Bergmann glial cells play critical roles in the structure and function of the cerebellum. During development, their radial processes serve as guides for migrating granule neurons and their terminal endfeet tile to form the glia limitans. As the cerebellum matures, Bergmann glia perform important roles in synaptic transmission and synapse maintenance, while continuing to serve as essential structural elements. Despite growing evidence of the diverse functions of Bergmann glia, the molecular mechanisms that mediate these functions have remained largely unknown. As a step toward identifying the molecular repertoire underlying Bergmann glial function, here we examine global gene expression in individual Bergmann glia from developing (P6) and mature (P30) mouse cerebellum. When we select for developmentally regulated genes, we find that transcription factors and ribosomal genes are particularly enriched at P6 relative to P30; whereas synapse associated molecules are enriched at P30 relative to P6. We also analyze genes expressed at high levels at both ages. In all these categories, we find genes that were not previously known to be expressed in glial cells, and discuss novel functions some of these genes may potentially play in Bergmann glia. We also show that Bergmann glia, even in the adult, express a large set of genes thought to be specific to stem cells, suggesting that Bergmann glia may retain neural precursor potential as has been proposed. Finally, we highlight several genes that in the cerebellum are expressed in Bergmann glia but not astrocytes, and may therefore serve as new, specific markers for Bergmann glia.
Citation: Koirala S, Corfas G (2010) Identification of Novel Glial Genes by Single-Cell Transcriptional Profiling of Bergmann Glial Cells from Mouse Cerebellum. PLoS ONE 5(2): e9198. doi:10.1371/journal.pone.0009198
Editor: Thomas A. Reh, University of Washington, United States of America
Received: October 6, 2009; Accepted: January 22, 2010; Published: February 12, 2010
Copyright: © 2010 Koirala, Corfas. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by National Institute of Neurological Disorders and Stroke (NINDS) grant R01 NS35884 (to G.C.), National Institutes of Health (NIH) Mental Retardation Research Center grant P30-HD 18655 (to G.C.), and a Fundamental Neurobiology National Institutes of Health (NIH) training grant T32 NS007484-06 (to S.K.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The Bergmann glial cell is a type of astroglia that performs a range of important functions in the cerebellum throughout the life of the animal. During development, the radial processes of Bergmann glia provide structural support to the expanding cerebellar plate, and endfeet of these cells adhere together to form a continuous glia limitans overlying the cerebellum , , . The radial fibers of Bergmann glia also act as essential guide rails for the migration of cerebellar granule cells , . Mice with Bergmann glial defects during development have severe abnormalities including pial rupture, disrupted neuronal migration and layering, and altered connectivity , . In addition, it has been proposed that Bergmann glia also contribute to the elaboration of Purkinje cell dendrites , ,  and the stabilization of synaptic connections onto these neurons .
After completion of cerebellar morphogenesis, Bergmann glia remain important for structural support , , but also serve additional roles in synapse maintenance, function and plasticity , , , . Bergmann glial processes ensheathe synapses on Purkinje cells, and play a supportive role in normal transmission by maintaining a physiological synaptic microenvironment. This involves buffering of ions, uptake of neurotransmitters, and production of glutamine, which neurons then convert to glutamate , , . Recent studies suggest that Bergmann glia may also play more active roles at synapses. These cells respond to synaptic activity and locomotor behavior with Ca++ elevations in vivo , and may in turn modulate synaptic transmission, synaptic plasticity, and blood perfusion changes around synapses , , .
Despite growing recognition of the diverse roles of Bergmann glia, the molecular mechanisms that mediate these roles remain largely unexplored. As a step toward addressing this gap in knowledge, we analyzed global gene expression in single acutely isolated Bergmann glia from the developing cerebellum (P6), during the peak of granule cell migration along glial fibers, and from the adult (P30), when developmental processes are complete and Bergmann glial processes have ensheathed synapses. We then compared gene expression profiles between Bergmann glia at these two ages, and also with data obtained previously from astrocytes  or stem cells . In situ hybridization and online gene expression atlases were used to validate Bergmann glial expression of genes of interest. Our analysis shows that although Bergmann glia robustly express many astroglial genes, as expected, there are also genes that appear to be Bergmann glia-specific, and may be useful as new markers and tools to manipulate this cell type. In addition, we identify a number of genes as potential candidates to mediate Bergmann glial roles in maintenance of cerebellar morphology, and in synaptic structure and function. Surprisingly, we also find that Bergmann glia express a large set of genes thought to be expressed specifically in stem cells, suggesting that this glial type may harbor progenitor potential. Together, this information should be useful to future studies of Bergmann glia and glial cells in general.
Single Cell Isolation and cDNA Synthesis from Bergmann Glia
To examine gene expression in Bergmann glia, we chose a single-cell cDNA library approach  using acutely isolated cells from the cerebella of mice expressing GFP under the Glial fibrillary acidic protein (Gfap) promoter . As shown Fig. 1Ai, Bergmann glial cell bodies could be easily visualized by their robust GFP expression in a live slice from a GFAP-GFP mouse cerebellum. After gentle dissociation of the tissue, four main types of cells were seen. Most cells were small with round cell bodies and lacked GFP signal (arrows in Fig. 1Aii), and most likely represent granule cells. Another frequent cell type had round cell bodies and relatively weak GFP fluorescence (arrowheads in Fig. 1Aii), almost certainly representing astrocytes. A third cell type, with large cell bodies and no fluorescence, we identified as Purkinje cells (not shown). Finally, a rather infrequent type (<1% of cells) could be readily distinguished from the others by a distinctive “bushy” unipolar morphology and strong GFP expression (arrowheads in Fig. 1Aiii). These putative Bergmann glia and some GFP- cells as controls were harvested individually using glass microelectrodes (Fig. 1Aiv). To minimize the possibility of contamination from other mRNAs, each cell was subjected to a rinse in a new dish with fresh buffer and picked with a new microelectrode before cDNAs were generated using protocols described before , , . In line with these protocols, single cell RT-PCR amplification generated cDNAs of 300–1000 base pairs (Fig. 1B, left panel). The single cell cDNAs were then subjected to rigorous quality control using PCR (Fig. 1B, right panels). Single cell cDNA libraries were considered to be of good quality if they were positive for high (β-actin or Actb) and low abundance (ornithine decarboxylase or Odc) markers, confirming that mRNAs of widely varying abundance were preserved during the RT-PCR amplification. High quality cDNAs of GFP+ and GFP− cells were then further characterized. GFP+ cells were positive for Gfap but negative for neurofilament light chain (Nefl), confirming that they were glia and that the samples were free of contaminating neuronal mRNA. GFP− cells were negative for Gfap and positive for Nefl, indicating they were most likely neurons. Finally, an additional quality control step was performed using Southern blot analysis (Fig. 1C). GFP+ and GFP− cells were positive for high, medium, and two low abundance transcripts (Actb, high; γ-actin or Actg, medium; Odc and protein phosphatase 1cα or Ppp1ca, low), indicating good amplification. GFP+ cells were positive for Gfap, fatty acid binding protein 7 or Fabp7 (also called brain lipid binding protein or BLBP), Sept4 (Septin 4) and the glutamate transporter, Slc1a3 (GLAST), confirming that they were astroglial cells. They were negative for the neuronal markers Nefl and microtubule-associated protein 2 (Mtap2). Conversely, GFP− cells were negative for glial markers but positive for neuronal ones, confirming their identity as neurons.
Ai, a live cerebellar slice obtained from an adult GFAP-GFP transgenic mouse imaged under phase contrast optics (left) and fluorescence illumination (right). The Bergmann glia (see arrowheads) are the cells with the most fluorescence. Aii, freshly dissociated cells include putative astrocytes, which are devoid of processes and show relatively weak GFP fluorescence (arrowheads), and putative granule neurons, which have small, round, GFP-negative cell bodies (arrows). Aiii, a freshly dissociated Bergmann glia (arrowhead in left panel) can be distinguished from other cells by the bushy processes that emanate from one side of the soma–these are the long Bergmann glial processes that have partially retracted or been sheared off during tissue dissociation. In addition, Bergmann glia display strong GFP fluorescence (arrowhead in right panel), with the mean GFP intensity of their cell bodies 2.9±0.8 fold that of astrocytes; n = 17 cells). Aiv, a single Bergmann glia being washed by placement in a new dish containing fresh buffer, before being picked again with a new microelectrode. This step is performed to exclude contaminating cells or mRNAs. Scale bar, 70 µm in Ai, 25 µm in Aii and Aiii, 40 µm in Aiv. B, left panel, agarose gel electrophoresis of cDNAs generated from single GFP+ and GFP− cells. The gels show that most of the cDNA lies between 300 and 1000 bases. Right panel, agarose gels showing PCR with primer pairs directed towards β-actin (Actb), ornithine decarboxylase (Odc), Gfap, and neurofilament light chain (Nefl). The results show that the single cell cDNAs from Bergmann glia (GFP+) and neurons (GFP−) contain both high and low abundance transcripts (Actb and Odc, respectively). Bergmann glia are positive for the astroglial marker Gfap and negative for the neuronal marker Nefl, whereas neurons are negative for Gfap and positive for Nefl. C. Southern blot analysis of cDNAs from two putative Bergmann glial cells (GFP+) and a putative neuron (GFP−) shows presence of the high, medium, and two low abundance markers (Actb, Actg, and Odc and Ppp1ca, respectively). In addition, the GFP+ cells are positive for Gfap, Fabp7 (BLBP), Sept4 and Slc1a3 (GLAST), confirming their glial identity, whereas the GFP− cell lacks all these markers. Conversely, the GFP+ cells are absent for the neuronal markers Nefl and Mtap2 whereas the GFP− cell is positive. These results confirm the preservation of low to high abundance transcripts after the single-cell RT-PCR amplification, and also confirm the cell identity of the Bergmann glia used for microarray analysis.
Purity and Accuracy of Expression Profiles of Individual Bergmann Glia
To examine the global transcriptional profiles of Bergmann glia, the amplified cDNA generated from five P6 and five P30 cells were individually hybridized to Affymetrix 430 2.0 Mouse Expression Arrays. On average, 31.7±1.5% of the 45101 probe sets per array showed positive expression at P6, and 27.5±2.3% at P30. Expression profiles of all the Bergmann glia demonstrated high levels of expression of known astrocyte-specific genes , consistent with their long-held classification as specialized astroglia (Fig. 2). In contrast, the expression of several genes considered to be markers for neurons, oligodendrocytes or microglia  was low or absent (Fig. 2), confirming that the samples were indeed free from contaminating mRNAs from other cell types.
Mean expression levels of well established markers for astroglia, neurons, oligodendrocytes and microglia were analyzed in the transcriptional profiles of ten Bergmann glia (five each from P6 and P30) using GeneSpring GX 7.3 software. For genes represented by multiple probe sets, the averaged expression of all probe sets were used. Astroglial genes were robustly expressed whereas markers of other cell types were absent or extremely low, confirming the astroglial identity of Bergmann glia and the absence of contaminating mRNAs from other cell types during cell harvesting. Error bars represent ± SEM.
To determine the level of heterogeneity between samples of the same and different ages, we performed two analyses. Pair-wise comparisons of all individual samples showed that cells from the same age are more similar than between ages (same age: mean correlation coefficient = 0.8 for P6, 0.82 for P30; between ages: mean correlation coefficient = 0.66; representative samples shown in Fig. 3A). The similarity between cells of the same age is comparable to that reported between individual cells of a glioblastoma cell line (mean correlation coefficient = 0.86; , using very similar techniques, suggesting that Bergmann glia from a particular age and sagittal location (vermis) are quite homogeneous. The similarity between the two ages (mean correlation coefficient = 0.66) is significantly higher than for disparate cell types (for example, olfactory epithelium neurons vs. heart cells, mean correlation coefficient = 0.42; , as expected for cells of the same type. Unsupervised hierarchical clustering of the samples indicated that cells were more distinct between ages than within each age (Fig. 3B). Taken as a whole, these results affirm the cell-type specificity and reproducibility of single cell expression profiling, and the validity of comparisons between ages.
A, Scatter plots of raw gene expression level compared between two cells from the same age (P6-1 vs. P6-2; left panel) and between ages (P6-1 vs. P30-1; right panel) (a.u.: arbitrary units). The samples from the same age show high similarity (mean within-age correlation coefficient = 0.81), which is comparable to what is reported between individual cells of a glioblastoma cell line (0.86) . The high level of concordance between samples of the same age increases the reliability of comparisons between ages. Samples of different ages are significantly more divergent (mean P6 vs. P30 correlation coefficient = 0.66). B, Dendrogram and sample clustering of individual Bergmann glia. Unsupervised hierarchical clustering based on overall gene expression profiles reveals two distinct clusters corresponding to the two ages, P6 and P30. This suggests that the samples from the two ages do indeed represent two statistically distinct populations suitable for valid comparison.
Identification of Genes that Are Developmentally Regulated in Bergmann Glia
Genes expressed differentially between P6 and P30 were identified using two criteria: those with a greater than 3-fold difference in normalized expression between ages, and those that were flagged “present” at one age and “absent” at the other. In total, 435 genes were found to fit the criteria for P6>P30 expression (Table S1), and 137 fit the criteria for P30>P6 expression (Table S2). Of these, the top twenty most differentially expressed probe sets determined using the two approaches are shown in Tables 1 and 2 for P6>P30 and in Tables 3 and 4 for P30>P6. Genes in all tables were categorized and annotated based on functional information obtained in online databases and/or previous studies.
We found that the set of genes expressed more highly at P6 is enriched in molecules known or predicted to be involved in cell growth and/or proliferation, cell cycle regulation, protein biosynthesis and other metabolic pathways, RNA processing and transport, and transcriptional regulation (including a large number of transcription factors) (Table S1). This suggests that at P6 Bergmann glia are in a state of active metabolism and growth, and some of them may potentially still be undergoing proliferation. The higher expression levels of ribosomal genes is also indicative of cells in a state of growth , , consistent with the postnatal extension and elaboration of Bergmann glial processes.
On the other hand, the genes enriched at P30 include a different set of functional categories, namely molecules known or predicted to be involved in maintaining synapse structure and function, regulating exocytosis, forming gap junctions, and mediating molecular transport (including ion, protein, and carbohydrate transporters) (Table S2). This result suggests that a key role of Bergmann glia in the adult cerebellum is to support/modulate synaptic function, and identifies some potential molecular players that may mediate this role. Interestingly, many of the differentially expressed genes in both sets (P6>P30 and P30>P6) have not previously been reported to be present in astroglial cells (Table 5), based on a search of the literature and exclusion of astrocyte-enriched genes listed by Cahoy and colleagues . Further study of these genes, some of which are discussed below, may provide new insight into molecular mechanisms underlying Bergmann glial functions.
In addition to individual genes, we also sought to identify signaling and metabolic pathways that are preferentially active in Bergmann glia at P6 or at P30. To do this, we utilized the Ingenuity Pathway Analysis (IPA) tool from Ingenuity Systems, a resource based on a curated list of all canonical signaling and metabolic pathways . IPA analysis of the differentially expressed gene sets in our data identified nine signaling or metabolic pathways statistically enriched at P6 and nine enriched at P30 (Table 6). The pathways enriched at P6 include the Notch, TGF-β and Wnt/β-catenin signaling pathways, which have key developmental roles in cell fate determination, cell growth, proliferation and maturation. The Notch pathway has been shown to be important for Bergmann glial specification and maturation , , , but the potential roles of the other pathways we have identified remain untested. At P30, the statistically enriched signaling pathways included the glutamate receptor signaling pathway, underscoring a role for Bergmann glia in synapse function/modulation. Unexpectedly, another enriched pathway is the embryonic stem cell pluripotency pathway, a finding that we discuss in greater detail below.
We were also surprised to find that at P6 Bergmann glia express a number of genes that have been traditionally thought of as neuron-specific (see Table S3). While we cannot completely exclude the possibility of contamination by some neuronal mRNAs, our quality control analysis (Fig. 2) argues against this, as does the finding that some of these are also enriched in cortical astrocytes . We therefore believe that these genes are expressed by Bergmann glia in addition to neurons, but that the level of expression in glia may be significantly lower than in neurons, causing them not to have been detected in the glia by other techniques such as in situ hybridization. We believe one of the strengths of our analysis is that we have not subtracted out genes thought to be specific to other cell types from our microarray data, and therefore we can detect genes that may be present in multiple cell types, even if they are more abundant in other cells than in Bergmann glia.
Analysis of Genes Abundantly Expressed in Postnatal Bergmann Glia (P6 and P30) and Identification of Novel Cell-Specific Markers
To gain further insight into the molecular makeup of Bergmann glia in the postnatal cerebellum, we also searched for genes that are expressed at both P6 and P30 at moderate to high levels (raw signal values >2000) (Table S4). As expected, this extensive list includes many astroglial markers such as vimentin, S100β, aquaporin 4 and aldehyde dehydrogenase 1 family, member L1 (Aldh1L1) . It also includes genes not previously known to be expressed in astroglia, some of which are listed in Table 7.
While examining the set of abundantly expressed genes, we found, surprisingly, that Bergmann glia express mRNA for the myelin protein peripheral myelin protein (Pmp22) and for myelin protein zero-like 1 (Mpzl1), which may be important for myelination  (Table S4). Furthermore, in the adult, Bergmann glia also express mRNAs for proteolipid protein 1 (Plp1), the predominant component of CNS myelin, and leucine-rich repeat LGI family, member 4 (Lgi4), which is important for myelination in the PNS  (Table S2). These findings were supported by in situ hybridization data from the Allen Brain Atlas (Fig. S1): although the strongest expression of Pmp22 and Plp1 in the adult is in putative oligodendrocytes (arrows in Fig. S1, top and second row), there is less intense but still significant staining in the Purkinje cell layer (PCL) consistent with expression in Bergmann glia (arrowheads). Lgi4 (Fig. S1, third row) and Mpzl1 (arrowheads in Fig. S1, bottom row) also show labeling consistent with expression in Bergmann glia. Since Bergmann glia play no known role in myelination, it is unknown what alternate function these genes may serve in these cells. Our data also indicated that Bergmann glia express agrin in vivo at P6 and P30 (Table S4), similar to what been shown in astrocytes in vitro . Whether agrin synthesized by Bergmann glia plays any role in synapse formation or maintenance in vivo merits further investigation.
Analysis of in situ hybridization data for genes we found expressed in adult Bergmann glia identified several candidates that we believe may serve as selective markers for Bergmann glia. Within the adult cerebellum, the expression of four genes, leucine zipper protein 2 (Luzp2), G protein-coupled receptor 89 (Gpr89), leucine-rich repeat LGI family, member 4 (Lgi4), and Growth and differentiation factor 10 (Gdf10) appears to be restricted solely to the Purkinje cell layer (Fig. S2); and the cellular expression patterns of these genes within the PCL closely match those of well established astroglial markers expressed by Bergmann glia (Fig. S3). However, unlike most currently used markers that also label cerebellar astrocytes, these four genes appear to be completely specific to Bergmann glia. Gdf10 is particularly noteworthy because its specific expression in the Purkinje cell layer had been reported before , but was thought, we believe in error, to be in Lugaro cells rather than in Bergmann glia.
Gpr126, an Adhesion GPCR Expressed Specifically in Developing Bergmann Glia
GPCRs were one of the gene families of particular interest to us in light of our recent finding that Gpr56, an adhesion GPCR, is essential for cortical and cerebellar development , . Moreover, few GPCRs have been studied in the context of glial function. Among the many GPCRs we found expressed in Bergmann glia, we were especially intrigued by the very specific spatiotemporal pattern of expression of Gpr126, another orphan receptor that is a close relative of Gpr56. At P6, Gpr126 was expressed specifically in the Purkinje cell layer (Fig. 4, center panels). Since RT-PCR analysis of single-cell cDNAs from Purkinje cells showed that Gpr126 was absent in these neurons (data not shown), the pattern of GPR126 mRNA is consistent with expression by Bergmann glial cells. As predicted by the microarray data, in situ hybridization signal for Gpr126 was no longer detectable in the adult cerebellum (Fig. 4, right panels). At E15, Gpr126 was present in the cerebellar ventricular zone (arrowheads in Fig. 4, left panels), where precursors of cerebellar neurons and glia–including Bergmann glia–are located. In contrast, Gpr126 was absent in the forebrain ventricular zone at E15 (Fig. 4, arrow in left panel), indicating that it does not play a role in cortical radial glia at this age. Based on a recent study in zebrafish, regulation of cyclic AMP by Gpr126 signaling plays a critical role in the initiation of myelination by Schwann cells . Whether Gpr126 in Bergmann glia also regulates second messenger pathways involved in cell differentiation, or instead regulates cell adhesion similar to other adhesion GPCRs such as Gpr56, Celsr2 and Celsr3 , ,  remains to be examined.
in situ hybridization with a 33P-labeled probe for Gpr126, a little known GPCR of the adhesion family, reveals signal specifically in the Purkinje cell layer at P7 (arrows, middle panels), consistent with expression in Bergmann glia. Gpr126 expression is developmentally regulated, and becomes undetectable in the adult (right panel). Unlike most classic Bergmann glial markers, which are also expressed by cortical radial glia, Gpr126 is specific to Bergmann glia and not detected in cortical radial glia at E15 (arrow in top left panel). Labeling is seen in the ventricular zone of the developing cerebellar anlage at E15 (arrowheads in left panels), suggesting that Gpr126 may be expressed in progenitors of Bergmann glia. All sections are oriented with rostral to the right. Scale bar, upper panels: left, 3 mm; center, 2 mm; right, 2.5 mm; lower panels: left, 100 µm; center, 1 mm; right, 1.4 mm.
Bergmann Glia Express Genes Typical of Neural Stem Cells
A striking finding of our microarray analysis is that Bergmann glia, even in the adult, express numerous genes thought to be expressed specifically by stem cells (Table 8), and show enrichment of the embryonic stem cell pluripotency pathway (Table 6). Of the 220 genes that  identified as a core set of “stemness” genes common to multiple types of stem cells but not found in differentiated cells, 26.8% (60/220) are expressed by Bergmann glia at P6 (Table 8). Remarkably, 18.3% (41/220) remain expressed at P30. In addition, 26.3% (647/2458) of neural stem cell genes identified in the same study are also present in Bergmann glia at one or both ages (data not shown). These include transcription factors such as Sox1, Sox2, Sox9, Hes1 and Hes5, which play important roles in cell proliferation and in maintaining neural stem cell identity. These findings support the intriguing hypothesis  that Bergmann glia may be a source of the newly described neural stem cells in the adult cerebellum , . Our observation in Bergmann glia is also consistent with reports of progenitor cell-like gene expression in Muller Glia of the retina , .
Despite growing evidence of the indispensable roles of glial cells in many aspects of nervous system development and function, much remains unknown about the molecules that mediate these roles, particularly in vivo. There is also a lack of specific markers for various subtypes of glial cells, and of tools to manipulate gene expression only in specific subtypes. This has hindered our understanding of the diverse roles of glial cells. Our results from single cell transcriptome analysis of cerebellar Bergmann glia identify numerous novel genes whose role in Bergmann glia, or glia in general, can now be tested in functional contexts. We also identify several genes that appear to be entirely Bergmann glia-specific in the cerebellum. Not only does this confirm that the GFP+ cells we harvested for this study were indeed Bergmann glia and did not inadvertently include astrocytes, but also offers new tools for understanding these important glial cells. In this study we focused mainly on genes not previously characterized in Bergmann glia or glial cells in general. For analysis of all genes that we found expressed in Bergmann glia, our complete data set can be viewed in Table S5, and the original Affymetrix. CEL files can be accessed at the Gene Expression Omnibus (GEO) repository (accession number GSE18617).
Genes identified here as being expressed in Bergmann glia fall into many classes, highlighting the diverse roles of this cell type throughout life. Some of the most salient with respect to cerebellar structure and function include retinoic acid signaling components–e.g. Cyp26b1, a regulator of retinoic acid activity - which may play a role in cerebellar patterning ; chemoattractive and chemorepulsive molecules (notably semaphorin 4B), which may regulate neuronal migration as well as dendritic outgrowth and synapse development , , ; growth factors and growth factor-like molecules such as Gdf10 ,  and meteorin , ; molecules for cell-cell communication, particularly gap junction proteins, which have been shown to play critical roles in radial glial proliferation  and neuronal migration along radial glial fibers in the cerebral cortex ; synapse-associated adhesion molecules such as neurexins, which are important for synapse formation and maintenance ; components of the D-serine pathway involved in modulation of NMDA receptor function , ; and molecules that mediate synaptic function and plasticity, such as glutamate receptors, transporters, and transmembrane AMPAR regulatory proteins (TARPs) . The expression of the enzymes glutamine synthetase and pyruvate carboxylase along with glial glutamate transporters Slc1a3 (GLAST) and Slc1a2 (GLT-1) provide further evidence for the involvement of Bergmann glia in the glutamate-glutamine cycle that supports synaptic activity , . Similarly, the expression of the lactate synthetic enzymes lactate dehydrogenase A and B (Ldha and Ldhb), and monocarboxylate transporter 1 (Mct1), the main lactate transporter responsible for rapid release of glial lactate, is consistent with the hypothesized glia-neuron lactate shuttle .
Our finding that neurexin III is expressed in Bergmann glia is particularly interesting in light of the proposed role of neurexins in the development and maintenance of functional synapses. Neurexins present on the presynaptic membrane are thought to bind neuroligins on the postsynaptic membrane, thereby forming a trans-synaptic link that helps maintain the close apposition of pre- and post-synaptic elements . Originally thought to be presynaptic , immuno-electron microscopy has now shown that neurexins are also present postsynaptically . What has been missing in this analysis is the consideration of glial processes, which are also integral components of most CNS synapses and maintained in close proximity to pre-and post-synaptic elements , . The expression of neurexin III in Bergmann glia raises the question of whether this molecule plays a role in anchoring glial processes to pre- and post-synaptic elements. Cell type-specific deletion of this gene in astroglial cells can be performed to address this hypothesis.
The role of glia in synaptic function has been reinforced by the finding that astrocytes in cell culture or brain slices can release glutamate , , which in turn can modulate synaptic transmission and plasticity , , . However, the actual mechanism of glial glutamate release has remained controversial . There is evidence that astrocytes in vitro express components of regulated vesicle exocytosis previously thought to be found only in neurons, including v-glut1/2, SNAP23, Munc18a, and synaptotagmin IV , suggesting that astrocytes release glutamate by vesicle exocytosis similar to neurons. However, a recent study  reported that acutely isolated mouse astrocytes do not express v-glut1/2, synaptotagmins or synapsin I, and therefore are unlikely to exhibit regulated vesicular glutamate release in vivo. Our finding that acutely isolated adult Bergmann glia do express some known or potential components of regulated vesicular exocytosis, including synapsin I, synaptotagmins XI and XVI, syntaxins, snapin , rim2 , and Lgi3  (a recently identified syntaxin interactor and potential regulator of exocytosis) suggests that, unlike cortical astrocytes, Bergmann glia in vivo may possess the machinery for regulated release of glutamate and/or possibly other neurotransmitters.
While much of the recent focus on glia has been on their novel roles, it is worth noting that our knowledge of the molecular mechanisms remains incomplete even for the oldest and most commonly-associated role of glia: serving as “nerve glue,” a term coined by Rudolf Virchow in 1859. In this context, it is interesting that one of the largest set of genes we found expressed at high levels in Bergmann glia consists of cell adhesion molecules and receptors known or hypothesized to mediate cell-cell or cell-ECM binding. These include well-established adhesion molecules and receptors such as brevican, tenascin C, integrin αv, and dystroglycan-1, some of which are critical for structural integrity of the glial scaffold . A less studied adhesion molecule that we find in Bergmann glia, Chl1, was also shown recently to be important for the guidance and stabilization of stellate cell arbors projecting onto Purkinje cell dendrites , highlighting the important role of glial cell adhesion molecules in the development and maintenance of neuronal connections. In light of these findings, we believe that the putative glial adhesion molecules we identify, including cadherin 22, CD164 and junction adhesion molecule 2 (Jam2), merit further investigation.
In addition to identifying possible molecular players in known functions of Bergmann glia, the genes emerging from our study also strengthen the possibility of novel roles of these cells. Recently, two studies identified putative neural stem cells in the postnatal cerebellum , . While the identities of these cells remain unknown, a hypothesis has emerged that perhaps Bergmann glia could be these stem cells . A study from the same lab found that two transcription factors that regulate neural stem cell identity, Sox1 and Sox2, are found in postnatal Bergmann glia . We now significantly expand this line of inquiry by examining the full expression profiles of Bergmann glia and identifying additional genes that have previously been implicated in “stem-ness” of neural stem cells . The molecular and morphological changes that Bergmann glia undergo in response to injury, granule cell death, or implantation of embryonic granule cell precursors indicate that they remain highly plastic . Whether they possess the latent genetic potential to serve as neural precursors and could do so in response to an appropriate stimulus remains a tantalizing possibility. Furthermore, the mechanisms that may normally repress this potential in the adult cerebellum merit further investigation. In this regard, our observation that the Bone Morphogenic Protein antagonist, Nbl1 (neuroblastoma, suppression of tumorigenicity 1) shows highly elevated expression in adult Bergmann glia compared to P6 is interesting, since this gene has been shown previously to repress maintenance of the precursor state and promote neuronal differentiation through its action on BMP7 .
Materials and Methods
Experiments were performed in accordance with National Institutes of Health guidelines for the care and use of laboratory animals, and with approval of the Animal Care and Use Committee of Children's Hospital Boston.
Isolation of Single Bergmann Glial Cells
Mice of ages P6 and P30 expressing GFP under the control of the GFAP promoter (GFAP-GFP mice) were used. From GFAP-GFP mouse brains, slices of the mid-sagittal third of the cerebellum were cut in cold Hank's Balanced Salt Solution (HBSS). By cutting only from the mid-sagittal region, where Bergmann glial processes run mostly parallel to the sagittal plane, damage to glial processes was minimized. Furthermore, isolation of Bergmann glia from a restricted region of the cerebellum should minimize developmental heterogeneity between individual cells. Slices were cut into smaller pieces in cold Ca++- and Mg++-free HBSS containing 10 mM HEPES; The tissues were incubated in papain (20 U/ml), and DNase I (20 U/ml) in Ca++- and Mg++-free HBSS on a shaker for ∼30 min at 37°C. The protease solution was then replaced with Hanks Balanced Salt Solution (HBSS) containing 1 mg/ml albumin ovomucoid protease inhibitor, and the tissue was gently triturated using fire-polished glass pipettes of decreasing bore diameter. Cells were pelleted by centrifugation and resuspended in cold HBSS. A small aliquot of the cell suspension was added to a Petri dish with cold Ca++- and Mg++-free HBSS and individual cells were harvested by mouth pipetting into pulled glass microcapillaries attached to a micromanipulator. Bergmann glia were recognized by their GFP fluorescence and morphology. Each picked cell was rinsed in a fresh dish with HBSS and re-picked with a new microcapillary. Harvested cells were immediately seeded into PCR tubes containing reverse transcription buffer, and placed on ice. In control experiments, single GFP-negative cells and putative astrocytes were also picked. Astrocytes from cerebella of GFAP-GFP mice showed weaker GFP staining compared to Bergmann glia and lacked the characteristic unipolar processes of the latter. The identity of the different cerebellar cells was always verified subsequently by PCR and Southern blot.
Single-Cell RT-PCR and Microarray Hybridization
Single-cell RT-PCR was performed as described previously , , . Briefly, amplified cDNA was synthesized by lysing the cell, reverse transcribing the cell RNA after oligo-dT priming, poly-A tailing the 5′end of the cDNA, and finally amplifying the cell cDNA with a unique poly-T primer (AL1: ATTGGATCCAGGCCGCTCTGGACAAAATATGAATTC(T)24). The reverse transcription was performed in limiting conditions of nucleotides and time in order to generate cDNAs of uniform size (∼0.5 to 1 kb), which are more likely to be uniformly amplified and to accurately reflect the relative abundances of various mRNAs in the cell. After 50 cycles of PCR, several micrograms of cDNA were generated from each cell. Five µl of the cDNA was run on a 1.5% agarose gel to verify the presence of a smear from ∼0.5 to 1 kb. Using this original cDNA as template, additional cDNA could be faithfully reamplified as necessary by PCR using the AL1 primer, as described previously . Southern blots for several ubiquitous and cell-specific marker genes were then performed as described ,  to assess the quality and representation of the single cell cDNA, and to verify cell identity. Actb and Actg were used as high and medium abundance markers, respectively; and Odc and Ppp1ca as low abundance markers . Gfap, Fabp7, Sept4 and Slc1a3 were used as glial markers. Finally, the presence of Nefl, Mtap2, and in some cases also tubulin β-III was checked to detect any contaminating neurons. Only the best single cell cDNAs (10 µg of each), as determined by RT-PCR and Southern blot, were selected for labeling and microarray hybridization. In total, over 150 putative Bergmann glia were harvested, of which ∼90 yielded good cDNA smears after RT-PCR. The activity of the reverse transcriptases appeared to be the most variable factor, with noticeable lot-to-lot differences. Seventy-four of the single cell cDNAs showed robust expression of Actb, and of these, 31 were positive for Actg, Odc and Ppp1ca as well. Sixteen cDNA samples passed all quality control criteria, including presence of all tested astroglial markers and absence of all tested neuronal markers. Of these, five P6 and five P30 samples that exhibited the best quality control parameters on Affymetrix Test3 arrays were then hybridized to Affymetrix GeneChip Mouse Genome 430 2.0 microarrays at the Harvard Biopolymers Facility using standard Affymetrix protocols.
Data Analysis of Affymetrix Gene Chips
Analyses of individual microarrays and comparisons between P6 and adult were performed using GeneSpring GX 7.3 (Agilent). Raw CEL files were processed using the RMA (Robust Multichip Average) normalization algorithm as implemented in GeneSpring GX 7.3. Normalization was performed using default settings, which included data transformation (RAW values of less than 0.01 were set to 0.01), per chip normalization to the median (each measurement was divided by the 50th percentile of all measurements in that sample), and per gene normalization (the raw expression level of each gene was divided by the median of its measurements in all samples). For statistical analysis, one-way ANOVA was performed with multiple testing correction using Benjamini and Hochberg false discovery rate (FDR) set at 0.05. To identify genes expressed at higher levels at P6 compared to P30, two separate analyses were performed. First, we selected for genes that showed over three-fold higher normalized expression at P6 compared to P30 (and additionally, met statistical criteria mentioned above and were present at RAW levels of >100 in at least four out of five P6 samples). Second, we selected for genes that were flagged “present” in P6 samples (at least four out of five) and “absent” in P30 samples (at least four out of five) (and, like above, were statistically significant and present at raw signal values >100 in P6 samples). Similar analyses were performed to identify genes that were expressed at higher levels at P30 than at P6. Finally, by using an expression level filter, highly expressed genes showing raw signal values of over 2000 in at least 4 out of 5 samples of each age (total 8 of 10) were also identified. Data for annotation and functional classification of genes was obtained through Genespring (Agilent), the Gene Ontology Consortium , Aceview (www.aceview.org) , and previous studies. Our microarray data is MIAME compliant and all raw data files have been deposited in the Gene Expression Omnibus (GEO) repository, a MIAME compliant database.
In Situ Hybridization
In situ hybridization was performed essentially as described previously . Briefly, DNA templates for transcribing cRNA probes were generated by PCR. The primers contained SP6 (in forward primers) and T7 (in reverse primers) RNA polymerase binding sequences. For Gpr126, the following two primer pairs were used, both of which yielded similar results:
1. Forward, 5′-ATTTAGGTGACACTATAGAAGTGAGTGGTGGAGTCCTATTCATGG-3′; reverse, 5′- TAATACGACTCACTATAGGGAGACTCTGCTGAGGTGAATCTTAGTC-3′.
2. Forward, 5′- ATTTAGGTGACACTATAGAAGTGATGGATCAGACTGTGGCATACAAG-3′; reverse, 5′- TAATACGACTCACTATAGGGAGAGTCCAGGTTGCTAAAGAATGAATG-3′. Underlined regions correspond to the SP6 (in forward primers) and T7 (in reverse primers) RNA polymerase binding sequences. 33P-labeled sense and antisense riboprobes were generated using SP6 and T7 polymerases respectively (Promega) and a reaction mix containing 33P-UTP (Perkin Elmer). Radioactive in situ hybridization was performed on 16 µm brain cryosections as described previously , .
Gene Expression Atlases
Several online reference atlases of mRNA expression were used to corroborate the expression of genes identified in the microarrays. These included the Allen Brain Atlas (ABA) (http://www.brain-map.org/) , the Brain Gene Expression Map (http://www.stjudebgem.org) , and GenePaint (http://www.genepaint.org) .
Developmentally-regulated Bergmann glia genes (P6>P30 by >3-fold, or present at P6 and absent at P30)
(0.11 MB XLS)
Developmentally-regulated Bergmann glia genes (P30>P6 by >3-fold, or present at P30 and absent at P6)
(0.04 MB XLS)
Presumptive neuron-specific genes seen in Bergmann glial cDNA samples
(0.02 MB XLS)
Genes expressed at moderate to high levels in Bergmann glia (P6 and P30) (mean raw signal values >2000)
(1.24 MB XLS)
Complete data set of 45,101 probe sets for all Gene Chips used in our single cell gene expression analysis
(9.52 MB ZIP)
Myelin-related genes are expressed in adult Bergmann glia. Mid-sagittal views of adult mouse cerebella with in situ hybridization images (left panels) and expression level analysis (right panels), as obtained from the Allen Brain Atlas. Insets in all images are from the dorso-rostral region of lobule V (asterisk in top panels). Top row, mRNA for Pmp22, a constituent of myelin, is expressed most strongly in putative oligodendrocytes in the white matter (arrows). Surprisingly, however, there is also signal in the Purkinje cell layer (PCL), in a pattern consistent with expression in Bergmann glia (arrowheads). Second row, Plp1, the major constituent of CNS myelin, exhibits a similar expression pattern. Although the strongest staining is in putative oligodendrocytes (arrows), there is also distinct signal in Bergmann glia (arrowheads). Third row, Lgi4, which has been shown to be important for myelination in the PNS, shows robust and Bergmann glia-specific expression in the cerebellum. Fourth row, Mpzl1, another gene thought to be involved in myelination, is expressed in at least a subset of Bergmann glia (arrowheads). Scale bar, 500 µm in all panels, 140 µm in insets.
(4.55 MB TIF)
Bergmann glia-specific gene expression in the adult cerebellum. Mid-sagittal views of adult mouse cerebella with in situ hybridization images (left panels) and expression level analysis (right panels), as obtained from the Allen Brain Atlas. Insets in all images are from the dorso-rostral region of lobule V as in Fig. S1. Luzp2, Gpr89, Lgi4 and Gdf10 all appear to be expressed very specifically in the Purkinje cell layer in the adult cerebellum. This restricted localization, along with a cellular expression pattern in the PCL that matches those of well established astroglial markers (Fig. S3), suggests that these genes are Bergmann-glia specific in the adult cerebellum and may serve as novel markers for these cells. Scale bar, 500 µm in all panels, 140 µm in insets.
(4.22 MB TIF)
Expression patterns of well established astroglial markers in the adult cerebellum. Mid-sagittal views of adult mouse cerebella with in situ hybridization images (left panels) and expression level analysis (right panels), as obtained from the Allen Brain Atlas. Insets in all images are from the dorso-rostral region of lobule V as in Fig. S1. The glial genes, Slc1a3, Fabp7, S100β and Sept4 are four widely used astroglial markers, and their expression in Bergmann glia (but not Purkinje cells or other cerebellar neurons) has been confirmed by previous studies. The cellular expression patterns of these genes are presented here to serve as controls against which the Bergmann glial expression of new genes (for example, as in Fig. S2) can be compared. Scale bar, 500 µm in all panels, 140 µm in insets.
(4.27 MB TIF)
We thank Dr. Catherine Dulac, Dr. Ian Tietjen, Dr.Ross Perak, and Dr. Jason Rihel for guidance with the single-cell RT-PCR technique.
Conceived and designed the experiments: SK GC. Performed the experiments: SK. Analyzed the data: SK GC. Wrote the paper: SK GC.
- 1. Hausmann B, Sievers J (1985) Cerebellar external granule cells are attached to the basal lamina from the onset of migration up to the end of their proliferative activity. J Comp Neurol 241: 50–62.
- 2. Sievers J, Mangold U, Berry M, Allen C, Schlossberger HG (1981) Experimental studies on cerebellar foliation. I. A qualitative morphological analysis of cerebellar fissuration defects after neonatal treatment with 6-OHDA in the rat. J Comp Neurol 203: 751–769.
- 3. Sievers J, von Knebel Doeberitz C, Pehlemann FW, Berry M (1986) Meningeal cells influence cerebellar development over a critical period. Anat Embryol (Berl) 175: 91–100.
- 4. Hatten ME (1999) Central nervous system neuronal migration. Annu Rev Neurosci 22: 511–539.
- 5. Rakic P (1990) Principles of neural cell migration. Experientia 46: 882–891.
- 6. Belvindrah R, Nalbant P, Ding S, Wu C, Bokoch GM, et al. (2006) Integrin-linked kinase regulates Bergmann glial differentiation during cerebellar development. Mol Cell Neurosci 33: 109–125.
- 7. Graus-Porta D, Blaess S, Senften M, Littlewood-Evans A, Damsky C, et al. (2001) Beta1-class integrins regulate the development of laminae and folia in the cerebral and cerebellar cortex. Neuron 31: 367–379.
- 8. Lippman JJ, Lordkipanidze T, Buell ME, Yoon SO, Dunaevsky A (2008) Morphogenesis and regulation of Bergmann glial processes during Purkinje cell dendritic spine ensheathment and synaptogenesis. Glia 56: 1463–1477.
- 9. Yamada K, Fukaya M, Shibata T, Kurihara H, Tanaka K, et al. (2000) Dynamic transformation of Bergmann glial fibers proceeds in correlation with dendritic outgrowth and synapse formation of cerebellar Purkinje cells. J Comp Neurol 418: 106–120.
- 10. Yamada K, Watanabe M (2002) Cytodifferentiation of Bergmann glia and its relationship with Purkinje cells. Anat Sci Int 77: 94–108.
- 11. Iino M, Goto K, Kakegawa W, Okado H, Sudo M, et al. (2001) Glia-synapse interaction through Ca2+-permeable AMPA receptors in Bergmann glia. Science 292: 926–929.
- 12. Altman J, Bayer SA (1997) Development of the Cerebellar System in Relation to its Evolution, Structure, and Functions. New York: CRC Press.
- 13. Cui W, Allen ND, Skynner M, Gusterson B, Clark AJ (2001) Inducible ablation of astrocytes shows that these cells are required for neuronal survival in the adult brain. Glia 34: 272–282.
- 14. Bellamy TC (2006) Interactions between Purkinje neurones and Bergmann glia. Cerebellum 5: 116–126.
- 15. Lopez-Bayghen E, Rosas S, Castelan F, Ortega A (2007) Cerebellar Bergmann glia: an important model to study neuron-glia interactions. Neuron Glia Biol 3: 155–167.
- 16. Muller T, Kettenmann H (1995) Physiology of Bergmann glial cells. Int Rev Neurobiol 38: 341–359.
- 17. Teichberg VI (1991) Glial glutamate receptors: likely actors in brain signaling. Faseb J 5: 3086–3091.
- 18. Ottersen OP, Chaudhry FA, Danbolt NC, Laake JH, Nagelhus EA, et al. (1997) Molecular organization of cerebellar glutamate synapses. Prog Brain Res 114: 97–107.
- 19. Ottersen OP, Zhang N, Walberg F (1992) Metabolic compartmentation of glutamate and glutamine: morphological evidence obtained by quantitative immunocytochemistry in rat cerebellum. Neuroscience 46: 519–534.
- 20. Nimmerjahn A, Mukamel EA, Schnitzer MJ (2009) Motor behavior activates Bergmann glial networks. Neuron 62: 400–412.
- 21. Bordey A, Sontheimer H (2003) Modulation of glutamatergic transmission by bergmann glial cells in rat cerebellum in situ. J Neurophysiol 89: 979–988.
- 22. Brockhaus J, Deitmer JW (2002) Long-lasting modulation of synaptic input to Purkinje neurons by Bergmann glia stimulation in rat brain slices. J Physiol 545: 581–593.
- 23. Cahoy JD, Emery B, Kaushal A, Foo LC, Zamanian JL, et al. (2008) A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for understanding brain development and function. J Neurosci 28: 264–278.
- 24. Ramalho-Santos M, Yoon S, Matsuzaki Y, Mulligan RC, Melton DA (2002) “Stemness”: transcriptional profiling of embryonic and adult stem cells. Science 298: 597–600.
- 25. Tietjen I, Rihel JM, Cao Y, Koentges G, Zakhary L, et al. (2003) Single-cell transcriptional analysis of neuronal progenitors. Neuron 38: 161–175.
- 26. Zhuo L, Sun B, Zhang CL, Fine A, Chiu SY, et al. (1997) Live astrocytes visualized by green fluorescent protein in transgenic mice. Dev Biol 187: 36–42.
- 27. Dulac C, Axel R (1995) A novel family of genes encoding putative pheromone receptors in mammals. Cell 83: 195–206.
- 28. Eberwine J, Yeh H, Miyashiro K, Cao Y, Nair S, et al. (1992) Analysis of gene expression in single live neurons. Proc Natl Acad Sci U S A 89: 3010–3014.
- 29. Mager WH (1988) Control of ribosomal protein gene expression. Biochim Biophys Acta 949: 1–15.
- 30. Tushinski RJ, Warner JR (1982) Ribosomal proteins are synthesized preferentially in cells commencing growth. J Cell Physiol 112: 128–135.
- 31. Jimenez-Marin A, Collado-Romero M, Ramirez-Boo M, Arce C, Garrido JJ (2009) Biological pathway analysis by ArrayUnlock and Ingenuity Pathway Analysis. BMC Proc 3: Suppl 4S6.
- 32. Eiraku M, Tohgo A, Ono K, Kaneko M, Fujishima K, et al. (2005) DNER acts as a neuron-specific Notch ligand during Bergmann glial development. Nat Neurosci 8: 873–880.
- 33. Komine O, Nagaoka M, Watase K, Gutmann DH, Tanigaki K, et al. (2007) The monolayer formation of Bergmann glial cells is regulated by Notch/RBP-J signaling. Dev Biol 311: 238–250.
- 34. Weller M, Krautler N, Mantei N, Suter U, Taylor V (2006) Jagged1 ablation results in cerebellar granule cell migration defects and depletion of Bergmann glia. Dev Neurosci 28: 70–80.
- 35. Roubelakis MG, Martin-Rendon E, Tsaknakis G, Stavropoulos A, Watt SM (2007) The murine ortholog of the SHP-2 binding molecule, PZR accelerates cell migration on fibronectin and is expressed in early embryo formation. J Cell Biochem 102: 955–969.
- 36. Bermingham JR Jr, Shearin H, Pennington J, O'Moore J, Jaegle M, et al. (2006) The claw paw mutation reveals a role for Lgi4 in peripheral nerve development. Nat Neurosci 9: 76–84.
- 37. Tournell CE, Bergstrom RA, Ferreira A (2006) Progesterone-induced agrin expression in astrocytes modulates glia-neuron interactions leading to synapse formation. Neuroscience 141: 1327–1338.
- 38. Zhao R, Lawler AM, Lee SJ (1999) Characterization of GDF-10 expression patterns and null mice. Dev Biol 212: 68–79.
- 39. Koirala S, Jin Z, Piao X, Corfas G (2009) GPR56-regulated granule cell adhesion is essential for rostral cerebellar development. J Neurosci 29: 7439–7449.
- 40. Li S, Jin Z, Koirala S, Bu L, Xu L, et al. (2008) GPR56 regulates pial basement membrane integrity and cortical lamination. J Neurosci 28: 5817–5826.
- 41. Monk KR, Naylor SG, Glenn TD, Mercurio S, Perlin JR, et al. (2009) A G protein-coupled receptor is essential for Schwann cells to initiate myelination. Science 325: 1402–1405.
- 42. Shima Y, Kengaku M, Hirano T, Takeichi M, Uemura T (2004) Regulation of dendritic maintenance and growth by a mammalian 7-pass transmembrane cadherin. Dev Cell 7: 205–216.
- 43. Tissir F, Bar I, Jossin Y, De Backer O, Goffinet AM (2005) Protocadherin Celsr3 is crucial in axonal tract development. Nat Neurosci 8: 451–457.
- 44. Alcock J, Scotting P, Sottile V (2007) Bergmann glia as putative stem cells of the mature cerebellum. Med Hypotheses 69: 341–345.
- 45. Klein C, Butt SJ, Machold RP, Johnson JE, Fishell G (2005) Cerebellum- and forebrain-derived stem cells possess intrinsic regional character. Development 132: 4497–4508.
- 46. Lee A, Kessler JD, Read TA, Kaiser C, Corbeil D, et al. (2005) Isolation of neural stem cells from the postnatal cerebellum. Nat Neurosci 8: 723–729.
- 47. Blackshaw S, Harpavat S, Trimarchi J, Cai L, Huang H, et al. (2004) Genomic analysis of mouse retinal development. PLoS Biol 2: E247.
- 48. Eichele G (1997) Retinoids: from hindbrain patterning to Parkinson disease. Trends Genet 13: 343–345.
- 49. Bagri A, Tessier-Lavigne M (2002) Neuropilins as Semaphorin receptors: in vivo functions in neuronal cell migration and axon guidance. Adv Exp Med Biol 515: 13–31.
- 50. Paradis S, Harrar DB, Lin Y, Koon AC, Hauser JL, et al. (2007) An RNAi-based approach identifies molecules required for glutamatergic and GABAergic synapse development. Neuron 53: 217–232.
- 51. Tran TS, Kolodkin AL, Bharadwaj R (2007) Semaphorin regulation of cellular morphology. Annu Rev Cell Dev Biol 23: 263–292.
- 52. Cunningham NS, Jenkins NA, Gilbert DJ, Copeland NG, Reddi AH, et al. (1995) Growth/differentiation factor-10: a new member of the transforming growth factor-beta superfamily related to bone morphogenetic protein-3. Growth Factors 12: 99–109.
- 53. Jorgensen JR, Thompson L, Fjord-Larsen L, Krabbe C, Torp M, et al. (2009) Characterization of Meteorin-An Evolutionary Conserved Neurotrophic Factor. J Mol Neurosci.
- 54. Nishino J, Yamashita K, Hashiguchi H, Fujii H, Shimazaki T, et al. (2004) Meteorin: a secreted protein that regulates glial cell differentiation and promotes axonal extension. Embo J 23: 1998–2008.
- 55. Weissman TA, Riquelme PA, Ivic L, Flint AC, Kriegstein AR (2004) Calcium waves propagate through radial glial cells and modulate proliferation in the developing neocortex. Neuron 43: 647–661.
- 56. Elias LA, Wang DD, Kriegstein AR (2007) Gap junction adhesion is necessary for radial migration in the neocortex. Nature 448: 901–907.
- 57. Craig AM, Kang Y (2007) Neurexin-neuroligin signaling in synapse development. Curr Opin Neurobiol 17: 43–52.
- 58. Mustafa AK, Kim PM, Snyder SH (2004) D-Serine as a putative glial neurotransmitter. Neuron Glia Biol 1: 275–281.
- 59. Panatier A, Theodosis DT, Mothet JP, Touquet B, Pollegioni L, et al. (2006) Glia-derived D-serine controls NMDA receptor activity and synaptic memory. Cell 125: 775–784.
- 60. Soto D, Coombs ID, Renzi M, Zonouzi M, Farrant M, et al. (2009) Selective regulation of long-form calcium-permeable AMPA receptors by an atypical TARP, gamma-5. Nat Neurosci 12: 277–285.
- 61. Schousboe A, Westergaard N, Waagepetersen HS, Larsson OM, Bakken IJ, et al. (1997) Trafficking between glia and neurons of TCA cycle intermediates and related metabolites. Glia 21: 99–105.
- 62. Magistretti PJ, Pellerin L (1999) Astrocytes Couple Synaptic Activity to Glucose Utilization in the Brain. News Physiol Sci 14: 177–182.
- 63. Dean C, Dresbach T (2006) Neuroligins and neurexins: linking cell adhesion, synapse formation and cognitive function. Trends Neurosci 29: 21–29.
- 64. Dean C, Scholl FG, Choih J, DeMaria S, Berger J, et al. (2003) Neurexin mediates the assembly of presynaptic terminals. Nat Neurosci 6: 708–716.
- 65. Taniguchi H, Gollan L, Scholl FG, Mahadomrongkul V, Dobler E, et al. (2007) Silencing of neuroligin function by postsynaptic neurexins. J Neurosci 27: 2815–2824.
- 66. Ventura R, Harris KM (1999) Three-dimensional relationships between hippocampal synapses and astrocytes. J Neurosci 19: 6897–6906.
- 67. Parpura V, Basarsky TA, Liu F, Jeftinija K, Jeftinija S, et al. (1994) Glutamate-mediated astrocyte-neuron signalling. Nature 369: 744–747.
- 68. Parpura V, Haydon PG (2000) Physiological astrocytic calcium levels stimulate glutamate release to modulate adjacent neurons. Proc Natl Acad Sci U S A 97: 8629–8634.
- 69. Araque A, Parpura V, Sanzgiri RP, Haydon PG (1998) Glutamate-dependent astrocyte modulation of synaptic transmission between cultured hippocampal neurons. Eur J Neurosci 10: 2129–2142.
- 70. Fellin T, Pascual O, Gobbo S, Pozzan T, Haydon PG, et al. (2004) Neuronal synchrony mediated by astrocytic glutamate through activation of extrasynaptic NMDA receptors. Neuron 43: 729–743.
- 71. Newman EA (2003) New roles for astrocytes: regulation of synaptic transmission. Trends Neurosci 26: 536–542.
- 72. Nedergaard M, Ransom B, Goldman SA (2003) New roles for astrocytes: redefining the functional architecture of the brain. Trends Neurosci 26: 523–530.
- 73. Volterra A, Meldolesi J (2005) Astrocytes, from brain glue to communication elements: the revolution continues. Nat Rev Neurosci 6: 626–640.
- 74. Pan PY, Tian JH, Sheng ZH (2009) Snapin facilitates the synchronization of synaptic vesicle fusion. Neuron 61: 412–424.
- 75. Schoch S, Mittelstaedt T, Kaeser PS, Padgett D, Feldmann N, et al. (2006) Redundant functions of RIM1alpha and RIM2alpha in Ca(2+)-triggered neurotransmitter release. Embo J 25: 5852–5863.
- 76. Park WJ, Lee SE, Kwon NS, Baek KJ, Kim DS, et al. (2008) Leucine-rich glioma inactivated 3 associates with syntaxin 1. Neurosci Lett 444: 240–244.
- 77. Moore SA, Saito F, Chen J, Michele DE, Henry MD, et al. (2002) Deletion of brain dystroglycan recapitulates aspects of congenital muscular dystrophy. Nature 418: 422–425.
- 78. Ango F, Wu C, Van der Want JJ, Wu P, Schachner M, et al. (2008) Bergmann glia and the recognition molecule CHL1 organize GABAergic axons and direct innervation of Purkinje cell dendrites. PLoS Biol 6: e103.
- 79. Sottile V, Li M, Scotting PJ (2006) Stem cell marker expression in the Bergmann glia population of the adult mouse brain. Brain Res 1099: 8–17.
- 80. Sotelo C, Alvarado-Mallart RM, Frain M, Vernet M (1994) Molecular plasticity of adult Bergmann fibers is associated with radial migration of grafted Purkinje cells. J Neurosci 14: 124–133.
- 81. Krizhanovsky V, Ben-Arie N (2006) A novel role for the choroid plexus in BMP-mediated inhibition of differentiation of cerebellar neural progenitors. Mech Dev 123: 67–75.
- 82. Dulac C (1998) Cloning of genes from single neurons. Curr Top Dev Biol 36: 245–258.
- 83. Church GM, Gilbert W (1984) Genomic sequencing. Proc Natl Acad Sci U S A 81: 1991–1995.
- 84. Dolter KE, Braman JC (2001) Small-sample total RNA purification: laser capture microdissection and cultured cell applications. Biotechniques 30: 1358–1361.
- 85. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, et al. (2000) Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25: 25–29.
- 86. Thierry-Mieg D, Thierry-Mieg J (2006) AceView: a comprehensive cDNA-supported gene and transcripts annotation. Genome Biol 7: Suppl 1S12 11–14.
- 87. Schaeren-Wiemers N, Gerfin-Moser A (1993) A single protocol to detect transcripts of various types and expression levels in neural tissue and cultured cells: in situ hybridization using digoxigenin-labelled cRNA probes. Histochemistry 100: 431–440.
- 88. Corfas G, Rosen KM, Aratake H, Krauss R, Fischbach GD (1995) Differential expression of ARIA isoforms in the rat brain. Neuron 14: 103–115.
- 89. Sassoon D, Rosenthal N (1993) Detection of messenger RNA by in situ hybridization. Methods Enzymol 225: 384–404.
- 90. Lein ES, Hawrylycz MJ, Ao N, Ayres M, Bensinger A, et al. (2007) Genome-wide atlas of gene expression in the adult mouse brain. Nature 445: 168–176.
- 91. Magdaleno S, Jensen P, Brumwell CL, Seal A, Lehman K, et al. (2006) BGEM: an in situ hybridization database of gene expression in the embryonic and adult mouse nervous system. PLoS Biol 4: e86.
- 92. Visel A, Thaller C, Eichele G (2004) GenePaint.org: an atlas of gene expression patterns in the mouse embryo. Nucleic Acids Res 32: D552–556.