Ectopic Expression of Nolz-1 in Neural Progenitors Promotes Cell Cycle Exit/Premature Neuronal Differentiation Accompanying with Abnormal Apoptosis in the Developing Mouse Telencephalon

Nolz-1, as a murine member of the NET zinc-finger protein family, is expressed in post-mitotic differentiating neurons of striatum during development. To explore the function of Nolz-1 in regulating the neurogenesis of forebrain, we studied the effects of ectopic expression of Nolz-1 in neural progenitors. We generated the Cre-loxP dependent conditional transgenic mice in which Nolz-1 was ectopically expressed in proliferative neural progenitors. Ectopic expression of Nolz-1 in neural progenitors by intercrossing the Nolz-1 conditional transgenic mice with the nestin-Cre mice resulted in hypoplasia of telencephalon in double transgenic mice. Decreased proliferation of neural progenitor cells were found in the telencephalon, as evidenced by the reduction of BrdU−, Ki67− and phospho-histone 3-positive cells in E11.5–12.5 germinal zone of telencephalon. Transgenic Nolz-1 also promoted cell cycle exit and as a consequence might facilitate premature differentiation of progenitors, because TuJ1-positive neurons were ectopically found in the ventricular zone and there was a general increase of TuJ1 immunoreactivity in the telencephalon. Moreover, clusters of strong TuJ1-expressing neurons were present in E12.5 germinal zone. Some of these strong TuJ1-positive clusters, however, contained apoptotic condensed DNA, suggesting that inappropriate premature differentiation may lead to abnormal apoptosis in some progenitor cells. Consistent with the transgenic mouse analysis in vivo, similar effects of Nozl-1 over-expression in induction of apoptosis, inhibition of cell proliferation and promotion of neuronal differentiation were also observed in three different N18, ST14A and N2A neural cell lines in vitro. Taken together, our study indicates that ectopic expression of Nolz-1 in neural progenitors promotes cell cycle exit/premature neuronal differentiation and induces abnormal apoptosis in the developing telencephalon.


Introduction
The nocA like C 2 H 2 zinc-finger gene-1 (Nolz-1), which is also known as Zfp503/Znf503 (Mouse Genome Informatics), is previously identified as a murine member of the noc/Nlz-elbowtlp-1 (NET) zinc-finger protein family [1]. The NET gene family has been shown to involve in control of a variety of developmental events. The C. elegans homologue Tlp1 regulates asymmetric cell fates of T blast cells [2]. The Drosophila homologues elB and noc specify the identity of dorsal-ventral branches of trachea [3]. The zebrafish homologue Nlz1 and Nlz2 regulate the rhombomere identity in developing hindbrain and the closure of optic fissure in zebrafish eye [4][5][6][7][8]. The chick homologue Nolz-1 controls the subtype identity of motor neurons in developing spinal cord [9].
Recent study shows that mouse Nolz-1 regulates the neurogenesis in neurosphere culture in vitro [10].
Nolz-1 expression is developmentally regulated in different mouse organs during embryogenesis [9,11,12]. In the developing telencephalon, Nolz-1 is preferentially expressed at high levels in the lateral ganglionic eminence (LGE, the primordium of striatum), and Nolz-1 expression is gradually down-regulated in the striatum after birth [10,11]. During striatal neurogenesis, neural progenitor cells reside in the ventricular zone (VZ) and the subventricular zone (SVZ). Upon differentiation, post-mitotic differentiating neurons migrate from the SVZ into the mantle zone (MZ) where differentiating neurons undergo terminal differentiation [13]. Therefore, unlike the VZ which contains proliferative neural progenitors, the SVZ contains both prolifer-ative progenitors as well as early differentiating post-mitotic neurons [13]. It is of interest that Nolz-1 is expressed at high levels in the SVZ of LGE, but not in the VZ. Nolz-1 expression level is lower in the differentiated MZ. Moreover, Nolz-1 is not expressed in Ki67-positive proliferating progenitor cells but is co-expressed with the early neuronal differentiation markers of TuJ1 and Isl-1 indicating that Nolz-1 is expressed in early post-mitotic striatal neurons [10,11,14]. Nolz-1 therefore serves as a developmental marker for differentiating striatal neurons.
Because Nolz-1 is not expressed by proliferative neural progenitors in the germinal zone [11], we investigated the effects of ectopic expression of Nolz-1 in neural progenitors. We generated the Nolz-1 conditional transgenic mice using the Cre/ loxP-mediated DNA recombination technology [15]. Ectopic expression of Nolz-1 in neural progenitors was achieved by intercrossing the Nolz-1 conditional transgenic mice with the nestin-Cre driver mice [16,17]. Telencephalic hypoplasia was found in the resulting double transgenic mice. Further examination showed that transgenic expression of Nolz-1 led to decreased proliferation and precocious differentiation of progenitor cells in developing telencephalon. Moreover, aberrant differentiation induced by transgenic Nolz-1 appeared to result in abnormal cell death. These findings suggest a role of Nolz-1 in promoting neuronal differentiation during neurogenesis in developing telencephalon.

Generation of Transgenic Mice
The transgenic plasmid of pLacZ-Nolz-1-eGFP (pZNG), for generating LacZ floxed Nolz-1 eGFP transgenic mice, was constructed using the pCCALL2 plasmid (Kindly provided by Dr. A. Nagy, University of Toronto) as backbone [18]. The coding sequence of C-terminal Flag-tagged Nolz-1 was generated by PCR, confirmed by DNA sequencing and then ligated with the downstream IRES-eGFP-polyA cassette which was composed by three DNA fragments released from pNTR-LacZPGKneolox, pCAGGS-eGFP and PGK-puro-lox2a plasmids (kindly provided by Dr. T.-F. Tsai, National Yung-Ming University), respectively. A direct repeat of a 1.2-kb fragment containing the DNase I-hypersensitivesite 4 of the chicken b-globin locus (HS4) fragment was released from SalI and XbaI sites of pJC13-1-59HS4 (kindly provided by Dr. T.-F. Tsai), and was then ligated downstream of the IRES-eGFP-polyA cassette in a reverse orientation to relieve position silence effects at transgene [19,20]. The transgenic fragment was released by ScaI and XhoI sites for C57/BL6 pronuclei microinjection (Level Biotechnology, Taipei, Taiwan). Potential founder mice were screened by PCR using tail DNA as template and HS4 specific primers (HS4-51:59-CCTCCTTGGGCAACCTGTTCAG-39; HS4-31:59-ATGTGGCACTGAGGGACATGGC-39). The PCR-positive transgenic lines were further confirmed by Southern blotting. Tail genomic DNAs from LacZ floxed Nolz-1 eGFP F1 offspring was digested with EcoRI, and the DNA probe containing the eGFP coding region was used in Southern blotting.

Ethics Statement
The transgenic mice were maintained in the Animal Centers of National Yang-Ming University (NYMU) and China Medical University (CMU). The animal protocols were approved by the Animal Care and Use Committees of NYMU and CMU.

In situ Hybridization and Immunostaining
In situ hybridization and immunostaining were performed as previously described [11,20,29]. The RNA probes included eGFP (coding region) and Nolz-1 (NM145459:2640-3744). The primary antibodies included rat-anti-BrdU (1:1,000, Accurate Chemical & Scientific, New York), sheep anti-BrdU (1:2,000, Biodesign International, Saco, ME), rabbit anti-phospho-histone 3 (PH3) (1:1,000, Millipore, Billerica, MA), rabbit anti-Ki67 (1:500, Novocastra, Newcastle, UK), rabbit anti-cleaved caspase-3 (1:500, Cell Signaling Technology, Danvers, MA) and rabbit anti-TuJ1 (1:1,000, Covance, Emeryville, CA). The Elite ABC kit (Vector Laboratories, Burlingame, CA) was used after biotinconjugated secondary antibodies for signal amplification. For fluorescence immunostaining, avidin-Texas red (1:1,000) and avidin-FITC (1:1,000) were used for signal detection. For the cell cycle exit assay, sequential double immunostaining of BrdU and Ki67 was performed as previously described [30] with the following modifications: the sections were incubated with the antigen-retrieval solution containing 10 mM citric acid (pH 6.0) at 95uC for 10 min followed by 1 N HCl treatment at 45uC for 15 min. Rat anti-BrdU and mouse anti-Ki67 antibodies, biotinylated rabbit anti-rat and goat anti-mouse antibodies were used with the Elite ABC procedure followed by TSA-FITC (for detecting BrdU) and TSA-Cy3 (for detecting Ki67). Avdin-biotin blocking was performed after the first BrdU immunostaining by incubating the sections with solution A (1:50, Vector laboratories) for 30 min and solution B (1:50) for 30 min followed by 1% H 2 O 2 treatment for 30 min. For triple immunostaining, the sections were sequentially immunostained for Ki67, TuJ1 and BrdU. The immunostaining of Ki67 and BrdU was as described above, and rabbit anti-TuJ1 antibody and goat ant-rabbit antibody conjugated with Alexa649 were used to detect TuJ1 signals in the triple immunostaining. The sections were counterstained with DAPI (1:10,000). For some sections, after Ki67 and TuJ1 immunostaining, the brain sections were then processed for TUNEL staining using the In Situ Cell Death Detection kit (Roche, Indianapolis, IN). Before the TUNEL reaction mixture application, the slides were treated with 0.1 M glycine in PBS buffer for 30 min and then incubated with 0.1% Triton X-100 in 0.1% sodium citrate buffer at 4uC for 2 min.

Image Analyses, Quantification and Statistics
Microscopic analysis was performed using Nikon E800M, Olympus BX63 and Leica DM6000B microscopes. Photomicrographs of X-gal stained embryos and dissected transgenic brains were taken by Nikon SMZ-1000 and NIS-element D3.0 software. Contrast and brightness adjustments for figures were made, and final plates were composed by using Adobe Photoshop CS (Adobe Systems, San Jose, CA). Forebrain sections, in which the borders of LGE, MGE and cortex were easy identified by morphology, were selected for quantification. Paired E12.5 brain sections of nCT and control T embryos from three independent littermates were analyzed for single PH3 and BrdU immunostaining. The numbers of PH3-and BrdU-positive cells in ventricular zone on left and right sides of brains were both calculated. Cell counting was performed under microscopy. After normalized with the lengths of ventricular zone or area of germinal zones (which were quantified by Image J software), the quantified data of PH3-and BrdU-positive cells in nCT brains were compared to that of control T littermate (PH3 cell counts) or T and C littermates (BrdU cell counts). For the cell cycle exit assay, telencephalic brain sections of 3-5 pairs of nCT, nC, T and WT embryos from four litters were analyzed. The numbers of BrdU+/Ki672 and total BrdU+ cells within 2 or 3 squares (50 mm650 mm) of the cortex, LGE and MGE were counted in the comparable brain sections from each genotype. Student's t test was used for statistical analysis.

Generation of the LacZ floxed -Nolz-1 eGFP Transgenic Mice
We generated the transgenic mice carrying Cre-dependent Nolz-1 allele. The transgenic construct contained a CMV enhancer/bactin (CAG) promoter followed by the loxP-flanked LacZ-stop-Neo (LacZ-Neo) cassette which contained the b-galactosidase/ neomycin resistance fusion gene and a triple polyA transcription termination (Stop) signal. Downstream of this cassette was the expression cassette of Nolz-1-Flag-IRES-eGFP-HS4, i.e., cDNAs encoding Flag-tagged Nolz-1, internal ribosomal entry site (IRES), enhanced green fluorescent protein (eGFP) reporter gene and two repeats of HS4 insulators [31] (Fig. 1A). The presence of an upstream ''Stop'' signal allowed the expression of b-galactosidase, but not the downstream Nolz-1 and eGFP. Upon the Cremediated DNA recombination, the loxP-flanked LacZ-Neo cassette was removed to permit the expression of Nolz-1 and eGFP by the CAG promoter (Fig. 1A).
Before performing the pronuclei microinjection in mice, we tested the expression of transgenic construct in HEK293T cells. Successful expressions of LacZ, Cre-mediated transgenic Nolz-1 (tNolz-1) and eGFP were confirmed in transfected HEK293T cells ( Fig. 1B-E).
The transgenic construct was microinjected into the female pronucleus of C57/BL6 mice for generating the LacZ floxed -Nolz-1 eGFP transgenic (Nolz-1 Tg) mice. The CAG promoter directed the transgenes expression presumably ubiquitously in the developing embryo. Among 23 lines of Nolz-1 Tg mice screened, #1, #3 and #9 were germ-line transmitted to F1 offspring. However, no transgenic F2 offspring was born from Nolz-1 Tg line #1. The X-gal staining analysis showed that the spatiotemporal expression pattern of LacZ in Nolz-1 Tg founder line #3 and #9 were different. LacZ expression was detected as early as embryonic day (E) 7.5 in Nolz-1 Tg line #9 (Fig. S1A, B). In contrast, LacZ expression was not observed until E13.5 in Nolz-1 Tg line #3 (data not shown). Although LacZ expression was detected in differentiated neurons in both line #3 and #9 Tg brains, LacZ expression was only detected in neural progenitor cells of the germinal zone in the brains of line #9 (data not shown).

Hypoplasia of Telencephalon in the nCT Double Transgenic Mice
The analysis of the embryonic and neonatal nCT brains of line #9 showed that the head size was dramatically reduced in the nCT double transgenic mice (Fig. 4). Apparent hypoplasia of the telencephalic region was observed in E12.5 nCT embryo (Fig. 4A). A prominent hypoplasia of the nCT brain was also found one day after birth (Fig. 4B). The overall structure of the nCT brain, including the forebrain, midbrain and hindbrain, was not significantly changed, though the brain size was reduced compared to that of the control littermates (Fig. 4B).

Reduction of Proliferative Neural Progenitor Cells in the nCT Brain
The reduction of brain size might be resulted from decreased proliferation of neural progenitors and/or abnormal cell death. As early as E11.5-E12.5, reduced structures, including the cortical primordia, the LGE and the medial ganglionic eminence (MGE) accompanying with the enlarged ventricle were evident in the nCT brain (Fig. 5). Pulse labeling of proliferating cells with bromodeoxyuridine (BrdU) for 1 hr demonstrated a significant reduction of BrdU-positive cells in the germinal zones of LGE and MGE in E11.5 and E12.5 nCT forebrains (Fig. 5A1, A2, B1, B2, C1, C2). BrdU-positive cells were reduced by 34% and 48% in the LGE and MGE of E12.5 nCT forebrain, respectively, compared to control littermates (Fig. 5D). Consistent with the results of BrdU pulse-labeled cells, immunostaining of Ki67, a marker for proliferating cells, also showed significant reduction of Ki67positive cells in E12.5 nCT forebrains (Fig. 5E1, E2, E1', E2'). Of

Induction of Abnormal Programmed Cell Death in the nCT Brain
To further test if abnormal programmed cell death occurred in the nCT brain, we performed the terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay. TUNELpositive apoptotic cells were increased in E12.5 nCT brains (Fig. 6A1, A2). Immunostaining of the activated caspase-3 (AC3) also showed increases of AC3-positive cells in E11.5 nCT brains (Fig. 6B1, B2). The abnormal cell death was found as early as E11.5, peaked at E12.5 and only a few AC3-positive cells were present in E15.5 nCT brains (Fig. S2). These findings indicated that ectopic expression of tNolz-1 induced abnormal programmed cell death in neural progenitor cells. It was of particular interest that the majority of TUNLE-positive apoptotic cells were found in the SVZ, the region between the VZ and the MZ (Fig. 6A1, A2). Notably, distinct clusters of DAPI-positive condensed nuclei were present in the SVZ, and these clusters were positive for TUNLE signals (Fig. 6C1-C3), indicating that apoptotic cells were inclined to aggregate themselves to form apoptotic clusters.

Reduction of Mitotic Neural Progenitors in the VZ of nCT Brain
An interesting question was whether ectopic expression of tNolz-1 in neural progenitors inhibited their proliferation. The concurrence of decreased proliferation and increased cell death in the SVZ of nCT brain made it difficult to resolve this issue. We then analyzed the mitotic cells in the VZ of nCT brain in which few TUNEL-positive apoptotic cells were found (Fig. 6C1-C3). Immunostaining of phospho-histone H3 (PH3), a mitotic marker, showed that PH3-positive cells undergoing the mitosis phase of cell cycle progression were lining along the wall of VZ (Fig. 7). PH3positive cells were significantly decreased in the VZ of E10.5 subpallium, E11.5 and E12. 5 LGE and MGE of nCT brains compared to the control brains ( Fig. 7A1-C2). The quantitative analysis showed that PH3-positive cells in the VZ of cortical primordium, LGE and MGE in E12.5 nCT brains were decreased to 6464.2%, 7863.2% and 7265.6% of the T control brains, respectively (Fig. 7D). Consistent with these findings, Ki67-positive proliferating cells lining along the surface of VZ were also decreased in the nCT brain (Fig. 6A1, A2). These results suggest that ectopic expression of tNolz-1 in neural progenitors may inhibit proliferation of progenitor cells.

Promotion of Cell Cycle Exit by tNolz-1 in the nCT Telencephalon
To further clarify the mechanism underlying tNolz-1 inhibited proliferation of neural progenitor cells, we performed the cell cycle exit assay. Considering the severe programmed cell death in nCT brains, the BrdU-pulsing period was shortened to 4 hours. Double labeling of BrdU and Ki67 showed that many BrdU+/Ki672 cells were present in the germinal zone of cortex, LGE and MGE of nCT brains (arrows, Fig. 8C, F, I), compared to few BrdU+/ Ki672 cells in the littermates control brains (Fig. 8A, B, D, E, G,  H). The quantification analysis showed that the cell cycle exit index of the cortex, LGE and MGE of nCT brain were 8.7%, 8.5 and 10.6%, respectively; whereas the cell cycle exit index in the littermates control T, nC and WT brains were 0?0.5% (Table 1). Note that some BrdU+/Ki672 cells appeared to line up along a radial line in the germinal zone of LGE as if they were derived from a clone of progenitors. Notably, similar Ki672negative radial-orientated domains were aforementioned (arrows, Fig. 5E2). Taken together with the finding of tNolz-1 and eGFP mRNAexpressing cells along radial lines in the LGE of E12.5 nCT brain (Fig. 3C, F) these results suggest that tNolz-1 over-expression in progenitor cells promotes cell cycle exit in the germinal zone of nCT telencephalon.

Precocious Neuronal Differentiation in the nCT Brain
Because endogenous Nozl-1 was expressed in differentiating neurons, another interesting question was whether ectopic expression of tNolz-1 in neural progenitors could promote neuronal differentiation. In the aforementioned cell cycle exit assay, we also performed triple immunostaining of BrdU, Ki67 and the neuron-specific class III beta-tubulin (TuJ1), a marker for early differentiating neurons [32]. Interestingly, a few cells coexpressing BrdU and TuJ1 without Ki67 that represented precociously differentiating neurons prematurely exiting cell cycle were found in the germinal zone of nCT telencephalon. (Fig. 8J-R).
Despite the abnormal cell death in the nCT brain, TuJ1 immunostaining showed an apparently general increase of TuJ1 immunoreactivity in E12.5 nCT brain compared to that of the control brain (Fig. 9A1-D2). The Western blotting further confirmed increases of TuJ1 expression in E12.5-E13.5 nCT cortical primordia compared to the control littermates (12364.5%, p,0.05; Fig. 9G).
TuJ1 immunostaining also showed that TuJ1-positive MZ appeared to expand toward the germinal zone (Fig. 9C1, C2, D1, D2). Scattered TuJ1-positive neurons were ectopically found in the VZ of nCT telencephalon, but not in the control telencephalon (Fig. 9D1, D2). Moreover, cell clusters with enhanced TuJ1 immunostaining were present in the germinal zone of nCT brain (Fig. 9E, F). These strong TuJ1-positive clusters contained condensed DAPI-positive nuclei, which was the sign of initiating apoptotic process (Fig. 9F').
Indeed, double-labeling of TuJ1 and TUNEL confirmed colocalization of TuJ1 and TUNEL signals in some ectopic TuJ1positive cells in the germinal zone of nCT forebrain (arrowhead, Fig. 10), though a few other TuJ1-positive germinal cells were negative for TUNEL signals (arrow, Fig. 10D). Interestingly, in some apoptotic cell clusters which appeared to line up as a radial line, tNolz-1 mRNA expression was detected as condensed particles (Fig. 10F, G; arrows). Taken together, these results suggest that inappropriate precocious differentiation triggered by tNolz1 in some neural progenitor cells may lead to abnormal apoptosis in the nCT brain.

Reduction of Cell Proliferation and Promotion of Neuronal Differentiation by Nolz-1 Over-expression in N2A Neural Cells
In addition to abnormal apoptosis, Nolz-1 over-expression also resulted in reduction of PCNA, a proliferating cell marker, to 76.866.4% of the mock control in N2A cells (n = 3, p,0.05; Fig. 11L). Interestingly, accompanying with the decrease of cell proliferation, many Nolz-1-transfected N2A cells morphologically differentiated into TuJ1-positive neurite-bearing neurons (Fig. 11M-Q). Significant numbers of Nolz-1 transfected cells extended longer TuJ1-positive neurites compared to the mock control. The quantitative analysis indicated that Nolz-1 overexpression induced a 15-fold increase of the number of neuritebearing neurons compared to mock control (15.461.38, n = 5, p,0.001; Fig. 11R).
Taken together with these results, it indicated that Nolz-1 overexpression induced abnormal apoptosis, suppressed cell proliferation and promoted neuronal differentiation in neural cell culture. These findings in cell culture in vitro are in good agreement with the findings in Nolz-1 Tg brains in vivo.

Discussion
In the present study, we investigated the function of Nolz-1 in developing telencephalon using conditional transgenic mouse approach. Because Nolz-1 is not expressed by proliferative WT: wild type brain; nC: nestin-Cre brain; T: LacZ floxed -Nolz-1 eGFP brain; nCT: nestin-Cre;LacZ floxed -Nolz-1 eGFP brain; LGE, lateral ganglionic eminence; MGE: medial ganglionic eminence. doi:10.1371/journal.pone.0074975.t001 neural progenitors in the germinal zone, we studied the effects of ectopic expression of Nolz-1 in neural progenitors using the nestin-Cre driver mice. Nestin-Cre dependent expression of tNolz-1 in neural progenitors of forebrain resulted in decreases of neural progenitors. There are two possibilities to account for the reduction of neural progenitors. These two possibilities are not exclusive to each other. One is that ectopic expression of tNolz-1 induces apoptosis of progenitor cells. Indeed, TUNELpositive and activated caspase-3-positive apoptotic cells were found in the germinal zone of the nCT brain. The other possibility is that ectopic expression of tNolz-1 promotes precocious neuronal differentiation, which in turn leads to depletion of progenitor pools. Consistent with this possibility, promotion of cell cycle exit and enhanced ectopic TuJ1 expression was found in the germinal zone of nCT brain. Notably, apoptotic cell clusters containing strong TuJ1-positive signals were found in the germinal zone of SVZ, suggesting that inappropriate premature differentiation of progenitors may cause abnormal cell death in some progenitor cells. In other progenitors, tNolz-1-promoted differentiating cells may survive, because there was a general increase of TuJ1 immunoreactivity in the nCT brain. The transgenic expression of Nolz-1 was driven by the CAG promoter that is likely to have different levels of activity in neural progenitor populations. It is plausible that the expression level of tNolz-1 may be varied in neural progenitors of the nCT brain. A Figure 9. Ectopic and precocious TuJ1 expression in the proliferative ventricular zone of E12.5 nCT brain. A1-D2: TuJ1 immunoreactivity was higher in the nCT forebrain (A2, B2, C2) than that in the control forebrain (A1, B1, D1). Ectopic and precocious TuJ1-positive neurons were found in the TuJ1-poor ventricular zone of the cerebral cortex (C2, arrows) and LGE (D2, arrows). The bracketed regions in D1 and D2 are shown at high magnification in the insets of D1 and D2, respectively. E: TuJ1-positive cell clusters are present in the germinal zone of the nCT brain. The bracketed region in E is shown at high magnification in F with confocal images. F, F': TuJ1-positive cell clusters contain apoptotic nuclei with condensed DAPI staining. The arrow indicates a TuJ1-positive neuron (F) whose nucleus was presumably under the process of DNA condensation as implicated by the condensed DAPI staining (F'). The arrowhead indicates a presumed degrading cell with condensed TuJ1 immunoreactivity without apparent DAPI staining. G: Western blotting and quantification of TuJ1 protein expression in E12.5-13.5 cortex. TuJ1 expression in nCT cortex is 12364.5% of control littermates (p,0.05). Scale bars in A1 (for A1, A2, B1, B2), 0. high level of tNolz-1 may promote aberrant differentiation which eventually causes cell death, whereas a low level of tNolz-1 may promote neuronal differentiation of progenitor cells without triggering apoptosis. If so, it implies that the physiological level of Nolz-1 expression must be under rigorous control during neurogenesis. Consistent with this hypothesis, as post-mitotic Nolz-1-expressiing neurons migrate from the SVZ to the differentiated MZ, the expression level of Nolz-1 goes through a transition from the high level in the SVZ to the low level in the mantle zone [10,11].

Aberrant Proliferation and Differentiation of Neural Progenitors Induced by Transgenic Nolz-1
The finding that PH3-positive mitotic cells were reduced in the ventricular zone in which few apoptotic cells were present suggests that transgenic Nolz-1 may negatively regulate proliferation of neural progenitors. Analyses of tNolz-1 and eGFP transgene expression showed that there were some tNolz-1and eGFPpositive cells lining up alone radial lines as if they were derived from a clone (Fig. 3C, F). This radial lines pattern was also observed in single Ki67 immunostaining and double BrdU/Ki67 immunostaining in the cell cycle exit assay (Fig. 8). These data provide evidence supporting that the promotion of cell cycle exit/ inhibition of cell proliferation in the germinal zone of telencephalic regions were associated with transgenic Nolz-1 expression. Moreover, some ectopic TuJ1 expression in the germinal zone was observed to extend along radial lines in nCT brains, particular in the cortex (Fig. 9C2). Notably, strong ectopic TuJ1 expression was more prominently found in many apoptotic cell clusters in the germinal zone of MGE where transgenic Nolz-1 mRNA was localized as condensed particles. Furthermore, triple immunostaining showed that a few cells co-expressing BrdU and TuJ1 without Ki67 which represented precociously differentiating neurons prematurely exiting cell cycle were found in the germinal zone of nCT telencephalon. These findings support the conclusion that transgenic Nolz-1 not only promotes cell cycle exit/inhibits cell proliferation, but also promotes premature neuronal differentiation.
In addition to the cells lining up as radial lines, tNolz-1 and eGFP mRNAs were also observed in scattered cells in the germinal zone of telencephalic nCT brains. This may explain the presence of some scattered BrdU+/Ki672 cells that were found in the cell cycle exit assay and the widely reduction of PH3-and Ki67positive proliferating cells. Other possible causes of the widely reduction of proliferating cells may be non-cell autonomous effects of apoptosis induced by Nolz-1 over-expression. Further study is required to elucidate these possibilities.
In our study, the findings of both tNolz-1/eGFP mRNAexpressing cells and BrdU+/Ki672 cells lining along radial lines in the LGE of E12.5 nCT brain suggest a cell-autonomous effect of tNolz-1 over-expression in promoting cell cycle exit/inhibiting proliferation of progenitor cells in the germinal zone of nCT telencephalon. Induction of premature neuronal differentiation may be a consequence of promoting cell cycle exit by tNolz-1 cellautonomously. Given the induction of differentiation of N2A cells into TuJ1-positive neurite-bearing neurons by Nolz-1 overexpression, the premature neuronal differentiation phenotype observed in telencephalic regions of nCT brains may be caused by a cell-autonomous effect of ectopic expression of tNozl-1 in neural progenitor cells. Notably, unlike the restricted expression of Nolz-1 to postmitotic neurons in developing striatum in?vivo [11,14], it has been reported that Nolz-1 is expressed in proliferative neural progenitors in neurosphere culture in which over-expression of hNolz-1 results in increases of cell cycle exit [10]. On the other hand, functional studies of elbow (elB) and no ocelli (noc), two Drosophila homologous genes of Nolz-1, indicate regulatory control of cell proliferation by elB and noc [33,34]. In the eye-antenna imaginal discs, which give rise to the adult head, eye and antenna, elB and noc are expressed in the proliferative region. In region anterior to the morphogenetic furrow, the elB-noc gene complex attenuates Notch-induced cell proliferation. Loss of elB and noc activities induces overgrowth of the head capsule [34]. Conversely, over-expression of elB by eyeless-GAL4 in proliferating cells of eye imaginal discs causes reduction of eye size. The small eye phenotype can be rescued by coexpression of the positive cell cycle regulator, cycE [33]. However, no significant phenotype is observed when elB is over-expressed in post-mitotic cells of the imaginal discs [33]. These findings implicate that elB acts as a negative cell cycle regulator in proliferating cells of Drosophila eye imaginal discs.
Our present study and previous studies from other groups suggest that Nolz-1 is involved in regulation of cell proliferation. In N2A cells, over-expression of Nolz-1 resulted in decreased PCNA proliferative signals accompanying with a prominent neuronal differentiation. Previous study has shown that proliferation of NIH3T3-A14 cells is regulated by cAMP, because the cells are arrested in G1 phase 24 and 48 hr after the forskolin/IBMX treatment. Interestingly, Nolz-1 is induced 2 hr after forskolin/ IBMX treatment, but not at other time points [35]. This suggests a transient role of Nolz-1 in regulating proliferation of NIH3T3-A14 cells. This Nolz-1 induction is MAPK-independent, because Nolz-1 expression is not affected by the MEK-inhibitor U0126 [35].
Our study also suggests that Nolz-1 promotes neuronal differentiation, because transgenic expression of Nolz-1 in neural progenitors enhanced TuJ1 expression in the germinal zone of nCT brain. Increases of neuronal, but not glial, differentiation by over-expression of hNolz-1 have previously been shown in a cell culture study [10].

Abnormal Apoptosis Induced by Transgenic Nolz-1
Our present study indicates that transgenic ectopic expression of Nolz-1 in neural progenitors induced abnormal apoptosis in the nCT brain. Consistent with our findings, previous study has shown that Hb9 promoter-driven over-expression of chicken homologue of Nolz-1 in newly differentiating motor neurons induces abnormal apoptosis in the developing spinal cord [9]. These results suggest that constitutive expression of high levels of Nolz-1 may eventually lead to abnormal cell death. In our study, over-expression of Nolz-1 in three neural cell lines of different origins all increased apoptotic cell death. Taken together with these findings, overexpression of Nolz-1 at high levels may have a common effect leading to programmed cell death in progenitor cells and differentiating neurons.
Previous studies have shown that the zebrafish and chick homologues of Nolz-1 function as transcription repressors [4,5,7,9]. The transcription repressor activity of zebrafish Nlz1 is dependent on binding to groucho and class I histone deacetylases (HDACs), HDAC1 and HDAC2 [7]. During embryogenesis, HDAC1 and HDAC2 are expressed in neural stem cells/ progenitors. Interestingly, aberrant apoptotic cell death and reduction of brain size, the phenotypes that were observed in the nCT brain, were also observed in HDAC1;HDAC2 double mutant mice [36]. Over-expression of Nolz-1 may act as gain-offunction by chelating HDAC1 and HDAC2 from other genes that they regulated, which may then consequently induces abnormal programmed cell death.
The hypoplasia phenotype occurred all over the nCT brain, including the forebrain, midbrain and hindbrain without changing the overall structure of the nCT brain, though the brain size was reduced compared to that of the control littermates. Similar effects of tNolz-1 on promoting cell cycle exit/inhibiting cell proliferation and inducing apoptotic cell death may also occurred in the regions of nCT brain other than telencephalon. It is likely that the induction of cell death by over-expression of tNolz-1 occurs in diverse types of neural progenitors, because active caspase 3positve cells were detected in several regions of the nCT brain. Indeed, in agreement with this finding, we found that Nolz-1induced apoptosis occurred in three lines of neural cells in vitro. It is notable that in the hindbrain region where endogenous Nolz-1 is also expressed [29], we also observed tNolz-1-positive apoptotic cell clusters and ectopically expressed TuJ-1 in the germinal zone of nCT hindbrain (data not shown). These results implicate that tNolz-1 over-expression-mediated promoting cell cycle exit and premature neuronal differentiation might also occur in the hindbrain region of nCT brain. Figure S1 Expression of LacZ transgene in developing transgenic mouse embryos and brains. A, B: X-gal positive signals were detected in the line #9 Nolz-1 Tg embryo at E7.5 (A), but not in the control wild type embryo (B) . C-E: In the neonatal (P1) and three-week old (3 wk) transgenic brains, many X-gal positive cells were detected in the striatum (ST; C, D) and cerebral cortex (CTX; E). (TIF) Figure S2 Immunostaining of active caspase 3 in E15.5 nCT telencephlon. Comparing to E12.5, there were only a few active caspase 3 (AC3)-positive cells (arrows) in the cortex and striatum (ST) of nCT brain at E15.5. AC3-positive cells were nearly absent in the cortex and ST of nC control brain. (TIF) Figure 11. Induction of abnormal apoptosis by over-expression of Nolz-1 in N18, ST14A and N2A neural cells. A-F: Nolz-1 overexpression led to an increase of cell death in N18 cells. N18 cells were co-transfected with pNolz-1 and pEYFP constructs. Co-localization of Nolz-1 protein (red, B) and GFP (green) were found in all GFP-positive N18 cells (arrows, C and D), but not in GFP-positive cells of mock control (A, B). E: The growth curve analysis of Nolz-1-transfected N18 cells. The number of Nolz-1-transfected N18 cells was progressively decreased with increasing culture times. F: Western blotting of cleaved caspase 3 shows a 1.4-fold increase of cleaved caspase 3 in N18 cells 24 hr after transfection. G,-J: Transfection of Nolz-1 (red, H) induced active-caspase 3 (AC3, red, J) in ST14A cells 24 hr after transfection. K: Quantitative analysis of AC3-positive ST14A cells with Nolz-1 over-expression. L: Induction of cleaved caspase 3 and reduction of PCNA in Nolz-1-transfeceted N2A cells. Upper panel shows the experimental procedure. The Western blotting shows increases and decreases of cleaved caspase 3 and PCNA, respectively, in Nolz-1-transfected N2A cells M-Q: Promotion of neuronal differentiation by Nolz-1 over-expression in N2A cells. N2A cells were co-transfected with Nolz-1 and GFP. Nolz-1 over-expression induced a 15-fold increase of the number of neurite-bearing neurons (arrows, N) compared to the mock transfected group (M) 72 hr after transfection. O-Q: Nolz-1-transfected cells expressed the neuronal differentiation marker of TuJ1. R: Quantitative analysis of neurite-bearing cells with Nolz-1 over-expression. *, p,0.05; **, p,0.01; ***, p,0.001, Student's t test. Scale bars in D (for A-D), in J (for G-J) and in Q (for O-Q), 20 mm. doi:10.1371/journal.pone.0074975.g011