Cend1 Is Expressed at Low Levels in Neural Stem/Progenitor Cells of the SVZ and at Higher Levels in Differentiated OB Neurons
Our initial immunohistochemical analysis revealed that along the adult mouse SVZ-RMS-OB pathway (Fig. 1A) Cend1 is expressed in low but detectable levels in the SVZ (Fig. 1A), where proliferating neuronal precursors reside, and is elevated in areas of differentiated neurons, such as the OB (Fig. 1A). To determine the cell type(s) expressing Cend1 in the SVZ, we performed double and triple labeling for Cend1 and characteristic markers of SVZ subtypes. To assess Cend1 expression in the type B cell population , we performed triple immunofluorescence labeling for Cend1, nestin, and glial fibrillary acidic protein (GFAP) (Fig. 1B). Numerous triple-labeled Cend1+/nestin+/GFAP+ cells were detected within the SVZ neurogenic niche (Fig. 1B, main panel and inset). In contrast to its colocalization with GFAP in the SVZ (Fig. 1B), Cend1 did not colocalize with GFAP+ astrocytes outside the SVZ area (data not shown; [8, 9]). Cend1 was also localized in (glutamate transporter) (GLAST+) cells lying near the ventricle (Fig. 1C). These cells, which express the radial glial marker GLAST, have previously been shown to coincide with the GFAP+ type B cell population of the SVZ [6, 21, 22]. To verify Cend1 expression in type B cells, as well as to quantify the extent of Cend1 localization in the GFAP+ cell population, we performed double immunofluorescence labeling in dissociated cells derived from adult SVZ, 3 hours after plating (Fig. 1F). Double labeling for Cend1 and GFAP demonstrated that Cend1 is detected in 56.0% ± 2.8% of GFAP+ cells of the SVZ at this age (Fig. 1F, arrows). Apart from its expression in type B cells, Cend1 is also present in Dlx2+ transit-amplifying type C cells (Fig. 1D) and in Pax6+ cells (Fig. 1E), which largely represent type A neuroblasts, but also some type C precursors .
Figure Figure 1.. Expression of Cend1 in the adult SVZ and OB. (A): Nissl-staining of a sagittal section of adult mouse brain, including the region where the SVZ-rostral migratory stream (RMS)-OB pathway localizes. (A1, A2): Following immunofluorescence staining for Cend1 in higher magnification (insets), note the low but still detectable levels of Cend1 in the SVZ (A2) and high expression in differentiated granule neurons of the OB (A1). (B–F, H–I): Maximum projections of confocal optical sections of the SVZ area and OB. (B): In the SVZ area, Cend1 is detected in the nestin+/GFAP+ NSC. The inset shows at higher magnification a triple-labeled Cend1+/nestin+/GFAP+ cell. (C): Cend1 also colocalizes with the embryonic radial glia and adult NSC marker GLAST. (D, E): Double immunofluorescence labeling for Cend1 and Dlx2 indicates that both proteins colocalize in the SVZ area (D), whereas double immunostaining of the anterior part of the SVZ and beginning of the RMS with antibodies to Cend1 and Pax6 indicate that Cend1 colocalizes with Pax6 (E). (F): In acute cultures derived from adult SVZ, Cend1 colocalizes with a subpopulation of GFAP+ cells (arrows). (G): In these acute cultures Cend1 expression levels (green) are lower in GLAST+ cells (arrowheads, left panel) and significantly higher in βIII-tubulin+ cells (arrows, right panel), appearing yellow from the merged green/red fluorescence of Cend1/βIII-tubulin colocalization. Quantification of Cend1 fluorescent intensity in individual cells, based on measurements from 50 independent cells for each condition, indicated that Cend1 expression is fivefold higher in differentiated neurons compared with precursor cells. (H, I): Coronal (H) and sagittal (I) sections of the OB depicting high expression of Cend1 in NeuN+ neurons and in GABAergic interneurons (arrows). Scale bars = 20 μm (A1, A2, B, G, H, I), 5 μm (D, E), 10 μm (C), and 40 μm (F). Abbreviations: GFAP, glial fibrillary acidic protein; OB, olfactory bulb; SVZ, subventricular zone.
Download figure to PowerPoint
In the OB Cend1 is expressed in mature NeuN+ neurons (Fig. 1H) and in GABAergic local interneurons (Fig. 1I). In accordance with previous observations [8, 9], the levels of Cend1 expression are higher in brain regions where differentiated neurons lie (for example, in Fig. 1A, 1B, compare the low expression of Cend1 in the SVZ with the immediately adjacent area, where Cend1 is highly expressed). This observation was further verified by double immunocytochemical staining of adult SVZ acute cultures (Fig. 1G), with antibodies to Cend1 (Fig. 1G, lower panel) and markers of either precursor cells, such as GLAST, or differentiated neurons, such as βIII-tubulin. Quantification of Cend1 fluorescence intensity in individual cells using the Cell-R software (Olympus, Tokyo, http://www.olympus-global.com) confirmed that Cend1 expression levels were, on average, fivefold higher in βIII-tubulin+ neurons compared with GLAST+ precursor cells. Thus, the expression pattern of Cend1 indicates that it marks both neuronal progenitors and differentiated neurons and that a positive correlation links its expression levels to the degree of neuronal maturation.
Cend1 Attenuates Proliferation of Neonatal SVZ Neural Precursors
To investigate whether Cend1 expression has functional significance for neuronal lineage progression during secondary neurogenesis, we used neurosphere cultures [1, 23] derived from postnatal day 5 (P5) SVZ as a model system to study neuronal precursor proliferation and differentiation. These cells self-renew when cultured in the presence of EGF/bFGF, either as free-floating neurospheres or after dissociation and plating as single cells, whereas they can differentiate into neurons, astrocytes, and oligodendrocytes upon growth factor withdrawal. Confocal analysis of dissociated neurosphere cells 2 days after plating showed that Cend1 is expressed by 90% ± 4% of nestin+ cells under proliferating conditions (Fig. 2A, 2F), whereas under differentiating conditions it is present in all βIII-tubulin+ neurons (Fig. 2D) but not in O4+ oligodendrocytes (Fig. 2E). Interestingly, Cend1 is also expressed by 48% ± 6% of GFAP+ cells after 2 days in culture in the absence of growth factors (Fig. 2B, arrow). Since a large percentage (approximately 80%) of GFAP+ cells in these cultures are also positive for nestin (Fig. 2C, 2F), a protein known to be expressed by immature astrocytes, it appears that Cend1 may also be expressed by immature astrocytes in vitro.
Figure Figure 2.. Immunofluorescence analysis of Cend1 expression in neurosphere cultures. Dissociated neurospheres were plated as single cells and cultured either in the presence (A) or in the absence (B–E) of epithelial growth factor and basic fibroblast growth factor. Shown are confocal images of cells stained for Cend1 and nestin (A), Cend1 and GFAP (B), nestin and GFAP (C), Cend1 and βIII-tubulin with the TuJ1 antibody (D), and Cend1 and O4 (E). TOPRO-3 was used for nuclear staining (A, B). Note that Cend1 is present in cells of the neuronal lineage (A, D), as well as in a subpopulation of GFAP+ cells (B). (F): Table indicating colocalization percentages of Cend1 and the principal lineage markers, as well as the percentage of nestin+ cells in the total GFAP+ population. Scale bar = 20 μm. Abbreviation: GFAP, glial fibrillary acidic protein.
Download figure to PowerPoint
To assess the contribution of Cend1 in neuronal lineage progression early postnatally, we genetically manipulated its expression in P5 SVZ-derived neurospheres, where Cend1 was either overexpressed or silenced using appropriate lentiviral vectors (Fig. 3). The protocol used for cell transduction with the lentiviral vectors is schematically presented in Figure 4A. Neurospheres were transduced as free-floating spheres but also, in one particular type of experiment, after dissociation to single cells that were then brought to clonal density and allowed to form spheres again (Fig. 4A). To study the effect of Cend1 overexpression, the Trip.Cend1 lentiviral vector, driving the expression of porcine Cend1 cDNA under the control of the CMV promoter, was used (Fig. 3A). The respective control vector was Trip.GFP driving expression of the GFP protein (Fig. 3A). Silencing of Cend1 was achieved via the Trip.shCend1 lentiviral vector, whereas as control, a Trip.shLuc vector was used (Fig. 3B). Transgene Cend1 expression was verified by semiquantitative reverse transcription (RT)-PCR using primers specific for both porcine and mouse Cend1-mRNA (Fig. 3C). Quantification of the intensity of the produced mRNA and protein bands relative to glyceraldehyde-3-phosphate dehydrogenase and β-tubulin, respectively, revealed a 2.7-fold increase of Cend1 mRNA and 2.9-fold increase of Cend1 protein levels, in Trip.Cend1-transduced compared with Trip.GFP-transduced neurospheres (Fig. 3D, 3E). Conversely, Cend1 silencing was verified by both real-time RT-PCR and Western blot, which revealed a 60% ± 4% and 85% decrease of Cend1 mRNA and protein levels, respectively, in Trip.shCend1-transduced compared with Trip.shLuc-transduced neurospheres (Fig. 3D, 3E). In addition, immunofluorescence staining and counting of the number of cells in which Cend1 was either overexpressed or silenced revealed that the transduction efficiency of the Trip.Cend1 vector was 85% ± 5%, and transduction efficiency of the Trip.shCend1 silencing vector was 74% ± 4% (Fig. 3F). Transduction of neurospheres with the corresponding control viral vectors (Trip.GFP and Trip.shLuc) did not cause any change in Cend1 expression levels, compared with nontransduced cells (data not shown).
Figure Figure 3.. Lentiviral vectors used for manipulating the expression levels of Cend1. (A): Representation of the proviral form of the pTrip vectors coding for Cend1 (upper schema) and GFP (lower schema) under the CMV promoter. (B): Representation of the proviral form of the TripSym vectors encoding the shRNA targeting Cend1 (upper schema) or Luciferase (lower schema). The shRNA-expressing cassettes are inserted into the U3 part of the LTR, and the internal part of the vector contains only the Ψ and RRE elements, with the central DNA FLAP. (C): Transgene Cend1 mRNA expression in neurosphere cultures 4 days after transduction determined by semiquantitative RT-PCR. Note that transgene Cend1 is highly expressed in cultures transduced with Trip.Cend1 compared with cultures transduced with Trip.GFP. GAPDH expression serves as control for loaded mRNA. (D): Transduction of neurospheres with the Trip.shCend1 lentiviral vector and determination of endogenous Cend1 mRNA levels by real-time RT-PCR 4 days after transduction revealed a 60% ± 4% downregulation of Cend1 mRNA in Trip.shCend1-transduced neurosphere cultures compared with Trip.shLuc-transduced cultures. (E): Western blot analysis of Cend1 protein expression in control neurospheres, as well as in neurospheres transduced with Trip.Cend1 or Trip.shCend1. β-Tubulin was used for estimation of loaded protein and normalization of Cend1 levels. (F): Immunofluorescence labeling of dissociated neurospheres for Cend1 protein, 4 days following lentiviral transduction with either Trip. Cend1 or Trip.shCend1 viral vector, showing the corresponding increase and decrease in the number of Cend1-overexpressing or -expressing cells, compared with control (nontransduced) neurosphere cultures. Measurement of total fluorescence intensity of Cend1 expression levels in all three conditions using Cell R̂ software (Olympus) indicated that upon Trip.Cend1 transduction, Cend1 levels rise 2.7 times compared with control, whereas following Trip.shCend1 transduction, Cend1 fluorescence has a twofold decrease compared with control. Abbreviations: CMV, cytomegalovirus; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GFP, green fluorescent protein; LTR, long terminal repeat; RT-PCR, reverse transcription-polymerase chain reaction; sh, short hairpin; RRE, RTA-responsive elements.
Download figure to PowerPoint
Figure Figure 4.. Effect of Cend1 overexpression and silencing on neurosphere proliferation. (A): Schematic representation of the experimental protocol used for neurosphere infection and culture. Third-passage postnatal day 5 SVZ neurospheres were transduced with Trip.Cend1 or the control Trip.GFP lentiviral vector. Three days after transduction, neurospheres were either left as free-floating spheres or dissociated to single cells and further grown for another 3 d, unless otherwise indicated, in the presence or absence of growth factors to assess proliferation or differentiation, respectively. The same strategy was used with the RNA interference set of lentiviral vectors, Trip.shCend1 and the control Trip.shLuc. (B–D): Effect of Cend1 overexpression on neurosphere size. Dissociated neurospheres were transduced and allowed to re-form floating spheres, and estimation of neurosphere size took place 4 d post-transduction. Note the smaller size of floating neurospheres upon Cend1 overexpression, compared with GFP-transduced ones, illustrated with fluorescence/phase contrast optics (B) and at a higher magnification with TOPRO-3 nuclear staining (C). (D): Estimation of neurosphere diameter indicates a 2.5-fold reduction in size in response to Trip.Cend1, compared with Trip.GFP transduction. *, p < .05 (n = 3). (E–J): Effect of Cend1 overexpression on neurosphere cycling properties. (E): Upon forced expression of Cend1, dissociated neurospheres cultured under proliferative conditions exhibit reduced BrdU incorporation compared with Trip.GFP-transduced ones, as revealed by double immunofluorescence labeling. (F): Quantification of BrdU+ cells following a 2-hour BrdU pulse in Cend1-overexpressing neurospheres, compared with control. Measurements were effectuated by dissociating the spheres and counting BrdU+ cells out of total DAPI+ nuclei. *, p < .05 (n = 4). (G): Real-time RT-PCR measurements indicate a 67% ± 5% reduction of Cyclin D1 mRNA levels upon Cend1 overexpression in neurosphere cultures. (H, I): Cend1 overexpression does not induce apoptosis, as indicated by immunohistochemistry for activated caspase-3 (H) and quantification of apoptotic profiles (I) that show 4% caspase-3+ cells in both Trip.Cend1- and Trip.GFP-transduced neurosphere populations (4% ± 0.5% for Trip.Cend1 vs. 3.9 ± 0.7 for Trip.GFP-transduced cells). (J): Western blot analysis also reveals a decrease, predominantly in Cyclin D1, but also B1 levels upon Cend1 overexpression. (J–M): Effect of Cend1 knock-down on neurospheres' cycling properties. (K): Silencing of Cend1 by Trip.shCend1 transduction enhances proliferation compared with control Trip.shLuc transduction, as revealed by Cend1/BrdU double immunostaining. (L): Counting of BrdU+ cells in Trip.shCend1-transduced versus control neurospheres indicates a 2.4-fold increase in BrdU incorporation (11.9% ± 0.8% BrdU index in Trip.sh.Cend1-transduced cultures vs. 4.8% ± 0.2% in Trip.shLuc-transduced cultures); p < .05 (n = 5). (M): Real-time RT-PCR measurements of Cyclin D1 mRNA in neurospheres transduced with Trip.shCend1 or Trip.shLuc showing a 53% ± 6.5% increase upon Cend1 knock-down. (J): Western blot analysis indicates elevated protein levels of Cyclin D1 and, to a lesser extent, B1 upon Cend1 silencing. Because of different total protein loading between the overexpression and silencing experiments, as revealed by β-tubulin levels, staining intensities for cycling protein bands are not comparable between the two control situations presented. Scale bars = 100 μm (B) and 40 μm (C, E, H, K). Abbreviations: bFGF, basic fibroblast growth factor; BrdU, 5-bromo-2′-deoxyuridine; d, days; DAPI, 4,6-diamidino-2-phenylindole; EGF, epidermal growth factor; GFP, green fluorescent protein; RT-PCR, reverse transcription-polymerase chain reaction; sh, short hairpin; SVZ, subventricular zone.
Download figure to PowerPoint
To determine the effect of Cend1 overexpression in cell proliferation, neurospheres were dissociated, lentivirally transduced at the single-cell level, and allowed to re-form floating spheres that were composed of progeny that also overexpressed Cend1. An important reduction in cell proliferation was observed in response to Cend1 overexpression (Fig. 4B–4D). Four days after transduction, neurospheres derived from Trip.Cend1-transduced cells had a statistically significant (p < .05) 2.5-fold reduction in sphere diameter compared with control neurospheres regrown from Trip.GFP-transduced cells (Fig. 4C, 4D). This size reduction was also obvious in secondary formed neurospheres overexpressing Cend1; however, no apparent changes were noted in the total number of secondary spheres generated (supplemental online Fig. 1). However, transduction at the single-cell level with the silencing Trip.shCend1 lentiviral vector resulted in formation of almost double the number (1.7-fold increase) of secondary-formed neurospheres compared with control (supplemental online Fig. 1), suggesting that Cend1 affects both neurosphere size and the number of neurospheres formed.
To further investigate the effect of Cend1 in cell proliferation, we estimated the 5-bromo-2′-deoxyuridine (BrdU) incorporation rate of cells infected in the sphere stage and then plated as single cells under proliferative conditions for 72 hours prior to analysis. Cells were given a 2-hour BrdU pulse before fixation, and estimation of the BrdU incorporation index out of the total number of 4,6-diamidino-2-phenylindole (DAPI)+ nuclei revealed a statistically significant reduction from 5.2% ± 0.5% in Trip.GFP-transduced cultures to 2.0% ± 0.6% in Trip.Cend1-transduced cultures (p < .05; Fig. 4E, 4F). No difference in apoptotic cell death was observed between Trip.Cend1-transduced and Trip.GFP cells as estimated by activated caspase-3 staining (Fig. 4H, 4I). These findings suggest that elevation of Cend1 expression induces a significant subpopulation of neurosphere cells to exit from the cell cycle and/or proliferate at a slower pace because of cell cycle elongation. Thus, we explored the levels of Cyclin D1, a protein playing fundamental role in cell cycle progression control. Real-time RT-PCR measurements indicated a 67% ± 5% decrease of Cyclin D1 mRNA levels in Trip.Cend1-transduced compared with Trip.GFP-transduced cultures (Fig. 4G). Since the levels of cell cycle mediators are regulated to a great extent post-transcriptionally, we also estimated the protein levels of Cyclins D1, B1, and E in Cend1-transduced neurospheres by Western blot. In accordance, a reduction of the protein levels of Cyclin D1 and, to a smaller extent, of Cyclin B1 was evident, whereas the protein levels of Cyclin E were not apparently affected (Fig. 4J).
To verify Cend1 participation in cell cycle progression/exit of proliferating precursors, we performed RNA interference experiments, using the lentiviral vectors Trip.shCend1 and Trip.shLuc as control (Fig. 3B). Silencing of Cend1 in proliferating neurospheres followed by dissociation and plating under proliferative conditions resulted in a statistically significant 2.4-fold increase in BrdU incorporation (Fig. 4K, 4L). Cend1 silencing also caused a 53% ± 6.5% increase in Cyclin D1 mRNA (Fig. 4M) and protein (Fig. 4J) levels compared with Trip.shLuc-infected cultures, whereas the protein levels of Cyclin E were not affected (Fig. 4J). To ensure that the effect of Cend1 on cell cycle progression observed in dissociated cells was an instructive rather than a selective postdissociation process because of elimination of apoptotic cells, we performed clonal analysis on whole-mount neurospheres transduced with either the overexpressing or the silencing Cend1 lentiviral vectors. First, we checked for apoptotic cells by immunostaining for activated caspase-3 and found that in both cases apoptosis levels were identical to those of the respective control conditions (supplemental online Fig. 2A, 2C). Then we assessed proliferation using antibodies to phosphorylated histone-3 and Ki67 proteins (supplemental online Fig. 2B, 2C). Both gain- and loss-of-function approaches in whole-mount neurospheres concluded that Cend1 overexpression results in reduced proliferative capacity, whereas its downregulation yields the opposite effect (supplemental online Fig. 2B, 2C). Taken together these findings indicate that Cend1 forms part of the machinery controlling the cell cycle properties of postnatal SVZ-derived NSCs.
Cend1 Promotes Neuronal Differentiation of Neonatal SVZ-Derived Neuronal Precursors
Cell cycle progression and differentiation are two tightly linked cellular events. We therefore sought to investigate Cend1 implication in the differentiation program of SVZ-derived precursors. To this end, lentivirally transduced neurospheres derived from P5 SVZ, either overexpressing or downregulating Cend1, were analyzed in relation to their fate choice. Neurospheres were infected with the corresponding lentiviral vectors while in the free-floating stage, left for 3 days, and then cultured for another 3 days in the absence of growth factors, either as whole-mount preparations or as dissociated cells. Immunohistochemical staining of whole-mount differentiating neurospheres revealed that, qualitatively upon forced Cend1 expression, neuronal markers such as doublecortin and βIII-tubulin were highly upregulated, whereas astrocytic markers such as GFAP were downregulated compared with control spheres (Fig. 5A). To quantify the number of cells expressing these molecular markers in Cend1- and GFP-transduced cultures, neurospheres were dissociated to single cells and further cultured for 3 days. Analysis of immunocytochemically stained cells revealed a statistically significant 68.7% ± 4% increase in doublecortin+ cells, a 100% ± 0.6% increase in βIII-tubulin+ cells, and a 32% ± 5% decrease in GFAP+ cells (Fig. 5B). There was no statistically significant decrease in O4 levels, suggesting that Cend1 is not involved in oligodendrocyte lineage progression (Fig. 5B). Accordingly, Cend1 knockdown had an opposite effect in the differentiation pathway of precursor cells (Fig. 5C). When Cend1 was silenced, neurospheres presented a less differentiated neuronal phenotype, as judged by the reduced expression of doublecortin and βIII-tubulin in whole-mount neurospheres (Fig. 5C). Quantification of this effect in dissociated neurospheres showed a statistically significant 44% ± 2.5% decrease in doublecortin+ cells, a 45.4% ± 0.3% decrease in βIII-tubulin+ neurons, and only a 13.9% ± 3.5% increase in GFAP-expressing astrocytes, whereas there was no statistically significant change in the percentages of O4+ cells in Trip.shCend1-transduced compared with Trip.shLuc-transduced neurospheres (Fig. 5D). Consistently, an increase and a corresponding decrease in βIII-tubulin+ neurons were estimated upon Cend1 overexpression and downregulation, respectively, when clonal analysis was performed in whole-mount neurospheres. Thus, Cend1 overexpression resulted in 26 ± 5 βIII-tubulin+ cells per sphere, compared with 15 ± 2 βIII-tubulin+ cells in control spheres and 8 ± 2 βIII-tubulin+ cells in Trip.shCend1-transduced spheres (Fig. 5Ae, 5Af, 5Ce, 5Cf; supplemental online Fig. 3). Apoptosis levels were identical in all situations (data not shown).
Figure Figure 5.. Qualitative and quantitative molecular phenotype analysis of neurosphere differentiation upon Cend1 overexpression and silencing. (A): Confocal images of differentiating neurospheres transduced with Trip.GFP or Trip.Cend1 and stained as whole-mount preparations for Cend1 (green), the indicated lineage (red), and TOPRO-3 (blue). Cend1 overexpression results in a qualitative increase of doublecortin+ and βIII-tubulin+ cells and a decrease of GFAP+ cells. (B): Graphs presenting the quantitative phenotypic changes upon Cend1 overexpression. In dissociated neurospheres, the percentage of DCX and βIII-tubulin+ cells is increased, whereas the percentage of GFAP+ cells is decreased. *, p < .05 (n = 4). The difference in the percentage of O4+ oligodendrocytes is not statistically significant. Measurements represent the percentage of lineage-marker+ cells out of DAPI+ nuclei in neurospheres dissociated to single cells and grown for 3 days in the absence of growth factors. (C): Confocal images of differentiating neurospheres transduced with Trip.shLuc or Trip.shCend1 and stained as whole-mount preparations for Cend1 (green), the indicated lineage markers (red), and TOPRO-3 (blue). Cend1 knockdown enhances astrocytic differentiation and reduces the neuronal phenotype. (D): Graphs presenting the quantitative phenotypic changes upon Cend1 silencing. Neuronal markers (DCX and βIII-tubulin) decrease, GFAP+ cells slightly increase, and O4+ cells are not affected. Note the opposite effect of Cend1 silencing on neurosphere molecular phenotype, compared with that obtained in the overexpression experiments. Measurements were performed as in (B). *, p < .05 (n = 4). Staining intensities for Cend1 are not comparable between the two control situations presented in (A) and (C). Scale bars = 40 μm (Cc, Cd, Ce, Cf, Cg, Ch) and 20 μm (Ca, Cb). Abbreviations: DCX, doublecortin+; GFAP, glial fibrillary acidic protein; GFP, green fluorescent protein; sh, short hairpin.
Download figure to PowerPoint
To investigate whether Cend1 forms part of a cell cycle exit/neuronal differentiation pathway, we investigated whether its overexpression in neurosphere cultures results in changes in the expression levels of Notch1 and its downstream effector Hes1. Western blot analysis revealed that Cend1-transduced neurospheres exhibit lower protein amounts of both Notch1 and Hes1 compared with GFP-transduced neurospheres (Fig. 6A). We next asked whether endogenous Cend1 expression is regulated by transcription factors inducing neurogenesis, such as basic helix-loop-helix (bHLH) proneural genes. Of these, we chose to examine the effect of Ngn1, since this proneural gene has previously been shown to transactivate the minimal Cend1 promoter in neural cell lines in vitro . In agreement, real-time RT-PCR measurements revealed that Ngn1 overexpression in the neurosphere system caused a 35% increase in Cend1 mRNA levels (Fig. 6B). To validate the Ngn1 effect on Cend1 expression levels, we checked the ability of Ngn1 to activate the Cend1 promoter directly. Computer-assisted analysis identified a consensus E-box sequence on the proximal Cend1 promoter, representing a putative bHLH protein binding site (Fig. 6C). Transactivation experiments were performed by cotransfecting the Ngn1 expression vector with either the E-box containing wild-type pGL(−283/+63) vector for expression of the 346-bp Cend1 minimal promoter, or the same vector carrying four point mutations at the consensus E-box (Fig. 6C). Luciferase reporter assays in both the N2A neuroblastoma cell line and in neurosphere cultures demonstrated a threefold increase in wild-type Cend1 promoter activity upon Ngn1 overexpression, which is abolished following mutation of the E-box (Fig. 6D). This indicates that Cend1 forms part of the neuronal differentiation pathway activated by proneural genes in the adult.
Figure Figure 6.. Cend1 participates in neuronal differentiation pathways. (A): Western blot analysis of neurospheres 4 days following transduction with the Trip.GFP and Trip.Cend1 lentiviral vectors reveals a 35% decrease of Notch1 and a 21% decrease of Hes1 protein levels upon Cend1 overexpression. (B): Real-time RT-PCR measurements of Cend1 mRNA in neurospheres transfected with the Neurogenin-1 (Ngn1) expression vector reveal a 35% increase in Cend1 mRNA levels. (C): Diagram illustrating the construct used for transactivation assays, in which a 346-base pair fragment of the human Cend1 proximal promoter is linked to the Firefly Luciferase gene, as well as the four point mutations made in the CAGATG E-box consensus sequence of the Cend1 promoter (−127/−132). (D): Measurement of Luciferase activity in Neuro 2a cells and neurospheres 2 days after cotransfection using the Ngn1 expression vector together with either the Cend1 wt or E-box-mutated promoter constructs indicates a threefold increase of Cend1 promoter activity in the presence of Ngn1 compared with wt promoter only. This increase in Luciferase activity is reduced to basal levels when Ngn1 is cotransfected with the mutated E-box Cend1 promoter. *, p < .05 (n = 3). Abbreviations: A, AvaI; B, BamHI; GFP, green fluorescent protein; Luc, Luciferase; mut, mutated; N, NcoI; P, PstI; RT-PCR, reverse transcription-polymerase chain reaction; S, SmaI; wt, wild-type.
Download figure to PowerPoint
Cend1 Affects Chain Formation in Explant Cultures Derived from Neonatal SVZ
To investigate whether the impact of Cend1 overexpression or silencing on cell cycle exit has also a potential effect on neuronal migration, we used P5 SVZ-explant cultures that mimic in vivo chain migration. Wild-type and Trip.GFP-transduced explants presented the normal pattern of chains migrating radially out of the explant, whereas following Trip.Cend1 lentiviral transduction, explants manifested a reduction in the number of chains migrating out of the explant core (Fig. 7A). Estimation of the number of DAPI+ nuclei lying out of the explant core in Trip.GFP and Trip.Cend1-transduced explants revealed a statistically significant 45% ± 7% decrease in response to Cend1 compared with GFP (Fig. 7B, 7C). In addition, the average chain length in Trip.Cend1-transduced explants was shorter by 21% ± 2% compared with wild-type or Trip.GFP-transduced explants (Fig. 7D). To try to distinguish whether reduced chain numbers are due to an effect on migration or proliferation, explants were given a 2-hour BrdU pulse before BrdU/Cend1 double immunofluorescence labeling (Fig. 7E). Quantification of BrdU+ cells per explant revealed a statistically significant threefold decrease in the proliferation rate of explants in response to Cend1 overexpression compared with control explants (Fig. 7E, 7F). Conversely, a statistically significant 2.3-fold increase in BrdU incorporation was evident as a consequence of Cend1 silencing (Fig. 7E, 7G). Thus, in explants, as in neurosphere cultures, Cend1 expression levels interfere with cell proliferation, thus contributing to the number of cells generated and consequently the number of chains formed. However, a migration defect per se cannot be ruled out as, in addition to the antiproliferative effect, a reduction in chain length was also noted in response to Cend1 overexpression.
Figure Figure 7.. Effect of Cend1 overexpression and silencing on postnatal day 5 subventricular zone explants. (A): Phase-contrast images of WT and Trip.GFP-transduced explants exhibit similar morphology, whereas Trip.Cend1-transduced explants exhibit reduced number of migrating chains. (B, C): 4,6-Diamidino-2-phenylindole (DAPI) nuclear staining of control and Cend1-transduced explants (B) and quantification of DAPI+ nuclei (C) indicate that the number of cells migrating outside the explant core is reduced in response to Cend1 overexpression. **, p < .01 (n = 5). (D): Estimation of the average chain length in WT, Trip.GFP-transduced, and Trip.Cend1-transduced explants reveals a statistically significant 21% ± 2% decrease in average chain length upon Cend1 overexpression. * p < .05 (n = 4). (E): Proliferation study following a 2-hour BrdU pulse in control (Trip.shLuc), Cend1-overexpressing (Trip.Cend1), and Cend1-silenced (Trip.shCend1) explants. Cend1 overexpression is correlated with a decrease in BrdU incorporation, whereas the opposite is observed upon Cend1 knockdown compared with the control situation. The Trip.GFP vector produced BrdU incorporation rates identical to those of the Trip.shLuc vector. (F, G): Graphs representing the changes in BrdU incorporation upon Cend1 overexpression (F) and downregulation (G). **, p < .01 (n = 5). Scale bars = 40 μm (A, B, F). Abbreviations: BrdU, 5-bromo-2′-deoxyuridine; GFP, green fluorescent protein; sh, short hairpin; WT, wild-type.
Download figure to PowerPoint