Six3, a homeodomain-containing transcriptional regulator belonging to the Six/so family, shows a defined spatiotemporal expression pattern in the developing murine telencephalon, suggesting that it may control the development of specific subsets of neural progenitors. We find that retrovirus-mediated misexpression of Six3 causes clonal expansion of isolated cortical progenitor cells by shortening their cell cycle and by prolonging their amplification period, while maintaining them in an immature precursor state. Our results show that the observed effects exerted by Six3 overexpression in mammalian brain depend strictly on the integrity of its DNA-binding domain, suggesting that Six3 action likely relies exclusively on its transcriptional activity. In vivo upregulation of Six3 expression in single progenitor cells of the embryonic telencephalon keeps them in an undifferentiated state. Our observations point to a role of Six3 in the control of the subtle equilibrium between proliferation and differentiation of defined precursor populations during mammalian neurogenesis.
One of the most crucial steps in the formation of the central nervous system (CNS) is the tight control of the balance between cell proliferation and differentiation of progenitor cells, which is necessary to correctly expand the early precursor pool before the onset of differentiation. A large body of evidence demonstrates a close interrelationship between factors governing cell proliferation and differentiation (see Ohnuma et al. 2001; Cremisi et al. 2003 for review). For instance, cell cycle regulators such as cyclin D2 and the cyclin-dependent kinase inhibitor p27 may influence cell fate choice (Huard et al. 1999; Caviness et al. 2003; Tarui et al. 2005; Itoh et al. 2007; Nguyen et al. 2006), whereas the products of patterning genes, such as the homeodomain-containing Pax6 and Emx2, may additionally control cell proliferation (Gangemi et al. 2001; Heins et al. 2001; Estivill-Torrus et al. 2002; Heins et al. 2002). Interestingly, it has been shown that several proteins capable of regulating both cell proliferation and differentiation exert their effects via 2 independent domains (for review see Zhu and Skoultchi 2001), suggesting that the 2 processes may require direct linkage through single proteins acting as bridging factors.
A candidate molecule for this bridging role in the vertebrate CNS is the transcription factor Six3, the product of one homolog of the Drosophila gene Optix (Kawakami et al. 2000). Six3 is known to be essential for the specification of the anterior neural plate, as well as for the proper formation of eyes and forebrain. Previous reports suggested that the gene could coordinate cell proliferation and differentiation following the observation that in both fish and amphibians its overexpression induces enlargement of the eyes and forebrain and coincidently induces ectopic optic cups and lenses (Oliver et al. 1996; Kobayashi et al. 1998; Loosli et al. 1999; Bernier et al. 2000; Seimiya and Gehring 2000; Lagutin et al. 2001; Ando et al. 2005; Gestri et al. 2005). It has been proposed that, during CNS development in fish (Orizyas latipes), Six3 and the DNA rereplication inhibitor Geminin may mutually inhibit each other by direct interaction (Del Bene et al. 2004). Geminin, in turn, is able to bind Cdt1, a DNA replication initiation factor, sequestering it and inhibiting the G1/S transition and cell proliferation (Wohlschlegel et al. 2000). These observations lend support to the notion that Six3 could indeed bridge the processes of proliferation and differentiation, regulating differentiation via its transcriptional activity and the cell cycle by sequestering Geminin.
Although a number of investigations have contributed to a thorough understanding of the role of Six3 during eye development, less is known about the function of the gene in other regions of the CNS and during later phases of neurogenesis. Especially obscure remains the function of the gene in the mammalian telencephalon, where it appears to be particularly abundant in the ganglionic eminences (GEs) (Oliver et al. 1995; Conte et al. 2005). At birth, the level of Six3 mRNA in this region is notably high in the subventricular zone of the GEs and in the dentate gyrus (Gray et al. 2004) where stem/progenitor cells reside (Alvarez-Buylla and Lim 2004). Because the function of Six3 in these neural regions is still unknown, we investigated the role of the gene in early murine neurogenesis by manipulating its expression in neural progenitor cells in vitro and in vivo. Our data suggest that Six3 may be a novel modulator of the maturation state of neural progenitor cells.
Materials and Methods
Neural progenitor cultures were prepared from embryonic telencephalic explants as previously described (Gotz et al. 1998). Dissociated cells were plated at a density of 2.5 × 105 cells/cm2 onto poly-D-lysine–coated coverslips and grown for 2–7 days in defined modified Bottenstein and Sato's medium. In some cases, cells were treated for 15 or 31 h with 10 μM BrdU, prior to fixation with 4% Paraformaldehyde for 15 min for further processing.
Mice were handled in agreement with guidelines conforming to the Italian current regulations regarding the protection of animals used for scientific purpose (D. lvo, 27 January 1992, no. 116). Procedures were specifically approved by the Ethical Committee for Animal Experimentation (Comitato per la Sperimentazione Etica sugli Animali) of the National Institute of Cancer Research and by the Italian Ministry of Health. All experiments have been performed on the C57/Bl6 mouse strain.
Retroviral Vectors and Transduction Procedures
A cDNA of human Six3, derived from the pCiNeo-hSix3 plasmid (Granadino et al. 1999), was kindly provided by Paola Bovolenta (Institute Cajal of Neuroscience, Madrid, Spain) and inserted into the SalI restriction site of the pCEG retroviral vector (Yoon et al. 2004), kindly provided by Gordon Fishell (The Skirball Institute of Biomolecular Medicine, New York), containing an internal ribosome entry sequence followed by the green fluorescent protein (GFP)–coding sequence. A mutated construct denoted Six3_t757c was generated from the pCiNeo-hSix3 using the QuickChange site-directed mutagenesis kit (Fermentas International Inc., Burlington, Canada) and similarly inserted into the SalI restriction site of the pCEG retroviral vector. A murine geminin cDNA was obtained from the pSPORT6-mgmn plasmid provided by the National Institutes of Health (NIH) Mammalian Gene Collection (clone MGC:28184 IMAGE:3988065) and cloned into the pCEG vector.
For Six3 downregulation, an engineered microRNA (miRNA-Six3; TGCTGTTACCGAGAGGATCGAAGTGCGTTTTGGCCACTGACTGACGCACTTCGCCTCTCGGTAA) was designed against the Six3-coding sequence following the BLOCK-iT Pol II miR RNAi Expression Vector kit guidelines (Invitrogen Corporation, Carlsbad, CA) and cloned in the pCEG vector. Control experiments were performed using either pCEG alone or the pCMMP retroviral GFP vectors (Hack et al. 2004), kindly provided by Cecilia Lundberg (Wallenberg Neuroscience Center, Lund, Sweden). Retroviral stocks were prepared by transiently transducing the plasmids into Phoenix packaging cells as described elsewhere (Pear et al. 1993). Primary cultures were infected using a low-titer viral supernatant to obtain a maximum of 30 individual clones per well and allow for a subsequent clonal analysis (Heins et al. 2002; Malatesta P, unpublished observations). In vivo retroviral transduction was performed through intrauterus injections of the retroviral particles into the telencephalic ventricles of E14 mouse embryos (the day of vaginal plug was considered as embryonic day 0, E0, Malatesta et al. 2001). At birth, brains of injected embryos were dissected, fixed for 4 h at 4 °C in 4% PFA, and cryoprotected in 20% sucrose. Transversal 14-μm thick sections were cut with a Leica CM1100 cryostat.
Immumostainings were performed using the following antibodies: mouse monoclonal antibodies against nestin (1:250 dilution, BD Pharmingen, Franklin Lakes, NJ), β-III-tubulin (1:100 dilution, Sigma-Aldrich), Glial Fibrillary Acidic Protein (1:200 dilution, Sigma-Aldrich), adenomatous polyposis coli (APC) (1:100 dilution, CC1, Calbiochem, Darmstadt, Germany), and bromodeoxyuridine (1:50 dilution, BioScience Products S.A. Emmenbrucke, Switzerland); rabbit polyclonal antisera against GFP (1:500 dilution, RDI, Concord, MA); and guinea pig polyclonal antisera against Glutamate Astrocyte-Specific Transporter (1:4000 dilution, Chemicon, Billerica, MA). Binding of primary antibodies was revealed with appropriate secondary fluorescein isothiocyanate– and tetramethyl-rhodamine-iso-thiocyanate-conjugated antibodies (1:50 dilution, Immucor, Inc., Norcross, GA) or biotinylated secondary antibodies (1:50 dilution, Dako, Glostrup, Denmark) that were revealed with streptavidin-conjugated 7-amino-4-methyl-coumarin-3-acetic acid (1:50 dilution, Vector Laboratories, Burlingame, CA). Nuclei were stained through 5 min incubation in 4′,6-diamidino-2-phenylindole solution (1 μg/ml, Sigma).
For clonal analyses, mean and standard errors (SEs) were calculated from the values obtained for the cell population contained in each coverslip. In nonclonal experiments, at least 100 cells per coverslip were analyzed and the mean and SEs were calculated from different coverslips. “nExp” was adopted to denote the total number of independent experiments, whereas “nCln” and “nCel” indicate the total number of clones and cells analyzed, respectively. The threshold of statistical significance was determined with a 2-tailed Student t-test and was considered as P < 0.05.
Fluororescence Activated Cell Sorter Analyses
To establish the proportion of Annexin-V–positive cells in transduced cultures, cells were harvested from dishes 2 days after transduction and incubated in Annexin-V-phycoerythrin (2.5 μg/ml, Alexis Biochemicals, Lausen Switzerland) according to the manufacturer's protocol. As positive control for apoptosis, we utilized cells transduced with the control virus and treated with 1 μM staurosporine for 6 h. To evaluate the distribution of transduced cells with respect to their nuclear DNA amount, cells were harvested 2 days after transduction, fixed in suspension with 4% PFA, incubated for 10 min with propidium iodide (PI) solution (50 μg/ml PI, 20 μg/ml RnaseA), and analyzed with a FacScan flow cytometer (BD Biosciences, San Jose, CA). To isolate transduced (GFP positive) cells for reverse transcription–polymerase chain reaction (PCR) analyses, cultures were harvested 2 or 4 days after transduction, resuspended in PBS, and sorted on the basis of the intensity of GFP fluorescence using a FacsAria Cell Sorter (BD Biosciences, Inc.).
Total RNA from tissues and from sorted cells was extracted using the RNeasy kit from Qiagen inc. (Valencia, CA). Retrotranscription was performed using 500 ng of RNA and the iScript cDNA Synthesis kit (Bio-Rad Laboratories, Hercules, CA). Quantitative real-time PCR was performed using the iQ SyBr green supermix (Bio-Rad Laboratories). The amplification was carried out for 40 cycles with 1/20 of the retrotranscribed RNA. As a housekeeping gene for data normalization was used the Rlp41 gene, coding for a ribosomal protein (Warrington et al. 2000). Data were analyzed by using the levels of Rlp41 as reference and individual calibration lines to quantify the amplification efficiency of each primer pair. The sequences of the primers are available on request.
Promoter Activity Assays
NIH3T3 cells were transfected using the Lipofectamine-2000 reagent (Invitrogen) with the following plasmids in different combination: pCEG-Six3, pCEG-Six3_t757c, pGL3-Six3–responsive plasmid containing the firefly luciferase–coding sequence (kindly provided by Jochen Wittbrodt, EMBL, Heidelberg, Germany), pCEG, and the Renilla luciferase–expressing phRL plasmid. Two days after transfection, luciferase assay was performed using the Dual-luciferase assay kit (Promega, Madison, WI) and following the manufacturer's instructions. Firefly luciferase values were normalized using Renilla luciferase as the internal control.
Six3 Affects the Cell Cycle and Proliferation Period of Cortical Progenitors
We used replication-incompetent retroviral vectors to induce the expression of Six3 in association with GFP in a small percentage of acutely dissociated E14 telencephalic precursor cells plated at high density. As previously described, this procedure allows to analyze the entire progeny of single precursor cells (Williams et al. 1991; Malatesta et al. 2003). Because Six3 is preferentially expressed in GEs and not in more dorsal regions of the developing brain (Supplementary fig. 1, Gray et al. 2004), we carried out separate overexpression experiments on cortical and GE progenitor cells. The effect of the retroviral transduction on Six3 mRNA levels was monitored by real-time PCR on RNA extracted from GFP-positive cells sorted with a Fluororescence Activated Cell Sorter (FACS; Fig. 1A). This analysis showed that Six3 transduction is able to increase the level of Six3 mRNA more than 300 times in cortical precursors to a level similar to that typical of the GEs. Given the already high Six3 expression in GEs cells, retroviral transduction in these precursors caused only a marginal increase (about three times) in their Six3 mRNA level (Fig. 1A).
Clonal analysis 7 days after in vitro transduction showed that Six3 overexpression induced a significant increase in the average clonal size: 6.2 ± 0.5 cells/clone in the control (nExp:8; nCln:757) versus 9.1 ± 0.5 cells/clone in the Six3-transduced cells (nExp:4; nCln:602; P < 0.01; Fig. 1B,C,F). On the contrary, transduction of Six3 did not significantly affect the clonal size in dissociated cells of the GEs (Supplementary fig. 2B), likely because of the modest increase of Six3 mRNA level compared with the endogenous one (Fig. 1A). Intriguingly, the size of the clones generated by wild-type GE precursor cells, 7 days after transduction with the control vector (9.8 ± 1.2 cells/clone), was larger than the clonal size of wild-type cortical precursors (6.2 ± 0.5 cells/clone) and similar to that of Six3-transduced cortical precursors (9.1 ± 0.5 cells/clone). Taken together, these results indicate that Six3 expression positively influences the proliferation of telencephalic progenitor cells.
We then downregulated Six3 in GE precursor cells with a vector expressing an engineered miRNA against its coding sequence (miRNA-Six3). Real-time PCR analysis showed that the transduction of miRNA-Six3 in GE precursor cells reduced Six3 mRNA to a quarter of its physiological level (Fig. 1A). Consequently, clones generated by these precursors showed, as early as 5 days after transduction, a significantly reduced size (3.6 ± 0.3 cells/clone; nExp:2; nCln:159) compared with control GE precursors after the same time in culture (5 ± 0.6 cells/clone; nExp:2; nCln:137; P < 0.05; Fig. 1G). This result corroborates the conclusion that Six3 plays a role in promoting precursor cell proliferation.
The analysis of the distribution of clonal sizes shows that the most visible effect of Six3 transduction is a dramatic reduction of single-cell and 2-cell clones in the cortex. A less pronounced reduction of the same size classes is shown in the GE precursors (Supplementary fig. 2A,D).
The observed increase in clonal size could be due to enhanced cell proliferation and/or to a reduced level of programmed cell death. We therefore assessed the number of Six3-transduced cells that exhibited picnotic nuclei 3 days post-transduction and did not find significant differences between Six3-transduced (5 ± 2% picnotic cells; nCell 330) and control transduced cultures (4 ± 1% picnotic cells; nCell 342). Annexin-V stainings further confirmed that the percentage of apoptotic cells among Six3- (4 ± 2%; nExp 2) and control transduced cells (3 ± 1%; nExp 2; Supplementary fig. 3) was similar. Moreover, these percentages were statistically indistinguishable from that observed in nontransduced (i.e., GFP negative) cells from the same cultures (2 ± 1%; nExp 4). These results are in good agreement with previous observations showing that the rate of cell death is very small in these culture models (Heins et al. 2001) and indicate that the size difference in clones generated by Six3-transduced cells is a result of enhanced cell proliferation.
The increase in average clonal size from cortical progenitor cells misexpressing Six3 could either be due to a shortened cell cycle or to an increase in the overall proliferation period of a precursor cell and its progeny. To discriminate between these possibilities, we assessed the average size of the clones after a shorter period of culture (i.e., 3 days post-transduction) before the majority of precursor cells stop dividing in such in vitro condition. Even in this case a significant difference in the average clonal size between Six3-transduced (3.0 ± 0.3 cells/clone; nExp:3; nCln:230) and control transduced progenitor cells (2.2 ± 0.1 cells/clone; nExp:4; nCln:168; P < 0.05; Fig. 1H) was observed. Since 3 days after transduction, the number of actively dividing cells in the Six3-transduced cultures was indistinguishable from that of control cultures, as shown by their cumulative labeling indexes with BrdU (80 ± 8% for Six3-transduced and 70 ± 5% for control transduced cells), we concluded that the increase in clonal size must have been due to a shortening of the cell cycle in the Six3-transduced cells. Accordingly, the labeling index after 15 h increased from 56 ± 2% (nExp:2; nCel 666) in the control to 64 ± 2% (nExp:2; nCel 702; P < 0.05) in the Six3-transduced cultures. Furthermore, FACS analysis of PI-stained cells showed that in the Six3-expressing population, the proportion of cells in S-phase raised at the expense of the percentage of cells in both G1 and G2 (nExp 2; Supplementary fig. 4). These findings suggest that Six3 affects the cell cycle of neural progenitor cells mostly by reducing the length of the G1- and G2-phases.
To assess whether, in addition to reducing the length of the cell cycle, Six3 also protracted the proliferation period, the cells were labeled with a 31-h BrdU pulse 6 days after transduction. By this time, the majority of the cells would normally be in a quiescent state. Accordingly, in the control transduced cultures, the percentage of BrdU-positive cells was 37 ± 3% (nExp:4, nCel:4198), whereas in the Six3-transduced cultures the percentage was still dramatically high (67 ± 4%, nExp:4; nCel: 2531; P < 0.01; Fig. 1D,E,I), not too far from that observed at 3 days after transduction. Such increase cannot exclusively depend on the shortening of the cell cycle because it would require a halving of the cell cycle time. In turn this would be reflected, even assuming all asymmetric divisions, by at least a doubling of the clonal size that is not compatible with what we found in the previous experiments. Thus, in addition to modulate the length of cell cycle, Six3 seems also to induce a lengthening of the overall proliferation period of these cells.
Six3 Maintains Neural Progenitors in an Immature State
In order to understand whether Six3 also governs the phenotypic fate choice of neural progenitors, we initially examined the composition of the clones generated by Six3-transduced cells by using β-III-tubulin as marker for neurons and GFAP to identify astroglial cells. Clones were classified herein as: “pure neuronal” when comprised only of β-III-tubulin–positive cells (Fig. 2A); “pure nonneuronal” if composed exclusively of β-III-tubulin–negative cells (Fig. 2B); and “mixed clones,” that is, constituted of at least one cell immunopositive for β-III-tubulin and at least another one immunonegative for this neuronal marker (Fig. 2C).
As shown in Figure 2D, transduction of Six3 in cortical precursors caused a significant decrease of the pure neuronal clones from 51 ± 3% (nExp:8; nCln:757) to 36 ± 3% (nExp:4; nCln:581; P < 0.01) and a parallel increase in the percentage of mixed clones from 35 ± 3% in the control (nExp:8; nCln:757) to 47 ± 4% (nExp:4; nCln:581; P < 0.05) in the Six3-transduced clones.
Notably, 64 ± 4% of the purely neuronal clones in the Six3-transduced population were composed of 1 or 2 cells (i.e., they very likely exited the cell cycle short after having been transduced). Because the retroviral expressed proteins requires several hours to reach a detectable level, such small pure neuronal clones are very likely derived from cells that, already committed to neuronal differentiation, underwent their last mitosis and terminally differentiated before transduced Six3 could induce any effect.
The transduction of Six3 in the GEs precursor cells did not significantly affect the composition of the clones (Supplementary fig. 2C). Remarkably, the composition of the clones generated by wild-type GEs precursors expressing endogenously high level of Six3, closely resembles that of Six3-transduced cortical precursors. The percentage of mixed clones is virtually identical (48 ± 5% vs. 47 ± 4%), and the percentage of pure neuronal clones is similar (30 ± 4% vs. 36 ± 3%).
We also measured the proportion of oligodendrocytes by double staining with the GFAP antibody and CC1, a monoclonal antibody that recognizes the product of the APC gene specifically in oligodendrocytes (Bhat et al. 1996) and, at a lower level, in astrocytes. Oligodendrocytes are positive for CC1 and negative for GFAP.
Misexpression of Six3 in the cortical precursor cells induced a significant decrease in the percentage of oligodendrocytes, as shown by the reduction in the proportion of CC1-positive/GFAP-negative cells from 25 ± 2% (nExp:2; nCel:1393) to 7 ± 2% (nExp:2; nCel:2105; P < 0.01, Fig. 2E–L). A reduction of β-III-tubulin–positive cells was also visible, although not significant (0.05 < P < 0.07). The reduction in the proportion of differentiated neural phenotypes was compensated by a significant increase in precursor cells showing exclusive positivity to nestin, a progenitor/stem cell marker (P < 0.05), and by the emergence of a population of cells, accounting for 16 ± 5% of the Six3-transduced cells, that were immunonegative for the markers of the 3 neural lineages as well as for nestin (Fig. 2G–L). This latter population might represent a subset of immature precursor cells as suggested by other authors (Kukekov et al. 1997).
Altogether these findings indicate that Six3 delays determination of neural precursor cells that tend to maintain a progenitor-like undifferentiated state.
Six3 and Geminin Exert Opposite Effects on Progenitor Cell Proliferation and Differentiation
In a first gene expression survey, we detected higher levels of geminin mRNA in the cortex and GEs than in the developing eye (Fig. 3A), which is known to correspond to the embryonic area where a direct interaction between Six3 and geminin has previously been proposed to occur during fish development (Del Bene et al. 2004). Coexpression of geminin and Six3 in the mammalian GEs suggests that a similar interaction may occur in the mammalian CNS. Upon retroviral transduction of geminin in both cerebral cortex and GEs progenitor cells, the size of the clones produced by the transduced precursor cells was significantly reduced. Cortical precursors transduced with geminin formed clones composed on average by 3.3 ± 0.5 cells (nExp:4; nCln:210), whereas control transduced cultures formed clones of 6.2 ± 0.5 cells (nExp:8; nCln:757; P < 0.01, Fig. 3B). A similar reduction in clonal size was observed in the precursor cells derived from the GEs, where the average clonal size of geminin-transduced cells was 4.3 ± 0.4 cells/clone (nExp:3; nCln:260) versus 9.8 ± 1.2 cells/clone (nExp:6; nCln:307; P < 0.01, Fig. 3B) in cultures transduced with the control viral construct.
Analysis of the composition of clones formed by transduced precursors showed that geminin only affected cells of the GEs, whereas the composition of similarly transduced cortical clones remained unchanged (data not shown). Geminin-expressing precursors derived from the GEs formed significantly fewer mixed clones (29 ± 3.1%; nExp:3; nCln:260) than control transduced cells (48 ± 4.7%; nExp:6; nCln:307; P < 0.01). Accordingly, they formed a significantly higher percentage of pure neuronal clones (42 ± 2.2%; nExp:3; nCln:260) than the control (30 ± 3.5%; nExp:6; nCln:307; P < 0.05, Fig. 3C). Thus, the observed effect of geminin on GEs-derived precursors (where Six3 is also expressed) seems virtually opposite to that of Six3 in the cortex. These results prompted us to verify whether the effect of Six3 is mediated by its transcriptional activity or by sequestering geminin.
The Effect of Six3 on Neural Progenitors Requires Its Transcriptional Activity
To answer this question, we investigated whether transcriptional activity is necessary for Six3 to elicit its effects on neural progenitor cell proliferation and differentiation. We therefore constructed a retroviral vector carrying a mutated form of Six3 Six3t757c. Six3t757c harbors the same amino acid substitution in the Six3 homeodomain previously introduced in the Medaka homolog in order to retain its geminin-binding ability while disrupting its DNA-binding activity (Del Bene et al. 2004). We then confirmed the lack of transcriptional activity of Six3t757c by a luciferase assay based on a reporter plasmid containing a Six3-responsive element (Goudreau et al. 2002). Although the expression of either human or murine Six3 in NIH3T3 cells inhibited the transcription of the luciferase gene from the luciferase reporter plasmid in a dose-dependent manner (Fig. 4A, Zhu et al. 2002), Six3t757c did not modulate the luciferase expression levels (Fig. 4B).
When Six3t757c was transduced into E14 mouse telencephalic progenitor cells, the proliferation and differentiation potential of the cells remained unchanged (Fig. 4C,D), despite the level of expression achieved was the same as its wild-type counterpart (data not shown). This finding demonstrates that the effects elicited by Six3 do require that its ability to bind to DNA is preserved.
Six3 Maintains Ventricular Progenitor Cells in an Immature State
The above observations indicate that Six3 delays the differentiation program of isolated cortical progenitors and maintains these cells in an immature state. We then asked whether the same occurs in vivo or whether the action of extrinsic factors could mask the cell-autonomous action of Six3 observed in culture. To address this problem, we modulated the expression of Six3 by injecting intrautero, into the telencephalic lateral ventricles of E14 mouse embryos, retroviral vectors coding for Six3. Because the injected virus infects only a small fraction of the dividing cells, this approach allowed us to examine the effect of the transduced gene at the single-cell level in situ. Analysis of the fate of Six3- versus control-transduced cells at birth (P0) revealed a significant increase in the percentage of cells that remained in contact with the ventricular surface within the Six3-overexpressing cell population, both in the cortex and in the basal ganglia (Fig. 5A). The percentage of cells in contact with the ventricular surface in the cortex increased from 3.7 ± 0.9% (nExp:3; nCel:1359) within the control population to 11.7 ± 1.9% in the Six3-transduced cells (nExp:3; nCel:253; P<0.05). In the basal ganglia, the effect was even more pronounced because the increase was from 3.7 ± 1.5% (nExp:3; nCel:2216) to 23.7 ± 1.8% (nExp:3; nCel:196; P < 0.01). Virtually all the cells in contact with the ventricular surface expressed the neural progenitor markers nestin and GLAST, typical of progenitor cells during neural development (Fig. 5C–E), and about half of them expressed the active proliferation marker Ki-67 (data not shown).
A higher number of nestin-positive cells were observed throughout the telencephalon in the Six3-transduced cell population when compared with the same population in control animals. In particular, their percentage increased from 32 ± 2% (nExp:7; nCel:766) to 44 ± 0.4% (nExp:7; nCel:212; P < 0.05) in the cortex and from 24 ± 2% (nExp:6; nCel:1098) to 57 ± 4% (nExp:7; nCel:190; P < 0.01) in the basal ganglia of the infected brains (Fig. 5B). Taken together, these findings suggest that Six3 overexpression biases the progenitor cells to preserve an undifferentiated state also in the presence of environmental and/or cell-autonomous factors that normally induce cell differentiation.
Gene Pathways Downstream of Six3
Because our data showed that Six3 acts on the proliferation and differentiation of mouse neural progenitor cells exclusively by acting as a transcriptional regulator, we surveyed the modulation of a set of candidate downstream genes by real-time PCR. Total RNA extracted from Six3-transduced cells sorted by FACS based on GFP fluorescence intensity was compared with a reference RNA from control transduced progenitors. Efficiency of cell sorting was verified by fluorescence microscopy and was found to be >90% for both sorted populations.
Consistent with its effect on cell proliferation, Six3 transduction induced the expression of cyclin D1 and D2, Cdk4, and Cdc37, which are known to be positive regulators of the cell cycle, whereas the level of the cycle inhibitor p27 was unchanged (Fig. 6). The Six3-transduced cell population showed an increased level of Sox2, a marker of stem and early progenitor cells (Graham et al. 2003) and a putative target of Six3 (Liu et al. 2006). Moreover Six3-transduced cells showed an increased expression of GLAST and, slightly, of Vimentin. During neurogenesis, both these molecules are markers of radial glial cells, which are known to be immature neural progenitors (Malatesta et al. 2000). Moreover, whereas Pax6 and Neurogenin-2 were not found to be modulated in cortical progenitor cells, Neurogenin-1 became undetectable in Six3-transduced cells, suggesting a strong downregulation (Fig. 6). Finally, we found increased levels of Smad1, previously shown to be a key bone morphogenetic protein signaling effector involved in the inhibition of neurogenesis (Nakashima et al. 2001; Takizawa et al. 2003).
We have investigated the role of the homeodomain-containing transcriptional regulator Six3 in the development of the mammalian telencephalon by misexpressing the gene in neural progenitor cells by retroviral transduction. Given the difference in the endogenous Six3 expression between cortex and GEs, the viral transduction caused a pronounced upregulation of Six3 mRNA only in cortical precursor cells, where the Six3 level raised from virtually zero to about the same level of the wild-type GEs precursors. On the contrary, only a minor modulation was seen in the GEs. Accordingly, Six3 transduction had significant effects only on cortical clones, whose size and composition became very similar to those of the wild-type GEs that are larger and more frequently mixed. We also downregulated the level of Six3 in GEs precursor cells by expressing an engineered miRNA and observed a reduction of clonal size, showing that correct cell proliferation in GEs is dependent on the level of Six3.
Our results show that Six3 influences maturation of neural precursors by modulating their cell cycle and maintaining them in an immature, multipotent state. Proliferation and differentiation are known to be intrinsically linked because neural progenitors start to differentiate into neurons only after exiting the cell cycle. It is therefore to be expected for a gene that delays the onset of differentiation to prolong the overall length of the proliferation period of the cells. Our data, however, show that this is not the only way by which Six3 acts on proliferation of neural precursors because it also affects the length of their cell cycle. This observation suggests that, during murine neurogenesis, Six3 may play a dual bridging role in the regulation of proliferation and differentiation.
In agreement with this view is the expansion of the neural territory at the expense of nonneural tissue, observed in Xenopus following misexpression of Six3 in the absence of cell proliferation (Gestri et al. 2005). In the mouse telencephalon, however, a possible intrinsic link between cell cycle length and differentiation fate has been suggested (Calegari and Huttner 2003). Therefore, our observations do not rule out the possibility that Six3 exerts its effects on the differentiation of precursor cells solely by affecting the length of their cell cycle.
Six3 Forces Neural Precursors to Maintain Progenitor/Stem Cell Characteristics
Most neural progenitor cells transduced in vivo with Six3 at a neurogenetic stage failed to differentiate into neurons and were mainly found in the near proximity of the telencephalic ventricles at postnatal stages, both in the ventral and in the dorsal telencephalon. In addition, the majority of Six3-transduced cells maintained the expression of the progenitor/stem cells marker nestin and the radial glia/astroglial marker GLAST. These data suggest that enhanced expression of Six3 forces progenitor cells to maintain more immature phenotypic traits. Interestingly, the transduction of Six3 in the developing brain had an effect more pronounced than in isolated precursors affecting also progenitors of the GEs. Because the main difference between the in vivo and in vitro conditions is the loss of cellular organization and paracrine stimuli, this finding suggests that extrinsic factors may act synergistically with Six3.
Putative Mechanism of Action of Six3 on Neural Progenitor Cells
During eye development in fish, Six3 may govern cell proliferation independently from its transcriptional activity by binding to and inhibiting geminin (Del Bene et al. 2004), a partner of Cdt1 that directly controls the initiation of DNA replication. This observation highlights the interesting possibility that the dual role of Six3 could be supported by 2 distinct functional domains of the molecule. Our data, however, indicate that the effects on proliferation and differentiation of Six3 misexpression are both abolished by a single-point mutation in the Six3 DNA-binding motif. Remarkably, in Medaka the same mutation does not influence the interaction with geminin (Del Bene et al. 2004). Thus, these findings show that in the mouse the effects of Six3 on proliferation and differentiation strictly rely on the integrity of its DNA-binding domain, suggesting, in contrast to what has been shown by Del Bene et al. in fish, that the Six3 activity in the mouse is likely due to its transcriptional activity. However, our observations are compatible with a possible role for geminin in inhibiting Six3, in agreement with what has been shown in Medaka (Del Bene et al. 2004). In fact, geminin overexpression in ventral telencephalic precursor cells, where Six3 is endogenously expressed, has an effect nearly opposite to that seen following overexpression of Six3 in cortical precursors.
The observation that the transcriptional activity of Six3 likely controls both proliferation and differentiation of neural progenitor cells prompted us to identify its possible downstream targets implicated in these effects. Many of the genes examined by real-time PCR were found to be upregulated in a seemingly unexpected manner considering that previous reports have shown that Six3 acts mainly as a transcriptional repressor (Kobayashi et al. 2001; Lengler et al. 2001; Zhu et al. 2002; Lopez-Rios et al. 2003; Gestri et al. 2005). However, our gene screening covered not only the direct targets of Six3 but also secondary targets that may be upregulated as a consequence of the downregulation of their repressors.
In agreement with the observed phenotype, genes found to be upregulated included crucial key regulators of the cell cycle, such as cyclin D1, cyclin D2, cdk4, and cdc37. The modulation of other genes supports the view that Six3 induces the maintenance of immature precursor cell characteristics. In particular, the neural stem cell marker Sox2 and the 2 markers for radial glia, GLAST and vimentin, were found to be upregulated, whereas downregulation was observed for the neuronal commitment–associated gene neurogenin-1. Thus, it may be concluded that the expression of Six3 may be pivotal during neurogenesis to ensure the correct expansion of pluripotent progenitor cells before they differentiate into more restricted lineage precursors.
Supplementary material can be found at: http://www.cercor.oxfordjournals.org/.
We would like to acknowledge Dr Paola Bovolenta, Dr Gordon Fishell, Dr Cecilia Lundberg, and Dr Jochen Wittbrodt for kindly providing plasmids that have been used for this work. Thanks to Dr Annette Gaertner for helpful comments on the article. This work was supported by the Ministry of the University (project PRONEURO and Fondo Integrativo per la Ricerca di Base) and by Associazione Italiana per la Ricerca sul Cancro grant NUSUG.
Conflict of Interest: None declared.