· web viewaxons from crossing (stenudd et al., 2015). in anamniotes, progenitor cells are...

102
TITLE: The spinal ependymal zone as a source of endogenous repair cells across vertebrates AUTHORS: Catherina G. Becker *1 , Thomas Becker *1 , Jean-Philippe Hugnot *2,3 ADDRESSES : 1-Centre for Discovery Brain Sciences, University of Edinburgh, Biomedical Sciences, The Chancellor’s Building, 49 Little France Crescent, Edinburgh EH16 4SB, UK 2-INSERM U1051, INM, Hopital Saint Eloi, 80 avenue Augustin Fliche, 34091 Montpellier, France 3-Université de Montpellier, Place Eugène Bataillon, 34095 Montpellier, France * equal contributions, listed alphabetically; correspondence to [email protected] , 1

Upload: others

Post on 13-Jan-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

TITLE: The spinal ependymal zone as a source of endogenous repair cells

across vertebrates

AUTHORS:

Catherina G. Becker*1, Thomas Becker*1, Jean-Philippe Hugnot*2,3

ADDRESSES :

1-Centre for Discovery Brain Sciences, University of Edinburgh, Biomedical

Sciences, The Chancellor’s Building, 49 Little France Crescent, Edinburgh

EH16 4SB, UK

2-INSERM U1051, INM, Hopital Saint Eloi, 80 avenue Augustin Fliche, 34091

Montpellier, France

3-Université de Montpellier, Place Eugène Bataillon, 34095 Montpellier,

France

* equal contributions, listed alphabetically; correspondence to

[email protected], [email protected], jean-

[email protected]

1

ABSTRACT:

Spinal cord injury results in the loss of neurons and axonal

connections. In mammals, including humans, this loss is permanent, but is

repaired in other vertebrates, such as salamanders and fishes. Cells in the

ependymal niche play a pivotal role for the outcome after injury. These cells

initiate proliferation and generate new neurons of different types in

regenerating species, but only glial cells, contributing to the glial scar, in

mammals. Here we compare the cellular and molecular properties of

ependymal zone cells and their environment across vertebrate classes. We

point out communalities and differences between vertebrates capable of

neuronal regeneration and those that are not. Comparisons like these may

ultimately lead to the identification of factors that tip the balance for

ependymal zone cells in mammals to produce appropriate neural cells for

endogenous repair after spinal cord injury.

2

1. INTRODUCTION:

After spinal injury or in neurodegenerative diseases, lost neurons are

not replaced in mammals, including humans. In anamniotes, particularly fishes

and salamanders, this is not the case. These species readily regenerate

neurons (Grandel and Brand, 2013; Becker and Becker, 2015; Alunni and

Bally-Cuif, 2016; Cardozo et al., 2017; Tazaki et al., 2017). Newly generated

neurons may replace those that are lost after an injury. This is significant,

because secondary neuron loss after spinal injury can be quite extensive in

mammals (Park et al., 2004). Moreover, in anamniotes, new neurons can

contribute to an axonal bridge that reconnects the spinal cord, potentially

acting as relay neurons for the injured spinal cord (Goldshmit et al., 2012). In

mammals, embryonic or induced neural progenitor cells transplanted into the

lesioned spinal cord can generate relay neurons. This leads to promising axon

growth from these cells and some functional recovery (Lu et al., 2012; Lu et

al., 2014). However, network integration of these cells is limited and immune

reactions to transplants, as well as tumour formation by transplanted cells

present additional problems to overcome (Sharp et al., 2014). Therefore,

finding and reprogramming endogenous spinal stem cells for neurogenesis

could be of great benefit to repairing spinal cord lesions. Endogenous spinal

stem cells exist in the adult mammalian spinal cord around the central canal

(Weiss et al., 1996). These cells start to proliferate after a lesion, but in situ

environmental factors prevent them from generating neurons and

oligodendrogliogenesis is rare. Instead, these cells generate mainly

astrocytes that contribute to a scar in the lesion site that prevents axons from

3

crossing (Stenudd et al., 2015). In anamniotes, progenitor cells are likewise

located at the central canal, proliferate after a lesion, express similar genes as

in mammals, but contribute to a very different regenerative outcome, namely

neurogenesis and functional spinal cord repair (Becker and Becker, 2015).

For these reasons, it is worthwhile to undertake a comparative description of

the importance of cells in the ependymal zone between regenerating and non-

regenerating vertebrate species. However, other cell types in the

parenchyma, such as oligodendrocyte progenitor cells and astrocytes that

react to a lesion with proliferation may also have stem cell potential, which is

reviewed elsewhere (Almad et al., 2011; Dimou and Gotz, 2014; Gotz et al.,

2015; Magnusson and Frisen, 2016).

Here we discuss mainly the zebrafish as a regenerating species, and

compare it to different mammalian species, namely rodents, monkeys, and

humans. We present evidence that ventricular cells across vertebrates are a

major source of new cells after injury and we describe some interesting

similarities between the likely spinal stem cells in species capable and not

capable of lesion-induced neurogenesis, in terms of cell morphology,

molecular signals, and transcription factors that are activated in these cells

after a lesion. We believe that from comparing the cellular and molecular

reactions of the spinal ventricular zone between anamniotes and mammals,

important principles of regenerative neurogenesis may be gleaned that could

be used in future experiments to boost neurogenesis from endogenous neural

stem cells in the injured spinal cord of mammals.

4

2. The spinal ependymal zone (EZ) in zebrafish and other species

showing regenerative neurogenesis in successful spinal cord regeneration

The spinal cord of species showing regenerative neurogenesis, such

as fishes and salamanders, consists of peripheral white matter tracts of large

diameter myelinated axons, ascending sensory axons that project dorsally

and descending axons that are located ventrally. Dorsal and ventral grey

matter areas are equivalent to dorsal and ventral horns in mammals and the

spinal cord contains motor neurons and a high diversity of interneurons,

including central canal contacting neurons, such as GABAergic Kolmer-

Agduhr cells that are located in the ependymal layer (Djenoune and Wyart,

2017). Other resident cell types are oligodendroglial cells, microglial cells, and

blood vessel-associated cells. A similar layout is found in all vertebrate

classes.

However, whereas the mammalian spinal cord contains ependymal

cells around the ventricle and free astrocytes in the parenchyma, fish and

salamander spinal cords contain only one astroglia-like cell type, sometimes

called “radial glia” or “ependymoglia”. In zebrafish, we call this cell type

ependymo-radial glia (ERG) (Becker and Becker, 2015), as a way to describe

that these cells have their soma at the central canal, thereby forming the

ependymal layer, and have radial processes that span the entire thickness of

the spinal cord to form endfeet at the pia. ERGs possess 1 to 2 cilia, as shown

by light and electron microscopy (Hui et al., 2015). These are likely to be

motile, as there is a constant flow of cerebrospinal fluid (Grimes et al., 2016)

and most ERGs express the transcription factor foxj1, which is involved in

motile cilia generation (Ribeiro et al., 2017). Hence, ERGs must fulfil a

5

number of homeostatic functions of ependymal cells and astrocytes. Much like

those of astrocytes, most ERG processes are positive for the glial fibrillary

acidic protein (GFAP), the glutamate transporter GLAST, brain lipid binding-

protein (BLBP), are highly branched and thus in a position to share the

homeostatic functions of astrocytes (Hui et al., 2015). Their ciliated

ependymal somata may sense signals from the central canal milieu. It has

been proposed that ERGs may not be a physically sustainable cell type in

species with thicker parenchyma, such as mammals (Reichenbach and

Wolburg, 2013). This correlates with the emergence of free astrocytes from

radial glial cells, which are only present during development in mammals.

2.1 Ependymo-radial glial cells as promoters of axonal regeneration

ERGs represent a unique cell type that also has multiple roles after

injury. An injury to the spinal cord in zebrafish leads to proliferation of ERGs in

the ependymal layer (Reimer et al., 2008; Ogai et al., 2014; Ribeiro et al., 2017),

again similar to mammals (Meletis et al., 2008). One role of new glial cells near

the lesion site is presumably to seal the blood-brain barrier, as it has been

shown that this proliferation of astrocytes (astrogliosis) is beneficial for lesion

outcome also in mammals (Sabelström et al., 2013; Anderson et al., 2016). In

mammals, astrogliosis is often considered inhibitory for the regeneration of

axons over the lesion site (Lang et al., 2014). In contrast, in zebrafish,

processes of GFAP+ astrocyte-like cells, derived from ERGs, elongate across

the lesion site and fasciculate with regenerating axons to reconnect the spinal

cord (Goldshmit et al., 2012; Mokalled et al., 2016; Wehner et al., 2017). Hence,

ERG-derived glia in zebrafish is permissive to axon regeneration by not

6

forming a barrier as in mammals. In addition, glial processes that connect the

two ends of the severed spinal cord, termed “glial bridge” (Goldshmit et al.,

2012) could provide a scaffold for regenerating axons. A number of

experimental studies show a correlation between disruption of the glial bridge,

axon crossing and functional recovery in zebrafish. For example, ERGs and

lesion site cells produce connective tissue growth factor a (ctgfa). Disrupting

ctgfa expression in a mutant impaired glial bridging and axonal regeneration,

which could both be restored by global over-expression of ctgfa or applying

recombinant human CTFG (Mokalled et al., 2016). ERGs/lesion site glia also

produce and are responsive to fibroblast growth factor (Fgf). Inhibition of Fgf

signalling disrupted both glial bridging and axon regrowth. Moreover,

experimentally increasing FGF pathway activity accelerated glial and axonal

regeneration (Goldshmit et al., 2012). Finally, activity of the Wnt pathway has

been reported in ERGs of lesioned larval and adult zebrafish and inhibition of

the pathway inhibited both glial bridging and axon regrowth (Briona et al.,

2015; Strand et al., 2016).

However, in larval zebrafish, live observations indicate dynamic growth

of axons into the lesion site, mostly independently of glial processes (Wehner

et al., 2017). Moreover, “axonal bridging” (continuity of axon labelling between

the spinal cord ends) is established earlier than glial bridging. Importantly,

when GFAP+ glia, and therefore most of the glial bridge, was conditionally

ablated, axonal bridging was unimpaired (Wehner et al., 2017). Similarly, spinal

cord transection in the adult eel led to bridging of the gap by glial processes

and axons, with axons “always the most rostral component in the bridge”

(Dervan and Roberts, 2003b; Dervan and Roberts, 2003a). These observations

7

suggest that the relative importance of the glial scaffold function may depend

on species and developmental stage of fishes. The glial scaffold function

could be particularly important in adult fish, in which distances that axons

have to bridge are larger than in larvae. In addition, ERG-derived growth

factors, such as Ctgfa and Fgf could promote axon regeneration directly. To

further elucidate the interdependency of axonal and glial bridging in the future,

there-dimensional whole-mount analyses of the adult lesion site using tissue

clearing methods, in combination with cell type-specific manipulations of the

above pathways will be informative.

It has to be kept in mind that a spinal lesion site shows high cellular

complexity. For example, cell type specific manipulation of the Wnt pathway

using the TetON system indicated that Wnt activity is most important in lesion-

site fibroblasts to induce deposition of regeneration-promoting Col XII.

(Wehner et al., 2017). How ERGs interact with important non-neural lesion

site cells remains to be elucidated.

2.3 Ependymo-radial glial cells as progenitor cells

Besides roles in re-establishing spinal cord integrity and continuity,

ERGs function as progenitor cells for neurons after a lesion. In mammals,

progeny of ependymal cells is restricted mostly to glial fates (Barnabe-Heider

et al., 2010). In contrast, early observations of histological preparations

indicated that the injured spinal cord of fishes is capable of regenerating

neurons from the ependymal zone (Kirsche, 1950). After a lesion of the spinal

cord in adult zebrafish, birth of new motor neurons, serotonergic interneurons,

V2-like interneurons, and other interneuron cell types has been observed

8

(Reimer et al., 2008; Kuscha et al., 2012a; Kuscha et al., 2012b). These cell

types are likely derived from ERGs based on the following evidence:

Proliferating cells at the central canal vastly outnumber parenchymal

proliferating cells and different types of neurons show wedge-shaped

distributions with the tip of the wedge in specific dorso-ventral positions of the

ependymal layer (Reimer et al., 2008). For example in the ventral spinal cord,

new serotonergic neurons are generated and serotonin-positive neurons can

be found within the ependymal zone during regeneration, but never in

unlesioned spinal cords (Kuscha et al., 2012a). In the case of motor neurons,

specific ventricular ERGs are labelled for both GFP, driven by the regulatory

sequences of the motor neuron progenitor gene olig2, and the transcription

factor mnx1 (also known as Hb9) during regeneration, but never in the

unlesioned spinal cord. A likely explanation for this double-labeling is that

differentiating motor neurons express the early marker mnx1, while they retain

the relatively stable GFP protein from the time the ventricular cell was still a

progenitor (Reimer et al., 2008). In this case, GFP protein acts as a short-term

lineage tracer. In injured larval zebrafish, genetic lineage tracing using

inducible Cre recombinase under the GFAP promoter, an ERG marker, was

used to demonstrate that HuC positive neurons are generated from spinal

ERGs (Briona et al., 2015). This matches similar experiments from other CNS

areas (Kroehne et al., 2011). In summary, ERGs are the only known source

for neurons after injury to the spinal cord so far.

9

2.4 Ependymo-radial glial cells retain their neural tube identities from

development

10

The observation that different cell types appear to derive from different

dorso-ventral positions of the ependymal layer is reminiscent of the

developing spinal cord, in which ventral and dorsal morphogen gradients set

up discrete progenitor domains that express different combinations of

transcription factors in the neural tube (Ferg et al., 2014). Indeed, though

apparently of homogeneous morphology, adult ERGs express low levels of

domain-defining transcription factor combinations along the dorso-ventral axis

in the intact late larval and adult spinal cord (Reimer et al., 2009; Briona and

Dorsky, 2014) (Fig. 1). The ERGs at the ventral midline express sonic

hedgehog (shh), a ventralising morphogen derived from the embryonic

floorplate. Adjacent domains express other transcription factors, similar to

those expressed during development of the ventral spinal cord, e.g. nkx6.1,

dbx1, as well as olig2 as the single defining transcription factor of the motor

neuron progenitor domain (Fig. 1). pax6 is not expressed in the very ventral

spinal cord, but is expressed around the rest of the central canal. This is a

difference to the developing neural tube, where pax6 is also not present in the

most dorsal region (Wilson and Maden, 2005). A likely explanation for this

discrepancy between the dorsal aspects of the developing and adult spinal

cord is that during conversion of the primitive neural tube lumen to the central

canal, dorsal regions of the neural tube are not included in the central canal

ependyma (Kondrychyn et al., 2013). Consistent with this idea, we failed to

detect markers for dorsal progenitor identity, such as pax7 and the dorsal

morphogens bmp 2 and 4 in the adult zebrafish spinal cord (Kuscha et al.,

2012b). In contrast, ERGs in the dorsal spinal cord of salamanders are

immuno-positive for Pax7 (Schnapp et al., 2005). Adult ERGs appear to

11

display a degree of lineage restriction, which depends on their developmental

dorso-ventral position of origin (Fig. 1). Therefore, the possibility exists that in

adult zebrafish, neuronal cell types that during development are derived from

very dorsal ventricular cells are not regenerated. Dorsal neuronal cell types

are mostly sensory and lack of regeneration of these could contribute to the

previously reported failure of re-innervation of sensory axon targets after

injury (Becker et al., 1997; Becker et al., 2005). Of note, after tail amputation,

salamanders regenerate the entire tail, including a fully reformed spinal cord

(Schnapp et al., 2005). This is also the case in the weakly electric fish

Sternarchus albifrons (Anderson and Waxman, 1981). Lineage analysis in

axolotls by grafting genetically labelled tissue early in development has shown

that the regenerated spinal cord is formed by cells of the pre-existing cord,

with clonal analysis showing plasticity of progenitors between dorsal and

ventral fates (McHedlishvili et al., 2007). However, it is not known whether

regenerated salamander or fish spinal cords contain the full complement of

(dorsal) neurons.

After a transection lesion in zebrafish, the spinal cord fuses at the

lesion site in an incomplete physical structure consisting mostly of myelinated

and unmyelinated axons crossing the lesion site (Becker et al., 1997).

However, the central canal adjacent to the lesion site widens considerably, a

feature shared with other regenerating and non-regenerating species

(Cardozo et al., 2017). Concomitant with this widening, progenitor domains

expand. At the same time, expression levels of signals, e.g. shh mRNA, and

transcription factors (nkx6.1, olig2, pax6), found at low levels in ERGs of the

unlesioned spinal cord, increase (Reimer et al., 2008). However, the spatial

12

relationships between expression domains of these factors are comparable in

the developing neural tube, the adult EZ, and the lesioned EZ. In other words,

adult ERGs are likely the direct progeny of the early neuroepithelial cells in

the same relative positions. Whether ERGs intrinsically retain their positional

identity and potential to generate specific neuronal cell types or whether this

has to be maintained by constant low-level homeostatic signalling, for

example by Shh (see below), remains to be determined.

All ERGs could represent spinal neural stem cells. Alternatively, true

stem cells may generate specific ERGs that act as transit amplifying cells and

are then lineage-restricted. ERGs express the neural stem cell marker sox2

and other stem cell-related markers, such as pou5f1 (also known as oct4),

which is essential for their lesion-induced proliferation (Ogai et al., 2014; Hui

et al., 2015). Evidence for the presence of distinct stem cells in the ventricular

zone comes from label-retaining experiments with the base-analog BrdU.

These experiments indicated that the ventricular zone of the lesioned spinal

cord contains a subpopulation of cells that retain BrdU labelling for a long time

after injury and are therefore slow-proliferating cells (Reimer et al., 2008).

Slow proliferation is one defining criterion for neural stem cells (Johansson et

al., 1999). Whether these potential stem cells are also lineage-restricted is

unknown.

ERGs may show some flexibility in their potential to generate different

cell types. For example, the pMN-like progenitor domain starts to generate

motor neurons during early development, followed by oligodendrogliogenesis

during later development (Ravanelli and Appel, 2015). Upon a spinal lesion in

adult zebrafish, this domain reverts back to generating motor neurons (Reimer

13

et al., 2008). In fact, we could demonstrate that at late larval stages, a spinal

lesion leads to re-initiation of motor neuron generation at the expense of

developmentally on-going oligodendrogliogenesis from the pMN-like domain

(Ohnmacht et al., 2016).

2.5 Signals in regenerative neurogenesis

A mechanical lesion of the spinal cord elicits regenerative proliferation

in ERGs (Reimer et al., 2008; Ogai et al., 2014; Hui et al., 2015; Ribeiro et al.,

2017). Genetically ablating motor neurons in the absence of a mechanical

lesion has the same effect. Both types of injury induce an immune response,

indicated by increased presence of microglial cells and macrophages (Becker

and Becker, 2001; Ohnmacht et al., 2016), important mediators of the

regenerative response. Indeed, inhibiting the innate immune response using

the drug dexamethasone, reduced microglia activation in larval zebrafish and

reduced numbers of regenerated motor neurons after injury (Ohnmacht et al.,

2016). It has been shown that in the stab-lesioned telencephalon of zebrafish,

the immune response is necessary and sufficient to induce regenerative

neurogenesis by a mechanism that involves Leukotriene C4 and its cognate

receptor Cysteinyl leukotriene receptor 1, which is present on progenitor cells

(Kyritsis et al., 2012). It is not known whether spinal ERGs receive the same

or different signals from the immune system.

Recently, a role for the adaptive immune system has been described

for zebrafish spinal cord regeneration. Regulatory T cells (Tregs) accumulate in

the adult spinal cord after injury and show secretion of Neurotrophin 3 (Nt3).

This is an organ-specific reaction, as Tregs secrete other factors in the injured

14

retina or heart. Lack of Tregs impaired ERG proliferation, neurogenesis, axon

regrowth and behavioural recovery, whereas application of recombinant Nt3

rescued proliferation and neurogenesis (Hui et al., 2017). This demonstrates a

direct promoting influence of Tregs on regenerative neurogenesis.

Spinal ERGs also react to classical developmental signals during

regeneration. For example, Shh, derived from ventral ERGs, stimulates

generation of serotonergic and motor neurons. This has been shown by

blocking Shh signalling with the plant-derived small molecule inhibitor

cyclopamine (Reimer et al., 2009; Kuscha et al., 2012a). In the future it will be

interesting to determine how this ventrally derived signal reaches the more

dorsal progenitor cells in the adult spinal cord, which is larger and more

complex than the developing neural tube, in which Shh acts in a comparable

fashion.

Notch signalling is a developmental short-range signal involved in

determining fate of adult stem cells and maintenance of “stemness” in

zebrafish (Alunni and Bally-Cuif, 2016). Ligands (deltaC, jagged1b), receptors

(notch1a/b) and down-stream effectors (her4.1, her4.5, her9) are strongly

upregulated from undetectable levels after an injury of the adult spinal cord.

These genes are expressed in partially non-overlapping domains of ERGs

and may reflect different notch activities in different progenitor domains.

Functional analyses using the gamma-secretase inhibitor DAPT to inhibit

notch signalling and heat-shock inducible over-expression of the active

intracellular domain of the notch receptor have shown that notch activity

attenuates lesion-induced neurogenesis of motor neurons (Dias et al., 2012).

15

In the unlesioned spinal cord, manipulations of notch had no effect, as

opposed to different brain regions, where notch inhibition leads to ERG

proliferation (Alunni et al., 2013; de Oliveira-Carlos et al., 2013). Thus, the

notch pathway is probably not constitutively active in the zebrafish spinal cord.

This indicates that adult ERGs in the spinal cord are not only different from

constitutively active ERGs in other CNS regions, but also from other quiescent

ERGs, inasmuch as quiescence in spinal ERGs is not dependent on notch

signalling.

Wnt signalling regulates proliferation of spinal ERGs after injury, as the

pathway is active in some ERGs and pharmacological or genetic global

interference with the Wnt pathway inhibited neurogenesis in injured larvae

(Briona et al., 2015). The effect on some ERGs could be indirect, as only a

minority of ERGs show pathway activity after injury in larval zebrafish (Wehner

et al., 2017). Future research will have to elucidate source and target cells of

Wnt signalling.

FGF signalling is upregulated in ERGs of the lesioned spinal cord, as

indicated by increased expression of the FGF receptor 2 and target genes

pea3, erm and spry4 in these cells. FGF controls proliferation of ERG-derived

glia, as shown by global pharmacological inhibition and over-expression of a

dominant-negative FGF receptor variant. Conversely, in a mutant for the FGF

downstream gene and inhibitor spry4, proliferation was enhanced. The source

for FGF ligands are ERGs themselves and neurons (Goldshmit et al., 2012). It

has not been analysed whether neurogenesis was altered after FGF

manipulations, but given the influence on proliferation this seems likely. Other

well-known developmental pathways, such as retinoic acid signalling,

16

components of which are upregulated after a lesion (Reimer et al., 2009), or

BMP signalling have not been functionally analysed.

Descending dopaminergic and serotonergic axons promote

regeneration of motor neurons, as shown by a reduction in motor neuron

regeneration following pharmacological ablation of serotonergic (by 5,7-

dihydroxytryptamine) or dopaminergic axons (by 6-hydroxydopamine). Under

these conditions, ERG proliferation and motor neuron regeneration was

impaired. ERGs express dopamine (drd4a) and serotonin receptors (hrt1aa

and hrt3a) (Reimer et al., 2013; Barreiro-Iglesias et al., 2015). Due to the fact

that almost all dopaminergic and serotonergic innervation of the spinal cord is

derived from the brain, this effect is only seen rostral to a lesion site. That is

because caudal to the lesion site, dopaminergic and serotonergic axons

degenerate by Wallerian degeneration even without toxin mediated ablation.

Conversely, addition of dopamine agonists or serotonin augments motor

neuron regeneration caudal to a lesion site that is normally deprived of these

molecules (Reimer et al., 2013; Barreiro-Iglesias et al., 2015). Both signals

also enhance developmental motor neuron generation, such that sensitivity to

dopamine and serotonin during regeneration can be considered a

recapitulation of a developmental mechanism. Interestingly, neither serotonin

nor dopamine injections into unlesioned animals led to any motor neuron

generation. Hence, while these signals modulate the regenerative

neurogenesis programme, they are insufficient to elicit regenerative

neurogenesis from quiescent spinal ERGs.

2.6 Gene expression changes in Ependymo-radial glial cells

17

The transitions from quiescence to proliferation and to neurogenesis

necessarily lead to changes in the gene expression programmes in ERGs.

Expression profiles of a spinal lesion site have revealed a number of

transcription factors and other genes that change expression after injury, such

as those involved in the inflammatory response, cell proliferation,

neurogenesis, patterning, and axon outgrowth-related genes (Hui et al.,

2014). Other genes found to be upregulated in spinal ERGs after injury are

HMGB1, syntenin-a, contactin-2, major vault protein, sox11b, mcam, and L1.2

(Guo, Y. et al., 2011; Lin et al., 2012; Pan et al., 2013; Yu and Schachner,

2013; Fang et al., 2014; Chen et al., 2016; Liu et al., 2016). Functional

studies, relying on anti-sense morpholino application to a spinal lesion site,

have found beneficial roles for regeneration of these genes, which might have

been due to morpholino uptake into axotomised neurons, ERGs or both.

Specific gene manipulation in ERGs will be needed to discern a role for these

genes in ERG function. We anticipate that expression profiling of highly

purified progenitor cell populations in combination with cell type specific

manipulations, e.g. using the TetON system (Wehner et al., 2017) or

conditional CRISPR interference (Yin et al., 2016) will clarify the contribution

of specific genes and pathways in ERGs to functional spinal cord regeneration

in the future.

3. The spinal ependymal zone in mammals

3.1 Rodents

18

3.1.1 Ependymal zone composition and regionalisation

The rodent ependymal zone (EZ) is a pseudo-epithelium originating

mainly from the ventral part of the developing neuroepithelium, as the dorsal

walls of the neuroepithelium fuse at approximately E15 in the mouse (Fu et

al., 2003). Cells from the roof plate in the developing spinal cord may also

contribute to dorsal EZ cells (Hugnot, unpublished) (Sevc et al., 2009). The

formation of the EZ zone is controlled by Shh signalling, as indicated by the

observation that in the absence of late Shh production it is not formed (Yu et

al., 2013). In the caudal part of the spinal cord, the EZ might be derived from

a secondary neurulation process through a mesenchymal-neuroepithelial

conversion (Lowery and Sive, 2004). In adults, the cervical EZ has a wide

lumen while in the thoracic and lumbar regions, the lumen is reduced to a slit

(Sturrock, 1981). The EZ contains different types of highly-polarized

ependymocytes which can be distinguished both morphologically and using

marker expression. As in zebrafish, dorsal and ventral midline cells are

characterized by a very long radial morphology. These cells maintain long

filament-rich processes extending to the pia and are sometimes referred as

tanycytes (Seitz et al., 1981; Sturrock, 1981; Mothe and Tator, 2005;

Rodriguez et al., 2005). At the lateral level, ependymocytes are either cuboid

or show a radial morphology with a process ending on a vessel. Compared to

the ependymocytes in the subventricular zone, in the spinal cord, these cells

are not multiciliated but mostly feature two motile cilia (occasionally one, three

or four cilia can be found) associated with large basal bodies but without

daughter centriole (Alfaro‐Cervello et al., 2012).

19

Two other non-ependymocyte cell types are also encountered in the EZ

region. 1/ cerebral-fluid contacting neurons (CSF-N) are cells equivalent to

Kolmer-Agduhr cells in zebrafish. These neurons have been observed in

many species, including lamprey and macaques. In rodents they are present

at the different levels of the spinal cord (Vigh et al., 2004; Djenoune et al.,

2014). Two types of CSF-N, lateral and ventral CSF-N, with different

developmental origin and properties have been described (Petracca et al.,

2016). These neurons have immature electrophysiological properties and

express Dcx, a marker expressed by immature neurons during development

and in adult neurogenesis (Shechter et al., 2007; Marichal et al., 2009;

Sabourin et al., 2009; Orts-Del’Immagine et al., 2014; Orts-Del'Immagine et

al., 2017). They also maintain expression of transcription factors found in the

developing spinal cord, such as Nkx6.1, Nkx2.2, FoxA2, Gata2/3 (Sabourin et

al., 2009; Petracca et al., 2016). However, CSF-N are not produced by

ongoing adult neurogenesis, but instead are produced late during

development from two ventral domains of the spinal cord (Kútna et al., 2014;

Petracca et al., 2016). These neurons express functional P2X2 ATP channels,

as well as high levels of the polycystic kidney disease-like channels (PKD2L1

and PKD1L2), which are very good markers for these cells (Huang et al.,

2006). CSF-N are mostly GABAergic and send axons toward the caudal part

of the spinal cord (Stoeckel et al., 2003). Their role in mammals remains

enigmatic, but they express ion channels responsive to pH, osmolarity, and

mechanical stimulation. This suggests that they are involved in the monitoring

of CSF composition and movement. These cells send a process into the

lumen and this process is terminated by a large vesicle-containing structure

20

called budge (several µm in size). Presence of this structure indicates the

potential for an intense secretory activity into the CSF. The role of these cells

in the control of cell proliferation in the adult ependyma remains unexplored.

In other CNS niches, GABA signalling controls neural stem cell quiescence

and proliferation (Alfonso et al., 2012; Pontes et al., 2013). Thus GABAergic

CSF-N may serve a similar function in the spinal cord. 2/ the second type of

non-ependymocyte cell found in the EZ region are central-canal contacting

astrocytes (Accs) which can be observed using GFAP-GFP reporter mice

(Sabourin et al., 2009), electron microscopy (Alfaro‐Cervello et al., 2012) and

marker expression (Sabourin et al., 2009; Fiorelli et al., 2013). These

infrequent cells, are mostly situated in the dorsal and ventral parts of the EZ

(80% of cells) and unlike ependymocytes, they have a single cilium with a

daughter centriole. In some cases, Accs send long processes along the

ventral and dorsal midline fissure (i.e. a fold of the pia mater into the

parenchyma) and appear to reach the pia or midline-located pial-lined

interstitial perivascular spaces that are fluid-filled (also known as Virchow-

Robin spaces). Accs represent a heterogeneous population, as only some of

these cells contain Nestin and Vimentin-containing intermediate filaments

(Alfaro‐Cervello et al., 2012). Accs do not proliferate (Alfaro‐Cervello et al.,

2012) or at very low levels (Fiorelli et al., 2013). However, Nestin+ proliferating

cells have been reported in the dorsal region of the EZ (Hamilton et al., 2009),

suggesting cellular heterogeneity in this part of the EZ.

With regard to marker expression, most ependymocytes have a Nestin-

Vimentin+, S100b+, CD24+, EphrinB1+, Sox2+, Sox9+, and CD133+ phenotype

(Sabourin et al., 2009; Pfenninger et al., 2011) and weakly express GFAP

21

(Alfaro‐Cervello et al., 2012). In comparison, radial midline glia cells located at

the roof and floor of the EZ, show lower Vimentin/CD24 expression and higher

levels of GFAP expression (Alfaro‐Cervello et al., 2012). Further marker

analysis has revealed that the adult EZ is in fact a mosaic of cells that show

different developmental gene expression, notably along the dorsal-ventral and

rostro-caudal axes. Almost all EZ cells express the ventral transcription factor

Nkx6.1 (Fu et al., 2003; Sabourin et al., 2009), but Pax6 is expressed only in

the dorsal part of the EZ (Yu et al., 2013) (Hugnot unpublished) and Nato3 is

expressed only in the ventral region (Khazanov et al., 2016). Nato3 is a

transcription factor found in the floor plate during development. Data from the

gene atlas and transgenic mice confirm that genes that during development

are specifically expressed in the dorsal EZ, e.g. MDK (Hugnot and Franzen,

2011), or its ventral part, e.g. Nkx2.2 (Yu et al., 2013) or Shh (Hugnot and

Franzen, 2011), are maintained at the adult stage. Radial cells found in the

roof of the EZ express the Ret receptor tyrosine kinase (Pfenninger et al.,

2011), as well as Zeb1, a transcription factor regulated by TGF-β signalling

(Sabourin et al., 2009). These cells form the dorsal median septum and may

have specific functions, which could also be shared by radial cells found in the

ventral part of the EZ (Alfaro‐Cervello et al., 2012).

During development, the spinal cord is regionalised along the rostro-

caudal axis by expression of different Hox transcription factors, the so-called

“Hox code”. Ependymocytes of the adult spinal cord maintain expression of

developmental Hox genes (Pfenninger et al., 2011). In addition,

neurospheres, derived from cervical, thoracic and lumbar spinal levels

maintain the expression of the combination of hox genes corresponding to the

22

segment they were isolated from over several passages in vitro (Sabourin et

al., 2009).

3.1.2 Proliferation in the adult spinal cord

EZ cells strongly proliferate after birth, corresponding to the elongation

of the spinal cord, which doubles its length over the first nine post-natal

weeks. During this period, proliferation is progressively reduced until the

spinal cord reaches its final size at around postnatal week 12 in mice

(Sabourin et al., 2009; Alfaro‐Cervello et al., 2012). However, EZ cells

maintain a low level of proliferation in adulthood which declines with age

(Gonzalez-Fernandez et al., 2016). The main other cell type to proliferate in

the adult spinal cord are oligodendrocyte progenitor cells in the parenchyma.

Proliferating cells in the EZ have been observed in its dorsal (Hamilton et al.,

2009) or dorsal and ventral region (Alfaro‐Cervello et al., 2012). Electron

microscopy indicates that bi-ciliated ependymocytes are the main source of

proliferative cells in the EZ (Alfaro‐Cervello et al., 2012). This low number of

cilia in spinal cord ependymocytes correlates with proliferative activity,

because ependymocytes in the brain are multiciliated and do not proliferate

(Spassky et al., 2005; Mirzadeh et al., 2008). Cilia are cellular organelles

controlling proliferation (Han and Alvarez-Buylla, 2010) and the presence of

multiple cilia might not be compatible with proliferation (Brooks and

Wallingford, 2014).

The growth factors and molecular mechanisms underlying

ependymocyte proliferation in the spinal cord remain ill-defined. Comparative

gene expression analysis revealed that in the spinal cord these cells are

23

enriched for expression of retinoic acid (RA)-responsive genes, including Hox

genes (Pfenninger et al., 2011). In contrast, ependymocytes at the lateral

ventricle of the brain show higher expression levels of Hey1, a transcription

factor downstream of Notch signalling, as well as expression of several genes

related to the TGF beta1 signalling (Pfenninger et al., 2011). Treatment of

purified spinal ependymocytes with RA increases their proliferation in vitro and

these cells also appear to respond to RA injection in vivo (Pfenninger et al.,

2011). Hox genes are controlled by RA and are involved in the proliferation of

many cell types (Foronda et al., 2009), thus the strong expression of Hox

genes in spinal cord ependymocytes may indicate the involvement of these

genes in the proliferative response to RA. Endogenous growth factors such as

FGF2 and EGF may also be involved in the proliferation of adult

ependymocytes, as infusion of these growth factors into the CSF increases

proliferation of these cells in vivo and in vitro (Weiss et al., 1996; Kojima and

Tator, 2002; Martens et al., 2002). FGF2 and EGF may be derived from the

vasculature, as some ependymocytes send radial extension to vessels. This

suggests intimate cross-talk between these cells, a situation reminiscent of

the neurovascular niche in the SVZ and hippocampus. Interestingly, exercise

increases ependymocytes proliferation (Cizkova et al., 2009) in the spinal

cord. Several growth factors have been shown to be increased in expression

after physical exercise, including FGF2 (Gómez-Pinilla et al., 1997) which

might contribute to ependymocyte proliferation. Ependymocytes also express

high levels of receptor B for endothelins (EDNRB), which are cytokines

expressed by vascular cells (Peters et al., 2003; Sabourin et al., 2009).

Endothelin 1 and 3 enhance the proliferation of embryonic spinal cord stem

24

cells in vitro, suggesting that these growth factors may also be involved in the

proliferation of EZ cells in adults (Sabourin et al., 2009). Finally, another

signal to be considered with respect to the control of adult EZ proliferation is

Wnt, as several ligands and receptors for this pathway are found to be

expressed in this region (Gonzalez-Fernandez et al., 2016) and Wnt-reporter

mice reveal Wnt pathway activity in some EZ cells (White et al., 2010; Garbe

and Ring, 2012; González-Fernández et al., 2014).

The vast majority of proliferating cells remain in the EZ and there is no

clear evidence for substantial migration into the parenchyma. Few cells

expressing oligodendrocytic markers (Olig2, Sox10) are produced during

postnatal development and adulthood in the EZ (Sevc et al., 2014) but their

contribution to formation of new oligodendrocytes in the parenchyma is still

undefined. Endogenous neurogenesis has been described in the dorsal horn

of the spinal cord (Shechter et al., 2007; Rusanescu, 2016). These new

neurons are involved in nociception (Rusanescu and Mao, 2015) and their

production is enhanced by mechanosensation (Shechter et al., 2011) and

peripheral nerve injury (Rusanescu and Mao, 2017). However, it is more likely

that these cells are produced by local progenitors rather than by EZ cells.

3.1.3 Neural stem cell functions of the adult ependymal zone

The EZ is a pseudo-epithelium in contact with the CSF and undergoes

mechanical stress. Like in most adult epithelia in the body, cells in the EZ

probably have to self-renew, at least to some extent. Most epithelial tissues

are renewed from resident multi-potent stem cells and/or uni-potent progenitor

cells (Blanpain et al., 2007). These stem cells can exist in a dormant or

25

activated state. This is controlled by the environment, physiological cycles or

pathological conditions. Once activated, they produce transiently amplifying

cells (TA) before differentiated cells are finally generated. This is the case in

the subventricular zone in the brain, where neural stem cells underlie adult

neurogenesis (Lim and Alvarez-Buylla, 2016). Stem cells in adults are often

maintained in specialised structures called niches that are found in specific

locations in the respective organs. These niches are highly organized

structures composed of stem and non-stem cells, in which activity of canonical

signalling pathways, such as Notch, Wnt and BMP is adequately maintained

to control stem cell status and fate. It is worth noting that some tissues do not

regenerate from stem cells, but by self-duplication of differentiated cells (for

instance liver and pancreas) (Dor et al., 2004; Yanger et al., 2014).

With regard to the EZ in the spinal cord, the situation appears complex,

as on the one hand self-duplication of ependymocytes (which show some

degrees of differentiation) has been observed using electron microscopy

(Alfaro‐Cervello et al., 2012), and on the other hand, in vitro studies provide

compelling evidence for the presence of sub-populations of stem cells in the

EZ. The presence of neural stem cells in the adult spinal cord EZ has been

demonstrated in 1996 using the neurosphere assay (Weiss et al., 1996) in

which cells are seeded at low density on a non-adherent substrate and

allowed to grow in suspension in the presence of a defined medium. Weiss et

al. found that compared to lateral ventricles, stem cells from the spinal cord

required both FGF2 and EGF to generate these neurospheres, which could

be expanded for several passages and generated oligodendrocytes, neurons,

and astrocytes when placed in differentiation conditions. Several groups have

26

confirmed that spinal EZ tissue from both mice and rats form neurospheres.

The demonstration that neurospheres are derived from the EZ has been

obtained using i/ microdissection of the central canal region (Martens et al.,

2002; Sabourin et al., 2009), ii/ animals in which the EZ cells are fluorescently

labelled by transgenes that are controlled by specific promoters (FoxJ1,

Nestin) (Meletis et al., 2008; Li et al., 2016), iii/ and by purification of EZ cells

based on surface protein presence (CD24, CD133) (Pfenninger et al., 2011).

Interestingly, the capacity of EZ cells to form neurospheres is much higher in

juveniles than in older animals (Li et al., 2016; Xu et al., 2017). At least two

cell types found in the EZ are able to form multipotent neurospheres in vitro.

Using hGFAP-CreERT2 and hGFAP-GFP mice, two groups (Sabourin et al.,

2009; Fiorelli et al., 2013) have demonstrated that GFAP+ cells found in the

EZ region generate neurospheres, which differentiate into astrocytes and

neurons. The presence of GFAP-expressing cells in primary neurospheres

was established using marker expression (Sabourin et al., 2009) and with

GFAP-GFP mice (Xu et al., 2017). However, only Sabourin et al. could

demonstrate long-term propagation of these cells, while Fiorelli et al. reported

limited self-renewal properties. Different culture conditions, i.e. sorted

(Sabourin et al., 2009) vs. non-sorted GFAP cells (Fiorelli et al., 2013), used

in these studies could account for this discrepancy. In addition to GFAP+ cells,

Fiorelli et al. found a second population of neurosphere-forming EZ cells, the

identity of which remains to be specified (Fiorelli et al., 2013). These cells

generate neurospheres that can be propagated for at least 7 passages.

Moreover, it is possible that additional rare neurosphere-forming cells are

present in the EZ region, as illustrated recently by the isolation of Oct4+

27

primitive neural stem cells (Xu et al., 2017). Using genetic lineage tracing in

two transgenic mice, Nestin-CreERT and Foxj1-CreERT mice (Meletis et al.,

2008), it has been found that FoxJ1+ and Nestin+ cells generate passageable

and multipotent neurospheres. FoxJ1 is expressed by most EZ cells including

GFAP+ cells (Hugnot, 2012 ). This is reminiscent of the brain subventricular

zone, where FoxJ1 is expressed by neurogenic GFAP+ cells (Jacquet et al.,

2009). Thus, it remains to be explored whether FoxJ1+/GFAP- and

FoxJ1+/GFAP+ EZ cells have distinct abilities to form neurospheres. In the

subventricular zone (Pastrana et al., 2009), both activated stem cell

astrocytes and transit amplifying type C cells can generate neurospheres.

Thus the situation could be similar in the spinal cord EZ region where, as

mentioned previously, several cell types co-exist. As these cells have different

morphologies and express different proteins, notably developmental genes,

e.g. Pax6, Zeb1, Nato3, new lines of transgenic mice should be designed to

further explore the diversity and properties of these EZ cells in vitro and in

vivo. The recent advances to perform single cell RNA profiling should also

shed light on the diversity of EZ cells and the presence of distinct types of

stem cells. This will show whether mammalian EZ cells have a potential to

generate different neurons that is similar to that observed for ERGs in

zebrafish.

While the presence of stem cells in the adult EZ has been reported by

many groups, the type of neurons they generate in vitro and the signals

regulating the gliogenic versus neurogenic choice remain ill-defined. The

default neuronal subtype appears to be GABAergic (Weiss et al., 1996;

Sabourin et al., 2009). However, it is very interesting that by treating rat spinal

28

cord neurospheres with developmental ventralizing and caudalizing factors,

Shh and retinoic acid, respectively, a high rate of HB9+ and Map2+

motoneuron-like cells with different levels of electrophysiological maturity can

be obtained (Moreno-Manzano et al., 2009). In addition, a high rate of

oligodendrocyte production in vitro has been reported after differentiation of

rat neurospheres, which reveals the potential of these cells for myelin repair

(Kulbatski et al., 2007; Kulbatski and Tator, 2009). However, neurospheres

derived from different spinal cord regions express specific combination of Hox

genes (Sabourin et al., 2009) and have different abilities to differentiate

(Kulbatski and Tator, 2009), thus the full capacity of these cells to generate

different types of spinal neurons remains to be established.

3.1.4 The ependymal zone, spinal cord injury and regeneration

The capacity of ependymal cells to rapidly react to spinal cord injury

has been observed around 30 years ago in rats and rabbits (Matthews et al.,

1979; Vaquero et al., 1981; Bruni and Anderson, 1987). Different types of

lesions (compression, contusion, dorsal funiculus incision or hemisection,

transection) have been shown to trigger intense cell proliferation within a few

days. For a review on differences observed in EZ proliferation in the different

models used for spinal cord injury, see (McDonough and Martínez-Cerdeño,

2012). Spinal cord injury appears to be accompanied by an increase in the

stem cell pool as the formation of neurospheres formed in vitro is increased

when the tissue is taken from the injured spinal cord (Xu et al., 2006; Moreno-

Manzano et al., 2009; Barnabe-Heider et al., 2010; Fiorelli et al., 2013; Li et

al., 2016; Xu et al., 2017). However, it cannot be excluded that a fraction of

29

these neurospheres were derived through a de-differentiation process of

reactive astrocytes situated in the parenchyma (Sirko et al., 2013).

Interestingly, after injury, these neurospheres generate more

oligodendrocytes suggesting that spinal cord injury affects spinal cord stem

cell fate (Li X et al, 2016). Indeed, modifications of EZ properties by spinal

cord injury are indicated by several changes in gene expression. A number of

studies have reported a significant increase in expression of the neural

progenitor cell marker Nestin in EZ cells (Frisén et al., 1995; Namiki and

Tator, 1999; Liu et al., 2002; Shibuya et al., 2002; Takahashi et al., 2003;

Orendáčová et al., 2004; Xu et al., 2006; Foret et al., 2010). In addition,

GFAP, a marker for astrocytes, but also for neural stem cells, is increased in

EZ cells (JP Hugnot, unpublished data)(Takahashi et al., 2003; Xu et al.,

2006). Expression of ICAM-1, an adhesion molecule involved in leukocyte

transmigration, is increased in EZ cells after spinal cord injury in rats

(Isaksson et al., 1999). Notably, cell fate regulators, like β1-integrin (North et

al., 2015), Bmp4, Msx2, Notch1, Numb, Pax6, and Shh show higher

expression levels after injury (Yamamoto et al., 2001; Chen et al., 2005),

suggestive of an incomplete recapitulation of a developmental program for

neurogenesis. The re-initiation of basic developmental processes after spinal

cord injury is further illustrated by the observation of asymmetric divisions of

EZ cells, involving the Notch1 receptor after injury (Johansson et al., 1999).

This is an evolutionarily conserved feature with zebrafish (Dias et al., 2012)

and indicates that notch signalling may fulfil essential functions in

regenerative neurogenesis.

30

The molecular mechanisms governing the activation of the EZ after

injury are not fully explored. EZ cells express several receptors, such as

purinergic receptors and Ednrb (Hugnot and Franzen, 2011) and may rapidly

detect molecules released by the lesion site, such as ATP (Franke and Illes,

2014) and inflammatory cytokines. In the uninjured state, some microglial cells

are located very close to the EZ (Alfaro‐Cervello et al., 2012) and could

rapidly relay lesion-induced signals to EZ cells. Apparently, ras signalling is

required, as in the “rasless” mouse, proliferation of EZ cells after spinal injury

is reduced (Sabelström et al., 2013). With regard to Wnt signalling, despite the

downstream effector beta-catenin being expressed by EZ cells, Wnt-signalling

reporter mice did not reveal increased activation of this pathway after spinal

cord injury (White et al., 2010; González-Fernández et al., 2014). This is

similar to observations in zebrafish (Wehner et al., 2017). Compared to

trauma, demyelinating lesions appear not to be able to trigger acute

proliferation in the EZ (Guo, F. et al., 2011; Lacroix et al., 2014) indicating that

a mechanical damage with disruption of the blood-spinal cord barrier (Maikos

and Shreiber, 2007) is probably needed.

In addition to proliferation, migration of EZ cells to the lesion site and

contribution to the glial scar have been reported using genetic linage-tracing

tools (Meletis et al., 2008) or EZ dye-labelling (Mothe and Tator, 2005).

Recent evidence indicates that the contribution of EZ-derived cells to the scar

depends on age, type, and severity of the injury (Li et al., 2016; Ren et al.,

2017). The migration of EZ cells to the lesion site could be analogous to an

epithelial-mesenchymal transition process, especially as these cells already

display mesenchymal traits such as the expression of the intermediate

31

filament Vimentin and the transcription factor Zeb1, both of which are involved

in migration and epithelial-mesenchymal transition (Cheng and Eriksson,

2017). EZ cells also express the polysialylated form of NCAM (Oumesmar et

al., 1995), as well as the CXCR4 receptor (Tysseling et al., 2011). These cell

surface proteins are both involved in neural precursor migration (Cremer et

al., 2000). Whether EZ-derived cells migrate freely through the parenchyma or

are associated with other cells such as vessels or astrocytes is currently

unknown. Using post-natal spinal cord sections maintained in vitro, dorsal and

ventral migration of cells from the EZ, named funicular migratory stream, has

been observed (Mladinic et al., 2014). This is associated with nuclear Atf3

expression. Whether a funicular migratory stream exists after spinal cord

injury in mice remains to be analysed.

After injury, the fate of EZ cells appears to be primarily glial. Using

genetic lineage-tracing tools and lesions of the dorsal funiculus, Meletis et al.

have shown that EZ-derived cells give rise to Sox9+/Vim+ astrocytes (80%,

only part of them express GFAP) and only few myelinating oligodendrocytes

(3%) (Meletis et al., 2008). These astrocytes, which contribute to the core of

the glial scar (for review on glial scar see (Cregg et al., 2014; Gregoire et al.,

2015)), are beneficial for recovery, as their reduction using genetic tools

provoked further axonal loss and secondary enlargement of the lesion volume

(Sabelström et al., 2013). This positive effect of ependymal cell–derived

astrocytes on the lesioned tissue could be mediated by their expression of

neurotrophic factors, such as CNTF, HGF, and IGF-1. Hence, whereas spinal

cord-derived neurospheres generate neurons in vitro, the spinal cord

environment after injury appears to restrict the fate of progenitor cells to glial

32

phenotypes. This notion is also supported by transplantation experiments

showing that multipotent progenitors of the adult spinal cord generate mostly

glial cells when grafted into the spinal cord, but generate neurons when

placed in the hippocampus, a neurogenic environment (Shihabuddin et al.,

2000). It is not clear whether this represents a true fate switch in EZ

progenitors, or instead if this is a selective effect on survival or differentiation

of a heterogeneous population of EZ progenitors. After spinal cord injury, the

strong inflammation and the presence of gliogenic cytokines such as BMPs

(Setoguchi et al., 2004) are likely to bias EZ cell progeny toward an astrocytic

fate at the expense of neurons and oligodendrocytes. Myelin-associated

inhibitors have also been demonstrated to promote differentiation of neural

progenitor cells into the glial lineage (Li et al., 2013).

Consequently, achieving extensive regeneration of the spinal cord and

production of new neurons from the EZ will need genetic (overexpression of

neurogenic transcription factors) or biochemical (infusion with neurogenic

molecules) intervention. For instance, Kojima and Tator (Kojima and Tator,

2002) have found that intrathecal administration of FGF2 and EGF boosts EZ

proliferation and leads to functional improvements, but not to production of

new neurons after spinal cord injury. In addition, Kim and others (Kim et al.,

2011) have found that treatment of exogenous progenitors in vitro or in vivo

with dibutyrylic cyclic AMP, a cell-permeable analog of cyclic AMP, enhances

their neuronal differentiation and survival after spinal cord injury. Combining

these two approaches might lead to new neurons being generated from the

EZ.

33

Migration of newly produced cells from the EZ is also an important

issue and enhancing migration could be beneficial for recovery. This could be

achieved by various strategies, such as chemokine application (Merino et al.,

2015), use of biomaterials, such as a collagen scaffold (Li et al., 2013), or

even by applying electric fields and electromagnetic waves. Indeed, following

pioneering work of Borgens in 1981, showing positive influence of applied

electric fields on spinal cord regeneration in lamprey (Borgens et al., 1981), it

has been shown that migration and differentiation of neural stem cells and

their progeny can be modulated by current (Huang et al., 2015; Yao and Li,

2016; Feng et al., 2017; Iwasa et al., 2017; Samaddar et al., 2017) or

electromagnetic field application (Cui et al., 2017).

3.6 The ependymal zone in primates

Cell types and their proliferative activity have been studied in the EZ

region of adult macaques (Alfaro‐Cervello et al., 2014). As in rodents, cellular

heterogeneity has been observed using marker analysis and electron

microscopy. One distinguishing feature of primate EZ cells compared to those

in rodents is the presence of lateral ependymocytes with multiple ciliae. These

cells do not express Nestin and do not proliferate. In comparison, dorsal and

ventral cells are uniciliated or biciliated, express Nestin and/or GFAP and

proliferate. Like in rodents, astrocytes contacting the central canal and CSF-

contacting neurons are also present in the macaque EZ.

Even though the human spinal cord EZ region has been studied since

1890 (von Lenhossék, 1891) using several histological techniques on post

mortem tissue, there is relatively little available data on it. As in rodents, this

34

region shows cellular heterogeneity during early and prenatal development.

Roof and floor plate cells differentiate first, express Nestin (Sakakibara et al.,

2007) and acquire a long radial morphology extending to the pial surface

(Sarnat, 1992). These cells constitute the midline dorsal and ventral septa

which might be involved in guiding descending and ascending axons, notably

by preventing their decussation (Sarnat, 1992). These cells could also have a

supply role for neurons before vessels appear. Indeed, ependymal cells from

the floor plate ensheath axons during spinal cord development (Gamble,

1968). Moreover, electron microscopy of the developing floor plate

neuroepithelium has also revealed that this structure has an intense secretory

activity, suggesting that, as in rodents, this region works as an organizing

centre (Tanaka et al., 1988; Wagner et al., 1990). In human infants (0.5-18

months), two subtypes of ependymocytes can be found showing either two

(exceptionally one) or more cilia (Alfaro‐Cervello et al., 2014). Astrocytes

contacting the lumen are also found. In adult humans, most ependymal cells

appear to be multi-cilialed, but a subpopulation with two cilia, resembling

macaque or rodent ependymal cells, are also present (Alfaro‐Cervello et al.,

2014). In contrast to rodents or macaques, no proliferating cells have been

found in the ependymal zone of adult humans using the Ki67 antibody

(Dromard et al., 2008; Alfaro‐Cervello et al., 2014; Garcia-Ovejero et al.,

2015). However, low levels of proliferation could have been missed. As the

spinal cord and ependyma are subjected to mechanical stress (twisting,

bending), it is likely that cells of the central canal very slowly self-renew,

similar to many epithelia of the body. Accumulation of mutations in these

proliferative cells may be at the origin of spinal cord ependymoma, an

35

infrequent tumour representing 2% of all primary brain tumours (Weller et al.,

2015).

One major structural difference between the human and rodent central

canal is that after the first decade of life, a central lumen is not patent and an

obliteration process is observed at different levels of the spinal cord in

humans. However, the central canal at the level of the medulla is maintained.

This observation was initially made by histology (Motavkin and Bakhtinov,

1973; Milhorat et al., 1994; Yasui et al., 1999) but was then confirmed using

MRI on 59 healthy volunteers (Garcia-Ovejero et al., 2015). The lumen is

replaced by a cellular mass, which is highly vascularized. This mass is

surrounded by radially arranged cells, giving a pseudo-rosette appearance.

More rarely, pseudo-canals, surrounded with epithelial cells with beta-catenin+

junctions, are found in these masses (Garcia-Ovejero et al., 2015). Multi-

ciliated VIM+ ependymocytes remain in this structure but astrocytes are also

found, suggesting a process of gliosis (Alfaro‐Cervello et al., 2014). Some

authors have proposed that this obliteration should not be regarded as an

entirely passive process but rather as formation of a secretory organ with

physiological functions (Motavkin and Bakhtinov, 1973). However, this

remains to be fully established.

In contrast to macaques, rodents, and fish, the presence of CSF-N has

not yet been reported for the human spinal cord (Dromard et al., 2008; Alfaro‐

Cervello et al., 2014; Djenoune et al., 2014). In zebrafish, these cells are

implicated in locomotor behaviour by responding to both passive and active

bending of the spinal cord and regulating the tail beat frequency (Böhm et al.,

2016). It is tempting to speculate that loss of the tail, which occurred during

36

human and chimpanzee evolution around 4-7 millions ago (Kumar et al.,

2005), or acquisition of bipedalism, was accompanied by a reduction or loss

of CSF-N. Despite the presumptive lack of CSF-N, the EZ receives

innervation which appears to increase with age (Motavkin and Bakhtinov,

1973). The role of this innervation, which is also present in rodents, is

unknown, but may be involved in the proliferation of EZ cells as in the SVZ,

where various nervous fibre subtypes (notably 5-HT and dopamine) control

proliferation and fate of neural stem and progenitor cells (Bond et al., 2015).

This innervation may also control some secretory functions of EZ cells, which

has been suggested in salamanders (Zamora, 1978). In zebrafish, this

innervation controls regenerative neurogenesis from the EZ (Reimer et al.,

2013; Barreiro-Iglesias et al., 2015).

With regard to expression of markers by adult human EZ cells, only a

few studies have been carried out and results are not always consistent. This

might be due to the quality of human samples (post-mortem delay), the

precise anatomical location of the studied tissues and the implemented

techniques (paraffin-embedding or cryosection). In addition, for some markers

with many isoforms, like GFAP with 10 isoforms (Hol and Pekny, 2015),

different antibodies may only detect specific forms of the protein. This could

lead to discrepancy between findings of different laboratories. However, the

available data indicate that in young adults (second decade), subsets of cells

express several markers typically found in immature neural cells such as

CD15, FOXA2, NESTIN, PAX6, SOX2, SOX11, and VIMENTIN (Hugnot,

unpublished results) (Dromard et al., 2008; Cawsey et al., 2015). There is

obvious cellular heterogeneity, as only a subset of EZ cells express GFAP or

37

NESTIN. In addition, some markers are mostly expressed in the dorsal part

(CD15, PAX6), while radial NESTIN+ cells are observed in the ventral and

dorsal parts of the EZ (Dromard et al., 2008; Cawsey et al., 2015) (Hugnot,

unpublished results). This indicates a dorsal-ventral regionalization of the

adult human spinal cord EZ, similar to that observed in the mouse (Sabourin

et al., 2009; Yu et al., 2013; Khazanov et al., 2016) and zebrafish (Figs 1, 2).

In older individuals, even after EZ obstruction, cells of the remaining mass

may retain some immature features, as indicated by the presence of CD15,

SOX2, PAX6, PAX7, and several HOX proteins. This has been detected by

tissue labeling and/or PCR (Garcia-Ovejero et al., 2015).

Evidence for the persistence of neural progenitor cells in the adult

human spinal cord also comes from in vitro studies. Dromard et al, were able

to generate SOX2+/NESTIN+ neurospheres from the gray matter central

region using defined media containing EGF and FGF2 (Dromard et al., 2008).

These neurospheres were multipotent and generated neurons and glial cells

after differentiation. However, the neurospheres could not be passaged,

suggesting that they were derived from proliferation-limited neural progenitors

rather than bona fide stem cells. Using different culture conditions (adherent

substrate and Matrigel) Mothe et al. were able to isolate SOX2+/NESTIN+

multipotent cells that could be passaged for at least 9 months (Mothe et al.,

2011). These cells generate glial and neuronal cells after transplantation into

spinal-injured rats and are thus of great interest with respect to understanding

the molecular mechanisms controlling their proliferation and fate.

There are several critical points when considering human EZ cells as a

suitable cellular source for regeneration. Human EZ cells appear to react to

38

spinal cord injury with an increase in NESTIN expression (Cawsey et al.,

2015), but whether these cells proliferate and migrate has not been reported.

In macaques, production of new glial cells and neurons has been observed

after injury (Yang et al., 2006; Vessal et al., 2007). However, whether these

cells originate from the EZ remains to be established.

An essential difference between rodents and primates is the presence of a

high level of GFAP intermediate filaments in a hypocellular layer, which

surrounds the EZ region in primates (Dromard et al., 2008; Alfaro‐Cervello et

al., 2014) (Fig. 2). A similar hypocellular GFAP+ layer is also found adjacent to

the SVZ neural stem cell niche in marmosets, macaques, and humans (Sanai

et al., 2004; Quiñones‐Hinojosa et al., 2006; Gil‐Perotin et al., 2009;

Sawamoto et al., 2011) and thus appears to be a distinctive feature of

primates. The role of this layer is currently unknown but its presence may

represent an obstacle for regeneration, as EZ cells would have to migrate

through this dense layer to reach lesioned areas. The large size of the spinal

cord might also be an obstacle to regeneration for three reasons. Firstly,

cellular regeneration in anamniotes is achieved notably by ependymo-radial

glia spanning from the EZ to the pial surface. While radial glial cells are clearly

present in rodents and macaque around the canal, it is not known whether

these cells reach the meninges, which are an important source of signals for

neural stem cell maintenance and fate determination during CNS

development (Siegenthaler and Pleasure, 2011). It can be speculated that

with the increase in size, radial cells around the EZ can no longer extend to

the pial surface in mammals and have lost some of their capacity for

39

regeneration. Secondly, longer spinal cords with a larger central canal cavity

could have favoured selection of multi-ciliated ependymocytes during

evolution (Nakayama and Kohno, 1974), so as to move the CSF fluid over a

greater distance. To support this hypothesis, it would be informative to study

the cellular composition of the EZ region in mammals with a long vertebral

column, such as giraffes and whales, to see if multi-ciliated ependymocytes

found in primates are also present. The presence of multi-ciliated

ependymocytes could have been selected for at the expense of proliferative

capacity and stemness, as multi-ciliated cells are post-mitotic (Brooks and

Wallingford, 2014). Thirdly, differences in myelin between anamniotes and

mammals may also account for the contrasting ability to regenerate neurons

from EZ cells. Whereas the basic structure of myelin is conserved between

fishes and mammals, some differences exist. Myelin Protein zero, a protein

possibly involved in zebrafish CNS regeneration (Schweitzer et al., 2003), is

present in both CNS and PNS myelin of fish, in contrast to the mammalian

homolog, which is expressed only in PNS myelin (Bai et al., 2011). Recent

results indicated that proteins of myelin, such as myelin basic protein, can

inhibit spinal cord neural stem cell proliferation in vitro and such mechanisms

could also operate in vivo to limit regeneration in mammals (Xu et al., 2017).

4. Comparing regenerating and non-regenerating species

From the above comparisons, far reaching similarities between EZ cells

in anamniotes and mammals become apparent (summarized in Table 1 and

Fig. 2). Morphologically, cells in the EZ are ciliated and have radial processes.

In anamniotes, these processes form endfeet at the pial surface. In mammals,

40

only EZ cells at the dorsal and ventral midline appear to have endfeet at the

pial surface, whereas the radial processes of most EZ cells are shorter and

end on parenchymal blood vessels. EZ cells in anamniotes and amniotes also

share expression of structural and functional genes, such as GFAP and

Vimentin.

Spinal EZ cells are not a homogeneous population, with dorso-ventral

and rostro-caudal differences in expression of developmental transcription

factors that confer regional identity. This may be related to a degree of

lineage restriction of progenitor cells. Remarkably, both anamniote and

mammalian EZ cells express transcription factors related to « stemness »,

such as Sox2. However, in anmniotes, the neuronal lineages of progenitor

cells are fully expressed, whereas in mammals environmental factors appear

to restrict this potential.

After a lesion, EZ cells in both anamniotes and mammals react to

lesion signals with strongly increased proliferation. Hence, any lesion stimulus

is adequate in anamniotes and mammals to bring EZ cells out of relative

quiescence to start proliferating. This suggests that lesion-induced

intracellular mechanisms may be similar. Indeed, as we discuss above, the

Notch pathway, which attenuates neurogenesis, is upregulated in EZ cells of

regenerating and non-regenerating species. Moreover, activity of the

hedgehog pathway is observed across vertebrates. It is therefore likely that

differences in intracellular signalling are not fundamental, but that different

levels of activity of these pathways lead to dramatically different net outcomes

in terms of gliogenesis versus neurogenesis.

41

However, fundamental differences are likely to exist between

progenitor cells in anamniotes and mammals. For example, ERGs in

anmniotes have to fulfil the functions of ependymal cells and astrocytes, as

there is little evidence for the presence of “free” astrocytes in anamniotes. In

mammals, there is a clear distinction between ependymal cells and

parenchymal astrocytes. Nevertheless, a closer comparison between cells in

the EZ between regenerating and non-regenerating species may be

informative in order to determine where lesion reactions diverge. Analyses

should also be performed in less-studies species such as turtles (Marichal et

al., 2009), birds, large or aquatic mammals and non-human primates such as

marmoset to bring additional molecular and cellular insights into the functions

of the EZ regions across species.

5. Outlook:

Based on transgenic marker expression, specific EZ cell populations

can be highly purified and subjected to gene expression profiling techniques.

This can be done in a systematic comparison between regenerating and non-

regenerating species and on an EZ domain-specific basis, in order to follow

up on potential lineage restrictions of progenitor domains. Moreover, single

cell sequencing has the capacity to identify unanticipated subpopulations of

progenitor cells (Cuevas-Diaz Duran et al., 2017). This has the potential to

unravel the gene expression programmes that are being switched on after a

lesion.

Three-dimensional ultrastructural analyses by serial block-face

scanning electron microscopy (Denk and Horstmann, 2004; Titze and

42

Genoud, 2016) can serve to describe the morphological diversity of cell types

in the EZ and be correlated with expression profiles. This can address

whether similar niches, with the presence of stem cells and transit amplifying

cells, exist as in neurogenic areas in the brain (Chang et al., 2016).

It will also be important to elucidate how EZ cells receive signals that

are crucial for regeneration. For example, the ventral source of Shh is much

more remote from target cells during regeneration of the adult spinal cord than

during neural tube development. A common denominator between cells in the

EZ is their close and often direct contact with the central canal, which may

transport signals (Zappaterra and Lehtinen, 2012). Hence, analysing the

composition of the CSF and receptors on EZ cells after injury should be

informative.

Finally, zebrafish are the foremost vertebrate in vivo model for

automated drug screening (White et al., 2016). Screening whole zebrafish

larvae for drugs that enhance lesion-induced neurogenesis (Reimer et al.,

2013) combined with those on spinal neurospheres from mammals (Cheng et

al., 2017) may yield important tools to tweak the lesion-induced gene

expression programmes in EZ cells towards a more pro-regenerative

phenotype. We conclude that a comparative approach to the analysis of

spinal EZ cells holds considerable promise to identify critical steps that decide

between inhibitory gliogenesis and pro-regenerative neurogenesis after spinal

cord injury.

43

ACKNOWLEDGEMENTS:

We thank the 2017 cohort of the Wellcome Trust 4 year PhD

programme in Tissue Repair at the University of Edinburgh for critically

reading the manuscript. This work was supported by BBSRC to CGB and TB;

INSERM, AFM, IRME and IRP to JPH’s team; ERA-NET NEURON CoFund

Consortium NEURONICHE (MRC, Spinal Research, ANR) to all authors.

44

REFERENCES:

Alfaro‐Cervello, C., Soriano‐Navarro, M., Mirzadeh, Z., Alvarez‐Buylla, A. and Garcia‐Verdugo, J. M. 2012 Biciliated ependymal cell proliferation contributes to spinal cord growth. J. Comp. Neurol. 520, 3528-3552.

Alfaro‐Cervello, C., Cebrian‐Silla, A., Soriano‐Navarro, M., Garcia‐Tarraga, P., Matías‐Guiu, J., Gomez‐Pinedo, U., Molina Aguilar, P., Alvarez‐Buylla, A., Luquin, M. R. and Garcia‐Verdugo, J. M. 2014 The adult macaque spinal cord central canal zone contains proliferative cells and closely resembles the human. J. Comp. Neurol. 522, 1800-1817.

Alfonso, J., Le Magueresse, C., Zuccotti, A., Khodosevich, K. and Monyer, H. 2012 Diazepam binding inhibitor promotes progenitor proliferation in the postnatal SVZ by reducing GABA signaling. Cell Stem Cell 10, 76-87.

Almad, A., Sahinkaya, F. R. and McTigue, D. M. 2011 Oligodendrocyte fate after spinal cord injury. Neurotherapeutics : the journal of the American Society for Experimental NeuroTherapeutics 8, 262-273.

Alunni, A. and Bally-Cuif, L. 2016 A comparative view of regenerative neurogenesis in vertebrates. Development 143, 741-753.

Alunni, A., Krecsmarik, M., Bosco, A., Galant, S., Pan, L., Moens, C. B. and Bally-Cuif, L. 2013 Notch3 signaling gates cell cycle entry and limits neural stem cell amplification in the adult pallium. Development 140, 3335-3347.

Anderson, M. A., Burda, J. E., Ren, Y., Ao, Y., O'Shea, T. M., Kawaguchi, R., Coppola, G., Khakh, B. S., Deming, T. J. and Sofroniew, M. V. 2016 Astrocyte scar formation aids central nervous system axon regeneration. Nature 532, 195-200.

Anderson, M. J. and Waxman, S. G. 1981 Morphology of regenerated spinal cord in Sternarchus albifrons. Cell Tissue Res 219, 1-8.

Bai, Q., Sun, M., Stolz, D. B. and Burton, E. A. 2011 Major isoform of zebrafish P0 is a 23.5 kDa myelin glycoprotein expressed in selected white matter tracts of the central nervous system. J. Comp. Neurol. 519, 1580-1596.

Barnabe-Heider, F., Goritz, C., Sabelstrom, H., Takebayashi, H., Pfrieger, F. W., Meletis, K. and Frisen, J. 2010 Origin of new glial cells in intact and injured adult spinal cord. Cell Stem Cell 7, 470-482.

Barreiro-Iglesias, A., Mysiak, K. S., Scott, A. L., Reimer, M. M., Yang, Y., Becker, C. G. and Becker, T. 2015 Serotonin Promotes Development and Regeneration of Spinal Motor Neurons in Zebrafish. Cell Rep 13, 924-932.

Becker, C. G. and Becker, T. 2015 Neuronal regeneration from ependymo-radial glial cells: cook, little pot, cook! Dev Cell 32, 516-527.

Becker, T. and Becker, C. G. 2001 Regenerating descending axons preferentially reroute to the gray matter in the presence of a general macrophage/microglial reaction caudal to a spinal transection in adult zebrafish. J. Comp. Neurol. 433, 131-147.

45

Becker, T., Lieberoth, B. C., Becker, C. G. and Schachner, M. 2005 Differences in the regenerative response of neuronal cell populations and indications for plasticity in intraspinal neurons after spinal cord transection in adult zebrafish. Mol Cell Neurosci 30, 265-278.

Becker, T., Wullimann, M. F., Becker, C. G., Bernhardt, R. R. and Schachner, M. 1997 Axonal regrowth after spinal cord transection in adult zebrafish. J. Comp. Neurol. 377, 577-595.

Blanpain, C., Horsley, V. and Fuchs, E. 2007 Epithelial stem cells: turning over new leaves. Cell 128, 445-458.

Böhm, U. L., Prendergast, A., Djenoune, L., Figueiredo, S. N., Gomez, J., Stokes, C., Kaiser, S., Suster, M., Kawakami, K. and Charpentier, M. 2016 CSF-contacting neurons regulate locomotion by relaying mechanical stimuli to spinal circuits. Nat Commun 7.

Bond, A. M., Ming, G.-l. and Song, H. 2015 Adult mammalian neural stem cells and neurogenesis: five decades later. Cell Stem Cell 17, 385-395.

Borgens, R. B., Roederer, E. and Cohen, M. J. 1981 Enhanced spinal cord regeneration in lamprey by applied electric fields. Science 213, 611-617.

Briona, L. K. and Dorsky, R. I. 2014 Radial glial progenitors repair the zebrafish spinal cord following transection. Exp Neurol 256, 81-92.

Briona, L. K., Poulain, F. E., Mosimann, C. and Dorsky, R. I. 2015 Wnt/beta-catenin signaling is required for radial glial neurogenesis following spinal cord injury. Dev Biol 403, 15-21.

Brooks, E. R. and Wallingford, J. B. 2014 Multiciliated cells. Curr. Biol. 24, R973-R982.

Bruni, J. and Anderson, W. 1987 Ependyma of the rat fourth ventricle and central canal: response to injury. Cells Tissues Organs 128, 265-273.

Cardozo, M. J., Mysiak, K. S., Becker, T. and Becker, C. G. 2017 Reduce, reuse, recycle - Developmental signals in spinal cord regeneration. Dev Biol.

Cawsey, T., Duflou, J., Weickert, C. S. and Gorrie, C. A. 2015 Nestin-positive ependymal cells are increased in the human spinal cord after traumatic central nervous system injury. J. Neurotrauma 32, 1393-1402.

Chang, E. H., Adorjan, I., Mundim, M. V., Sun, B., Dizon, M. L. and Szele, F. G. 2016 Traumatic Brain Injury Activation of the Adult Subventricular Zone Neurogenic Niche. Front Neurosci 10, 332.

Chen, J., Leong, S. Y. and Schachner, M. 2005 Differential expression of cell fate determinants in neurons and glial cells of adult mouse spinal cord after compression injury. Eur J Neurosci 22, 1895-1906.

Chen, T., Yu, Y., Hu, C. and Schachner, M. 2016 L1.2, the zebrafish paralog of L1.1 and ortholog of the mammalian cell adhesion molecule L1 contributes to spinal cord regeneration in adult zebrafish. Restor Neurol Neurosci 34, 325-335.

46

Cheng, C., Fass, D. M., Folz-Donahue, K., MacDonald, M. E. and Haggarty, S. J. 2017 Highly Expandable Human iPS Cell-Derived Neural Progenitor Cells (NPC) and Neurons for Central Nervous System Disease Modeling and High-Throughput Screening. Current protocols in human genetics 92, 21 28 21-21 28 21.

Cheng, F. and Eriksson, J. E. 2017 Intermediate filaments and the regulation of cell motility during regeneration and wound healing. Cold Spring Harb Perspect Biol 9, a022046.

Cizkova, D., Nagyova, M., Slovinska, L., Novotna, I., Radonak, J., Cizek, M., Mechirova, E., Tomori, Z., Hlucilova, J. and Motlik, J. 2009 Response of ependymal progenitors to spinal cord injury or enhanced physical activity in adult rat. Cell. Mol. Neurobiol. 29, 999-1013.

Cregg, J. M., DePaul, M. A., Filous, A. R., Lang, B. T., Tran, A. and Silver, J. 2014 Functional regeneration beyond the glial scar. Exp Neurol 253, 197-207.

Cremer, H., Chazal, G., Lledo, P.-M., Rougon, G., Montaron, M., Mayo, W., Le Moal, M. and Abrous, D. 2000 PSA-NCAM: an important regulator of hippocampal plasticity. Int. J. Dev. Neurosci. 18, 213-220.

Cuevas-Diaz Duran, R., Wei, H. and Wu, J. Q. 2017 Single-cell RNA-sequencing of the brain. Clin Transl Med 6, 20.

Cui, M., Ge, H., Zhao, H., Zou, Y., Chen, Y. and Feng, H. 2017 Electromagnetic Fields for the Regulation of Neural Stem Cells. Stem Cells Int 2017, 9898439.

de Oliveira-Carlos, V., Ganz, J., Hans, S., Kaslin, J. and Brand, M. 2013 Notch receptor expression in neurogenic regions of the adult zebrafish brain. PLoS One 8, e73384.

Denk, W. and Horstmann, H. 2004 Serial block-face scanning electron microscopy to reconstruct three-dimensional tissue nanostructure. PLoS Biol 2, e329.

Dervan, A. G. and Roberts, B. L. 2003a The meningeal sheath of the regenerating spinal cord of the eel, Anguilla. Anat Embryol (Berl) 207, 157-167.

Dervan, A. G. and Roberts, B. L. 2003b Reaction of spinal cord central canal cells to cord transection and their contribution to cord regeneration. J Comp Neurol 458, 293-306.

Dias, T. B., Yang, Y. J., Ogai, K., Becker, T. and Becker, C. G. 2012 Notch signaling controls generation of motor neurons in the lesioned spinal cord of adult zebrafish. J. Neurosci. 32, 3245-3252.

Dimou, L. and Gotz, M. 2014 Glial cells as progenitors and stem cells: new roles in the healthy and diseased brain. Physiol Rev 94, 709-737.

Djenoune, L. and Wyart, C. 2017 Light on a sensory interface linking the cerebrospinal fluid to motor circuits in vertebrates. J Neurogenet 31, 113-127.

Djenoune, L., Khabou, H., Joubert, F., Quan, F. B., Figueiredo, S. N., Bodineau, L., Del Bene, F., Burcklé, C., Tostivint, H. and Wyart, C. 2014 Investigation of spinal cerebrospinal fluid-contacting neurons expressing PKD2L1: evidence for a conserved system from fish to primates. Front Neuroanat 8.

47

Dor, Y., Brown, J., Martinez, O. I. and Melton, D. A. 2004 Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature 429, 41.

Dromard, C., Guillon, H., Rigau, V., Ripoll, C., Sabourin, J. C., Perrin, F. E., Scamps, F., Bozza, S., Sabatier, P., Lonjon, N. et al. 2008 Adult human spinal cord harbors neural precursor cells that generate neurons and glial cells in vitro. J Neurosci Res 86, 1916-1926.

Fang, P., Pan, H. C., Lin, S. L., Zhang, W. Q., Rauvala, H., Schachner, M. and Shen, Y. Q. 2014 HMGB1 contributes to regeneration after spinal cord injury in adult zebrafish. Mol Neurobiol 49, 472-483.

Feng, J. F., Liu, J., Zhang, L., Jiang, J. Y., Russell, M., Lyeth, B. G., Nolta, J. A. and Zhao, M. 2017 Electrical Guidance of Human Stem Cells in the Rat Brain. Stem cell reports 9, 177-189.

Ferg, M., Armant, O., Yang, L., Dickmeis, T., Rastegar, S. and Strahle, U. 2014 Gene transcription in the zebrafish embryo: regulators and networks. Brief Funct Genomics 13, 131-143.

Fiorelli, R., Cebrian‐Silla, A., Garcia‐Verdugo, J. M. and Raineteau, O. 2013 The adult spinal cord harbors a population of GFAP‐positive progenitors with limited self‐renewal potential. Glia 61, 2100-2113.

Foret, A., Quertainmont, R., Botman, O., Bouhy, D., Amabili, P., Brook, G., Schoenen, J. and Franzen, R. 2010 Stem cells in the adult rat spinal cord: plasticity after injury and treadmill training exercise. J. Neurochem. 112, 762-772.

Foronda, D., de Navas, L. F., Garaulet, D. L. and Sánchez-Herrero, E. 2009 Function and specificity of Hox genes. Int. J. Dev. Biol. 53, 1409-1419.

Franke, H. and Illes, P. 2014 Nucleotide signaling in astrogliosis. Neurosci. Lett. 565, 14-22.

Frisén, J., Johansson, C. B., Török, C., Risling, M. and Lendahl, U. 1995 Rapid, widespread, and longlasting induction of nestin contributes to the generation of glial scar tissue after CNS injury. J Cell Biol 131, 453-464.

Fu, H., Qi, Y., Tan, M., Cai, J., Hu, X., Liu, Z., Jensen, J. and Qiu, M. 2003 Molecular mapping of the origin of postnatal spinal cord ependymal cells: evidence that adult ependymal cells are derived from Nkx6.1+ ventral neural progenitor cells. J Comp Neurol 456, 237-244.

Gamble, H. 1968 Axon ensheathing by ependymal cells in the human embryonic and foetal spinal cord. Nature 218, 182-183.

Garbe, D. S. and Ring, R. H. 2012 Investigating tonic Wnt signaling throughout the adult CNS and in the hippocampal neurogenic niche of BatGal and ins-TopGal mice. Cell. Mol. Neurobiol. 32, 1159-1174.

Garcia-Ovejero, D., Arevalo-Martin, A., Paniagua-Torija, B., Florensa-Vila, J., Ferrer, I., Grassner, L. and Molina-Holgado, E. 2015 The ependymal region of the adult human spinal cord differs from other species and shows ependymoma-like features. Brain 138, 1583-1597.

48

Gil‐Perotin, S., Duran‐Moreno, M., Belzunegui, S., Luquin, M. R. and Garcia‐Verdugo, J. M. 2009 Ultrastructure of the subventricular zone in Macaca fascicularis and evidence of a mouse‐like migratory stream. J. Comp. Neurol. 514, 533-554.

Goldshmit, Y., Sztal, T. E., Jusuf, P. R., Hall, T. E., Nguyen-Chi, M. and Currie, P. D. 2012 Fgf-dependent glial cell bridges facilitate spinal cord regeneration in zebrafish. J. Neurosci. 32, 7477-7492.

Gómez-Pinilla, F., Dao, L. and So, V. 1997 Physical exercise induces FGF-2 and its mRNA in the hippocampus. Brain Res. 764, 1-8.

Gonzalez-Fernandez, C., Arevalo-Martin, A., Paniagua-Torija, B., Ferrer, I., Rodriguez, F. J. and Garcia-Ovejero, D. 2016 Wnts Are Expressed in the Ependymal Region of the Adult Spinal Cord. Mol. Neurobiol., 1-14.

González-Fernández, C., Fernández-Martos, C. M., Shields, S. D., Arenas, E. and Javier Rodríguez, F. 2014 Wnts are expressed in the spinal cord of adult mice and are differentially induced after injury. J. Neurotrauma 31, 565-581.

Gotz, M., Sirko, S., Beckers, J. and Irmler, M. 2015 Reactive astrocytes as neural stem or progenitor cells: In vivo lineage, In vitro potential, and Genome-wide expression analysis. Glia 63, 1452-1468.

Grandel, H. and Brand, M. 2013 Comparative aspects of adult neural stem cell activity in vertebrates. Dev. Genes Evol. 223, 131-147.

Gregoire, C. A., Goldenstein, B. L., Floriddia, E. M., Barnabe-Heider, F. and Fernandes, K. J. 2015 Endogenous neural stem cell responses to stroke and spinal cord injury. Glia 63, 1469-1482.

Grimes, D. T., Boswell, C. W., Morante, N. F., Henkelman, R. M., Burdine, R. D. and Ciruna, B. 2016 Zebrafish models of idiopathic scoliosis link cerebrospinal fluid flow defects to spine curvature. Science 352, 1341-1344.

Guo, F., Maeda, Y., Ma, J., Delgado, M., Sohn, J., Miers, L., Ko, E. M., Bannerman, P., Xu, J. and Wang, Y. 2011 Macroglial plasticity and the origins of reactive astroglia in experimental autoimmune encephalomyelitis. J. Neurosci. 31, 11914-11928.

Guo, Y., Ma, L., Cristofanilli, M., Hart, R. P., Hao, A. and Schachner, M. 2011 Transcription factor Sox11b is involved in spinal cord regeneration in adult zebrafish. Neuroscience 172, 329-341.

Hamilton, L., Truong, M., Bednarczyk, M., Aumont, A. and Fernandes, K. 2009 Cellular organization of the central canal ependymal zone, a niche of latent neural stem cells in the adult mammalian spinal cord. Neuroscience 164, 1044-1056.

Han, Y.-G. and Alvarez-Buylla, A. 2010 Role of primary cilia in brain development and cancer. Curr. Opin. Neurobiol. 20, 58-67.

Hol, E. M. and Pekny, M. 2015 Glial fibrillary acidic protein (GFAP) and the astrocyte intermediate filament system in diseases of the central nervous system. Curr. Opin. Cell Biol. 32, 121-130.

49

Huang, A. L., Chen, X., Hoon, M. A., Chandrashekar, J., Guo, W., Trankner, D., Ryba, N. J. and Zuker, C. S. 2006 The cells and logic for mammalian sour taste detection. Nature 442, 934-938.

Huang, Y., Li, Y., Chen, J., Zhou, H. and Tan, S. 2015 Electrical Stimulation Elicits Neural Stem Cells Activation: New Perspectives in CNS Repair. Frontiers in human neuroscience 9, 586.

Hugnot, J.-P. and Franzen, R. 2011 The spinal cord ependymal region: a stem cell niche in the caudal central nervous system. Frontiers in bioscience (Landmark edition) 16, 1044-1059.

Hugnot, J. P. 2012 The spinal cord neural stem cell niche. In Neural Stem Cells and Therapy, pp. 71-92: InTech.

Hui, S. P., Nag, T. C. and Ghosh, S. 2015 Characterization of Proliferating Neural Progenitors after Spinal Cord Injury in Adult Zebrafish. PLoS One 10, e0143595.

Hui, S. P., Sengupta, D., Lee, S. G., Sen, T., Kundu, S., Mathavan, S. and Ghosh, S. 2014 Genome wide expression profiling during spinal cord regeneration identifies comprehensive cellular responses in zebrafish. PLoS One 9, e84212.

Hui, S. P., Sheng, D. Z., Sugimoto, K., Gonzalez-Rajal, A., Nakagawa, S., Hesselson, D. and Kikuchi, K. 2017 Zebrafish Regulatory T Cells Mediate Organ-Specific Regenerative Programs. Dev Cell 43, 659-672 e655.

Isaksson, J., Faroque, M., Holtz, A., Hillered, L. and Olsson, Y. 1999 Expression of ICAM-1 and CD11b after experimental spinal cord injury in rats. J. Neurotrauma 16, 165-173.

Iwasa, S. N., Babona-Pilipos, R. and Morshead, C. M. 2017 Environmental Factors That Influence Stem Cell Migration: An "Electric Field". Stem Cells Int 2017, 4276927.

Jacquet, B. V., Salinas-Mondragon, R., Liang, H., Therit, B., Buie, J. D., Dykstra, M., Campbell, K., Ostrowski, L. E., Brody, S. L. and Ghashghaei, H. T. 2009 FoxJ1-dependent gene expression is required for differentiation of radial glia into ependymal cells and a subset of astrocytes in the postnatal brain. Development 136, 4021-4031.

Johansson, C. B., Momma, S., Clarke, D. L., Risling, M., Lendahl, U. and Frisen, J. 1999 Identification of a neural stem cell in the adult mammalian central nervous system. Cell 96, 25-34.

Khazanov, S., Paz, Y., Hefetz, A., Gonzales, B. J., Netser, Y., Mansour, A. A. and Ben-Arie, N. 2016 Floor plate descendants in the ependyma of the adult mouse Central Nervous System. Int. J. Dev. Biol. 61, 257-265.

Kim, H., Zahir, T., Tator, C. H. and Shoichet, M. S. 2011 Effects of dibutyryl cyclic-AMP on survival and neuronal differentiation of neural stem/progenitor cells transplanted into spinal cord injured rats. PLoS One 6, e21744.

Kirsche, W. 1950 Die regenerativen Vorgänge am Rückenmark erwachsener Teleostier nach operativer Kontinuitätstrennung. Z. Mikrosk Anat. Forsch. 77, 190-265.

50

Kojima, A. and Tator, C. H. 2002 Intrathecal administration of epidermal growth factor and fibroblast growth factor 2 promotes ependymal proliferation and functional recovery after spinal cord injury in adult rats. J Neurotrauma 19, 223-238.

Kondrychyn, I., Teh, C., Sin, M. and Korzh, V. 2013 Stretching morphogenesis of the roof plate and formation of the central canal. PLoS One 8, e56219.

Kroehne, V., Freudenreich, D., Hans, S., Kaslin, J. and Brand, M. 2011 Regeneration of the adult zebrafish brain from neurogenic radial glia-type progenitors. Development 138, 4831-4841.

Kulbatski, I. and Tator, C. H. 2009 Region-specific differentiation potential of adult rat spinal cord neural stem/precursors and their plasticity in response to in vitro manipulation. J. Histochem. Cytochem. 57, 405-423.

Kulbatski, I., Mothe, A. J., Keating, A., Hakamata, Y., Kobayashi, E. and Tator, C. H. 2007 Oligodendrocytes and radial glia derived from adult rat spinal cord progenitors: morphological and immunocytochemical characterization. J. Histochem. Cytochem. 55, 209-222.

Kumar, S., Filipski, A., Swarna, V., Walker, A. and Hedges, S. B. 2005 Placing confidence limits on the molecular age of the human–chimpanzee divergence. Proc. Natl. Acad. Sci. USA 102, 18842-18847.

Kuscha, V., Barreiro-Iglesias, A., Becker, C. G. and Becker, T. 2012a Plasticity of tyrosine hydroxylase and serotonergic systems in the regenerating spinal cord of adult zebrafish. J Comp Neurol 520, 933-951.

Kuscha, V., Frazer, S. L., Dias, T. B., Hibi, M., Becker, T. and Becker, C. G. 2012b Lesion-induced generation of interneuron cell types in specific dorsoventral domains in the spinal cord of adult zebrafish. J. Comp. Neurol. 520, 3604-3616.

Kútna, V., Ševc, J., Gombalová, Z., Matiašová, A. and Daxnerová, Z. 2014 Enigmatic cerebrospinal fluid-contacting neurons arise even after the termination of neurogenesis in the rat spinal cord during embryonic development and retain their immature-like characteristics until adulthood. Acta Histochem. 116, 278-285.

Kyritsis, N., Kizil, C., Zocher, S., Kroehne, V., Kaslin, J., Freudenreich, D., Iltzsche, A. and Brand, M. 2012 Acute inflammation initiates the regenerative response in the adult zebrafish brain. Science 338, 1353-1356.

Lacroix, S., Hamilton, L. K., Vaugeois, A., Beaudoin, S., Breault-Dugas, C., Pineau, I., Levesque, S. A., Gregoire, C.-A. and Fernandes, K. J. 2014 Central canal ependymal cells proliferate extensively in response to traumatic spinal cord injury but not demyelinating lesions. PLoS One 9, e85916.

Lang, B. T., Cregg, J. M., DePaul, M. A., Tran, A. P., Xu, K., Dyck, S. M., Madalena, K. M., Brown, B. P., Weng, Y. L., Li, S. et al. 2014 Modulation of the proteoglycan receptor PTPsigma promotes recovery after spinal cord injury. Nature.

Li, X., Floriddia, E. M., Toskas, K., Fernandes, K. J. L., Guerout, N. and Barnabe-Heider, F. 2016 Regenerative Potential of Ependymal Cells for Spinal Cord Injuries Over Time. EBioMedicine 13, 55-65.

51

Li, X., Xiao, Z., Han, J., Chen, L., Xiao, H., Ma, F., Hou, X., Li, X., Sun, J., Ding, W. et al. 2013 Promotion of neuronal differentiation of neural progenitor cells by using EGFR antibody functionalized collagen scaffolds for spinal cord injury repair. Biomaterials 34, 5107-5116.

Lim, D. A. and Alvarez-Buylla, A. 2016 The adult ventricular–subventricular zone (V-SVZ) and olfactory bulb (OB) neurogenesis. Cold Spring Harb Perspect Biol 8, a018820.

Lin, J. F., Pan, H. C., Ma, L. P., Shen, Y. Q. and Schachner, M. 2012 The cell neural adhesion molecule contactin-2 (TAG-1) is beneficial for functional recovery after spinal cord injury in adult zebrafish. PLoS One 7, e52376.

Liu, C. J., Xie, L., Cui, C., Chu, M., Zhao, H. D., Yao, L., Li, Y. H., Schachner, M. and Shen, Y. Q. 2016 Beneficial roles of melanoma cell adhesion molecule in spinal cord transection recovery in adult zebrafish. J Neurochem 139, 187-196.

Liu, K., Wang, Z., Wang, H. and Zhang, Y. 2002 Nestin expression and proliferation of ependymal cells in adult rat spinal cord after injury. Chin. Med. J. 115, 339-341.

Lowery, L. A. and Sive, H. 2004 Strategies of vertebrate neurulation and a re-evaluation of teleost neural tube formation. Mech Dev 121, 1189-1197.

Lu, P., Woodruff, G., Wang, Y., Graham, L., Hunt, M., Wu, D., Boehle, E., Ahmad, R., Poplawski, G., Brock, J. et al. 2014 Long-distance axonal growth from human induced pluripotent stem cells after spinal cord injury. Neuron 83, 789-796.

Lu, P., Wang, Y., Graham, L., McHale, K., Gao, M., Wu, D., Brock, J., Blesch, A., Rosenzweig, E. S., Havton, L. A. et al. 2012 Long-distance growth and connectivity of neural stem cells after severe spinal cord injury. Cell 150, 1264-1273.

Magnusson, J. P. and Frisen, J. 2016 Stars from the darkest night: unlocking the neurogenic potential of astrocytes in different brain regions. Development 143, 1075-1086.

Maikos, J. T. and Shreiber, D. I. 2007 Immediate damage to the blood-spinal cord barrier due to mechanical trauma. J. Neurotrauma 24, 492-507.

Marichal, N., Garcia, G., Radmilovich, M., Trujillo-Cenoz, O. and Russo, R. E. 2009 Enigmatic central canal contacting cells: immature neurons in "standby mode"? J Neurosci 29, 10010-10024.

Martens, D. J., Seaberg, R. M. and Van Der Kooy, D. 2002 In vivo infusions of exogenous growth factors into the fourth ventricle of the adult mouse brain increase the proliferation of neural progenitors around the fourth ventricle and the central canal of the spinal cord. Eur. J. Neurosci. 16, 1045-1057.

Matthews, M., Onge, M. S. and Faciane, C. 1979 An electron microscopic analysis of abnormal ependymal cell proliferation and envelopment of sprouting axons following spinal cord transection in the rat. Acta Neuropathol 45, 27-36.

McDonough, A. and Martínez-Cerdeño, V. 2012 Endogenous proliferation after spinal cord injury in animal models. Stem Cells Int. 2012.

52

McHedlishvili, L., Epperlein, H. H., Telzerow, A. and Tanaka, E. M. 2007 A clonal analysis of neural progenitors during axolotl spinal cord regeneration reveals evidence for both spatially restricted and multipotent progenitors. Development 134, 2083-2093.

Meletis, K., Barnabé-Heider, F., Carlén, M., Evergren, E., Tomilin, N., Shupliakov, O. and Frisén, J. 2008 Spinal cord injury reveals multilineage differentiation of ependymal cells. PLoS Biol. 6, e182.

Merino, J. J., Bellver-Landete, V., Oset-Gasque, M. J. and Cubelos, B. 2015 CXCR4/CXCR7 molecular involvement in neuronal and neural progenitor migration: focus in CNS repair. J Cell Physiol 230, 27-42.

Milhorat, T. H., Kotzen, R. M. and Anzil, A. P. 1994 Stenosis of central canal of spinal cord in man: incidence and pathological findings in 232 autopsy cases. J Neurosurg 80, 716-722.

Mirzadeh, Z., Merkle, F. T., Soriano-Navarro, M., Garcia-Verdugo, J. M. and Alvarez-Buylla, A. 2008 Neural stem cells confer unique pinwheel architecture to the ventricular surface in neurogenic regions of the adult brain. Cell Stem Cell 3, 265-278.

Mladinic, M., Bianchetti, E., Dekanic, A., Mazzone, G. L. and Nistri, A. 2014 ATF3 is a novel nuclear marker for migrating ependymal stem cells in the rat spinal cord. Stem cell research 12, 815-827.

Mokalled, M. H., Patra, C., Dickson, A. L., Endo, T., Stainier, D. Y. and Poss, K. D. 2016 Injury-induced ctgfa directs glial bridging and spinal cord regeneration in zebrafish. Science 354, 630-634.

Moreno-Manzano, V., Rodriguez-Jimenez, F. J., Garcia-Rosello, M., Lainez, S., Erceg, S., Calvo, M. T., Ronaghi, M., Lloret, M., Planells-Cases, R., Sanchez-Puelles, J. M. et al. 2009 Activated spinal cord ependymal stem cells rescue neurological function. Stem Cells 27, 733-743.

Motavkin, P. and Bakhtinov, A. 1973 Postnatal development of human spinal cord ependymal innervation. Neurosci. Behav. Physiol. 6, 253-259.

Mothe, A. and Tator, C. 2005 Proliferation, migration, and differentiation of endogenous ependymal region stem/progenitor cells following minimal spinal cord injury in the adult rat. Neuroscience 131, 177-187.

Mothe, A. J., Zahir, T., Santaguida, C., Cook, D. and Tator, C. H. 2011 Neural stem/progenitor cells from the adult human spinal cord are multipotent and self-renewing and differentiate after transplantation. PLoS One 6, e27079.

Nakayama, Y. and Kohno, K. 1974 Number and polarity of the ependymal cilia in the central canal of some vertebrates. J. Neurocytol. 3, 449-458.

Namiki, J. and Tator, C. H. 1999 Cell proliferation and nestin expression in the ependyma of the adult rat spinal cord after injury. Journal of neuropathology and experimental neurology 58, 489-498.

53

North, H. A., Pan, L., McGuire, T. L., Brooker, S. and Kessler, J. A. 2015 β1-Integrin alters ependymal stem cell BMP receptor localization and attenuates astrogliosis after spinal cord injury. J. Neurosci. 35, 3725-3733.

Ogai, K., Nakatani, K., Hisano, S., Sugitani, K., Koriyama, Y. and Kato, S. 2014 Function of Sox2 in ependymal cells of lesioned spinal cords in adult zebrafish. Neurosci Res 88, 84-87.

Ohnmacht, J., Yang, Y., Maurer, G. W., Barreiro-Iglesias, A., Tsarouchas, T. M., Wehner, D., Sieger, D., Becker, C. G. and Becker, T. 2016 Spinal motor neurons are regenerated after mechanical lesion and genetic ablation in larval zebrafish. Development 143, 1464-1474.

Orendáčová, J., Račeková, E., Kuchárová, K., Poušová, B., Ondrejčák, T., Martončíková, M., Daxnerová, Z. and Maršala, J. 2004 Ependyma as a possible morphological basis of ischemic preconditioning tolerance in rat spinal cord ischemia model: nestin and Fluoro-Jade B observations. Cell. Mol. Neurobiol. 24, 477-489.

Orts-Del'Immagine, A., Trouslard, J., Airault, C., Hugnot, J. P., Cordier, B., Doan, T., Kastner, A. and Wanaverbecq, N. 2017 Postnatal maturation of mouse medullo-spinal cerebrospinal fluid-contacting neurons. Neuroscience 343, 39-54.

Orts-Del’Immagine, A., Kastner, A., Tillement, V., Tardivel, C., Trouslard, J. and Wanaverbecq, N. 2014 Morphology, distribution and phenotype of polycystin kidney disease 2-like 1-positive cerebrospinal fluid contacting neurons in the brainstem of adult mice. PLoS One 9, e87748.

Oumesmar, B. N., Vignais, L., Duhamel‐Clérin, E., Avellana‐Adalid, V., Rougon, G. and Evercooren, A. 1995 Expression of the highly polysialylated neural cell adhesion molecule during postnatal myelination and following chemically induced demyelination of the adult mouse spinal cord. Eur. J. Neurosci. 7, 480-491.

Pan, H. C., Lin, J. F., Ma, L. P., Shen, Y. Q. and Schachner, M. 2013 Major vault protein promotes locomotor recovery and regeneration after spinal cord injury in adult zebrafish. Eur J Neurosci 37, 203-211.

Park, E., Velumian, A. A. and Fehlings, M. G. 2004 The role of excitotoxicity in secondary mechanisms of spinal cord injury: a review with an emphasis on the implications for white matter degeneration. J Neurotrauma 21, 754-774.

Pastrana, E., Cheng, L.-C. and Doetsch, F. 2009 Simultaneous prospective purification of adult subventricular zone neural stem cells and their progeny. Proc. Natl. Acad. Sci. (USA) 106, 6387-6392.

Peters, C. M., Rogers, S. D., Pomonis, J. D., Egnazyck, G. F., Keyser, C. P., Schmidt, J. A., Ghilardi, J. R., Maggio, J. E. and Mantyh, P. W. 2003 Endothelin receptor expression in the normal and injured spinal cord: potential involvement in injury-induced ischemia and gliosis. Exp. Neurol. 180, 1-13.

Petracca, Y. L., Sartoretti, M. M., Di Bella, D. J., Marin-Burgin, A., Carcagno, A. L., Schinder, A. F. and Lanuza, G. M. 2016 The late and dual origin of cerebrospinal fluid-contacting neurons in the mouse spinal cord. Development 143, 880-891.

Pfenninger, C. V., Steinhoff, C., Hertwig, F. and Nuber, U. A. 2011 Prospectively isolated CD133/CD24‐positive ependymal cells from the adult spinal cord and lateral

54

ventricle wall differ in their long‐term in vitro self‐renewal and in vivo gene expression. Glia 59, 68-81.

Pontes, A., Zhang, Y. and Hu, W. 2013 Novel functions of GABA signaling in adult neurogenesis. Frontiers in biology 8, 496-507.

Quiñones‐Hinojosa, A., Sanai, N., Soriano‐Navarro, M., Gonzalez‐Perez, O., Mirzadeh, Z., Gil‐Perotin, S., Romero‐Rodriguez, R., Berger, M. S., Garcia‐Verdugo, J. M. and Alvarez‐Buylla, A. 2006 Cellular composition and cytoarchitecture of the adult human subventricular zone: a niche of neural stem cells. J. Comp. Neurol. 494, 415-434.

Ravanelli, A. M. and Appel, B. 2015 Motor neurons and oligodendrocytes arise from distinct cell lineages by progenitor recruitment. Genes Dev 29, 2504-2515.

Reichenbach, A. and Wolburg, W. 2013 Astrocytes and ependymal glia. In Neuroglia (eds H. Kettenmann and B. R. Ransom), pp. 35-49: Oxford University Press.

Reimer, M. M., Sorensen, I., Kuscha, V., Frank, R. E., Liu, C., Becker, C. G. and Becker, T. 2008 Motor neuron regeneration in adult zebrafish. J Neurosci 28, 8510-8516.

Reimer, M. M., Kuscha, V., Wyatt, C., Sorensen, I., Frank, R. E., Knuwer, M., Becker, T. and Becker, C. G. 2009 Sonic hedgehog is a polarized signal for motor neuron regeneration in adult zebrafish. J Neurosci 29, 15073-15082.

Reimer, M. M., Norris, A., Ohnmacht, J., Patani, R., Zhong, Z., Dias, T. B., Kuscha, V., Scott, A. L., Chen, Y. C., Rozov, S. et al. 2013 Dopamine from the Brain Promotes Spinal Motor Neuron Generation during Development and Adult Regeneration. Dev Cell 25, 478-491.

Ren, Y., Ao, Y., O’Shea, T. M., Burda, J. E., Bernstein, A. M., Brumm, A. J., Muthusamy, N., Ghashghaei, H. T., Carmichael, S. T. and Cheng, L. 2017 Ependymal cell contribution to scar formation after spinal cord injury is minimal, local and dependent on direct ependymal injury. Sci. Rep. 7.

Ribeiro, A., Monteiro, J. F., Certal, A. C., Cristovao, A. M. and Saude, L. 2017 Foxj1a is expressed in ependymal precursors, controls central canal position and is activated in new ependymal cells during regeneration in zebrafish. Open Biol 7.

Rodriguez, E. M., Blazquez, J. L., Pastor, F. E., Pelaez, B., Pena, P., Peruzzo, B. and Amat, P. 2005 Hypothalamic tanycytes: a key component of brain-endocrine interaction. Int Rev Cytol 247, 89-164.

Rusanescu, G. 2016 Adult spinal cord neurogenesis: A regulator of nociception. Neurogenesis 3, e1256853.

Rusanescu, G. and Mao, J. 2015 Immature spinal cord neurons are dynamic regulators of adult nociceptive sensitivity. J. Cell. Mol. Med. 19, 2352-2364.

Rusanescu, G. and Mao, J. 2017 Peripheral nerve injury induces adult brain neurogenesis and remodelling. J. Cell. Mol. Med. 21, 299-314.

55

Sabelström, H., Stenudd, M., Réu, P., Dias, D. O., Elfineh, M., Zdunek, S., Damberg, P., Göritz, C. and Frisén, J. 2013 Resident neural stem cells restrict tissue damage and neuronal loss after spinal cord injury in mice. Science 342, 637-640.

Sabourin, J. C., Ackema, K. B., Ohayon, D., Guichet, P. O., Perrin, F. E., Garces, A., Ripoll, C., Charité, J., Simonneau, L. and Kettenmann, H. 2009 A Mesenchymal‐Like ZEB1+ Niche Harbors Dorsal Radial Glial Fibrillary Acidic Protein‐Positive Stem Cells in the Spinal Cord. Stem Cells 27, 2722-2733.

Sakakibara, A., Aoki, E., Hashizume, Y., Mori, N. and Nakayama, A. 2007 Distribution of nestin and other stem cell‐related molecules in developing and diseased human spinal cord. Pathol. Int. 57, 358-368.

Samaddar, S., Vazquez, K., Ponkia, D., Toruno, P., Sahbani, K., Begum, S., Abouelela, A., Mekhael, W. and Ahmed, Z. 2017 Transspinal direct current stimulation modulates migration and proliferation of adult newly born spinal cells in mice. J Appl Physiol (1985) 122, 339-353.

Sanai, N., Tramontin, A. D., Quinones-Hinojosa, A. and Barbaro, N. M. 2004 Unique astrocyte ribbon in adult human brain contains neural stem cells but lacks chain migration. Nature 427, 740.

Sarnat, H. B. 1992 Regional differentiation of the human fetal ependyma: immunocytochemical markers. J. Neuropathol. Exp. Neurol. 51, 58-75.

Sawamoto, K., Hirota, Y., Alfaro‐Cervello, C., Soriano‐Navarro, M., He, X., Hayakawa‐Yano, Y., Yamada, M., Hikishima, K., Tabata, H. and Iwanami, A. 2011 Cellular composition and organization of the subventricular zone and rostral migratory stream in the adult and neonatal common marmoset brain. J. Comp. Neurol. 519, 690-713.

Schnapp, E., Kragl, M., Rubin, L. and Tanaka, E. M. 2005 Hedgehog signaling controls dorsoventral patterning, blastema cell proliferation and cartilage induction during axolotl tail regeneration. Development 132, 3243-3253.

Schweitzer, J., Becker, T., Becker, C. G. and Schachner, M. 2003 Expression of protein zero is increased in lesioned axon pathways in the central nervous system of adult zebrafish. Glia 41, 301-317.

Seitz, R., Lohler, J. and Schwendemann, G. 1981 Ependyma and meninges of the spinal cord of the mouse. A light-and electron-microscopic study. Cell Tissue Res 220, 61-72.

Setoguchi, T., Nakashima, K., Takizawa, T., Yanagisawa, M., Ochiai, W., Okabe, M., Yone, K., Komiya, S. and Taga, T. 2004 Treatment of spinal cord injury by transplantation of fetal neural precursor cells engineered to express BMP inhibitor. Exp Neurol 189, 33-44.

Sevc, J., Daxnerova, Z. and Miklosova, M. 2009 Role of radial glia in transformation of the primitive lumen to the central canal in the developing rat spinal cord. Cell Mol Neurobiol 29, 927-936.

Sevc, J., Matiasova, A., Kutna, V. and Daxnerova, Z. 2014 Evidence that the central canal lining of the spinal cord contributes to oligodendrogenesis during postnatal development and adulthood in intact rats. J Comp Neurol 522, 3194-3207.

56

Sharp, K. G., Yee, K. M. and Steward, O. 2014 A re-assessment of long distance growth and connectivity of neural stem cells after severe spinal cord injury. Exp Neurol 257, 186-204.

Shechter, R., Ziv, Y. and Schwartz, M. 2007 New GABAergic interneurons supported by myelin‐specific T cells are formed in intact adult spinal cord. Stem Cells 25, 2277-2282.

Shechter, R., Baruch, K., Schwartz, M. and Rolls, A. 2011 Touch gives new life: mechanosensation modulates spinal cord adult neurogenesis. Mol. Psychiatry 16, 342.

Shibuya, S., Miyamoto, O., Auer, R., Itano, T., Mori, S. and Norimatsu, H. 2002 Embryonic intermediate filament, nestin, expression following traumatic spinal cord injury in adult rats. Neuroscience 114, 905-916.

Shihabuddin, L. S., Horner, P. J., Ray, J. and Gage, F. H. 2000 Adult spinal cord stem cells generate neurons after transplantation in the adult dentate gyrus. J Neurosci 20, 8727-8735.

Siegenthaler, J. A. and Pleasure, S. J. 2011 We have got you ‘covered’: how the meninges control brain development. Curr. Opin. Genet. Dev. 21, 249-255.

Sirko, S., Behrendt, G., Johansson, P. A., Tripathi, P., Costa, M. R., Bek, S., Heinrich, C., Tiedt, S., Colak, D. and Dichgans, M. 2013 Reactive glia in the injured brain acquire stem cell properties in response to sonic hedgehog. Cell Stem Cell 12, 426-439.

Spassky, N., Merkle, F. T., Flames, N., Tramontin, A. D., García-Verdugo, J. M. and Alvarez-Buylla, A. 2005 Adult ependymal cells are postmitotic and are derived from radial glial cells during embryogenesis. J. Neurosci. 25, 10-18.

Stenudd, M., Sabelstrom, H. and Frisen, J. 2015 Role of endogenous neural stem cells in spinal cord injury and repair. JAMA neurology 72, 235-237.

Stoeckel, M. E., Uhl-Bronner, S., Hugel, S., Veinante, P., Klein, M. J., Mutterer, J., Freund-Mercier, M. J. and Schlichter, R. 2003 Cerebrospinal fluid-contacting neurons in the rat spinal cord, a gamma-aminobutyric acidergic system expressing the P2X2 subunit of purinergic receptors, PSA-NCAM, and GAP-43 immunoreactivities: light and electron microscopic study. J Comp Neurol 457, 159-174.

Strand, N. S., Hoi, K. K., Phan, T. M., Ray, C. A., Berndt, J. D. and Moon, R. T. 2016 Wnt/beta-catenin signaling promotes regeneration after adult zebrafish spinal cord injury. Biochem Biophys Res Commun 477, 952-956.

Sturrock, R. 1981 An electron microscopic study of the development of the ependyma of the central canal of the mouse spinal cord. J. Anat. 132, 119.

Takahashi, M., Arai, Y., Kurosawa, H., Sueyoshi, N. and Shirai, S. 2003 Ependymal cell reactions in spinal cord segments after compression injury in adult rat. J. Neuropathol. Exp. Neurol. 62, 185-194.

57

Tanaka, O., Yoshioka, T. and Shinohara, H. 1988 Secretory activity in the floor plate neuroepithelium of the developing human spinal cord: morphological evidence. The Anatomical Record 222, 185-190.

Tazaki, A., Tanaka, E. M. and Fei, J. F. 2017 Salamander spinal cord regeneration: The ultimate positive control in vertebrate spinal cord regeneration. Dev Biol 432, 63-71.

Titze, B. and Genoud, C. 2016 Volume scanning electron microscopy for imaging biological ultrastructure. Biol Cell 108, 307-323.

Tysseling, V. M., Mithal, D., Sahni, V., Birch, D., Jung, H., Miller, R. J. and Kessler, J. A. 2011 SDF1 in the dorsal corticospinal tract promotes CXCR4+ cell migration after spinal cord injury. J Neuroinflammation 8, 16.

Vaquero, J., Ramiro, M., Oya, S. and Cabezudo, J. 1981 Ependymal reaction after experimental spinal cord injury. Acta neurochirurgica 55, 295-302.

Vessal, M., Aycock, A., Garton, M. T., Ciferri, M. and Darian‐Smith, C. 2007 Adult neurogenesis in primate and rodent spinal cord: comparing a cervical dorsal rhizotomy with a dorsal column transection. Eur. J. Neurosci. 26, 2777-2794.

Vigh, B., Frank, C., Vincze, C., Czirok, S., Szabó, A., Lukáts, A. and Szél, A. 2004 The system of cerebrospinal fluid-contacting neurons. Its supposed role in the nonsynaptic signal transmission of the brain. Histology and histopathology 19, 607-628.

von Lenhossék, M. 1891 Zur Kenntnis der Neuroglia des menschlichen Rückenmarkes. (Verh. Anat. Ges. 5, 193–221.

Wagner, M., Thaller, C., Jessell, T. and Eichele, G. 1990 Polarizing activity and retinoid synthesis in the floor plate of the neural tube. Nature 345, 819.

Wehner, D., Tsarouchas, T. M., Michael, A., Haase, C., Weidinger, G., Reimer, M. M., Becker, T. and Becker, C. G. 2017 Wnt signaling controls pro-regenerative Collagen XII in functional spinal cord regeneration in zebrafish. Nat Commun 8, 126.

Weiss, S., Dunne, C., Hewson, J., Wohl, C., Wheatley, M., Peterson, A. C. and Reynolds, B. A. 1996 Multipotent CNS stem cells are present in the adult mammalian spinal cord and ventricular neuroaxis. J. Neurosci. 16, 7599-7609.

Weller, M., Wick, W., Aldape, K., Brada, M., Berger, M., Pfister, S. M., Nishikawa, R., Rosenthal, M., Wen, P. Y., Stupp, R. et al. 2015 Glioma. Nature reviews. Disease primers 1, 15017.

White, B. D., Nathe, R. J., Maris, D. O., Nguyen, N. K., Goodson, J. M., Moon, R. T. and Horner, P. J. 2010 β‐Catenin Signaling Increases in Proliferating NG2+ Progenitors and Astrocytes during Post‐Traumatic Gliogenesis in the Adult Brain. Stem Cells 28, 297-307.

White, D. T., Eroglu, A. U., Wang, G., Zhang, L., Sengupta, S., Ding, D., Rajpurohit, S. K., Walker, S. L., Ji, H., Qian, J. et al. 2016 ARQiv-HTS, a versatile whole-organism screening platform enabling in vivo drug discovery at high-throughput rates. Nat Protoc 11, 2432-2453.

58

Wilson, L. and Maden, M. 2005 The mechanisms of dorsoventral patterning in the vertebrate neural tube. Dev Biol 282, 1-13.

Xu, W., Sachewsky, N., Azimi, A., Hung, M., Gappasov, A. and Morshead, C. M. 2017 Myelin basic protein regulates primitive and definitive neural stem cell proliferation from the adult spinal cord. Stem Cells 35, 485-496.

Xu, Y., Kitada, M., Yamaguchi, M., Dezawa, M. and Ide, C. 2006 Increase in bFGF-responsive neural progenitor population following contusion injury of the adult rodent spinal cord. Neurosci. Lett. 397, 174-179.

Yamamoto, S.-i., Nagao, M., Sugimori, M., Kosako, H., Nakatomi, H., Yamamoto, N., Takebayashi, H., Nabeshima, Y.-i., Kitamura, T. and Weinmaster, G. 2001 Transcription factor expression and Notch-dependent regulation of neural progenitors in the adult rat spinal cord. J. Neurosci. 21, 9814-9823.

Yang, H., Lu, P., McKay, H. M., Bernot, T., Keirstead, H., Steward, O., Gage, F. H., Edgerton, V. R. and Tuszynski, M. H. 2006 Endogenous neurogenesis replaces oligodendrocytes and astrocytes after primate spinal cord injury. J. Neurosci. 26, 2157-2166.

Yanger, K., Knigin, D., Zong, Y., Maggs, L., Gu, G., Akiyama, H., Pikarsky, E. and Stanger, B. Z. 2014 Adult hepatocytes are generated by self-duplication rather than stem cell differentiation. Cell Stem Cell 15, 340-349.

Yao, L. and Li, Y. 2016 The Role of Direct Current Electric Field-Guided Stem Cell Migration in Neural Regeneration. Stem cell reviews 12, 365-375.

Yasui, K., Hashizume, Y., Yoshida, M., Kameyama, T. and Sobue, G. 1999 Age-related morphologic changes of the central canal of the human spinal cord. Acta Neuropathol 97, 253-259.

Yin, L., Maddison, L. A. and Chen, W. 2016 Multiplex conditional mutagenesis in zebrafish using the CRISPR/Cas system. Methods Cell Biol 135, 3-17.

Yu, K., McGlynn, S. and Matise, M. P. 2013 Floor plate-derived sonic hedgehog regulates glial and ependymal cell fates in the developing spinal cord. Development 140, 1594-1604.

Yu, Y. and Schachner, M. 2013 Syntenin-a promotes spinal cord regeneration following injury in adult zebrafish. Eur J Neurosci 38, 2280-2289.

Zamora, A. J. 1978 The ependymal and glial configuration in the spinal cord of urodeles. Anat Embryol (Berl) 154, 67-82.

Zappaterra, M. W. and Lehtinen, M. K. 2012 The cerebrospinal fluid: regulator of neurogenesis, behavior, and beyond. Cell Mol Life Sci 69, 2863-2878.

59

TABLES:

Table 1: Evolutionary comparison of cells in the ependymal zone of the spinal cord

Cilia CSF-N Dorsol-ventral

regionalisation

Rostro-caudal

regionalisa-tion

Retention of developmental

gene expression

Shared ependyma

specific gene

expression

Expression of other genes

Cell types generated

after lesion

Zebrafish ✓ ✓ ✓ Not known FoxA2 ?, Nkx6.1,

Pax6, Shh, Sox2, Sox11

FoxJ1 GFAP, Vim, nestin, CD15 ?

Motor neurons,

interneurons, astrocyte-like bridging glia

Rodents ✓ ✓ ✓ ✓ FoxA2, Nkx6.1,

Pax6, Shh, Sox2, Sox9

FoxJ1 GFAP, Vim, Nestin, CD15

Astrocytes, Oligodendro-

cytes

Primates ✓(presence of multi-ciliated cells)

✓(macaque)

No evidence in

human

Not known FoxA2, Nkx6.1,

Pax6, Sox2

FoxJ1 GFAP, Vim, Nestin, CD15

Not known

60

FIGURE LEGENDS:

Fig. 1: Distinct progenitor domains give rise to different types of neurons after

spinal cord transection of adult zebrafish. A schematic cross-section through

the adult spinal cord is shown, indicating domains of ependymo-radial glial

cells (ERGs) that are defined by expression of the indicated transcription

factors. These ERG domains give rise to the indicated neuronal cell types in

response to recapitulated developmental signals and injury-related signals.

See sections 2.3 and 2.4 for relevant references.

Fig. 2: The organisation of the ependymal zone is complex across

vertebrates. A schematic representation of the EZ is shown for different

species. Note that for humans, the EZ is represented for a young adult before

obstruction of the central canal (see text). The presence of radial cells

contacting vessels in humans is extrapolated from the macaque EZ, but has

not been directly shown.

Figure References:

1-Sabourin JC, Ackema KB, Ohayon D, Guichet PO, Perrin FE, Garces A,

Ripoll C, Charité J, Simonneau L, Kettenmann H (2009) A Mesenchymal‐Like

ZEB1+ Niche Harbors Dorsal Radial Glial Fibrillary Acidic Protein‐Positive

Stem Cells in the Spinal Cord. Stem cells 27: 2722-2733

61

2-Yu K, McGlynn S, Matise MP (2013) Floor plate-derived sonic hedgehog

regulates glial and ependymal cell fates in the developing spinal cord.

Development 140: 1594-1604

3-Gensat Atlas (http://www.gensat.org/imagenavigator.jsp?imageID=5671)

4-Khazanov S, Paz Y, Hefetz A, Gonzales BJ, Netser Y, Mansour AA, Ben-

Arie N (2016) Floor platedescendants in the ependyma of the adult mouse

Central Nervous System. International Journal of Developmental Biology 61:

257-265

5-Dromard C, Guillon H, Rigau V, Ripoll C, Sabourin JC, Perrin FE, Scamps

F, Bozza S, Sabatier P, Lonjon N, Duffau H, Vachiery-Lahaye F, Prieto M,

Tran Van Ba C, Deleyrolle L, Boularan A, Langley K, Gaviria M, Privat A,

Hugnot JP, Bauchet L. Adult human spinal cord harbors neural precursor cells

that generate neurons and glial cells in vitro

J Neurosci Res. 2008 Jul;86(9):1916-26.

6-Reimer, M.M., Kuscha, V., Wyatt, C., Sorensen, I., Frank, R.E., Knuwer, M.,

Becker, T., Becker, C.G., 2009. Sonic hedgehog is a polarized signal for

motor neuron regeneration in adult zebrafish. J Neurosci 29, 15073-15082.

7-Reimer, M.M., Sorensen, I., Kuscha, V., Frank, R.E., Liu, C., Becker, C.G.,

Becker, T., 2008. Motor neuron regeneration in adult zebrafish. J Neurosci 28,

8510-8516.

62

Fig. 1

63

Fig. 2

64