Neuroscience 151 (2008) 82–91

CHARACTERIZATION OF HETEROGENEOUS GLIAL CELL POPULATIONS INVOLVED IN DEHYDRATION-INDUCED PROLIFERATION IN THE ADULT RAT NEUROHYPOPHYSIS I. VIRARD, O. GUBKINA, F. ALFONSI AND P. DURBEC*

internal changes, either in physiological or pathological conditions. It is now clearly established that plasticity occurs in the adult mammal CNS (for a review, see Lledo et al., 2006). For instance, the hypothalamo-neurohypophysial system, which secretes oxytocin and vasopressin, shows remarkable plasticity in response to appropriate stimulation, in particular in the posterior lobe of the pituitary or neurohypophysis (NH). During physiological stimulation, such as dehydration, parturition and lactation, the NH undergoes an important structural reorganization involving both neurons and pituicytes (for a review see Hatton, 1997). Pituicytes are specialized glial cells expressing astrocyte markers such as glial fibrillary acidic protein (GFAP), vimentin and S100B (Cocchia and Miani, 1980; Salm et al., 1982; Velasco et al., 1982; Marin et al., 1989). In the resting state, the pituicytes enclose oxytocin- and vasopressin-secreting axons from hypothalamic neurons, while they release the engulfed neurosecretory axons during prolonged physiological stimulation. This active retraction of pituicyte processes allows a greater contact between the neurosecretory endings and blood vessels and is thought to favor oxytocin and vasopressin secretion into the general circulation (Tweedle and Hatton, 1980; Hatton, 1990). Furthermore, an earlier study showed that prolonged dehydration induced by saline substitution of drinking water also results in an intense cell proliferation in the adult NH (Murugaiyan and Salm, 1995). However, from this report, it was unclear whether the dividing cell population derives exclusively from pituicytes induced to reenter the cell cycle or from the stimulation of resident glial precursors or stem cells maintained in a resting state under basal conditions. Interestingly, we had previously demonstrated that, despite the complete absence of oligodendrocytes in the adult rodent NH, oligodendrocyte precursor cells (OPCs) can be identified in the developing structure. These OPCs generate pituicytes instead of myelinating oligodendrocytes and some OPC markers are maintained in a cell population in the adult NH (Virard et al., 2006). This observation highlighted the fact that, in the adult NH, the glial population could be more heterogeneous than previously thought. Therefore, we wanted to investigate further the cellular composition of the adult rat NH, to establish whether it contains immature glial cells and/or stem cells and to determine how these cells participate in the plasticity of the structure. In this work, we used a variety of glial lineage markers and functional in vitro assays to investigate the heteroge-

Université de la Méditerranée, CNRS-UMR6216, Institute of Developmental Biology of Marseille-Luminy, Case 907, Campus de Luminy, 13288 Marseille Cedex 9, France

Abstract—The adult neurohypophysis (NH) is a well-established site of CNS plasticity: its glial cells, the pituicytes, reorganize their structure and undergo increased proliferation in response to stimulations such as dehydration. However, it remains to be clarified whether the newly-formed cells derive from pituicytes re-entering the cell cycle or from glial precursors or stem cells. Here, we first analyze the expression of several glial markers in the adult rat NH and demonstrate that the pituicytes constitute a heterogeneous population. In particular, we identify a distinct subtype of glial cells expressing the oligodendrocyte precursor marker plateletderived growth factor receptor alpha (pdgfr␣). In addition, adult NH explants can give rise to migratory precursors able to differentiate into mature oligodendrocytes, unlike NH cells in vivo. This led us to hypothesize that the adult NH could contain immature cells, therefore we used a neurosphereforming assay to test for the presence of stem or progenitor cells. Adult NH cells can generate bipotent primary neurospheres but not secondary ones, suggesting that the structure contains glial progenitors but probably not stem cells. Finally, when the NH is stimulated by dehydration, we observe an increase in cell proliferation associated with an increase in cell death. By identifying the cells incorporating bromodeoxyuridine (BrdU) or positive for Ki67, we demonstrate that this increased proliferation concerns all glial cell types in the adult NH, including the pdgfr␣ⴙ cells. Our study shows that the NH is a complex structure composed of multiple glial subtypes, which all participate in the physiological response to dehydration. © 2008 IBRO. Published by Elsevier Ltd. All rights reserved. Key words: CNS, gliogenesis, glial progenitors, stem cells, pituicytes.

Brain plasticity refers to the structural and/or functional modifications that allow the brain to adapt to external or *Corresponding author. Tel: ⫹33-491-26-97-46; fax: ⫹33-491-82-06-82. E-mail address: [email protected] (P. Durbec). Abbreviations: BrdU, bromodeoxyuridine; DIG, digoxigenin; DMEM, Dulbecco’s modified Eagle’s medium; FCS, fetal calf serum; FGF, fibroblast growth factor; GalC, galactocerebroside; GFAP, glial fibrillary acidic protein; HBSS, Hanks’ balanced salt solution; ILB4, isolectin B4; MBP, myelin basic protein; NH, neurohypophysis; NSCs, neural stem cells; NT3, neurotrophin-3; OPCs, oligodendrocyte precursor cells; PBlec, phosphate-buffered saline 1% Triton X-100 with 0.1 mM CaCl2, 0.1 mM MgCl2, 0.1 mM MnCl2; PBS, phosphate-buffered saline; pdgfr␣, platelet-derived growth factor receptor alpha; PF, paraformaldehyde; SVZ, subventricular zone; TSA, tyramide signal amplification; 3R, 6R, 3 or 6 days of rehydration following 9 days of dehydration; 5D, 9D, either 5 or 9 days of dehydration, respectively.

0306-4522/08$32.00⫹0.00 © 2008 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2007.10.035

82

I. Virard et al. / Neuroscience 151 (2008) 82–91

neity of NH glial cells and analyzed how these populations respond to dehydration in terms of proliferation.

EXPERIMENTAL PROCEDURES Animals We used adult male Sprague–Dawley rats (IFFA CREDO, SaintGermain-sur-l’Arbesle, France). All procedures involving the use of animals were performed in accordance with the European Community Council Directive of 24 November 1986 on the protection of animals used for experimental purposes (86/609/EEC). The experimental protocols were carried out in compliance with institutional Ethical Committee guidelines for animal research. All efforts were made to minimize the number of animals used and their suffering.

Neurohypophysial explant cultures Solutions for cell culture were purchased from Gibco (Invitrogen, Cergy-Pontoise, France) and other chemicals from Sigma (SaintQuentin-Fallavier, France) if not stated otherwise. Neurohypophysial culture was performed as described earlier (Wang et al., 1994; Virard et al., 2006). Briefly, pituitaries from adult rats were dissected in Hanks’ balanced salt solution (HBSS). Meninges were stripped off and the neural lobes carefully separated from the anterior and intermediate lobes. Each NH was sectioned into four to six pieces in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal calf serum (FCS). The pieces were then explanted in the presence of 0.4% methylcellulose onto poly-Llysine-treated glass coverslips and maintained in serum-free medium (DMEM-F12 complemented with 100 ␮g/ml human transferrin, 5 ␮g/ml insulin, 100 ␮M putrescine, 20 nM progesterone, 30 nM sodium selenite), supplemented with fibroblast growth factor (FGF) -2 (10 ng/ml), PDGF-AA (10 ng/ml) and neurotrophin-3 (NT3) (10 ng/ml) as indicated in the text. Cultures were incubated at 37 °C in a 5% CO2 and 95% air atmosphere and fed with fresh medium every 2–3 days.

83

Dehydration and bromodeoxyuridine (BrdU) incorporation experiments For our dehydration experiments, 6-week-old rats were maintained either with normal drinking water (controls) or with a 2% NaCl solution (dehydrated) for either 5 or 9 days (respectively termed 5D and 9D). After 9 days of dehydration, groups of rats were allowed a rehydration period of 3 or 6 days (3R and 6R). At least three rats were used in each group. Cumulative in vivo labeling of cell proliferation was performed by repetitive BrdU (Sigma) injections. A daily injection (50 mg/kg) of BrdU (10 mg/ml in phosphate-buffered saline (PBS), pH 7.4) was given intraperitoneally during the entire dehydration period. There were no BrdU injections during the rehydration period.

Tissue processing Anesthetized rats were perfused with cold PBS followed by cold 4% paraformaldehyde (PF), or a 3% PF solution containing sodium periodate and lysine monohydrochloride for optimal staining with the OX42 antibody (McLean and Nakane, 1974). Their pituitaries and brain were post-fixed with 4% PF for 1–2 h at 4 °C, then cryoprotected with 20% sucrose, included in OCT (Sakura Finetek, Bayer Diagnostic) and kept at ⫺80 °C until used. Cryostat sections (12 ␮m) were collected on SuperFrost slides (VWR, Fontenay-sous-Bois, France).

In situ hybridization on sections In situ hybridization was done as previously described (Virard et al., 2006) on adult rat brain and NH cryosections using digoxigenin (DIG) -labeled antisense riboprobes for rat cnp2 (Gravel et al., 1994), olig1, olig2 (Lu et al., 2000), platelet-derived growth factor receptor alpha (pdgfr␣) (Mudhar et al., 1993), or mouse plp/dm20 (Peyron et al., 1997). To detect pdgfr␣ by fluorescent in situ hybridization, we used an anti-DIG antibody coupled to horseradish peroxidase, then labeled it with a tyramide signal amplification (TSA) Plus Cyanine 3 kit (NEL744B001KT; Perkin-Elmer, Courtaboeuf, France) according to the manufacturer’s protocol (6 min incubation in TSA).

Neurosphere cultures To obtain primary neurospheres, 5-week-old rats were killed by decapitation and their brains and hypophyses were dissected out. Brains were sectioned into 400 ␮m thick slices using a vibratome (Leica Microsystème, Rueil-Malmaison, France). The subventricular zone (SVZ) from the lateral wall of the anterior lateral ventricle horn was dissected in HBSS and cut into 100 ␮m diameter explants. NHs were dissected as described above and cut into 100 ␮m diameter pieces. NH and SVZ fragments were dissociated mechanically after enzymatic treatment (trypsin, 5 mg/ml). After rinsing, the cells were plated at a concentration of 15,000 cells/ml and cultured in the serum-free medium described above, complemented with B-27 (1/50), FGF-2 (10 ng/ml) and epithelial growth factor (10 ng/ml). The medium was changed every 2–3 days. Secondary spheres were obtained by chemical (Accumax, Sigma) and mechanical dissociation of primary neurospheres after 7 days in vitro and cultured in the same medium. Direct counting of primary and secondary spheres was performed using a Leica DM IRB inverted microscope after 7 and 14 days in vitro, respectively. The primary and secondary sphere diameter was measured based on an image analysis approach using the ImageJ 1.33u software (Wayne Rasband, NIH, Bethesda, MD, USA). Primary sphere differentiation was induced by plating the neurospheres on poly-L-lysine-coated coverslips in the serum-free medium complemented with 0.3% FCS. Sphere multipotentiality was tested by immunofluorescence 5–7 days after plating, using anti-GFAP, Tuj1 and O4 antibodies to stain astrocytes, neurons and oligodendrocytes, respectively.

Immunolabeling We used the following primary antibodies and lectin: mouse A2B5, galactocerebroside (GalC) and O4 (pure supernatant ATCC, Manassas, VA, USA), mouse anti-BrdU (1/100, DAKO, France), mouse anti–myelin basic protein (MBP) (1/500, Euromedex), rabbit anti-Olig2 (1/20,000, after heat antigen retrieval, gift from D. Rowitch, Boston, MA, USA), rabbit anti-NG2 (1/100, Chemicon), mouse anti-CD11b, clone OX42 (1/1000, Cymbus Biotechnology, UK), mouse anti-GFAP (1/1000, Sigma), rabbit anti-GFAP (1/ 1000, Sigma), isolectin B4 from Griffonia simplicifolia (ILB4, 1/100, Sigma), rabbit anti-Ki67 (1/400 after heat antigen retrieval, Abcam, UK), mouse anti-Nkx2.2 (1/800, Developmental Studies Hybridoma Bank, USA), rabbit anti-S100 (1/300, Dako), mouse antiS100, clone SHB1 (1/1000, Sigma), mouse anti-␤III tubulin (Tuj1, 1/5000, BAbCO, Richmond, CA, USA) and mouse anti-vimentin (1/200, Sigma). For immunolabeling, the tissue sections were incubated overnight at 4 °C in a primary antibody solution containing 10% FCS, as well as 0.3% Triton-X100 for internal markers. The sections were subsequently washed and incubated with the appropriate fluorescently-labeled secondary antibodies (1/200; Jackson Immunoresearch Laboratories, West Grove, PA, USA) for 1–2 h at room temperature. In the case of ILB4, the sections were first rinsed twice with PBlec (phosphate-buffered saline 1% Triton X-100 with 0.1 mM CaCl2, 0.1 mM MgCl2, 0.1 mM MnCl2), then incubated overnight with ILB4 1/100 in PBlec and washed. Finally, we detected the biotinylated ILB4 using streptavidin coupled to

84

I. Virard et al. / Neuroscience 151 (2008) 82–91

either Texas Red or Oregon Green (Molecular Probes, 1/1000 in saturation buffer). For immunostaining on explant and neurosphere cultures, the cells were fixed with 4% PF, rinsed, incubated in primary antibodies for 30 – 60 min, rinsed again, and incubated for 1 h in a secondary antibody solution. Cell nuclei were stained with Hoechst 33342 (1/1000, Sigma). For BrdU labeling, the sections were incubated for 15–30 min at 37 °C in 2 N HCl and 0.5% Triton X-100. After three washes in 0.1 M sodium tetraborate, they were incubated overnight with the anti-BrdU antibody at 4 °C. After several washes, we covered the sections with a biotinylated anti-mouse IgG antibody solution (1: 200) for 5 h, washed them, then incubated them with Oregon Green– coupled streptavidin for 2 h. In the case of BrdU stainings combined with another immunolabeling, this other labeling was performed first, then the tissues were fixed with 4% PF and we proceeded with the anti-BrdU protocol.

Cell apoptosis assay Apoptotic cells were identified on adult rat hypophysis sections by TUNEL immunohistochemistry using the ApopTag kit (QBiogene, Illkirch, France) according to the manufacturer’s instructions.

Fluorescent microscopy, quantification and statistical analysis Sections were examined using Zeiss confocal or ApoTome microscopes and cell counting was done on digital pictures. The number of positive cells for a given marker was counted on at least three randomly-chosen NH sections and compared with the total number of cells identified with the nuclear marker. To calculate surface areas, we used the Visiolab 2000 software (Biocom), the area of one field being 250 ␮m2. The presented values are means⫾standard error of the mean (S.E.M.). For two-group comparisons, the data were statistically processed with Student’s t-test when normality passed. Pⱕ0.05 was considered significant.

RESULTS Glial cell diversity in the adult rat NH In order to characterize the pituicyte population from the adult NH, we performed an immunohistochemical analysis using different markers for the astroglial lineage, namely GFAP, S100 and vimentin. GFAP⫹ cells could clearly be identified on sections of adult rat NH as process-bearing cells (Fig. 1A). Although GFAP is usually considered as the main pituicyte marker, we observed that S100 was more widely expressed in pituicytes: antibodies to GFAP (Fig. 1A), S100 (Fig. 1B) and vimentin (Fig. 1C) labeled respectively 10⫾1%, 43⫾3% and 12⫾2% of the total cells within the structure. Using double immunostaining and confocal analysis, we showed that 10⫾1% of the total population co-expressed S100 and GFAP (Fig. 1D), 7⫾1% S100 and vimentin (Fig. 1E), and 4⫾1% GFAP and vimentin (Fig. 1F). In addition to pituicytes, it is well known that the NH contains blood vessels, onto which terminate axons from hypothalamic nuclei. These vessels are made of numerous endothelial cells as well as perivascular cells (fibroblasts, pericytes, microglia, mastocytes) (Paterson and Leblond, 1977). Most of these cells can be labeled using ILB4 (Laitinen, 1987; Streit and Kreutzberg, 1987), which marks 23⫾1% of the total NH cells (Fig. 1G). Microglia represents

about half of the ILB4⫹ cells, since the microglial marker OX42 labeled 9% of the total NH cells (Fig. 1H). We did not observe any cell co-labeled with ILB4 and pituicyte markers S100 or GFAP (not shown). Overall, our data showed that adult rat NH cells include approximately 43% S100⫹ pituicytes and revealed heterogeneity within these cells, with subpopulations expressing various combinations of astroglial markers (Fig. 1M). Identification of pdgfr␣-expressing cells in the adult NH We recently showed that pituicytes derive from OPCs and that some OPC markers are expressed in the adult structure (Virard et al., 2006). In order to uncover the identity of this cell population, we performed an analysis with markers of the oligodendrocyte lineage. As demonstrated before (Virard et al., 2006), immunofluorescence on sections using antibodies against mature oligodendrocyte markers such as GalC or MBP revealed no labeling in the NH (not shown), thus indicating the absence of late stages of the oligodendrocyte lineage. However, we detected the expression of NG2, an integral membrane chondroitin sulfate proteoglycan expressed by early committed OPCs (Nishiyama et al., 1997), in 19⫾1% of the adult NH cells (Fig. 1J green; for interpretation of the references to color in this figure, the reader is referred to the Web version of this article). Double labeling showed that the NG2⫹ cells were negative for vimentin, S100 and GFAP, but that about two thirds were co-labeled with ILB4, indicating that they were blood vessel or associated cells (not shown). The pdgfr␣ transcript, another early OPC marker (Hall et al., 1996), was detected in numerous cells scattered throughout the NH (Fig. 1I), representing 6.2⫾ 0.4% of the total cells. These pdgfr␣⫹ cells are distinct from GFAP⫹ (not shown) or S100⫹ pituicytes (Fig. 1K) and from ILB4⫹ blood vessel cells (Fig. 1L). Only occasionally did pdgfr␣⫹ cells appear to express NG2 (Fig. 1J). No staining could be detected using other OPC markers, specifically probes for olig1, olig2, plp/dm20 or cnp2, or antibodies against Nkx2.2 or Olig2, whereas they stained numerous cells in control tissues like the adult corpus callosum or the developing optic nerve (not shown). Overall, these results indicate that the adult pituicytes include a separate subpopulation of cells positive for the OPC marker pdgfr␣ (Fig. 1M) but expressing no other oligodendrocyte lineage marker. Adult NH progenitors can generate oligodendrocytes in culture We then wanted to determine the functional properties of the pdgfr␣⫹ putative progenitors found in the adult rat NH and therefore turned to in vitro assays. Unfortunately, due to the lack of known cell surface markers resistant to strong enzymatic treatment necessary to dissociate adult NH tissues, it was not technically possible to isolate the various NH cell populations to examine separately their differentiation potential. Therefore, we resorted to using a cruder approach of adult NH explants to establish the functional characteristics of the NH cells.

I. Virard et al. / Neuroscience 151 (2008) 82–91

85

Fig. 1. Cell diversity in the adult rat NH. (A–F) Pituicyte subpopulations express heterogeneous combinations of the astroglial markers GFAP, S100 and vimentin. The numbers indicate the percentage of the total NH cells immunopositive for each marker (A–C) or pair of markers (D–F). Arrows point to doubly labeled cells. (G, H) Blood vessel cells stained with ILB4 represent 23% of the NH cells (G), including OX42⫹ microglia (H). (I–L) In situ hybridization with a pdgfr␣ probe identifies a new glial cell population (I) which rarely co-expresses the OPC marker NG2 (J) and is distinct from S100⫹ pituicytes (K) and ILB4⫹ blood vessel cells (L). Small panels show enlargements of the boxed areas in J–L, with separate color channels and a merged image. (M) Recapitulative scheme of the different NH cell populations. Results obtained with NG2 were not included due to localization of this marker on various cell populations. Scale bars⫽25 ␮m; 10 ␮m in J–L insets.

First, when NHs were dissected out and cultured in defined medium in the absence of any trophic factor, no cellular migration was observed around the explants (not shown). Second, when FGF-2, PDGF-AA and NT3, known to favor migration, proliferation and survival of optic nerve OPCs (McKinnon et al., 1990; Barres et al., 1994), were added to the defined medium, occasional cell migration

was observed. After 3– 8 days, rare motile cells with a bipolar morphology (round cell bodies and two long processes) and positive for the OPC marker A2B5 (Fig. 2A) were seen around the explants. These cells were also immunopositive for O4, a marker of early committed oligodendrocytes (Fig. 2B). After 15–21 days, we found scarce cells with the typical ramified morphology of mature oligo-

86

I. Virard et al. / Neuroscience 151 (2008) 82–91

Fig. 2. Oligodendrocyte differentiation of adult NH-derived cells in vitro. After 8 days in the presence of FGF-2, PDGF-AA and NT3, rare bipolar cells exit adult rat NH explants. These cells express the oligodendrocyte lineage markers A2B5 (A) and O4 (B). After 15 days, they are multipolar and immunopositive for GalC (C) and MBP (D), indicating their differentiation into mature oligodendrocytes. Scale bar⫽58 ␮m.

dendrocytes, expressing GalC (Fig. 2C) and showing MBP labeling (Fig. 2D). Despite the presence of GFAP⫹ cells within the explants, no GFAP or ␤III-tubulin labeling was ever observed around these explants, indicating respectively the absence of astrocytes and neuronal cells in the migrating cells (not shown). These results indicate that a progenitor cell population able to give rise to oligodendrocytes persists in the adult NH. However, in their normal environment, these progenitors do not naturally generate oligodendrocytes, since no myelin or mature oligodendrocyte marker can be detected in the NH. Adult NH cells can generate primary but not secondary neurospheres The oligodendrocytes generated in our in vitro assays could actually derive from neural stem cells (NSCs) or multipotent progenitors. To determine whether the NH contains such cells, we examined the neurosphere-forming potential of NH cells. Indeed, testing the ability to form neurospheres is widely used to demonstrate the presence of NSCs in the CNS (Reynolds and Weiss, 1992; Weiss et al., 1996). Thus, we dissociated the NHs from adult rats and put them in culture at a low density. In parallel, as a positive control for culture conditions, we also plated SVZ cells from the same rats. Primary neurospheres were generated from cells from both tissues (Fig. 3A, B), indicating that the adult NH contains progenitors capable of proliferation. However, the NH cells gave rise to a much smaller number of primary spheres (Fig. 3E): 21⫾3 per 10,000 plated cells compared with 160⫾2 for the SVZ after 7 days in culture. Also, the NH spheres were slightly smaller, with an average diameter of 42⫾1 ␮m instead of 58⫾1 ␮m (Fig. 3F, P⫽0.001). In addition, when we dissociated the primary spheres, we were able to obtain secondary neurospheres only from SVZ cells, not from NH cells (Fig. 3C, D). This suggests that NH-derived neurospheres are not able to proliferate indefinitely. Primary spheres were allowed to differentiate and their multipotentiality was tested by triple immunofluorescence 5–7 days after plating (Fig. 3G, H). As expected, all SVZderived neurospheres gave rise to GFAP⫹ astrocytes, O4⫹ oligodendrocytes and ␤III tubulin⫹ neuronal cells (Fig. 3G), indicating that they are multipotent. In comparison, NH-derived spheres only generated GFAP⫹ astrocytes and O4⫹ oligodendrocytes (Fig. 3H) and did not give rise to neurons. Therefore, the NH contains progenitors that differ from SVZ stem cells in their limited capacity to self-renew and

their ability to generate only glial cells. This implies that the oligodendrocytes obtained in our NH explant assay most likely did not derive from NSCs but rather from glial precursors. All glial cell types examined participate in dehydration-induced proliferation Having identified different cell subtypes in the adult NH, we were highly interested in understanding how these cell populations react to a physiological stimulation of the structure. A way to stimulate the NH is a prolonged dehydration, which induces a strong proliferation in NH cells (Paterson and Leblond, 1977; Kawamoto and Kawashima, 1984; Murugaiyan and Salm, 1995). Therefore, we submitted adult rats to dehydration by saline substitution of their drinking water. In a first set of experiments, we wanted to study how NH cells globally respond to stimulation over a dehydration–rehydration period in terms of proliferation and cell death. Control or 9D adult rats received daily BrdU injections during the dehydration period. We monitored cell proliferation by counting BrdU⫹ cells either at the end of dehydration (9D) or after a rehydration period of 3 (3R) or 6 (6R) days. We observed that the density of BrdU⫹ cells was considerably increased in dehydrated animals compared with controls (Fig. 4A, B). Indeed, after 9D, 29.4⫾2% of the NH cells were BrdU⫹ compared with 11.2⫾0.7% in the control situation (Fig. 4C). Cell proliferation persisted during the rehydration period (Fig. 4C): the proportion of BrdU⫹ cells was 40.9⫾0.9% after 3R compared with 11.4% in controls. The subsequent decrease in the number of BrdU⫹ cells observed on the 6th day of rehydration could be explained by a persistence of proliferation that causes BrdU dilution among the progeny of dividing cells. To examine further the levels of cell proliferation, we used Ki67 as a marker for cells undergoing cell cycle during the dehydration period, at the end of dehydration, and during rehydration. The anti-Ki67 antibody labeled 2.4⫾0.4% of the total NH cells in control rats compared with 8.5⫾0.9%, 9.4⫾1.1% and 6.9⫾1.4% in dehydrated rats after 5D, 9D and 3R, respectively (not shown). Therefore, there is a highly significant difference between control and both 5D and 9D values (one-way ANOVA followed by a Holm-Sidak analysis; P⫽0.001 and P⫽0.003 respectively). This confirms that proliferation is indeed elevated in NH cells over the dehydration period studied. The difference between control and 3R values is less important (P⫽0.014), a sign that proliferation has started returning to normal levels during rehydration.

I. Virard et al. / Neuroscience 151 (2008) 82–91

Fig. 3. Neurosphere-forming potential of adult NH cells compared with SVZ cells. (A–D) Primary neurospheres from adult rat SVZ (A) and NH (B) cells. Dissociated SVZ neurospheres give rise to secondary neurospheres (C) but NH neurospheres do not (D). Though plated in the same conditions, the NH primary spheres are much less numerous (E) and slightly smaller (F) than SVZ-derived spheres. Cells from adult rat SVZ (G) and NH (H) primary neurospheres were allowed to differentiate for 6 days and tested for their expression of GFAP (turquoise), ␤III-tubulin (red) and O4 (green). Scale bars⫽100 ␮m in A–D; 25 ␮m in G, H. For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.

87

Although stimulating the NH induced a strong proliferative response, we did not observe any increase in the size of the NH (Fig. 4D), nor in its cell density (Fig. 4E). Thus, increased cell proliferation is not associated with a concomitant detectable increase in the total cell number. Consequently, we hypothesized that there must be an augmentation of cell death. Indeed, we could observe a peak in cell apoptosis in rat NH during rehydration as monitored by TUNEL staining (Fig. 4F). Our next objective was to determine how the different cell subpopulations of the adult NH respond to dehydration, in particular the S100⫹ and GFAP⫹ pituicyte subpopulations, the ILB4⫹ blood vessel cells and the newly identified pdgfr␣⫹ glial cells. We chose to study them after 3R since it is the time point at which we observed the largest numbers of BrdU⫹ cells (Fig. 4C) and TUNELstained cells (Fig. 4F). First, by counting the number of cells positive for S100, GFAP, ILB4 or pdgfr␣, we could not detect any significant difference in their proportion between control and stimulated rats (Fig. 5A). Then, double stainings was used to determine the phenotype of the BrdU⫹ cells (Fig. 5C, D). Among the S100⫹ pituicytes, BrdU⫹ cells represented 1.4⫾0.4% of the cells in controls and 20.3⫾ 5.7% in 3R rats (Fig. 5B, D). Within the GFAP⫹ pituicyte population, BrdU stained 1.3⫾0.3% and 9.0⫾1.8% of cells in controls and 3R animals respectively (Fig. 5B). BrdU was incorporated in 5.3⫾0.7% out of the total in pdgfr␣⫹ glial cells in controls, versus 11.8⫾2.8% in rehydrated rats (Fig. 5B, C). Overall, in the rats rehydrated for 3 days, there was a significantly larger number of BrdU⫹ cells among each of the three glial subpopulations we examined. Concerning the ILB4⫹ blood vessel cells, 4.8⫾1.4% of them presented detectable levels of BrdU in controls and 9.0⫾1.5% in stimulated rats but this difference was not significant (Fig. 5B). Finally, because cumulative BrdU incorporation cannot be used to determine which cells actually undergo division, we sought to examine the phenotype of the Ki67⫹ cells during (5D) and at the end (9D) of dehydration, as well as during rehydration (3R). Among the Ki67⫹ positive cells, we observed only a small number of cells also positive for glial cell markers. This precluded any reliable quantification and statistical analyses. Nevertheless, we found GFAP⫹, S100⫹ and pdgfr␣⫹ cells among the Ki67⫹ cells. This indicates that all glial subpopulations self-renew during dehydration and rehydration, as well as in control conditions. Our findings suggest that all resident NH glial cell populations are slowly proliferating in normally-hydrated animals and that all are induced to divide during stimulation of the NH, including the pdgfr␣⫹ cells. Also, this increase in cell proliferation is accompanied by a parallel increase in cell death.

DISCUSSION In the present study, we established that the pituicyte population is heterogeneous based on the expression of the glial markers S100, GFAP and vimentin, and we identified a separate cell type expressing pdgfr␣. Using in vitro assays, we showed that the adult NH contains glial pro-

88

I. Virard et al. / Neuroscience 151 (2008) 82–91

Fig. 4. BrdU incorporation, NH size, cell density and apoptosis during dehydration and rehydration. (A, B) Anti-BrdU staining illustrates that NH cells proliferate at a basal level in control NH (A) that is greatly increased after 9D (B). (C) When BrdU is incorporated during 9D, the number of BrdU⫹ cells is approximately tripled in the NH. It increases further after 3R and starts decreasing but remains elevated after 6R. (D, E) At either 9D, 3R or 6R, there is no significant change in the size (D) or cell density (E) of any of the hypophysis lobes (IL: intermediate lobe; AH: adenohypophysis). (F) TUNEL assays indicate that there is increased cell death in the NH after dehydration. Scale bar⫽500 ␮m.

genitors capable of differentiating into oligodendrocytes and/or astrocytes but that the structure does not harbor NSCs. Finally, in response to prolonged dehydration, increased proliferation was observed in all the different glial cell populations we had identified in the adult NH. Pituicyte heterogeneity Our work demonstrates that the glial cells present in the adult NH do not all express the same combination of astrocyte markers (Fig. 1): pituicytes express S100 but only fractions of them co-express vimentin or GFAP. Pituicyte heterogeneity had already been reported based on morphological criteria in earlier electron and conventional light microscopy studies (Bucy, 1932; Dellmann and Owsley, 1969; Takei et al., 1980). However, these results were variable and several studies described only one type of pituicytes, especially in rodents (Krsulovic and Bruckner, 1969; Galabov and Schiebler, 1978). Also, since pituicytes are known to undergo morphological changes during NH

stimulation, it could not be excluded that this heterogeneity was linked to simple shape differences. Here, we clearly established based on marker expression that pituicytes are heterogeneous in the resting state and that this heterogeneity is maintained during stimulation: the relative proportions of cells expressing different glial markers do not change after dehydration (Fig. 5A). An interesting possibility would be that these glial subpopulations could constitute maturation stages of a single pituicyte lineage. It has already been suggested that the low expression of intermediate filaments (GFAP, vimentin) in pituicytes compared with astrocytes could be linked to their morphological plasticity (Miyata and Hatton, 2002). Therefore, vimentin⫹ or GFAP⫹ pituicytes could represent less plastic, more mature pituicytes while those expressing only S100 could be more plastic cells. We also identified a distinct subpopulation of cells expressing pdgfr␣, with an unusual identity. Indeed, pdgfr␣⫹ cells, often called synantocytes or adult OPCs,

I. Virard et al. / Neuroscience 151 (2008) 82–91

89

et al., 1999). Also, a recent study on adult OPCs in the mouse spinal cord indicates that the oligodendrocyte lineage marker Olig2 is detected in a great majority of NG2⫹ cells (Kitada and Rowitch, 2006), while we cannot observe any staining in the adult NH using the same antibody (data not shown). Likewise, no staining could be detected in the adult NH using other OPC markers, namely olig1, plp/ dm20, cnp2, or Nkx2.2 (not shown). Although some heterogeneity has been reported in the expression of certain markers in adult OPCs (Horner et al., 2002; Polito and Reynolds, 2005), our results suggest that the pdgfr␣⫹ cells in the adult NH are radically different from most NG2⫹ adult OPCs. Yet, these pdgfr␣⫹ cells are not blood vessel cells (Fig. 1L), and we consider them as glial cells of the NH. In other words, we believe that the pdgfr␣⫹ cells are a distinct subgroup within the heterogeneous pituicyte population. In vitro characteristics of progenitors in the adult NH

Fig. 5. Cell subtype proportions, BrdU incorporation and Ki67 expression after NH stimulation. (A) At 3R, the proportions of cells expressing pdgfr␣, S100, GFAP or ILB4 are not affected by dehydration. (B, C) BrdU is detected in different cell subpopulations, including pdgfr␣⫹ cells (B) and S100⫹ pituicytes (C). (D) All four cell subtypes examined show increased BrdU incorporation at 3R. (E, F) The cell cycle marker Ki67 can be detected in cells of all examined subpopulations, for instance in pdgfr␣⫹ (E) and GFAP⫹ (F) cells. Small panels are enlargements of the boxed areas, with separate color channels and a merged image. Scale bars⫽25 ␮m; 10 ␮m in insets.

are present in other regions of the adult CNS (Dawson et al., 2000; Butt et al., 2002) but they express markers which are not found in the adult NH pdgfr␣⫹ cells. In particular, pdgfr␣ and NG2 are almost never co-expressed in the NH (Fig. 1L) whereas NG2 is expressed in the majority of pdgfr␣⫹ cells in the adult brain (Nishiyama

Our explant experiments show that adult NH cells can give rise to rare A2B5⫹ precursors that later differentiate into myelinating MBP⫹ oligodendrocytes (Fig. 2). In addition, these cells only migrate out of NH explants when put in the presence of trophic factors known to favor OPC motility (McKinnon et al., 1990; Barres et al., 1994). Therefore, the adult NH explants have the potential to generate cells which share some important functional properties with OPCs. Moreover, we were able to grow primary neurospheres from adult NH cells and these spheres differentiated exclusively into astrocytes and oligodendrocytes, which confirms the presence of glial progenitors (Fig. 3). Furthermore, the absence of secondary spheres and the fact that the sphere-forming cells in the adult NH are not multipotent point to the absence of NSCs in the structure. However, there remains the question of the identity of these precursors in vivo. Unfortunately, the very small number of progenitors present in the adult NH and the absence of adequate cell surface markers preclude any cell sorting analysis which could give more insight into the identity of the NH progenitors. We can nevertheless put forward the hypothesis that the glial NH progenitors could be the pdgfr␣⫹ cells. They do not express most classical adult OPC markers but this could be due to specific environmental cues present in the NH which could downregulate oligodendrocyte lineage markers and thus prevent oligodendrocyte differentiation, similar to what is thought to occur during NH development (Virard et al., 2006). Glial cell division in the stimulated adult NH Next, we sought to determine how the different NH cell populations respond to stimulation by dehydration in terms of proliferation. We first noticed that the number of BrdU⫹ cells greatly increased after 9D (Fig. 4A, B), consistent with a previous report (Murugaiyan and Salm, 1995). We observed a peak of BrdU⫹ cells after 3R (Fig. 4C), which would indicate that proliferation persists at the beginning of the rehydration period. The subsequent decrease is possibly due to BrdU dilution, since the cell cycle marker Ki67

90

I. Virard et al. / Neuroscience 151 (2008) 82–91

still labels a high proportion of cells at 3R. Nonetheless, we cannot completely exclude the possibility that some BrdU⫹ cells disappear as a result of cell death (see below). Our BrdU incorporation experiments also show that, after 9D followed by 3R, BrdU is detectable in a larger proportion of each glial cell population (pdgfr␣⫹, S100⫹ and GFAP⫹ cells) (Fig. 5B). This indicates that all these cell types participate in the dehydration-induced increase in gliogenesis. Of course, one must bear in mind that the BrdU has accumulated over the entire dehydration period. Therefore if, for example, an S100⫹ pituicyte is labeled with BrdU (Fig. 5D), it can either mean that the S100⫹ cell has undergone cell cycle or that a S100⫺ precursor has divided and differentiated into an S100⫹ cell. However, the experiments using Ki67 seem to indicate that all cell types can proliferate. Incidentally, we have previously shown that, during development, S100⫹ pituicytes derive from olig1⫹ OPCs, which also express pdgfr␣. We can then hypothesize that the pdgfr␣⫹ cells present in the adult NH could be part of the pituicyte lineage and represent progenitors maintained with the purpose of generating new pituicytes. This would be a similitude with neurogenic areas, in which stem or progenitor cell populations participate in cell renewal. In the hippocampus for example, neuronal progenitors with limited self-renewal capacity locally proliferate and differentiate into postmitotic granule neurons of the dentate gyrus, as can be monitored by BrdU incorporation coupled to immunocytochemical staining with neuronal lineage markers (Kuhn et al., 1996). In the NH, many of the newly-formed cells are glial cells that can continue dividing even in basal conditions (Fig. 5D), therefore it is not possible to use BrdU in order to identify the progenitor cells and to trace their progeny: BrdU⫹ cells would not necessarily be progenitors and, conversely, newly-formed cells could be BrdU- because of BrdU dilution. Still, we can postulate that the various pituicyte subpopulations we have identified could represent different maturation stages of the pituicyte lineage and thus be a sign of constant cell renewal in the mature NH. In order to unquestionably demonstrate a lineage relationship from pdgfr␣⫹ to S100⫹ cells, further studies are needed using for instance a genetic approach to follow the fate of the adult pdgfr␣⫹ cells. Cell number regulation during NH stimulation A remarkable feature of CNS regions where neural progenitors or stem cells generate new cells is the regulation of cell number (Biebl et al., 2000). For example, stem cells in the well-studied SVZ constantly give rise to neuroblasts that migrate to the olfactory bulb, where they differentiate into new interneurons, and this adult neurogenesis is at least partly compensated for by cell death, including apoptosis of newly-generated neurons (Winner et al., 2002). Here we showed that there is no apparent change in either the NH size or cell density (Fig. 4B, C) over stimulation, thus the increased proliferation is not linked with a concomitant increase in the total NH cell number. We showed that NH cells are eliminated over the time course of stimulation by an increase in apoptosis (Fig. 4D). To our knowl-

edge, this is the first demonstration that there is a balance between newly-formed NH cells and cell death in response to physiological stimulation in the NH. Cell death could affect young pituicytes or there could be a preferential elimination of more mature pituicytes that are possibly less plastic in terms of process retraction.

CONCLUSION In summary, our study demonstrates that the NH glial cell population is more complex than expected, comprising multiple cell subtypes that are all involved in the NH response to dehydration. In particular, it reveals the presence of pdgfr␣⫹ cells, which do not express the same combination of markers as other pdgfr␣⫹ cells in the rest of the CNS. Their relationship with other pituicytes remains a matter of debate that will need to be examined using genetic tools. Acknowledgments—The authors are grateful to G. Monti for her expert assistance with the cultures; to Drs. B. Zalc and R. Lu for providing probe templates; and to Dr. D. Rowitch for the anti-Olig2 antibody. Many thanks to all laboratory members for constant input and helpful comments on the manuscript. I.V. was supported by a fellowship from the Fondation Recherche Médicale (FRM), and F.A. by the Association Française contre les Myopathies (AFM).

REFERENCES Barres BA, Raff MC, Gaese F, Bartke I, Dechant G, Barde YA (1994) A crucial role for neurotrophin-3 in oligodendrocyte development. Nature 367:371–375. Biebl M, Cooper CM, Winkler J, Kuhn HG (2000) Analysis of neurogenesis and programmed cell death reveals a self-renewing capacity in the adult rat brain. Neurosci Lett 291:17–20. Bucy P (1932) The hypophysis cerebri. In: Cytology and cellular pathology of the nervous system, Vol. II (Penfield W, ed), pp 707– 738. New York: Paul B. Horber Inc. Butt AM, Kiff J, Hubbard P, Berry M (2002) Synantocytes: new functions for novel NG2 expressing glia. J Neurocytol 31:551–565. Cocchia D, Miani N (1980) Immunocytochemical localization of the brain-specific S-100 protein in the pituitary gland of adult rat. J Neurocytol 9:771–782. Dawson MR, Levine JM, Reynolds R (2000) NG2-expressing cells in the central nervous system: are they oligodendroglial progenitors? J Neurosci Res 61:471– 479. Dellmann HD, Owsley PA (1969) Investigations on the hypothalamoneurohypophysial neurosecretory system of the grass frog (Rana pipiens) after transection of the proximal neurohypophysis. II. Light- and electron microscopic findings in the disconnected distal neurohypophysis with special emphasis on the pituicytes. Z Zellforsch Mikrosk Anat 94:325–336. Galabov P, Schiebler TH (1978) The ultrastructure of the developing neural lobe. Cell Tissue Res 189:313–329. Gravel M, DeAngelis D, Braun PE (1994) Molecular cloning and characterization of rat brain 2=,3=-cyclic nucleotide 3=-phosphodiesterase isoform 2. J Neurosci Res 38:243–247. Hall A, Giese NA, Richardson WD (1996) Spinal cord oligodendrocytes develop from ventrally derived progenitor cells that express PDGF alpha-receptors. Development 122:4085– 4094. Hatton GI (1990) Emerging concepts of structure-function dynamics in adult brain: the hypothalamo-neurohypophysial system. Prog Neurobiol 34:437–504. Hatton GI (1997) Function-related plasticity in hypothalamus. Annu Rev Neurosci 20:375–397.

I. Virard et al. / Neuroscience 151 (2008) 82–91 Horner PJ, Thallmair M, Gage FH (2002) Defining the NG2-expressing cell of the adult CNS. J Neurocytol 31:469 – 480. Kawamoto K, Kawashima S (1984) Ultrastructural changes and proliferation of pituicytes in mouse posterior lobe during water deprivation and rehydration. Acta Anat (Basel) 119:136 –141. Kitada M, Rowitch DH (2006) Transcription factor co-expression patterns indicate heterogeneity of oligodendroglial subpopulations in adult spinal cord. Glia 54:35– 46. Krsulovic J, Bruckner G (1969) Morphological characteristics of pituicytes in different functional stages. Light- and electron microscopy of the neurohypophysis of the albino rat. Z Zellforsch Mikrosk Anat 99:210 –220. Kuhn HG, Dickinson-Anson H, Gage FH (1996) Neurogenesis in the dentate gyrus of the adult rat: age-related decrease of neuronal progenitor proliferation. J Neurosci 16:2027–2033. Laitinen L (1987) Griffonia simplicifolia lectins bind specifically to endothelial cells and some epithelial cells in mouse tissues. Histochem J 19:225–234. Lledo PM, Alonso M, Grubb MS (2006) Adult neurogenesis and functional plasticity in neuronal circuits. Nat Rev Neurosci 7:179 –193. Lu QR, Yuk D, Alberta JA, Zhu Z, Pawlitzky I, Chan J, McMahon AP, Stiles CD, Rowitch DH (2000) Sonic hedgehog-regulated oligodendrocyte lineage genes encoding bHLH proteins in the mammalian central nervous system. Neuron 25:317–329. Marin F, Boya J, Lopez-Carbonell A (1989) Immunocytochemical localization of vimentin in the posterior lobe of the cat, rabbit and rat pituitary glands. Acta Anat 134:184 –190. McKinnon RD, Matsui T, Dubois-Dalcq M, Aaronson SA (1990) FGF modulates the PDGF-driven pathway of oligodendrocyte development. Neuron 5:603– 614. McLean IW, Nakane PK (1974) Periodate-lysine-paraformaldehyde fixative. A new fixation for immunoelectron microscopy. J Histochem Cytochem 22:1077–1083. Miyata S, Hatton GI (2002) Activity-related, dynamic neuron-glial interactions in the hypothalamo-neurohypophysial system. Microsc Res Tech 56:143–157. Mudhar HS, Pollock RA, Wang C, Stiles CD, Richardson WD (1993) PDGF and its receptors in the developing rodent retina and optic nerve. Development 118:539 –552. Murugaiyan P, Salm AK (1995) Dehydration-induced proliferation of identified pituicytes in fully adult rats. Glia 15:65–76. Nishiyama A, Chang A, Trapp BD (1999) NG2⫹ glial cells: a novel glial cell population in the adult brain. J Neuropathol Exp Neurol 58:1113–1124.

91

Nishiyama A, Yu M, Drazba JA, Tuohy VK (1997) Normal and reactive NG2⫹ glial cells are distinct from resting and activated microglia. J Neurosci Res 48:299 –312. Paterson JA, Leblond CP (1977) Increased proliferation of neuroglia and endothelial cells in the supraoptic nucleus and hypophysial neural lobe of young rats drinking hypertonic sodium chloride solution. J Comp Neurol 175:373–390. Peyron F, Timsit S, Thomas JL, Kagawa T, Ikenaka K, Zalc B (1997) In situ expression of PLP/DM-20, MBP, and CNP during embryonic and postnatal development of the jimpy mutant and of transgenic mice overexpressing PLP. J Neurosci Res 50:190 –201. Polito A, Reynolds R (2005) NG2-expressing cells as oligodendrocyte progenitors in the normal and demyelinated adult central nervous system. J Anat 207:707–716. Reynolds BA, Weiss S (1992) Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255:1707–1710. Salm AK, Hatton GI, Nilaver G (1982) Immunoreactive glial fibrillary acidic protein in pituicytes of the rat neurohypophysis. Brain Res 236:471– 476. Streit WJ, Kreutzberg GW (1987) Lectin binding by resting and reactive microglia. J Neurocytol 16:249 –260. Takei Y, Seyama S, Pearl GS, Tindall GT (1980) Ultrastructural study of the human neurohypophysis. II. Cellular elements of neural parenchyma, the pituicytes. Cell Tissue Res 205:273–287. Tweedle CD, Hatton GI (1980) Evidence for dynamic interactions between pituicytes and neurosecretory axons in the rat. Neuroscience 5:661– 671. Velasco ME, Roessmann U, Gambetti P (1982) The presence of glial fibrillary acidic protein in the human pituitary gland. J Neuropathol Exp Neurol 41:150 –163. Virard I, Coquillat D, Bancila M, Kaing S, Durbec P (2006) Oligodendrocyte precursor cells generate pituicytes in vivo during neurohypophysis development. Glia 53:294 –303. Wang C, Rougon G, Kiss JZ (1994) Requirement of polysialic acid for the migration of the O-2A glial progenitor cell from neurohypophyseal explants. J Neurosci 14:4446 – 4457. Weiss S, Reynolds BA, Vescovi AL, Morshead C, Craig CG, van der Kooy D (1996) Is there a neural stem cell in the mammalian forebrain? Trends Neurosci 19:387–393. Winner B, Cooper-Kuhn CM, Aigner R, Winkler J, Kuhn HG (2002) Long-term survival and cell death of newly generated neurons in the adult rat olfactory bulb. Eur J Neurosci 16:1681–1689.

(Accepted 11 October 2007) (Available online 12 November 2007)