineage restricted in their potential and divide symmetrically. In the adult rodent brain intermediate neural progenitor cells are generated from the neural stem cells in the SVZ and SGZ. Additionally, intermediate oligodendrocyte progenitors are found in the brain parenchyma. NG2-expressing oligodendrocyte progenitors are distributed throughout the white and gray matter of the entire brain (Noble,2000; Aguirre et al.,2007; Barres,2008; Rivers et al.,2008). The developmental origin of oligodendrocyte progenitors is not completely understood but at least some of them are derived from the radial glia and in the adult from the SVZ (Levison and Goldman,1993; Nait-Oumesmar et al.,1999; Menn et al.,2006). Interestingly, parenchymal astrocytes and oligodendrocyte progenitors rapidly react to brain injury by proliferation, although they strictly remain within their lineage in vivo (Buffo et al.,2005; Buffo et al.,2008; Burn et al.,2009). Furthermore, parenchymal astrocytes and oligodendroyte progenitors have been induced to function as neuronal progenitors in vitro and in vivo by permissive conditions in culture or overexpression of transcription factors (Kondo and Raff, 2000; Heins et al.,2002; Buffo et al.,2005; Berninger et al.,2007; Heinrich et al.,2010; KronenBerg et al.,2010). Taken together, these results identify and delineate astrocytes and oligodendrocyte progenitors as interesting endogenous progenitor cell populations that have the potential to act as progenitors, which could be coaxed to participate in the neuronal regeneration.

In general little is known about different cellular subtypes of macroglia and their function in nonmammalian vertebrates. Stellate like astrocytes are scarce or absent cells in the brain of fish, amphibians and reptiles (Jacobson,1991; Kalman,2002). In the teleost forebrain stellate like astrocytes are absent (Kalman et al.,1998; Grandel et al.,2006; Pellegrini et al.,2007; Ganz et al.,2010; März et al., 2010a), although some astroglia-like cells have been detected in the spinal cord (Kalman et al.,1998; Kawai et al., 2001).

It is well known that radial glia (also known as ependymoglia) persists in the adult brain of nonmammalian vertebrates (Garcia-Verdugo et al.,2002; Kalman,2002). Radial glia have been proposed to be the progenitors for adult neurogenesis in the CNS of nonmammalian vertebrates because of the obvious link to their role as neural stem cells during development and their presence in the adult brain (Alvarez-Buylla et al.,2001; Götz et al., 2002; Pinto and Götz, 2007). In song-birds retroviral lineage-tracing and in depth analysis of the telencephalic ventricular zone has suggested that ventricular radial glia function as neuronal progenitors (Goldman et al.,1996; Alvarez-Buylla et al.,1998). Furthermore, radial glia also contribute to the constitutive neurogenesis and neuronal regeneration in zebrafish, reptiles and amphibians (Margotta et al., 1991; Perez-Canellas et al., 1997; Font et al.,2001; Lopez-Garcia et al., 2002; Weissman et al., 2003; Romero-Aleman et al., 2004; Berg et al.,2010, unpublished results). Although, radial glia have been suggested to function as neural and glial progenitors in the zebrafish brain (Grandel et al.,2006; Adolf et al., 2006; Pellegrini et al.,2007; Lam et al.,2009; Ganz et al.,2010; März et al., 2010a) also other populations—glia marker negative cells and neuroepithelial-like cells—with neural and glial progenitor characteristics have been identified (Raymond and Easter,1983; Kaslin et al.,2009; Ganz et al.,2010; Ito et al.,2010). Thus, it remains to be determined (1) if radial glia is the only type of neural stem cells in non-mammalian vertebrates, (2) the link between glia and progenitor state in non-mammalian vertebrates since it is not clear if all or just specific subsets of radial glia serve as gliogenic and neurogenic progenitors, and (3) the relative abundance and role of transit amplifying cells in the brain of nonmammalian vertebrates.

Examples of the Neurogenic Zones of the Adult Zebrafish Brain

Telencephalon and the Olfactory Bulb.

The progenitor zones in the telencephalon are the most-studied niches in adult zebrafish. Several groups have reported findings on the cellular composition of the niche, genes expressed in the progenitors, signaling mechanisms involved in adult neurogenesis, and which neuronal types are produced (Huang and Sato,1998; Zupanc et al.,2005; Grandel et al.,2006; Adolf et al., 2006; Kapsimali et al., 2007; Berberoglu et al.,2009; Lam et al.,2009, 2010; Chapouton et al.,2010; Ganz et al.,2010; März et al., 2010a,b). The telencephalic progenitors are heterogeneous and are located at the ventricular region in two distinct major domains—one ventral and one dorsal domain (Adolf et al., 2006; Grandel et al.,2006) [Fig. 2(B)]. These two progenitor domains are further demarcated by their differential expression of glial markers (Ganz et al.,2010).

Figure 2.

Stem cell niches of the adult zebrafish telencephalon and the cerebellum. A: Schematic summary of the zebrafish telencephalic progenitor niche. Dorsal and ventral telencephalon show distinct stem cell niche characters. Ventral telencephalon contains Nestin-positive neuroepithelial-like progenitors (gray), the majority of which are proliferating (red). These progenitors give rise to HuC-positive post-mitotic differentiating neurons (pink). Dorsal telencephalon contains progenitors that express glial markers S100β, GFAP and Vimentin (green) and have radial glial morphology. A proportion of these cells are proliferating and based on BrdU-label retention, they are likely to give rise to newborn neurons expressing HuC. Dorsal telencephalon also contains cells at the ventricular region that are negative for any glial marker but can be found proliferating. The nature of these cells and whether they contribute to neurogenesis is unclear. Arrows represent the direction of neurogenesis and migration. B: Schematic summary of the zebrafish cerebellar progenitor niche. Neural progenitors are maintained in the dorsomedial part of the cerebellum around a remnant of the IVth ventricle (the cerebellar recessuss). The progenitors give rise to granule neurons in a distinct outside-in fashion. (1) Polarized neuroepithelial-like progenitors (green) are restricted to the midline of the dorsal cerebellum. The progenitors give rise to rapidly migrating granule precursors (dark green) that initially migrate dorsolaterally. During this initial phase the granule precursors still proliferate. After reaching the meninge the granule precursors change to a unipolar morphology and start to migrate in ventrolateral direction to the GL. The granule precursors migrate into the GL and differentiate in to granule neurons. (2) A few glia with a radial morphology (light blue) are found close to the midline and they are used as scaffolds during the initial dorsal migration of granule precursors. (3) Bergmann glia-like cells are interspersed in the PL (dark blue) and generated from the lateral portion of the progenitor niche. CR, cerebellar recessus; GL, granule cell layer; ML, molecular layer; PL, Purkinje cell layer.

The ventral progenitor zone is a prominent proliferation domain that extends longitudinally from the olfactory bulb to the anterior commissure and dorsoventrally positioned between the dorsal and ventral nuclei of the ventral telencephalon [Fig. 2(A)]. The ventral telencephalic ventricular zone consists of progenitors with neuroepithelial characteristics such as Nestin expression, low or absent glia marker expression, interkinetic nuclear migration and a polarized morphology (Ganz et al.,2010). The presumptive stem cells, the label retaining and slow-cycling cells are primarily found in the dorsal part of the domain while a large bulk of fast-cycling cells is confined to the ventral part. The ventral domain with fast-cycling cells express PSA-NCAM, a marker known to be involved in structural plasticity and expressed by neuroblasts in the rodent SVZ (Grandel et al., 2000; Adolf et al., 2006; Kaslin et al.,2008; Ganz et al.,2010; März et al., 2010a). Brdu and IdU pulse chase experiments showed that neurons produced from the ventral progenitor domain migrate into all major ventral telencephalic nuclei and the olfactory bulb (Byrd and Brunjes1995, 1998, 2001; Grandel et al.,2006). Furthermore, scattered proliferating cells are detected throughout the olfactory bulb. The olfactory bulb is also a neurogenic zone in the mammals (Altman,1969; Corotto et al., 1993; Lois and Alvarez-Buylla,1994). Neurons generated in the SVZ of the lateral ventricle migrate through the rostral migratory stream (RMS) to the olfactory bulb throughout the adult stages (Doetsch and Alvarez-Buylla,1996; Alvarez-Buylla et al.,2008). Furthermore, different types of neurons are derived from heterogeneous progenitors that are located in different subregions of the SVZ (Merkle et al.,2007; Kriegstein and Alvarez-Buylla,2009). Although, the ventral progenitor domain in zebrafish show some similarities with the rodent SVZ and the RMS—e.g. presence of slowly cycling and fast cycling cells, Nestin expression in fast cycling cells, PSA-NCAM expression, presence of migrating neuroblasts in the olfactory bulb, there are some notable differences. For example, the majority of the produced cells remain in within the telencephalon. Indeed, BrdU and IdU pulse experiments suggest that newborn cells in the olfactory bulb primarily originate from the very rostral part of the ventral progenitor zone or from progenitors within the olfactory bulb (Grandel et al.,2006). In fact, the neurogenic and migratory patterns in zebrafish telencephalon share more similarities with ones in songbird than the ones in rodents. Similar to the zebrafish, neuroblasts in the songbird telencephalon migrate through different modes and routes to many diverse nuclei in the telencephalon (Alvarez-Buylla et al., 1994; Alvarez-Buylla and Kirn, 1997; VelleMa et al.,2010). Brdu pulse-chase experiments showed that presumptive dopaminergic (tyrosine hydroxylase immunopositive) and GABAergic cells (gad67 expressing) are produced in the ventral telencephalon and olfactory bulb (Kaslin et al.,2001; Adolf et al., 2006; Grandel et al.,2006; Mueller and Guo,2009). The transcription factors pax6a-b, ascl1a-b, dlx2a, dlx5a are involved in the specification of dopaminergic and GABAergic interneurons in the SVZ of the rodent brain (Kempermann, 2005; Kriegstein and Alvarez-Buylla,2009). In agreemenet with this in zebrafish the expression patterns of these transcription factors overlap with the domains where the ventral telencephalic progenitors and their progeny are found (Adolf et al., 2006; Mione et al., 2008; Ganz et al.,2010; März et al., 2010b, Ganz et al., in review). However, further lineage analysis of the ventral telencephalic progenitors and their progeny is needed in order to reveal specific migratory patterns and to explore how hetergenous the progenitors are in terms of specific glial and neuronal progeny.

The dorsal telencephalic progenitor domain contains scattered clusters of proliferating cells, along the whole ventricular surface (Adolf et al., 2006; Grandel et al.,2006; Ganz et al.,2010). Radial glia that is labeled with markers such as GFAP, Gluthamine synthtase, Vimentin, S100B, Aromatase-B, BLBP are present through the whole dorsal progenitor domain (Adolf et al., 2006; Grandel et al.,2006; Pellegrini et al.,2007; Lam et al.,2009; Ganz et al.,2010; März et al., 2010a). However, two distinct populations of proliferating cells have been found in the dorsal progenitor domain, radial glia-marker positive and glia-marker negative cells (Ganz et al.,2010; März et al., 2010a). The ratio between proliferating radial glia and proliferating glia-marker negative cells is region dependent. Typically, 50-75% of the proliferating cells are radial glia. Interestingly, two different populations of the presumptive stem cells (label retaining actively cycling cells) are detected: radial glia-marker positive and glia-marker negative cells (Ganz et al.,2010). Interestingly, the actively cycling glia-marker negative label retaining cells are notably more abundant than the label retaining radial glia. Lineage analysis of the putative progeny in proximity to the label retaining cells did not show any significant difference between glia-marker negative and positive label retaining cells. Analysis of different BrdU chase times show that label retaining radial glia and glia-marker negative cells have a similar proliferation rate. Although, a small population of proliferating ventricular olig2:gfp positive cells are found in the telencephalon, label retention experiments do not identify the glial marker negative cells as oligodendrocyte progenitors (Ganz et al.,2010). Thus, the nature and purpose of the glia-marker negative progenitors is unclear. Because clusters of PSA-NCAM cells are detected in the dorsal progenitor domain it is appealing to suggest that the proliferating glia negative cells serve as intermediate progenitors or neuroblasts (Kaslin et al.,2008; März et al., 2010a). This explanation may be true for some the cells but fits poorly with the proliferation characteristics, lineage analysis of progeny in proximity to presumptive stem cells and the notion of glia-marker negative label retaining actively cycling cells. It is possible that the glial marker negative progenitors represent a distinct class of neural progenitors in the dorsal telencephalon. Interestingly, heterogeneity and behavior of progenitor populations in embryonic and adult mammals has been linked to differential expression of glial markers (Hartfuss et al., 2001; Malatesta et al., 2000, 2003; Kriegstein and Alvarez-Buylla,2009). However, further marker analysis and lineage tracing is needed in zebrafish to understand the hierarchy and relationship among the different progenitors in the dorsal telencepahon. In contrast to the ventral telencephalon neurons are slowly added in a laminar fashion in the dorsal telencephalon (Grandel et al.,2006). Little is know about which types of neurons that are produced except that some of them contain parvalbumin (Grandel et al.,2006). However, several pallial transcription factors such as: neurog1, emx1, emx3, neurod, fezf2, tbr1, tbr2 are expressed in or in close proximity to the dorsal progenitor domain (Adolf et al., 2006; Berberoglu et al.,2009, Ganz et al., unpublished). In rodents tbr1-2, neurod and fez2 are involved in the specification of glutamergic neurons during development and in adult neurogenesis (Hevner et al., 2006; Shimizu and Hibi, 2008; Brill et al., 2009). Therefore it is tempting to presume that mainly glutamergic neurons are produced in the dorsal telencephalon of zebrafish. Although, heteregenous progenitors in terms of glial marker expression, progenitor characteristics and expression of molecular markers are present in the zebrafish telencephalon it is so far unclear if the progenitors display other regional characteristics.

Optic Tectum and the Isthmic Proliferation Zone.

The coordinated growth of the retina and optic tectum in fish and amphibians is a well documented phenomenon (Straznicky and Gaze,1971; Straznicky and Gaze,1972; Johns and Easter,1977; Beach and Jacobson,1979; Raymond and Easter,1983; Marcus et al.,1999). However, little is still known about cellular and molecular mechanisms that regulate tectal growth. The proliferating cells in the zebrafish optic tectum are found in the periventricular gray zone along the peripheral rim of the optic tectum (Marcus et al.,1999; Grandel et al.,2006; Ito et al.,2010). The majority of the progenitor acticity in the optic tectum of adult zebrafish takes place in the caudal part (Ito et al.,2010). The proliferating progenitors do not express glial markers and thus do not belong to the population of radial glia that are found beside the ventral surface of the optic tectum that is facing the tectal ventricle. However, the progenitors in the optic tectum display neuroepithelial properties and express progenitor markers such as Sox2 and Musashi1 (Ito et al.,2010). BrdU pulse chase experiments show that GABAergic and glutamergic neurons, oligodenrocytes and radial glia continuously are generated in the optic tectum of adult zebrafish (Grandel et al.,2006; Ito et al.,2010).

A prominent progenitor zone, the isthmic proliferation zone is found at the posterior mesencephalic lamina at the junction between the mid- and hindbrain in the in adult zebrafish (Adolf et al., 2006; Chapouton et al.,2006; Grandel et al.,2006). Progenitors and their progeny in this progenitor domain are nicely labeled by her5:gfp expression (Chapouton et al.,2006). The her5:gfp positive cells constitute a heterogenous population of progenitors and are broadly distributed around the ventricular region at the isthmus. A subpopulation of the her5:gfp positive cells are proliferating and some of the proliferating her5:gfp expressing cells are found in domains where glial and progenitor markers such as GFAP, BLBP and sox2 are expressed (Chapouton et al.,2006). However, proliferating her5-expressing cells with that do not express glial markers and proliferating her5 negative cells are also found in the isthmic proliferation zone.

BrdU pulse chase experiments show that neurons, glia, and oligodendrocytes are generated from the isthmic proliferation zone and partially overlaps with her5:gfp expressing cells (Chapouton et al.,2006). Further studies are needed to determine the role of her5 in the isthmic proliferation zone and characteristics of the isthmic progenitors.


Cerebellar progenitor activity and neurogenesis continue into adulthood in zebrafish (Zupanc et al.,2005; Grandel et al.,2006; Kaslin et al.,2009). Notable progenitor activity is detected along the IVth ventricle and its remnant, the cerebellar recessus. Interestingly, the neural progenitors persist in a specialized niche in the zebrafish cerebellum. Label retaining, distinctly polarized neuroepithelial-like cells that express the neural stem cell and progenitor markers nestin, Sox2, Meis, and Musashi are found along the IVth ventricle and its dorsomedial and lateral outpockets, the cerebellar recessus (Kaslin et al.,2009). Interestingly, the nestin positive progenitors do not express radial or astroglial markers such as S100β, GFAP, Vimentin, BLBP, Gluthamine synthase, or Aromatase-B (Kaslin et al.,2009). The cerebellar stem cell niche consists of polarized neuroepithelial-like progenitors that inhabit the dorsal part of the cerebellar recessus while epithelial-like glia line the ventral part of the recessus [Fig. 2(B)]. Additionally, the neuroepithelial-like progenitors are flanked medially by radial glia and laterally by Bergmann glia. The neuroepithelial-like progenitors give rise to a population of dividing intermediate progenitors. The neuroepithelial-like cells and the intermediate progenitors express characteristic upper rhombic lip and granule cell markers such as atoh1a-c, zic1, zic3, reelin, neurod1, Pax6, Meis1, and GAP-43 (Costagli et al.,2002; Kaslin et al.,2009; Chaplin et al.,2010; Kani et al.,2010). The intermediate progenitors rapidly migrate in a distinct outside-in fashion into the granule cell layer where they differentiate into granule cells. In the adult cerebellum the granule precursors migrate into the granule cell layer within three days. Although several subtypes of inhibitory and excitatory cells are found in the zebrafish cerebellum, mainly granule cells and are produced in the adult (Kaslin et al.,2009).

In vivo imaging of the cerebellar stem cell niche from embryonic to adult stages showed that the niche is generated in a two-step process that initially involves morphogenetic movements and later tissue growth. Because of these processes rhombic lip progenitors and a small portion of the ventricle, the cerebellar recessus, are displaced deep into the cerebellar tissue (Kaslin et al.,2009). In contrast, the cerebellar midline and the ventricle are lost as the two cerebellar hemispheres fuse in chick and rodents (Altman and Bayer1997; Louvi et al.,2003; Sgaier et al.,2005). Upper rhombic lip progenitors are maintained dorsal to the cerebellar recesus in the adult zebrafish while ventricular zone derived progenitors and glia are found ventral to the recessus. The ventricular zone derived progenitors are almost quiescent in the adult zebrafish which is consistent with the notion that few inhibitory neurons and glia are produced in the adult.

During the development of the amniotes granule cell precursors migrate from the URL to the cerebellar surface where they transiently form a highly proliferative second germinal zone, the external granule layer. Sonic hedgehog secreted from Purkinje neurons act as the major mitogenic signal for granule precursors in the external granule layer (Dahmane etal.,1999; Wechsler-Reya et al.,1999). In contrast to amniotes, sonic hedgehog signaling and a prominent external granule cell layer is lacking in the developing and adult zebrafish cerebellum (Kaslin etal.,2009; Chaplin et al.,2010). Interestingly, the advent of a secondary zone of transient amplification (e.g., the external granule cell layer) seems to be an amniote specific developmental adaptation (Chaplin et al.,2010). However, dividing intermediate granule precursors can be continuously detected in the zebrafish cerebellum although the amplification rate is very low. The production of granule cells is more likely to be controlled by the primary progenitors (Kaslin et al.,2009). Not much is known about signals that control the proliferation and differentiation of the cerebellar progenitors. However, Fgf signaling is required for the maintenance and proliferation of the adult cerebellar progenitors (see below, Kaslin et al.,2009). In addition, sex differences in proliferation have been detected in the adult cerebellum (Ampatzis et al.,2007).


Injury Methods to Induce Regeneration in the Brain

Different methods have been developed in mammals, amphibians and fish for inducing an acute injury. These applications range from surgical removal of big parts of the brain to specifically killing small population of neurons. These assays all have their own strengths and weaknesses, but they can serve as good tools for asking specific questions about regeneration.

Physical Lesions.

The most common injury paradigm to study brain regeneration is physical lesions (Kirsche,1967a,b; Maeda et al.,1977; Stevenson and Yoon,1980; Oka and Ueda,1981; Bohn and Reier,1985; Roberts and Alford,1986; Alonso and Privat,1993; Cameron et al.,1993; Zupanc et al.,1998; Zupanc and Ott,1999; Kalman and Ajtai,2000; Hampton et al.,2004; Buffo et al.,2005; Shavit et al.,2005; Zupanc, 2008; Kroehne et al.,2009; Gomez et al.,2010; Serrano-Perez et al., 2011). Mechanical injury methods comprise damaging brain parts to assess their functional relevance, as well as the removal of only small regions with nanosurgery methods. Traumatic physical lesions are characterized by a pronounced complex multi-cellular response that involves apoptosis, inflammation, proliferation of glial cells and increased progenitor cell activity can lead to an increase in neurogenesis depending on the severity of the lesion, site of the trauma and competency of the progenitor cells (Eccles, 1976; Zupanc,2001; Bazan et al.,2005; Leker et al.,2007; Endo et al.,2007; Kaslin et al.,2008; Fitch and Silver,2008). However, side effects of these traumatic lesion paradigms, such as increased cell death through secondary degeneration, inflammation and breach of the blood-brain barrier, can complicate the interpretation regarding regenerative events. Recent advancements in nanosurgery methods, which reduce the traumaticity and increase the precision of the injury, seems promising (Yanik et al.,2004; Steinmayer et al., 2010; Samara et al.,2010) and will probably serve as better tools to alleviate degenerative effects of physical lesions in the future.

Chemical Injuries.

Chemical compounds can act as neurotoxins either nonspecifically or particularly on neuronal or glial cells. Several different substances such as triethyltin, somatostatin, 6-hydroxydopamine (6-OHDA), methylmercury, 3-acetylpyridine (3AP), 1,3-dinitrobenzene (DNB), 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) and 1-methyl-4-phenylpyridinium (MPP+) were used to elicit neuronal damage in the brain (Balaban et al.,1988; O'Callaghan,1988; Aschner and LoPachin,1993; Font et al.,1997; Philbert et al.,2000; Dorman, 2000; Anichtchik et al.,2004; ReaLi et al.,2005; Mavroudis et al.,2006; Parish et al.,2007; Sallinen et al.,2009). Provided that specific neurotoxins could be found to kill desired cell types effectively, the major advantage of using these methods would be effective targeting. However, besides the difficulty in achieving the specificity of the chemical, toxicological treatments often display unwanted side effects or the optimal doses to kill cells might cause unwanted systemic effects in animals. A promising approach is generation of transgenic animals that express toxic proteins or enzymes that catalyze the conversion of otherwise harmless compounds to toxic products in a tissue- and cell-specific manner (see below).

Transgenic Approaches.

Zebrafish is a useful organism for generating transgenic animals that drive expression of tissue/cell-specific reporters, altered gene products, toxic proteins, and regulatory RNAs. Using different transgenesis approaches (Jessen et al.,1998; Kawakami,2005; Balciunas et al.,2006; Kawakami,2007; Yang et al.,2009) many transgenic lines have been generated to drive expression of reporters under specific promoters or enhancer elements in diverse regions of the animal, including the brain (Asaoka et al.,2002; Picker et al.,2002; Dorsky et al.,2002; Nakada et al.,2004; Gao et al.,2005; Bernardos and Raymond,2006; Jin et al.,2006; Zhao et al.,2006; Kotani et al.,2006; Bai et al.,2007; Yeo et al.,2007; Park et al.,2007; Scott et al.,2007; Kleinjan et al.,2008; Asakawa et al.,2008; Chen et al.,2009; Kaslin et al.,2009; Scott and Baier,2009). Furthermore, these constructs can be easily modified to allow promoter-specific cell ablations in transgenic animals. For instance, in this manner, Diphteria toxin A has been used to suppress the lens growth (Kurita et al.,2003) and a bacterial toxin (Kid) was used to kill germ cells (Slanchev et al.,2005). One major drawback of using promoters and enhancers directly to drive expression of toxic factors is nonconditionality; namely, the presence of the promoter activity during the embryonic development, which allows the ablation agent and toxic compound to be expressed continuously in a way that compromise the survival of the animal. Therefore, a new method for conditional cell ablation was developed using an enzyme—nitroreductase—that catalyzes the reduction of a pro-drug metronidazole into a cytotoxic product that kills the cells (Pisharath et al.,2007; Curado et al.,2008). With this system, temporal control can be achieved by addition of the pro-drug and the spatial control is attained with the promoter driving the nitroreductase expression. Although this method has been used in several studies so far (Davidson et al., 2007; Zhao et al.,2009; Montgomery et al.,2010; Hu et al.,2010), the enzyme nitroreductase may be more efficient in some cell types than others, creating a technical hurdle for particular cell types.

Another system has been developed in zebrafish for conditional gene expression and cell ablation relying on the bacteriophage P1 site-specific DNA recombinase Cre and its cognate site loxP (Abremski et al.,1983; Abremski and Hoess,1984) to generate conditional genetic alterations induced by Cre-mediated recombination (Thummel et al.,2005; Le et al., 2007; Yoshikawa et al.,2008; Hans et al.,2009; Hesselson et al.,2009; Hans et al.,2011). This two-component system in essence has (1) a driver cassette with a promoter expressing Cre recombinase in a tissue/cell-specific manner and (2) an effector cassette which after recombination leads to expression of the gene of interest whose expression is driven by an inducible promoter such as the heat shock promoter (unpublished). This dual control allows expressing a gene of interest in a spatially and temporally regulated manner. One pitfall for the Cre-lox system may arise when the promoter expressing Cre shows leakiness. In this case, the recombination happens ectopically or prematurely which may hinder the interpretation of results. To overcome the leakiness, a modified version of the system in which the Cre is fused to the ligand binding domain of the estrogen receptor is adapted (Metzger et al.,1995; Hans et al.,2009) (see Fig. 3). This version of the Cre is restricted to the cytoplasm unless the animals are treated with Tamoxifen or 4-hydroxy-Tamoxifen, which leads to nuclear translocation of the Cre and the recombination (Hans et al.,2009; Hans et al.,2011). Therefore, this system allows precise manipulation of gene expression or cell ablation in zebrafish. The conditional cre-loxP system is also a powerful tool for lineage tracing experiments.4

Figure 3.

Cre/lox system for conditional gene expression in zebrafish. Cre/lox system in zebrafish for regulating gene expression in a spatial and temporal manner. A: This system is composed of two transgenic cassettes. Spatial control is achieved by a tissue specific promoter (TSP) driving the expression of mCherry reporter and a modified version of the Cre recombinase (CreERT2). These two proteins are translated from a single transcript but the T2A self-cleaving peptide sequence eventually generates two non-fused proteins. The temporal control is fulfilled by a conditionally inducible promoter (CIB), mostly the heat-inducible, zebrafish hsp70 gene promoter. Temporal cassette contains two loxP sites flanking the first reporter, DsRed, which is followed by a stop codon. This way, transgenes downstream to the DsRed are not expressed. This cassette also contains a second reporter, EGFP which is linked to the open reading frame of the gene of interest (GOI) or any cell ablation agent (AA) to be used for conditional cell ablations. B: The modified version of Cre allows a third way of control. When there is no Tamoxifen, Cre is kep in the cytoplasm with the activity of the Estrogen receptor (ER) domain. This way, recombination does not happen even if the Cre is present. When Tamoxifen is applied, CreERT2 translocates into the nucleus and performs the recombination of the temporal transgenes by excising the DsRed and the stop codon. Therefore, EGFP and other transgenes can now be expressed. C: As ablation agents, transgenic lines containing Nitroreductase, Diphteria Toxin A and Killer Red are available. See text for details.

Figure 4.

A method to identify the genes involved in progenitor cell behavior during regeneration response in adult zebrafish brain. After a brain injury that elicits a regeneration response, brains can be dissected and dissociated into single cells. Transgenic lines can be used to label different cell populations in the adult brain. Also, proliferating and non-proliferating cells can be distinguished with the use of a DNA dye. The RNA isolated from the sorted cells can be used for gene expression profiling using microarray or next-generation sequencing technologies. Results can be analyzed by different bioinformatic tools and the candidates can be verified by in situ hybridization analyses.

Based on the stochastic recombination of several fluorophore genes using a Cre-lox strategy, the “brainbow” system has been established, which allows labeling of different cells and cell lineages in distinct color combinations (Livet et al.,2007). With the establishment of the conditional Cre-lox system in adult zebrafish (Hans et al.,2009), such a high-resolution labeling method will be a powerful tool for analyzing the progeny of stem cells at cellular resolution, which may help in better defining the topological organization and cellular behavior of the progenitor regions in the zebrafish brain. Additionally, in order to be used for conditional cell ablation strategies, several Cre-effector lines of transgenic zebrafish using a variety of ablation agents such as nitroreductase (Curado et al.,2008), Diphteria toxin A (Kurita et al.,2003) and Killer Red protein (Bulina et al.,2006; Del Bene et al.,2010) can be generated. These tools will help tissue and cell-type specific killing of neurons to initiate a regeneration response, which will presumably involve activation of lineage-specific regeneration programs.

Regeneration Response

A remarkable regenerative potential after severe physical injuries, including ablation of whole brain parts, has been found along the whole rostro-caudal neuraxis in many teleost fishes (Richter, 1965; Segaar, 1965; Richter, 1970; reviewed by Kirsche, 1965; Becker,2008; Kaslin,2008; Zupanc, 2009). However, the classic literature also includes reports about absent or aberrant reconstitution of the fish brain after injury (Kranz, 1975; Pflugfelder, 1954; Bernstein, 1967). This varied data suggests that many parameters, like the age, the species, the type and size of the injury and the affected brain area determine to which extent regeneration occurs in the teleost brain.

The requirement of reactive ventricular progenitor zones (also called matrix zones) for successful brain regeneration in bony fishes was elegantly demonstrated by a series of optic tectum lesion experiments in juvenile carps, Carassium carassium L. (Kirsche and Kirsche, 1961). The initial cellular reactions after injury are characterized by widespread traumatic degeneration of neuronal cells and processes. To determine if newborn cells derive from the constitutive ventricular progenitor zones, the authors performed lesions sparing different proportions of the “matrix zones”. Interestingly the authors find a direct correlation between the size of the remaining matrix zone and the degree of tissue architecture restoration (Kirsche and Kirsche, 1961). These experiments indicated the requirement of “matrix zones”—which we now refer to as progenitor zones, or ventricular zones, with proliferation capacity (Grandel et al.,2006; Kaslin et al.,2008)—for successful brain regeneration.

An interesting series of experiments analyzed the morphological restoration of the telencephalon of juvenile and adult guppies (Lebistes reticulatus) after dorsal stab lesions (Richter, 1970). In juveniles, reactive proliferation is seen after the lesion in the dorsal ventricular zone, coinciding with the elevated numbers of cells that migrate from the ventricle towards the inner parenchyma during the course of regeneration. After six months, the lesion cannot be detected anymore and the tissue architecture seems to be almost completely restored. In adult animals the regeneration process is very similar to what is seen in juveniles. However, early cellular degeneration processes and cerebral edema seem to be more severe and result in elevated cell death in adults compared to juveniles. Furthermore, the lesion canal is filled with newly generated connective tissue, probably derived from the meninges, from one month to one year after the injury. This study suggested that the teleost brain has the capacity to regenerate and that this capacity might be age-dependent (Richter, 1970; Kranz, 1975).

More recently, the cerebellum of the weakly electric fish Apteronotus leptorhynchus was introduced as a model to study regeneration after traumatic injuries in bony fishes (Zupanc et al.,1998, 2006; Zupanc, 2009). These studies suggested that following an injury, a cascade of physiological events take place such as apoptotic response, removal of dead cells, and formation of new neurons. Also, a proteomic analyses was performed to find the proteins involved in the regeneration process, and several housekeeping proteins and extracellular matrix components were found to be highly upregulated (Zupanc et al., 2006).

Although several studies have descriptively analyzed the regeneration response in teleost fish, there is very little information on the molecular and physiological aspects of the brain regeneration of adult zebrafish. Apart from few studies that identified and characterized some of the stem cell niches in adult fish brain (Adolf et al., 2006; Grandel et al.,2006; Chapouton et al.,2006; Kaslin et al.,2009; Ganz et al.,2010), we do not know many fundamental aspects of the progenitors involved in the adult neurogenesis and regeneration. For instance, the cell types that give rise to new neurons, regional differences in regenerative capacity of the brain and the molecular pathways involved in any of those processes are not known. To fill the gaps in our understanding of the adult neurogenesis and regeneration in the zebrafish brain, it is of great importance (1) to understand the molecular programs governing regeneration, and (2) to characterize in more detail the stem cell niches of the adult brain, (3) to determine the origin and identity of the neurogenic progenitor cells for instance by genetic lineage-tracing experiments, (4) to better define the physiological and molecular regeneration response by detailed histological analyses, and (5) to identify new regulators of progenitor cell activity. Already, several laboratories contributed significantly towards this understanding (Adolf et al., 2006; Chapouton et al.,2006; Grandel et al.,2006; Ninkovic and Götz, 2007; Pelegrini et al., 2007; Kaslin et al.,2008; Stigloher et al.,2008; Rieger et al., 2008; Kaslin et al.,2009; Brown and Brown, 2009; Berberoglu et al.,2009; Lam et al.,2009; Chapouton et al.,2010; Ganz et al.,2010; März et al., 2010; unpublished results); however, much remains to be done to elucidate the cellular and molecular basis of regeneration in adult zebrafish brain.

Regulation of the Progenitor Cell Activity

The complex nature of stem cells in the brain is also reflected in the signaling cascades deployed in animals (Panchision and McKay,2002; Alvarez-Buyylla et al., 2008; Suh et al.,2009). The large majority of our current knowledge on how the stem cells are regulated in terms of cellular behavior such as quiescence, proliferation, migration, survival, and differentiation in adult zebrafish brain comes from studies examining adult neurogenesis. In this section, we will discuss these findings in comparison to what is known in mammalian brains.

Notch Signaling.

In animals, Notch signaling regulates fundamental binary cell fate decisions of juxtaposing cells such as glia and neurons in the CNS (Campos-Ortega,1995; Artavanis-Tsakonas et al.,1999; Cau and Blader,2009; Fortini,2009). During neurogenesis, Notch receptor and its ligands Delta and Jagged have pleiotropic and context-dependent roles in regulating neuronal and glial differentiation (Fortini and Artavanis-Tsakonas,1993; Nye et al.,1994; Louvi and Artavanis-Tsakonas,2006). Also during zebrafish embryonic neurogenesis, Notch signaling is involved in fate decision mechanism of neural progenitors (Jiang et al.,1996; Haddon et al.,1998; Takke et al.,1999; Park and Appel,2003; Shin et al.,2007; Tallafuss et al.,2009).

A recent study examined the role of Notch in adult neurogenesis of the zebrafish telencephalon and suggested that the presence of Notch signaling is associated with keeping the progenitor cells quiescent (Chapouton et al.,2010). High levels of Notch signaling activity has been suggested to be required for maintaining the quiescent state of the progenitors in the telencephalon (Chapouton et al.,2010). Inhibiting Notch signaling by DAPT treatment increased the proliferation of the progenitors while enhancing Notch signaling by overexpressing Notch intracellular domain (NICD) which is a constitutively active form of the Notch receptor forced glial cells to reduce their proliferation rate (Chapouton et al.,2010).

Several Notch-target genes, bHLH transcription factors, are also expressed during neurogenesis in embryonic and adult zebrafish (reviewed in Stigloher etal., 2007). Although the roles of these factors in early neurogenesis have been studied (Takke et al.,1999; Wang et al.,2001; Geling et al., 2001,2004; Scholpp et al.,2009), their roles in adult neurogenesis and regeneration remain unexplored. However, the promoter region of the her4 gene (a presumptive orthologue of mammalian hes5) has been isolated and transgenic zebrafish line driving reporter expression under this promoter serves now as a useful tool to mark ventricular progenitor cells (Yeo et al.,2007; Ganz et al.,2010). This line is responsive to Notch signaling as it senses high-levels of Notch activity (Yeo et al.,2007; Ganz et al.,2010). This suggests that the radial glial cells have high levels of Notch signaling but not all of them are proliferative. Therefore, the relation between the active Notch signaling and the proliferation status of the radial glial cells is not completely clear. Further studies are needed to elaborate on the relation of Notch activity with proliferation at single-cell resolution to determine whether Notch signaling regulates progenitor behavior directly or indirectly.

Regarding Notch signaling, other aspects that remain unclear are whether the nonproliferative Notch-positive cells are permanently quiescent or if there are any fluctuations in Notch signaling to synchronize the progenitor cells similar to what is seen in the presomitic mesoderm during zebrafish embryogenesis (Holley,2007; Riedel-Kruse et al.,2007). Lineage tracing experiments using the conditional Cre-lox system for these quiescent cells in adult neurogenesis and regeneration may help addressing the long-term potential of those cells.

Growth Factor and Morphogen Signaling.

Several growth factors and morphogens have been shown to play roles on maintenance and regulation of stem cell proliferation in animals (Palmer et al.,1995; Kuhn et al.,1997; Machold et al.,2003; Tureyen et al.,2005; Palma et al.,2005; Jacobs et al.,2006; Pozniak and Pleasure,2006; Wang et al.,2007; Colak et al.,2008; Favaroe et al., 2009; Ma et al.,2009; Kuwabara et al.,2009; Wittko et al.,2009; Prajerova et al.,2010; Krampert et al.,2010; Mira et al.,2010; Zang et al., 2010). However, the current data about growth factor and morphogenic signaling in adult zebrafish brain are scarce, with some exceptions that we will discuss in the following subsections.

Fgf Signaling

Fgf signaling is an essential pathway for mitogenic activity (Paris and Pouysségur, 1991; Boilly et al., 2000; Wills et al.,2008). Two recent functional analyses showed the involvement of Fgf signaling as a major cue for cell proliferation in the progenitor regions of the cerebellum and the telencephalon (Kaslin et al.,2009; Ganz et al.,2010). In the cerebellum, Fgf signaling is required for proliferation of the stem cells. When heat-inducible transgenic line Tg(hsp:dnfgfr1-gfp) (Lee et al.,2005) is used to conditionally block Fgf signaling, the number of BrdU-incorporating cells are reduced in the cerebellar progenitor niche (Kaslin et al.,2009). In the adult zebrafish cerebellum Fgf signaling is primarily mediated through fgfr2 and fgfr3. Expression of downstream targets of the Fgf signaling pathway and tyrosine kinase receptor activity suggest that the most active Fgf signaling takes place in glia and granule cells. Although expression of fgfr2 is found in the progenitor niche markers for high levels of Fgf signaling activity are not found in the proliferating cells. This is in agreement with a recent study that suggests that Fgf activity is not associated with proliferation in the adult zebrafish brain (Topp et al., 2008). Interestingly, Fgfs (fgf3, fgf8a, and fgf8b) are expressed in the adult zebrafish cerebellum suggesting that diverse Fgfs are involved in the regulation of cerebellar progenitors and neurogenesis (Topp et al., 2008; Kaslin et al.,2009). Similarly in the telencephalon, Fgf activity regulates proliferation of adult telencephalic progenitors (Ganz et al.,2010). Fgf receptors are hetergenously expressed by ventricular glia in the telencephalon (Topp et al., 2008; Ganz et al.,2010). High levels of fgfr2 and lower levels of fgfr1 and fgfr3 are found in the telencephalic progenitor domains while fgfr4 expression mainly is confined to the dorsolateral parts of the telencephalon (Ganz et al.,2010). Blocking Fgf signaling in the telencephalon with the Tg(hsp:dnfgfr1-gfp) line leads to a reduction of BrdU-incorporation of the progenitors in the ventral telencephalon, while overexpressing Fgf8 increases mitogenic activity in the whole telencephalon (Ganz et al.,2010), suggesting a strong link between progenitor activity and Fgf signaling in the zebrafish telencephalon. The higher sensitivity to Fgf signaling of the progenitors in the ventral telenephalic domain correlates with the expression of Fgf ligands fgf3 and fgf8a and the Fgf downstream target genes dusp6, spyr2, and pea3 that is ventral telencephalic progenitor domain (Ganz et al.,2010). In agreement with this the ventral telencephalon in zebrafish depends on Fgf3 and Fgf8a signaling already during embryogeneis (Shanmugalingam et al., 2000; Furthauer et al., 2001), indicating that progenitors in this region are exposed to and might require higher levels of Fgf pathway activation. Taken together, this suggests that Fgf signaling is an important regulator of progenitor activity and neurogenesis in the ventral telecephalon of zebrafish (Ganz et al.,2010).

Bmp Signaling

Bmp signaling has been suggested to inhibit neurogenesis in adult mammalian brains and expression of Noggin—an antagonist of the Bmp—is required for promoting neurogenesis (Lim et al.,2000; Bonaguidi et al.,2005; Colak et al.,2008; Bonaguidi et al.,2008; Mira et al.,2010). Although several studies in zebrafish showed involvement of Bmp signals in regulating neural precursors in embryonic development (Barth et al.,1999; Guo et al.,1999; Londin et al.,2005; Jia et al.,2009), the roles of Bmp signaling in adult neurogenesis of the fish are uncharted. In our lab, we observed that ectopic expression of bmp2b showed a significant reduction of proliferation in the telencephalon, suggesting that Bmp signaling plays a role in negatively regulating telencephalic progenitor cell proliferation (Ganz,2009). Analyses of which cells receive Bmp signals, the proliferative status of those cells, and overlap of Bmp signaling with Fgf or Notch signals will be informative for refining the molecular architecture of the stem cell niches in the adult zebrafish brain.

Wnt Signaling

Previous studies in mammals indicated that Wnt signaling is required for progenitor cell proliferation, migration of precursor cells and the differentiation of neurons in the adult hypothalamus (Lee et al.,2000; Chenn and Walsh,2002; Lie et al.,2005; Wexler et al.,2009; Qu et al.,2010; Inestrosa and Arenas, 2010; Zhang et al., 2010). Consistent with the embryonic neurogenesis in the hypothalamus (Lee et al., 2006), the adult zebrafish hypothalamus contains high-levels of Wnt signaling activity (Dorsky et al.,2002; Wang et al., 2009). The zebrafish hypothalamus contains a broad zone of proliferating cells that are located in the vicinity of the cells with high-levels of Wnt signaling activity (Grandel etal.,2006). The cells that express high-levels of Wnt signaling activity in the hypothalamus do not co-localize with proliferating cells, Hu-C/D-positive neurons or GFAP-positive radial glia in the adult (Wang et al., 2009). One possibility is that these cells are tanycyte-like cells of the paraventricular organ (Kaslin et al.,2001; unpublished data). In addition, neurons in the optic tectum and reticular formation also show high-levels of Wnt signaling activity (Wang et al., 2009; unpublished data).

Thus, the regulation of progenitor cells of the adult zebrafish brain may not be dependent on canonical Wnt signaling, or Wnt signals emanating from endothelial cells or other cells with unclear identity might play a role. Additionally, the role of noncanonical Wnt signaling has not been addressed in zebrafish adult brain yet. Therefore, the regulation of progenitor cell activity by Wnt signaling pathways is a fertile research area for zebrafish brain during adult neurogenesis and regeneration.

Other Morphogens

Shh activity is associated with progenitor cell maintenance and proliferation in mammalian brains during adult neurogenesis and injury response (Bambakidis et al.,2003; Machold et al.,2003; Ahn and Joyner,2005; Han et al.,2008; Favaro et al.,2009; Prajerova et al.,2010). However, Shh signaling in adult zebrafish brain has not been documented. Since Shh signaling is required for development of dopaminergic and serotonergic neurons of the zebrafish brain (Lam et al.,2003; Holzschuh et al.,2003; Tang et al.,2010), investigating the role of this signaling pathway during regeneration in fish would be interesting for therapeutic applications for some human neurodegenerative disorders.

Retinoic acid (RA) is a differentiation cue for neuronal progenitors (Takahashi et al.,1999; Denisenko-Nehrbass et al.,2000; Crandall et al.,2004; Jacobs et al.,2006; Popa et al., 2011). However, so far, the role of RA signaling has not been investigated in adult zebrafish brain. Using conditional transgenic zebrafish with enzymes that modify the RA production such as raldh2 and cyp26a1 (Grandel et al.,2002; Begemann et al.,2004; Emoto et al.,2005) would be instrumental for investigating the requirements of RA for progenitor cell activity in zebrafish adult brain.

Hormonal Regulation.

Hormones as systemic factors regulate adult neurogenesis in different ways. They can positively or negatively affect the neurogenesis in adult brains. For instance steroid hormones increases formation of granule cells in the cerebellum and hippocampus (Bohn and Lauder,1980; McEwen,1996; Montaron et al.,2003), while thyroid hormone has a complex involvement in progenitor cell proliferation in the brain as it can both suppress and enhance it (Lemkine et al.,2005; Desouza et al.,2005; Kapoor et al.,2011). Other hormones such as gonadal hormone, prolactin or estrogen were suggested to have a positive effect on progenitor cell proliferation (Brown et al.,1993; Shingo et al.,2003; Barha et al.,2009; Kordower et al.,2010). Additionally, the knock-out mice for 11-β-hydroxysteroid dehydrogenase gene (hsd11b1)—an enzyme that catalyzes the conversion of the active glucocorticoids corticosterone and cortisol to inert products and therefore convert glucocorticoids to activated ligands that can bind to their receptors—shows long-term potentiation of hippocampal neurogenesis, suggesting a negative effect on neurogenesis by glucocorticoids (Seckl et al.,2002; Holmes and Seckl,2006; Seckl,1997).

In zebrafish, the gene cyp19a1b—an aromatase that converts androgens to estrogen (Trant et al.,2001)—is expressed in the ventricular cells of the adult zebrafish telencephalon with radial glial properties (Goto-Kazeto et al.,2004; Menuet et al.,2005; Pellegrini et al.,2007; Tong et al.,2009; Ganz et al.,2010; März et al., 2010a). Although the exact role and the cells responsive to estrogen are not clear due to lack of functional studies, these results suggest that estrogen signaling might have a role in regulation of progenitor cell activity in adult zebrafish brain.

Cytokine Signaling.

Cytokines are secreted molecules that are used for cell-to-cell communications in diverse cells and tissue types such as the immune system and the nervous system (Nakashima and Taga,2002; Thomson and Lotze,2003). In adult neurogenesis and regeneration, several cytokines have been shown to take part in diverse roles such as cell proliferation, modulation of neuroinflammation and cell survival (Das and Basu,2008; Ekdahl et al.,2009; Seuntjens et al.,2009; Molina-Holgado and Molina-Holgado,2010). Most of the cytokines are expressed in immune cells such as macrophages. One important cytokine for the nervous system is Tumor necrosis factor (TNF). TNF has a negative role on adult hippocampal neurogenesis (Liu et al.,2005; Iosif et al.,2006; Gonzalez-Perez et al.,2010). In zebrafish, the biological relevance of cytokines in the adult brain has not been addressed yet.

Cell migration is a hallmark of adult neurogenesis and regeneration in the brain, and therefore cues that regulate migration are important factors for these processes. Chemokines, as secreted molecules that influence the chemotactic migration of recipient cells, are expressed in a variety of cell types in the adult mammalian brains to modulate the neuron-neuron and neuron-immune system interactions (reviewed in Cartier et al.,2005; de Haas et al.,2007). One of the most studied chemokine signaling component is Cxcr4, the receptor for the Chemokine Sdf1/Cxcl12. It is required for neuronal migration, pathfinding and survival (Stumm and Hollt,2007; Liapi et al.,2008; Kolodziej et al.,2008; Bhattacharyya et al.,2008; Kokovay et al., 2010). In zebrafish, Cxcl12 and Cxcr4 are required for many processes such as lateral line migration (David et al.,2002; Gilmour et al.,2004), guidance of germ cells (Doitsidiou et al., 2002; Knaut et al.,2003), retinal development (Li et al.,2005), muscle formation (Chong et al.,2007), and forebrain development (Palevitch et al.,2010). Recently, Cxcr4 and Cxcl12 were suggested to be expressed in the radial glial cells of the adult zebrafish telencephalon (Diotel et al.,2010). However, based on the discrepancies between the immunohistochemistry and the in situ hybridization results the detailed identity of the Cxcr4-positive cells remains yet unclear.

Other Regulators of Neural Progenitor Cells.

Several other regulators of stem cells in the adult vertebrate brain have been implicated, such as Nitric oxide (Gibbs,2003; Moreno-López and González-Forero,2006), glutamate and γ-aminobutryic acid (Schlett,2006; Jagasia et al.,2006; Gakhar-Koppole,2008; Markwardt and Overstreet-Wadiche,2008; Mattson,2008; Platel et al.,2010), dopamine (Cave and Baker,2009), vascular endothelial growth factor (VEGF) (Jin et al.,2002; Sun et al.,2003; Zhu et al.,2003; Sun and Guo,2005; Wittko et al.,2009), pigment epithelium derived growth factor (PEDF) (Ramirez-Castillejo et al.,2006), glial-cell derived neurotrophic factor (GDNF) (Chen et al.,2005; Kobayashi et al.,2006; Wei et al.,2007), brain-derived neurotrophic factor (BDNF) (Benraiss et al.,2001; Pencea et al.,2001; Scharfman et al.,2005; Rossi et al.,2006; Chan et al.,2008; Lee and Son,2009), and epigenetic modifications by histone methylation (Ma et al.,2009), adding further degrees of complexity to the regulation of adult stem cells in the vertebrate brain. The roles of these regulators in zebrafish brain have not been addressed, which offers an interesting research realm.


The molecular mechanisms and the signaling programs underlying regenerative events in the zebrafish brain have barely been examined. Therefore, understanding such mechanisms is of utmost importance for shedding light onto the nature of zebrafish brain regeneration.

As we have discussed before, a successfully mounted regeneration response is likely to involve different steps: induction of progenitor cell proliferation, commitment and differentiation of progenitors, migration, survival, pathfinding, synaptogenesis and integration into the circuitry. Our knowledge on how these steps happen in time and space during regeneration, how they are regulated and which factors are important is quite limited. Yet, data from other organisms and from other organs of zebrafish suggest that different steps of the regeneration response are regulated in a highly complex and strict manner.

One hallmark of a regeneration response in the brain is the increase in cell proliferation. The proliferating cells are a combination of inflammatory cells and neural progenitors, which in a context-dependent manner have the capacity to give rise to new neurons and glia. Therefore, understanding how those progenitor cells react to injuries is an intriguing question because the difference in the regenerative potential of different animals may possibly lie in the initial mode of activation of the progenitors—a branching point that manifests itself in whether they will generate new neurons or will form glial scar tissue.

To understand the molecular programs required for stem cell activity in adult neurogenesis and regeneration, we performed a high-throughput gene expression analyses of progenitor cell populations in the telencephalon, diencephalon and the cerebellum using microarray technology (unpublished data). Since the progenitor cells are not too numerous in the brain, a pre-isolation of those cells was necessary. We used different transgenic zebrafish lines that mark progenitor cell populations with a GFP reporter: Tg(her4.1:GFP) that labels the radial glial cells, and Tg(nestin:GFP), which labels the neuroepithelial cells of the ventral telencephalon, diencephalons and the cerebellum. To add a further degree of comparison and separation, we also used a dye that labels DNA, which allowed separating cells that have doubled their amount of DNA, and hence are proliferating. After FACS sorting these cells we undertook a DNA microarray analysis which yielded a multitude of genes that are differentially expressed during homeostatic adult neurogenesis and brain regeneration in the progenitor cells that are proliferating or non-proliferating. This sorting technique is efficient to sort the desired population of cells, because the expression levels of known glial and neuroepithelial markers correlated well with the sorted cell populations (unpublished results). After verification of the genes differentially expressed in a statistically significant manner, we found many genes that are specifically localized to the progenitor cells during adult neurogenesis and regeneration (unpublished data). Interestingly, in the telencephalon, we found several genes that are significantly upregulated in the lesioned hemisphere but not to the same degree in the unlesioned hemisphere. Expression of some of these genes has not been previously reported in adult zebrafish brain (unpublished data). This general experimental approach is likely to yield many new genes that are involved in progenitor cell activity during adult neurogenesis and during the regeneration response.

The Cre-loxP system is a powerful tool for conditionally manipulating gene expression in specific cells of interest also in zebrafish (Hans et al.,2009; Hans et al., 2010; unpublished results). The combination of the sorting and gene expression profiling mentioned above with the Cre-loxP system should allow identification and functional studies of genes that are expressed in specific cells under specific conditions (such as proliferation, commitment, misexpression of a specific factor, cellular damage or ablation, etc.). Because the Cre-loxP system can be also used for lineage tracing, a designed combination of lineage tracing and expression analyses experiments may also give insights into the maturation programs of a progenitor cell after it stops proliferation. In the future, innovative technologies such as optogenetics (Del Bene and Wyart, 2012) will further increase our understanding of the molecular basis of progenitor cell activity in the adult zebrafish brain.


The interest in understanding to what extent the animals can regenerate and why some can do it better than others dates back to ancient times. The oldest known disguised reference to the effects of CNS injury and the failure of the human CNS to regenerate is the Edwin Smith surgical papyrus, an ancient egyptian note that dates back to approximately 1500-2500 BC (Breasted, 1980; Case, 2005). The long-known inability of the CNS in mammals stimulated a multitude of researchers to study regenerative events in a variety of species. Although many studies of the 19th and 20th century have opposing results and conflicting conclusions, most authors concurred that a sturdy regenerative response in the CNS (i.e.: the spinal cord in most studies) is found in many amphibians, reptiles and bony fishes (reviewed by Kirsche, 1965; Kaslin,2008). Recent studies that analyzed the regeneration response in different organisms more in detail confirmed that the capacity of amphibians, reptiles and of teleost fish to regenerate different organs and tissues, including the central nervous system, exceeds that of mammals by far (Tanaka and Ferretti,2009; Poss,2010).

Zebrafish offers a unique opportunity to apply molecular biology and genetics tools to studying the high regenerative capacity in a vertebrate, and therefore is a promising model organism for exploring the uncharted fields of molecular signaling of stem cells during adult neurogenesis and regeneration of the brain. However, our current understanding of adult neurogenesis and brain regeneration in nonmammalian vertebrates clearly lags behind the knowledge of the mammalian CNS, there are many quite intriguing questions that need answers. As a final section in this review, we will point to some of them, which are important for guiding the future research.

Stimulating Questions

What Is So Special About Zebrafish That It Can Regenerate Its CNS While We Mammals Cannot.?

As we pointed out before throughout the text, different steps of the regeneration response may have different characteristics between mammals and zebrafish. This can range from the availability and identity of the progenitor cells to the modes of activation of proliferation, from the role of inflammation after the injury to the permissiveness of the brain environment. At present, we cannot pinpoint the reasons for the distinctive regenerative properties of the zebrafish. Yet, one intriguing feature of zebrafish regeneration is the lack of overt scarring (Poss et al.,2002; Akimenko et al.,2003; Redd et al.,2004; Sanchez-Alvarado and Tsonis,2006). Mammals typically form persistent scars after lesion, which prevent a complete restoration of the injured tissue. One special form of this is the glial scar in the central nervous system, which is impenetrable to any axonal growth and tissue regeneration (Rolls et al.,2009; Sofroniew et al.,2009). This type of scar is formed by reactive astrocytes, which are not neurogenic in adult brain after injury, but can give rise to multipotent neurospheres in culture (Buffo et al.,2008; Costa et al.,2010; Heinrich et al.,2010). Indeed, Olig2 was identified as a non-permissive factor for neurogenesis by astrocytes in adult mammalian brain (Buffo et al.,2005). Similarly, a recent study also indicated that after injury, normally quiescent regions of the adult newt midbrain can be induced to generate neurons (Berg et al.,2010), suggesting a repressive environment for proliferation and neurogenesis under homeostatic conditions also in newt. These results suggest that in mammalian and amphibian adult brains a possible restrictive mechanism may be hampering the constitutive neurogenesis, while zebrafish can effectively circumvent this repression and thus has widespread neurogenic capacity throughout its life. Another possibility why zebrafish brain has a regenerative capacity might lie in the ability to induce proliferation of progenitor cells. Because in mammals, the major population of the astrocytes that react to injury is located in the parenchyma and they do not express high-levels of GFAP, Vimentin or S100β (Horner et al.,2002; Pekny and Nilsson,2005; Buffo et al.,2008; Robel et al.,2011). Although after the injury these mature astrocytes start expressing glial markers, they might not be affected by the stimulatory microenvironment of the stem cell niche, and therefore are not neurogenic (Buffo et al.,2008). However, zebrafish seems to be activating proliferation in cells that are at the ventricle and are hence subjected to the as yet-elusive cues that make adult neurogenesis and constitutive proliferation possible (Grandel et al.,2006; Adolf et al., 2006; Kaslin et al.,2009; Ganz et al.,2010). More detailed lineage tracing and genetic experiments will provide some answers to these questions. Understanding the molecular mechanisms and in particular, the signals that reprogram the progenitor cells in zebrafish brain to give rise to new cells rather than forming scar tissue after the injury response would be a way to explore the difference between zebrafish and mammals.

Is Zebrafish Activating Molecular Programs During Regeneration That Are Different From Developmental Programs.?

It is often assumed that regeneration may recapitulate the developmental programs. Although there is substantial evidence for this hypothesis, several studies have shown that during organ and tissue regeneration of the adult zebrafish, genes that are normally not present during the development of those organs can be expressed (Raya et al.,2003; Kizil et al.,2009; Stewart et al.,2009; Millimaki et al.,2010). We also identified a gene, the expression of which is injury-dependent in a specific region of the adult zebrafish brain (unpublished data). This suggests that zebrafish regenerative capacity may also involve the ability to activate special programs under extraordinary conditions of injury- and employ them to regenerate lost tissues. The notion of “injury-induced regeneration-specific molecular programs” may receive more support as we better understand the genes that are involved in the zebrafish regeneration response, and what these genes do following an injury in non-regenerating organisms.

Can We Use What We Learn From Zebrafish For Therapeutic Applications in Humans.?

Every year hundreds of thousands of people experience acute nervous system injuries or develop neurodegenerative diseases (Kandel et al.,2000; Flint Beal et al.,2005; Morimoto,2006). Therapies that would enable controlled nerve regeneration could help to restore nervous system function back to what it was before the tissue insult. Several methods, such as neurotrophic factor administration, transplantation of neural stem cells into adult tissues, gene therapy or RNA interference have been and are being tried (Gage et al.,1987; Beck et al.,1995; Aebischer and Ridet,2001; Corti et al., 2003; Gill et al.,2003; Storch et al.,2004; Forman et al.,2004; Thuret et al.,2006; Skowronsky et al., 2006; Gonzalez-Alegre,2007; Meyer et al.,2010). However, as yet these approaches have not yet yielded the desired results, mostly due to either unwanted side effects, or an inability to initiate formation of new neurons rather than anti-apoptotic effects or gliogenesis, or due to scarcity of transplantable cells, poor efficiency of transplantations, pathogenicity, immune activity, or ethical reasons. At this point, therapies aiming to initiate a regeneration response for neurons with the patients' own cells can be considered as an ultimate goal.

Zebrafish has an extensive ability to induce a regeneration response in all regions of the brain, and we can use this as a tool and advantage to analyze brain regions that are non-neurogenic in adult mammals. Findings can help us to answer fundamental questions such as the molecular programs dictating the stem/progenitor cells to proliferate and replenish lost neurons. Such studies have a great potential to help understand the regeneration biology of the vertebrate central nervous system, and to open up new avenues for designing therapies for human neurodegenerative disorders or acute neural injuries.


We gratefully acknowledge grant support by the DFG (SFB 655) and European Union (ZF Health) to M.B.