Molecular and Cellular Neuroscience 42 (2009) 208–218
Contents lists available at ScienceDirect
Molecular and Cellular Neuroscience j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y m c n e
NG2 cells differentiate into astrocytes in cerebellar slices Giampaolo Leoni a, Marcus Rattray b, Arthur M. Butt a,⁎ a b
Institute of Biology and Biomedical Sciences, School of Pharmacy and Biomedical Sciences, University of Portsmouth, UK Reading School of Pharmacy, University of Reading, UK
a r t i c l e
i n f o
Article history: Received 19 January 2009 Revised 3 June 2009 Accepted 3 July 2009 Available online 17 July 2009 Keywords: Glia NG2 Astrocyte OPC Synantocyte Cerebellum
a b s t r a c t NG2-glia are an abundant population of glial cells that have been considered by many to be oligodendrocyte progenitor cells (OPCs). However, growing evidence suggests that NG2-glia may also be capable of differentiating into astrocytes and neurons under certain conditions. Here, we have examined NG2-glia in cerebellar slices, using transgenic mice in which the astroglial marker glial specific protein (GFAP) drives expression of the reporter gene enhanced green fluorescent protein (EGFP). Immunolabelling for NG2 shows that NG2-glia and GFAP-EGFP astroglia are separate populations in most areas of the brain, although a substantial population of NG2-glia in the pons also express the GFAP-EGFP reporter. In the cerebellum, NG2glia did not express EGFP, either at postnatal day (P)12 or P29–30. We developed an organotypic culture of P12 cerebellar slices that maintain cytoarchitectural integrity of Purkinje neurons and Bergmann glia. In these cultures, BrdU labelling indicates that the majority of NG2-glia enter the cell cycle within 2 days in vitro (DIV), suggesting that NG2-glia undergo a ‘reactive’ response in cerebellar cultures. After 2 DIV NG2-glia began to express the astroglial reporter EGFP and in some cases the respective GFAP protein. However, NG2glia did not acquire phenotypic markers of neural stem cells or neurons. The results suggest that NG2-glia are not lineage restricted OPCs and are a potential source of astrocytes in the cerebellum. © 2009 Elsevier Inc. All rights reserved.
Introduction The NG2 chondroitin sulphate proteoglycan (CSPG) was first shown to identify a major population of glial cells in the CNS by Levine et al. (1986) and Stallcup and Beasley (1987). From the outset, antibodies for NG2 were shown to label bipotential oligodendrocytetype-2 astrocyte (O-2A) progenitor cells (Stallcup and Beasley, 1987), which had been isolated by Raff et al. from optic nerves and shown to differentiate into astrocytes and oligodendrocytes, depending on culture conditions (Raff et al., 1983). In vivo, antibodies to NG2 labelled a substantial population of novel cells in developing and adult brain (Levine et al., 1986), which here we refer to as NG2-glia. NG2glia have been shown to have the antigenic phenotype of oligodendrocyte progenitor cells (OPCs), characterised by expression of markers such as platelet-derived growth factor alpha receptors (PDGFαR), but not markers for mature astrocytes or oligodendrocytes (Nishiyama et al., 1996a,b). Many researchers considered NG2-glia to be OPCs, however there is also a body of evidence which shows that NG2-glia are not simply precursors of other cells but represent an important class of functional cells in the adult CNS with properties typical of differentiated cells, which we have termed synantocytes (e.g. Butt et al., 1999, 2005; Wigley et al., 2007).
⁎ Corresponding author. School of Pharmacy and Biomedical Sciences, University of Portsmouth, St Michael's Building, White Swan Road, Portsmouth PO1 2DT, UK. E-mail address:
[email protected] (A.M. Butt). 1044-7431/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.mcn.2009.07.007
Genetic fate tracing of NG2 glia in NG2-cre/CAGG-Z/EG doubletransgenic mice (Zhu et al., 2008), and PDGFRα/NG2 glia in PdgfracreERT2/Rosa26-YFP double-transgenic mice (Rivers et al., 2008) clearly show that NG2-expressing cells generate oligodendrocytes during development and in the adult. In addition, studies in transgenic mice in which expression of enhanced green fluorescent protein (EGFP) was driven by the glial fibrillary acidic protein (GFAP) promoter indicated low expression of the reporter gene in a small subpopulation of NG2-glia in the hippocampus and respiratory area of the pons (Matthias et al., 2003; Grass et al., 2004). However, evidence that NG2-glia can differentiate into astrocytes in vivo is conflicting. In gray matter, some Olig2-expressing progenitors have been reported to generate astrocytes (Dimou et al., 2008), and the study in NG2-cre/ CAGG-Z/EG double-transgenic mice showed that some GFP-labelled protoplasmic astrocytes developed from NG2 cells, but never neurons (Zhu et al., 2008). In contrast, in PdgfracreERT2/Rosa26-YFP doubletransgenic mice there was no evidence for YFP-labelled astrocytes in gray or white matter, although small numbers of projection neurons were generated in the forebrain (Rivers et al., 2008). Given the lack of clear evidence on whether NG2 cells can become astrocytes, to examine this issue, we developed an ex vivo organotypic culture of cerebellar slices from TgN(GFAPEGFP) GFEC-FKi transgenic mice in which EGFP is expressed by astroglia (Nolte et al., 2001). The results show that, while there is no overlap of EGFP and NG2 in the cerebellum in vivo, cerebellar NG2-glia begin to express the astroglial EGFP reporter gene but not markers for neural stem cells or differentiated neurons within 2 days in organotypic slice cultures.
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
Results Relations between NG2-glia and astroglia in GFAP-EGFP mice We examined whether NG2-glia express the GFAP-EGFP reporter in the line of mice used in this study and found no overlap between
209
NG2 immunolabelling and EGFP expression (Fig. 1), with the exception of the pons (Fig. 2). Astroglia in different areas of the brain were readily identified by their expression of the GFAP-EGFP reporter (Fig. 1), although not all astroglia express EGFP in the TgN (GFAPEGFP)GFEC-FKi transgenic mouse line used (Nolte et al., 2001; Wigley et al., 2007; Wigley and Butt, in press). For the most part, astroglia and NG2-glia could be distinguished as discrete cell types by their morphology and differential expression of NG2 and EGFP, as illustrated in the cerebral cortex (Fig. 1A), hippocampus (Fig. 1B), and striatum (Fig. 1C). NG2-glia and protoplasmic astroglia had overlapping process domains and in many cases had closely apposed cell somata (Fig. 1B). Within a discrete region of the pons, we observed numerous NG2-glia that expressed the astroglial reporter EGFP (Fig. 2). In these EGFP+ NG2-glia, the expression of EGFP was confined to the cell body (Figs. 2B–G), whereas EGFP was expressed even in the finest processes of neighbouring astrocytes (Fig. 2A). EGFP+ NG2-glia in the pons were morphologically distinct from EFGP+ protoplasmic astrocytes, and there was no apparent morphological difference between EGFP+ NG2-glia and EGFP− NG2-glia. In the pons, EGFP+ NG2-glia were observed to form clusters of two or three daughter cells (Fig. 2E); such clones of NG2-glia were observed elsewhere but were not EGFP+. NG2-glia in the cerebellum NG2 immunolabelling of the cerebellum from GFAP-EGFP mice shows that the distribution and orientation of NG2-glia is distinct from that of Bergmann glia (Fig. 3). Bergmann glia have their somata in the Purkinje cell layer (PCL), and extend their long primary processes through the molecular layer (ML) to the pia (Fig. 3A), but otherwise we did not observe EGFP+ astroglial somata in the normal ML. NG2-glia are arranged individually, with cell somata regularly distributed throughout the GCL, PCL and ML of the cerebellum, and have radial process domains that traverse the layers (Fig. 3A). NG2glia in the PCL are often directly apposed to Bergmann glia, and their processes enwrap Bergmann glia somata (Fig. 3B). NG2-glia with cell somata in the ML extend processes radially through the ML to cover the vertical process domains of many Bergmann glia (Fig. 3C). NG2glia in the ML extend processes predominantly in two planes, horizontal and vertical to the primary processes of the Bergmann glia (Fig. 3D), and the vertical processes of NG2-glia are closely related to the primary Bergmann glial cell processes and appear to run together towards the pia (Fig. 3E). Integrity of neuron-glial relations in cerebellar slice cultures
Fig. 1. NG2-glia and astrocytes are distinct populations. Confocal microscopic images of 50 μm thick coronal brain sections from P29 GFAP-EGFP mice immunolabelled for NG2 (red). (A) Cerebral cortex. (B) Hippocampus; insets are individual channels illustrating the directly apposed EGFP-positive astrocyte (asterisk) and NG2-glial cell (star). (C) Striatum. We did not observe expression of the EGFP reporter by NG2-glia within these brain regions. NG2-glia and astrocytes were distinct populations with domains that overlapped considerably.
Neuron-glial relations in cerebellar slices from P12 GFAP-EGFP mice were examined by immunohistochemistry in acute slices and in slices maintained in interface cultures for 2DIV (Fig. 4). The distribution of astroglia clearly demarcated the cytoarchitecture in acute P12 cerebellar slices (Fig. 4A), which appeared equivalent to that described above for mature P29 mice (Fig. 3). Bergmann glial cell somata delineated the GCL and PCL, and their primary radial processes defined the ML to the pia (Fig. 4A). There was little overall change in the cytoarchitecture of the cerebellum after 2 DIV (Fig. 4B), and closer examination of Bergmann glia (Figs. 4C, D) and Purkinje neurons (Figs. 4E, F) shows that their characteristic distribution and process radiation observed in acute slices (Figs. 4C, E) were retained at 2 DIV (Figs. 4D, F). Bergmann glia in the cerebellar cultures retained their vertical primary processes aligned with the dendritic tress of the Purkinje neurons and passing the breadth of the ML (Figs. 4D, F). The distribution and morphology of NG2-glia in acute P12 slices (Fig. 4G) were the same as described above for mature cerebellum (Fig. 3), and were not markedly altered after 2 DIV (Fig. 4H). NG2-glia were distributed throughout the PCL and ML and extended processes radially to traverse multiple cerebellar layers and extend along the
210
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
Fig. 2. Expression of the astroglial reporter EGFP in NG2-glia of the pons. Confocal microscopic images of 50 μm thick coronal sections of the pons from P29 GFAP-EGFP mice immunolabelled for NG2 (red). (A) Projection of a 38 μm thick z-stack illustrating numerous NG2-glia expressing the astroglial reporter protein EGFP (b–g, arrowheads). (B–G) Single plane optical sections corresponding to cells b–g illustrated in (A), showing EGFP expression localised to cell somata of NG2-glia. Note the clone of daughter cells in (E).
primary vertical processes of Bergman glia (Figs. 4G, H). As in the mature cerebellum (Fig. 3), NG2-glia in acute P12 slices were not observed to express the astroglial GFAP-EGFP reporter gene (Fig. 4G). In contrast, EGFP+ NG2-glia were evident throughout the ML after 2 DIV (Fig. 4H). Expression of GFAP-EGFP by NG2-glia in cerebellar slice cultures In normal brain, a small number of NG2-glia were observed to express the EGFP reporter but only in the pons (Fig. 2), and not in the cerebellum (Fig. 3). In contrast, in cerebellar slice cultures there were abundant NG2-glia that expressed the EGFP reporter protein within the ML (Figs. 5 and 6). These EGFP+ NG2-glia were morphologically indistinguishable from NG2-glia in the normal cerebellum and brain (Figs. 1–3). BrdU labelling shows that the majority of NG2-glia in cerebellar slice cultures were in cell cycle (Fig. 5A, insets), and a number of EGFP+ NG2-glia were duplets and triplets, indicating recent cell division (Fig. 5B). Our cerebellar slice cultures were normally maintained in medium containing serum, which has been shown to induce astrocyte differentiation from cultured bipotential O-2A cells, which express NG2 in vitro (Raff et al., 1983; Stallcup and Beasley, 1987). We therefore examined cerebellar slices maintained in serum-free conditions and observed abundant NG2-glia expressing the EGFP reporter protein (Fig. 6), indicating serum did not affect the fate choice of NG2-glia in our experiments. To determine whether any EGFP+ NG2-glia also expressed the GFAP protein, we performed double immunofluorescence labelling for NG2 and GFAP in cerebellar slices from GFAPEGFP mice (Fig. 7). The vast majority of NG2-glia examined that expressed the EGFP astroglial reporter (N50 cells from 6 slices) did not exhibit clear positive evidence of immunolabelling for GFAP (Fig. 7A), but evidence of coexpression of GFAP and NG2 immunolabelling was observed in some cells (Fig. 7B). In this case, four closely
related cells have the same morphology and differentially express NG2 and the EGFP reporter (Fig. 7B). Three of the cells are NG2immunopositive (Fig. 7Bi) and three express EGFP (Fig. 7Bii); two are NG2+EGFP+ (Fig. 7Biv, filled asterisks), one is NG2+EGFP− (Fig. 7Biv, open asterisk), and a fourth is NG2−EGFP+ (Fig. 7Biv, star). All the EGFP+ cells expressed GFAP within their cell processes (Fig. 7Bv), and this included the NG2+/EGFP+ cells (Fig. 7Bvi, arrowheads). Examination of expression of neural cell markers by NG2-glia We investigated the possible co-expression of NG2 with neural stem cell (NSC) markers nestin (Figs. 8A, B) and PSA-NCAM (Figs. 8C, D), and the neuronal marker NeuN (Figs. 8E–G). Process-bearing NG2glia did not express nestin in acute slices (Fig. 8A) or at 3DIV (Fig. 8B), but vascular pericytes did co-express NG2 and nestin (Fig. 8A, arrows). Nestin+ cells were not numerous in either acute or cultured slices, and where present displayed a rounded cell body with few processes, and were often closely associated to NG2-glia and blood vessels (Fig 8A, insets i and ii). In cerebellar slice cultures, the blood vessels degenerated and appeared as tubular structures that expressed NG2, nestin and αSMA (Fig. 8B, and inset i). NG2-glia and pericytes were mitotically active, as resolved by BrdU and NG2 double labelling in acute and cultured cerebellar slices (Fig. 8A, inset iii; Fig. 8B, insets ii and iii). There was abundant immunolabelling for PSA-NCAM (Fig. 8C) and NeuN (Figs. 8E–G) in 3DIV slices, but there was no evidence of coexpression with NG2; apparent co-expression indicated from zstack images was found to be closely apposed cells on examination of single z-sections (Fig. 8D). Discussion NG2-glia have generally been considered to be OPC (Dawson et al., 2000), but recent studies indicate they may be able to generate all
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
211
Fig. 3. Relations between NG2-glia and Bergmann glia in the cerebellum. Confocal microscopic images of 50 µm thick coronal sections of the cerebellum from P29 GFAP-EGFP mice immunolabelled for NG2 (red). (A) Astroglia can be subdivided into EGFP-positive astrocytes in the granule cell layer (GCL), and Bergmann glia with cell somata in the Purkinje cell layer (PCL) and primary processes running in parallel through the molecular layer (ML) to end at the pial surface (Pia). NG2-glia are distributed throughout the different cerebellar layers, and they were morphologically homogeneous. (B) High magnification of the PCL-GCL interface illustrating NG2-glial cell body closely apposed to those of Bergmann glia, and extending its processes along the PCL to enwrap Bergman glial cell somata, and into the GCL and ML. (C) In the ML, the domains of individual NG2-glia overlap with vertical process domains of multiple Bergmann glia. (D) NG2-glial cells in the ML extending primary processes in two main directions, running vertical and horizontal in relation to the pia and the orientation of Bergmann cell processes. (E) NG2-glial cell processes running vertically extend along the primary processes of Bergmann glia and form intimate contacts with their lateral tufts. Scale bars = 20 μm in A, C and 5 μm in B, D, E.
three major kinds of neural cells, namely neurons, astrocytes and oligodendrocytes (Aguirre et al., 2004; Dimou et al., 2008; Rivers et al., 2008; Zhu et al., 2008). However, the evidence is not entirely consistent, with one study showing NG2-glia generate neurons but not astrocytes (Rivers et al., 2008) and another showing NG2-glia generate astrocytes but not neurons (Zhu et al., 2008). In the present study, we show that NG2-glia express the GFAP-EGFP reporter and in some cases the GFAP protein in cerebellar slice cultures. NG2-glia were not observed to express the GFAP-EGFP reporter in age-matched normal cerebellum. We did not find evidence that NG2-glia in our culture conditions expressed markers of NSC or neurons. In addition, the vast majority of NG2-glia in cerebellar slice cultures incorporated BrdU, indicating they were in cell cycle. NG2-glia are known to proliferate rapidly in response to many different types of brain injury (Levine et al., 2001). Some EGFP+ NG2-glia may be newly generated, but our observations indicate that intrinsic NG2-glia begin to express the EGFP reporter after injury. Our results are consistent with NG2glia undergoing a rapid response to injury in the cerebellar slices, and indicate that NG2-glia may generate astrocytes in this injury paradigm.
A subpopulation of NG2-glia in the pons express the GFAP-EGFP reporter GFAP-EGFP mice are a valuable tool for visualising astroglia both in vivo and ex vivo, since EGFP is distributed throughout the whole astrocyte and distinguishes the morphological complexity of even the finest astroglial processes (Nolte et al., 2001; Wigley et al., 2007; Wigley and Butt, in press). The morphology of different classes of astroglia were recognised in brain sections from P29 and P12 GFAPEGFP mice — protoplasmic and fibrous astrocytes in the gray and white matter, respectively, and velate astrocytes and Bergmann glia in the cerebellum. At both ages, using immunolabelling for NG2, we found no co-expression of the transgene protein EGFP and NG2 in most areas of the brain, supporting the concept that NG2-glia and astrocytes represent distinct populations of glia. In contrast to a previous study (Matthias et al., 2003), we did not find evidence of NG2-glia expressing low levels of EGFP in the hippocampus, although the previous studies used a rat monoclonal antibody against the mouse antigen AN2, and the different specificity of the antibody used in our studies might account for the different results (Karram et al., 2008). In addition, hippocampal NG2-glia have been shown to co-
212
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
213
NG2+/EGFP+ cells was confined in the cell bodies, whereas their cell processes were EGFP−, in stark contrast to neighbouring protoplasmic astrocytes. This suggests transient activity of the GFAP promoter in the NG2-glia, resulting in very low levels of the reporter protein. In addition, these NG2+/EGFP+ cells were often found in doublets and triplets, indicating recent cell division. BrdU-positive NG2-glia were observed throughout the normal brain and with the exception of those in the pons did not express the EGFP reporter, indicating that there was not a direct relationship between EGFP expression and cell division per se. Nonetheless, in light of the evidence in cerebellar slices, it is tempting to speculate that dividing EGFP+/NG2-glia in the pons may generate astrocytes. NG2-glia and astroglia are separate populations in the cerebellum We did not observe NG2+/EGFP+ cells in normal cerebellum, and NG2-glia and astroglia were morphologically and spatially distinct populations. The process domains of Bergmann glia and NG2-glia overlapped, and both formed multiple associations with the dendritic trees of Purkinje neurons. Purkinje neurons form numerous synapses with climbing and parallel fibres within the molecular layer and Bergmann glia and NG2-glia have both been shown to contribute to these synapses (Grosche et al., 1999; Lin et al., 2005). However, while Bergmann glia have a characteristic polarised shape, extending processes along the Purkinje cell dendritic tree, NG2-glia extend processes radially through the different layers and interconnect different neuronal and astroglial domains (Wigley et al., 2007; Wigley and Butt, in press). In fact, NG2-glia processes encompass multiple Bergmann glial cell domains, forming intricate associations with their cell bodies and processes. Neuronal activity at this level may affect NG2-glia and Bergmann glia differently. Bergmann glia express glutamate transporters and have been shown to play an essential role in preventing glutamate spill-over, and ensure the functional oneto-one relationship between climbing fibres and Purkinje neurons (Grosche et al., 1999; Takayasu et al., 2006). In contrast, NG2-glia appear to lack glutamate transporters (Matthias et al., 2003), but do express AMPA receptors and respond to neuronally released glutamate, therefore sensing neuronal activity (Lin et al., 2005). Fig. 5. NG2-glia develop expression of the astroglial reporter EGFP in cerebellar slices maintained in organotypic culture. Confocal micrographs of cerebellar slice cultures from P12 GFAP-EGFP mice examined at 2 DIV. (A) Immunolabelling for NG2 (red) shows that most NG2-glia in the ML expressed the EGFP reporter (green) within the cell bodies (curved arrows); insets are slices double labelled for NG2 (green) and BrdU (red) illustrating the most NG2-glia in the ML were mitotically active. (B) Projection of a z-stack illustrating a dividing NG2-glial cell (red), stained with the nuclear Hoescht dye (blue) and expressing the EGFP reporter (green); insets show the individual channels in a single z-plane, illustrating the process-bearing NG2-immunopositive cell, which expresses the EGFP reporter within the cell body and root of the primary processes, and Hoechst staining demonstrating two directly apposed cell nuclei (asterisks), suggesting recent cell division. Scale bars = 20 μm in all panels.
express the astroglial marker S100β (Matthias et al., 2003; Karram et al., 2008), although S100β is expressed by oligodendrocyte lineage cells in the mouse (Hachem et al., 2005; Rivers et al., 2008). We did find a large number of NG2-glial cells to express the GFAP-EGFP reporter in the pons, confirming a previous study on the respiratory centre in the pons (Grass et al., 2004). EGFP expression in these
NG2-glia express the GFAP-EGFP reporter in cerebellar slice cultures We used the interface method to maintain cerebellar slices in short-term cultures, which has been described extensively in the literature (Stoppini et al., 1991; Gogolla et al., 2006). Numerous brain regions have been used, including hippocampus, cerebellum, thalamus, cortex, spinal cord, striatum and others (Gahwiler et al., 1997), and these show the age of the donor animals is critical. Survival of adult tissues ex vivo is poor and the cytoarchitecture is not maintained (Gahwiler et al., 1997; Stoppini et al., 1991), and so we used P12 mice, because, as we confirm, the cytoarchitecture and neuron-glial relations of the cerebellum are largely established. In this system, the general cytoarchitecture of Purkinje neurons, Bergmann glia and NG2-glia was substantially preserved. Our results revealed that after 2 days in culture, the majority of NG2-glia in the cerebellar slices developed expression of the EGFP reporter in their cell bodies, and in their primary processes proximal to the cell bodies. This was not dependent on the presence of serum in the medium, which in vitro
Fig. 4. Cerebellar slices maintained in organotypic cultures maintain the integrity of their neuron-glial relations. Confocal micrographs of cerebellar slices from P12 mice, examined acutely (left-hand panels) and following organotypic culture for 2 days using the interface method (right-hand panels). (A) Acute slice from P12 GFAP-EGFP mouse illustrating the characteristic cerebellar structure delineated by Bergmann glia. (B) Slice from P12 GFAP-EGFP at 2 DIV, showing that the cerebellar structure is maintained in organotypic culture conditions. (C, D) Comparison of Bergmann glia in acute P12 slices (C) and at 2 DIV (D), shows astroglial integrity is maintained, although the primary process appear thicker and have less colateral ‘tufts’. (E, F) Calbindin labelling of Purkinje neurons in acute slices (E) and after 2 DIV (F) showing that the gross neuronal cytoarchitecture is preserved, with Purkinje neurons lying with their cell bodies in regular rows of cells and extensive dendritic arborisations within the ML. (G, H) Immunolabelling for NG2 (red) in slices from GFAPEGFP mice, showing that the distribution of NG2-glia and their relations with Bergmann glia appear the same in acute slices (G) and slices maintained for 2 DIV (H); note that some NG2-glia in the ML express the EGFP reporter (curved arrow). Scale bars = 5 μm in A, B, and 20 μm in B–H.
214
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
Fig. 6. NG2-glia in cerebellar slice cultures develop expression of the astroglial reporter EGFP in serum-free medium. Confocal micrographs of cerebellar slice cultures maintained in serum-free medium, from P12 GFAP-EGFP mice examined at 2 DIV by immunolabelling for NG2 (red). (A) NG2-glia expressing the EGFP reporter within their cell bodies and primary processes (curved arrows), interspersed amongst NG2+/EGFP- cells (asterisks) and NG2−/EGFP+ astrocytes (small arrows); Aii and Aiii show the individual channels. (B) Projection of a small z-series illustrating an NG2-glial cell strongly expressing the EGFP reporter (curved arrow), and two further NG2+ cells weakly expressing the EGFP reporter (small arrows); Bii and Biii show the individual channels. Scale bars = 10 μm in all panels.
induces the differentiation of bipotential O-2A cells into astrocytes (Raff et al., 1983; Stallcup and Beasley, 1987). We also observed GFAP immunolabelling in NG2+/EGFP+ cells, albeit rarely. Since GFAP-EGFP promoter activity is seen during development several days before the GFAP protein can be detected (Brenner and Messing, 1996), the lack of strong GFAP expression in NG2+/EGFP+ cells is not surprising. In addition, NG2 is rapidly down-regulated as NG2-glia differentiate into astroglia and express GFAP (Zhu et al., 2008), so GFAP immunopositive NG2-glia are likely to represent cells in transition between NG2glia and astrocytes. Our data does not exclude the possibility of misexpression of GFAP transcriptional activity in NG2-glia, but our finding that more or less all NG2-glia in the cerebellar ML begin to express the GFAP reporter gene, and that some of these express the GFAP protein within 2 days indicates a lineage relationship between NG2-glia and astroglial in situ.
In cerebellar slice cultures the majority of NG2-glia incorporated BrdU, which is a characteristic response of NG2-glia to almost any insult in the adult CNS (Levine et al., 2001; Dawson et al., 2003). In addition, we show that dividing cells express the GFAP-EGFP reporter, suggesting that NG2-glia may contribute to reactive astrogliosis. Selective ablation of proliferating astrocytes has been shown not to significantly affect astrogliosis in an animal model of amyotrophic lateral sclerosis (Lepore et al., 2008), indicating another source of reactive astroglia (Alonso, 2005). Longer term ex vivo experiments are required to prove lineage progression, but our results support the possibility that NG2-glia are activated after CNS injury and generate new astrocytes, possibly contributing to glial scar formation and isolation of the lesion. In vivo experiments are required to test this, but an important implication is that the inhibition of differentiation of
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
215
Fig. 7. NG2-glia rarely express GFAP protein in cerebellar slices maintained in organotypic culture. Confocal micrographs of cerebellar slice cultures from P12 GFAP-EGFP mice examined at 2 DIV by immunolabelling for GFAP (red) and NG2 (blue). (A) Single z-plane confocal micrograph shows numerous NG2-glial cells (blue) coexpressing the astroglial EGFP reporter protein (curved arrows); there was little clear evidence of co-expression of NG2 and GFAP, whereas co-expression of EGP and GFAP was evident and appears yellow. (B) Projection of a small z-series of four morphologically equivalent neighbouring cells, which differentially express NG2 (blue), the astroglial reporter gene EGFP (green) and the astroglial marker GFAP (red), together with individual channels (i–iii) and partially merged images (iv–vi). Three cells express NG2 (i) and three cells express EGFP (ii), and the partially merged image (iv) shows that two of the cells were NG2+/EGFP+ (filled asterisks), one was NG2+/EGFP− (open asterisk), and the fourth was NG2−/EGFP+ (star). The partially merged images for EGFP and GFAP (v) and NG2 and GFAP (vi) show expression of GFAP in some processes of NG2+/EGFP+ cells (arrowheads).
NG2-glia into reactive astroglia could be beneficial for axonal regeneration. NG2-glia did not express markers for neurons in cerebellar slice cultures We found no evidence that NG2-glia expressed nestin, a marker for neural stem cells, nor PSA-NCAM, a marker for migratory neuroblasts, nor NeuN, a marker for differentiated neurons. We did find that
perivascular pericytes, identified by expression of αSMA, expressed both NG2 and nestin. It has been recently suggested that adult brain capillaries contain a population of NG2+/nestin+ pericytes that can differentiate into neurons in vitro (Dore-Duffy, 2008; Dore-Duffy et al., 2006). Our results support the possibility that pericytes have a neural stem cell phenotype in situ, although we did not observe complex process-bearing nestin+ or PSA-NCAM+ NG2+ cells. This is consistent with gene fate-map studies on NG2-glia progeny which found no
216
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
Fig. 8. Lack of expression of neural cell markers in NG2-glia in cerebellar slice cultures. Confocal micrographs of cerebellar slice cultures double immunofluorescence labelled for NG2, with nestin (A, B), PSA-NCAM (C, D), or NeuN (E–G). (A) NG2-glia did not express nestin in acute slices, but coexpression of NG2 (red) and nestin (green) was observed in some pericytes directly apposed to blood vessels (arrows); insets in A illustrate (i) simple nestin+ cells (green) surrounding an NG2-glial cell (red), (ii) a process-bearing nestin+ cell (green) closely apposed to NG2+ pericytes on a blood vessel (red), and (iii) triplet of dividing BrdU+(red)/NG2+(green) cells (asterisks). (B) At 2 DIV, the number of nestin+ cells (green) was decreased, and co-expression of nestin (green) and NG2 (red) was only observed in vascular pericytes; insets in B illustrate double αSMA (red) and NG2 (green) double immunolabelling, to identify pericytes (i), and BrdU (red) and NG2 (green) double immunolabelling of a triplet of dividing NG2-glia (ii, asterisks), and BrdU+/NG2+ pericyte (iii). (C) There was abundant expression of PSA-NCAM (green) in cerebellar slices, and double labelling with NG2 (red) suggested some colocalization of the two markers (square), but analysis of single z-sections and orthogonal projections in x- and y-planes of the region indicates close apposition of NG2+ (red) and PSA-NCAM+ (green) cells (D). (E–G) Projections of small z-series illustrate no evidence of co-expression of NG2 (red) and NeuN (green). Scale bars = 20 μm in A, B, C, E and 10 μm in D, E, F and all insets.
evidence they generate neurons in the developing forebrain in vivo (Zhu et al., 2008). In contrast, neural precursors in the developing SVZ that expressed NG2 and a Cnp-GFP transgene have been reported to generate GABAergic hippocampal neurons (Aguirre et al., 2004). Furthermore, a recent study using gene fate-mapping of adult YFP+ PDGFαR/NG2 glia indicates they generate small numbers of neurons throughout the ventral forebrain, but mainly in the primary olfactory cortex (Rivers et al., 2008). However, the same study noted that neurogenesis occurred over an extended period and was unable to
identify YFP+ cells expressing PSA-NCAM, nor BrdU+/NeuN+/YFP+ neurons, which suggested to the authors that neurons may be formed by trans-differentiation of post-mitotic PDGFαR/NG2 glia (Rivers et al., 2008). Hence, our finding that NG2-glia do not express nestin, PSA-NCAM or NeuN in cerebellar slices, does not exclude the possibility that NG2-glia are neurogenic, and future studies will require genetic fate-mapping of NG2-glia over an extended period to examine whether they are capable of neurogenesis in the cerebellum. The possibility that adult neurogenesis occurs outside the accepted
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
neurogenic regions has now been reported in a number of different brain regions (Gould et al., 1999; Zhao et al., 2003; Dayer et al., 2005; Kokoeva et al., 2005; Shapiro et al., 2007; Rivers et al., 2008). Conclusions The implication of our findings is that NG2-glia, which are considered to be OPCs and have been shown to generate oligodendrocytes in the developing and adult brain (Dimou et al., 2008; Rivers et al., 2008; Zhu et al., 2008), can also generate astrocytes in situ. The proliferative response of NG2-glia and their expression of the GFAPEGFP reporter in cerebellar slice cultures may reflect a rapid activation of NG2-glia in response to injury. This implies that NG2-glia may be an important source of reactive astrogliosis. However, in light of the findings in the pons, the generation of astrocytes by NG2-glia may be a more general phenomenon in the adult CNS. Experimental methods Animals and tissue Experiments were performed on brains from non-transgenic mice (CFLP or C57BL strains), and from two transgenic lines: (1) GFAP-EFGP mice in which the expression of enhanced green fluorescent protein (EGFP) was under the control of the human glial fibrillary acidic protein (GFAP) promoter [line TgN(GFAPEGFP) GFEC-FKi], fromFrank Kirchhoff (University of Goettingen), via Alex Verkhratsky (University of Manchester), and characterised by Nolte et al. (2001). Postnatal mice aged postnatal day (P)12 were used for cerebellar slices, and mature P29–30 mice were also used for immunohistochemistry. Mice were killed humanely in accordance with the Home Office Animals (Scientific) Act 1986 (UK), and brains were prepared fresh for cerebellar slice cultures or immersion fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) for immunohistochemistry. Cerebellar slice cultures Brains from P12 mice were rapidly removed and placed in chilled Gey's balanced salt solution (GBSS, Invitrogen), supplemented with 25 mM D-glucose. Cerebellar slices were cut at 300 μm using a McIlwain tissue chopper and slices were placed on inserts (Millipore, Bedford, MA, USA) and cultured using the interface method, as first described by Stoppini et al. (1991) with little modification. The inserts were transferred into 6-well culture plates with 1 ml culture medium per well. The culture medium was composed of 50% Opti-MEM™ (Invitrogen), 25% horse serum (Invitrogen), 25% Hank's Balanced Salt Solution (HBSS, Sigma), supplemented with 25 mM D-glucose (Sigma), and with a solution of Penicillin–Streptomycin (Penicillin G sodium 10 000 U/ml, Streptomycin sulphate 1000 μg/ml, Invitrogen) diluted 1:500. Alternatively, serum-free modified media was used, composed of Neurobasal™ (Invitrogen), B27 supplement (Invitrogen) diluted 1:50, 25 mM D-Glucose and antibiotics as above. Slices were cultured at 37 °C in an incubator with 5% CO2 and 95% air for 2–3 days in vitro (DIV), and the culture medium was replaced every two or three days. Slices were regularly examined and discarded if contamination occurred. At the end of the experiments, slices were immersion fixed in 4% PFA for 30 min at RT for immunohistochemistry. In some cases, prior to fixation, cell proliferation was examined by incubating cerebellar slice cultures for 2h with 0.1 μM of 5-bromo2′-deoxyuridine (BrdU, Sigma), diluted directly in culture media. Immunohistochemistry In the case of brains, following fixation overnight at 4 °C, tissue was washed in PBS and sections were cut at 50 μm using a VT1000 S vibrating blade microtome (Leica Microsystems) and collected in PBS.
217
For cultures, cerebellar slices were carefully removed from the inserts, immersion fixed in 4% PFA for 30 min at RT, and then washed in PBS. Cerebellar slice cultures and brain sections were then treated in the same way. Tissues were washed in PBS, and a blocking stage was performed by incubation at 1 h at RT in blocking buffer composed of 10% normal goat serum (NGS) and 1% tween in phosphate buffered saline (PBST), prior to incubation overnight at 4 °C with primary antibodies diluted in blocking buffer: rabbit anti-NG2, either from Dr Stallcup (1:500) or Chemicon (1:200); chicken anti-GFAP (1:200; Chemicon); rabbit anti-calbindin D-28K (1:300; Chemicon); mouse anti-BrdU (1:300; Chemicon); mouse anti-nestin (1:200; BD Biosciences); mouse anti-PSA-NCAM (1:400; AbCys); mouse anti-NeuN (1:200; Chemicon). After three washes in PBST, tissues were incubated for 1h at RT with the appropriate secondary antibodies conjugated with 488Alexafluor, 567Alexafluor or 405Alexafluor (1:500, Molecular Probes). For double immunofluorescence labelling, primary antibodies of different origin were diluted together in blocking buffer and co-dilutions of the appropriate secondary antibodies were used. In some cases, nuclei were labelled with Hoechst 33342 (1:1000, Molecular Probe). After final washes in PBST, tissue was mounted on poly-lysine coated glass slices with Vectashield mounting media (Vector Laboratories), and images were acquired using a LSM 5 Pascal Axioskop2 confocal microscope (Zeiss). Acknowledgments This work was supported by the BBSRC. Giampaolo Leoni was in receipt of an Anatomical Society PhD Studentship. We thank Professor William Stallcup for the NG2 antibody, and Frank Kirchoff for the GFAP-GFP mice. References Aguirre, A.A., Chittajallu, R., Belachew, S., Gallo, V., 2004. NG2-expressing cells in the subventricular zone are type C-like cells and contribute to interneuron generation in the postnatal hippocampus. J. Cell Biol. 165, 575–589. Alonso, G., 2005. NG2 proteoglycan-expressing cells of the adult rat brain: possible involvement in the formation of glial scar astrocytes following stab wound. Glia 49, 318–338. Brenner, M., Messing, A., 1996. GFAP transgenic mice. Methods 10, 351–364. Butt, A.M., Duncan, A., Hornby, M.F., Kirvell, S.L., Hunter, A., Levine, J.M., Berry, M., 1999. Cells expressing the NG2 antigen contact nodes of Ranvier in adult CNS white matter. Glia 26, 84–91. Butt, A.M., Hamilton, N., Hubbard, P., Pugh, M., Ibrahim, M., 2005. Synantocytes: the fifth element. J. Anat. 207, 695–706. Dawson, M.R., Levine, J.M., Reynolds, R., 2000. NG2-expressing cells in the central nervous system: are they oligodendroglial progenitors? J. Neurosci. Res. 61, 471–479. Dawson, M.R., Polito, A., Levine, J.M., Reynolds, R., 2003. NG2-expressing glial progenitor cells: an abundant and widespread population of cycling cells in the adult rat CNS. Mol. Cell. Neurosci. 24, 476–488. Dayer, A.G., Cleaver, K.M., Abouantoun, T., Cameron, H.A., 2005. New GABAergic interneurons in the adult neocortex and striatum are generated from different precursors. J. Cell Biol. 168, 415–427. Dimou, L., Simon, C., Kirchhoff, F., Takebayashi, H., Gotz, M., 2008. Progeny of Olig2expressing progenitors in the gray and white matter of the adult mouse cerebral cortex. J. Neurosci. 28, 10434–10442. Dore-Duffy, P., 2008. Pericytes: pluripotent cells of the blood brain barrier. Curr. Pharm. Des. 14, 1581–1593. Dore-Duffy, P., Katychev, A., Wang, X., Van Buren, E., 2006. CNS microvascular pericytes exhibit multipotential stem cell activity. J. Cereb. Blood Flow Metab. 26, 613–624. Gahwiler, B.H., Capogna, M., Debanne, D., McKinney, R.A., Thompson, S.M., 1997. Organotypic slice cultures: a technique has come of age. Trends Neurosci. 20, 471–477. Gogolla, N., Galimberti, I., DePaola, V., Caroni, P., 2006. Preparation of organotypic hippocampal slice cultures for long-term live imaging. Nat. Protoc. 1, 1165–1171. Gould, E., Reeves, A.J., Graziano, M.S., Gross, C.G., 1999. Neurogenesis in the neocortex of adult primates. Science 286, 548–552. Grass, D., Pawlowski, P.G., Hirrlinger, J., Papadopoulos, N., Richter, D.W., Kirchhoff, F., Hulsmann, S., 2004. Diversity of functional astroglial properties in the respiratory network. J. Neurosci. 24, 1358–1365. Grosche, J., Matyash, V., Moller, T., Verkhratsky, A., Reichenbach, A., Kettenmann, H., 1999. Microdomains for neuron-glia interaction: parallel fiber signaling to Bergmann glial cells. Nat. Neurosci. 2, 139–143. Hachem, S., et al., 2005. Spatial and temporal expression of S100B in cells of oligodendrocyte lineage. Glia 51, 81–97.
218
G. Leoni et al. / Molecular and Cellular Neuroscience 42 (2009) 208–218
Karram, K., Goebbels, S., Schwab, M., Jennissen, K., Seifert, G., Steinhauser, C., Nave, K.A., Trotter, J., 2008. NG2-expressing cells in the nervous system revealed by the NG2EYFP-knockin mouse. Genesis 46, 743–757. Kokoeva, M.V., Yin, H., Flier, J.S., 2005. Neurogenesis in the hypothalamus of adult mice: potential role in energy balance. Science 310, 679–683. Lepore, A.C., Dejea, C., Carmen, J., Rauck, B., Kerr, D.A., Sofroniew, M.V., Maragakis, N.J., 2008. Selective ablation of proliferating astrocytes does not affect disease outcome in either acute or chronic models of motor neuron degeneration. Exp. Neurol. 211, 423–432. Levine, J.M., Beasley, L., Stallcup, W.B., 1986. Localization of a neurectodermassociated cell surface antigen in the developing and adult rat. Brain Res. 392, 211–222. Levine, J.M., Reynolds, R., Fawcett, J.W., 2001. The oligodendrocyte precursor cell in health and disease. Trends Neurosci. 24, 39–47. Lin, S.C., Huck, J.H., Roberts, J.D., Macklin, W.B., Somogyi, P., Bergles, D.E., 2005. Climbing fiber innervation of NG2-expressing glia in the mammalian cerebellum. Neuron 46, 773–785. Matthias, K., Kirchhoff, F., Seifert, G., Huttmann, K., Matyash, M., Kettenmann, H., Steinhauser, C., 2003. Segregated expression of AMPA-type glutamate receptors and glutamate transporters defines distinct astrocyte populations in the mouse hippocampus. J. Neurosci. 23, 1750–1758. Nishiyama, A., Lin, X.H., Giese, N., Heldin, C.H., Stallcup, W.B., 1996a. Co-localization of NG2 proteoglycan and PDGF alpha-receptor on O2A progenitor cells in the developing rat brain. J. Neurosci. Res. 43, 299–314. Nishiyama, A., Lin, X.H., Giese, N., Heldin, C.H., Stallcup, W.B., 1996b. Interaction between NG2 proteoglycan and PDGF alpha-receptor on O2A progenitor cells is required for optimal response to PDGF. J. Neurosci. Res. 43, 315–330.
Nolte, C., MAtyash, M., Pivneva, T., Schipke, C.G., Ohlemeyer, C., Hanisch, U.K., Kirchhoff, F., Kettenmann, H., 2001. GFAP promoter-controlled EGFP-expressing transgenic mice: a tool to visualize astrocytes and astrogliosis in living brain tissue. Glia 33, 72–86. Raff, M.C., Miller, R.H., Noble, M., 1983. A glial progenitor cell that develops in vitro into an astrocyte or an oligodendrocyte depending on culture medium. Nature 303, 390–396. Rivers, L.E., Young, K.M., Rizzi, M., Jamen, F., Psachoulia, K., Wade, A., Kessaris, N., Richardson, W.D., 2008. PDGFRA/NG2 glia generate myelinating oligodendrocytes and piriform projection neurons in adult mice. Nat. Neurosci. 11, 1392–1401. Shapiro, L.A., et al., 2007. Origin, migration and fate of newly generated neurons in the adult rodent piriform cortex. Brain Struct. Funct. 212, 133–148. Stallcup, W.B., Beasley, L., 1987. Bipotential glial precursor cells of the optic nerve express the NG2 proteoglycan. J. Neurosci. 7, 2737–2744. Stoppini, L., Buchs, P.A., Muller, D., 1991. A simple method for organotypic cultures of nervous tissue. J. Neurosci. Methods 37, 173–182. Takayasu, Y., Iino, M., Shimamoto, K., Tanaka, K., Ozawa, S., 2006. Glial glutamate transporters maintain one-to-one relationship at the climbing fiber-Purkinje cell synapse by preventing glutamate spillover. J. Neurosci. 26, 6563–6572. Wigley, R., and Butt, A.M., in press. Integration of NG2-glia (synantocytes) into the neuroglial network. Neuron Glial Biol. Wigley, R., Hamilton, N., Nishiyama, A., Kirchhoff, F., Butt, A.M., 2007. Morphological and physiological interactions of NG2-glia with astrocytes and neurons. J. Anat. 210, 661–670. Zhao, M., et al., 2003. Evidence for neurogenesis in the adult mammalian substantia nigra. Proc. Natl. Acad. Sci. U. S. A. 100, 7925–7930. Zhu, X., Bergles, D.E., Nishiyama, A., 2008. NG2 cells generate both oligodendrocytes and gray matter astrocytes. Development 135, 145–157.
