All animals were housed with access to food and water ad libitum in a colony room maintained at a constant temperature (19 to 22°C) and humidity (40 to 50%) on a 12:12 h light:dark cycle. Animal treatments and handling procedures were approved by the Local Committees, in accordance with the European Community Council Directive (86/609/EU).
B6CBA wild-type mice (from Charles River Laboratories, Les Oncins, France), Raldh3-deficient mice , and Gsx2-deficient mice  were used in this study. For embryonic ages, time of pregnancy was determined by first detection of a vaginal sperm plug by daily inspection and considered as E0.5. For postnatal studies, the day of birth was considered as postnatal day 0 (P0).
To induce vitamin A deficiency in mice, pregnant mice were fed with the vitamin A-deficient diet TD.86143 (Harlan Laboratories Inc., Indianapolis, IN, USA).
E12.5 or E14.5 fetal brains were excised and placed in sterile phosphate-buffered saline pH 7.4, and the LGEs were dissected bilaterally, pooled and gently dissociated with a fire-polished Pasteur pipette.
Mixed neuron-glial primary cultures were obtained by plating the cells onto 24-well plates containing glass coverslips precoated with 0.1 mg/ml poly-D-lysine (Sigma Chemical Co., St Louis, MO, USA) at a density of 150,000 cells/cm2 in Eagle's minimum essential medium (Invitrogen SA, Prat de Llobregat, Barcelona, Spain) supplemented with 7.5% fetal bovine serum (FBS; Invitrogen SA), 0.6% D-(+)-glucose (Sigma Chemical Co.), 100 U/mL of penicillin and 100 mg/mL streptomycin (both obtained from Invitrogen SA). Three or 5 days after seeding, cultures were fixed with 4% paraformaldehyde solution (PFA; Merck Biosciences Ltd, Nottingham, UK) in 0.1 M phosphate buffer pH 7.4 and processed for immunocytochemistry.
LGE-derived neurosphere cultures were obtained by seeding 50,000 cells/cm2 in medium containing Dubelcco's Modified Eagle's Medium (DMEM; Sigma Chemical Co.):F12 (Invitrogen SA) (1:1); supplemented with 0.3% glucose (Sigma Chemical Co.), 0.3 mg/ml glutamine (Invitrogen SA), 5 mM HEPES (Invitrogen SA), 100 U/ml penicillin, and 100 mg/ml streptomycin (Invitrogen SA), 4 μg/ml heparin (Sigma Chemical Co.), 4 mg/ml bovine serum albumin (Sigma Chemical Co.), 1× N2 supplement (Invitrogen SA), 20 ng/ml fibroblast growth factor (Sigma Chemical Co.) and 10 ng/ml epidermal growth factor (Invitrogen SA). Every 5 days neurospheres were collected, dissociated by pipetting approximately 40 times with a P100 micropipette and re-plated in fresh media at a density of 10,000 cells/cm2.
For cell differentiation, 12,500 cells/cm2 were seeded. The day after plating, cells were collected and incubated onto Matrigel-treated coverslips (Growth Factor Reduced Matrigel Matrix, BD Biosciences, Clontech-Takara Bio Europe, Saint-Germain-en-Laye, France). Media was changed to medium supplemented with only 20 ng/ml fibroblast growth factor (Sigma Chemical Co.) and cells were allowed to differentiate in this medium for 2 more days. Then, media was changed again to medium supplemented with 2% FBS (Invitrogen SA) and cells were grown for 3 more days (until a total of 6 days of in vitro differentiation). Cell pellets for each time point (0, 3 and 6 days of differentiation) were obtained and frozen at -80°C for RNA or protein extraction.
In the present study, we used the mouse embryonic stem cell line R1 obtained from Dr Andras Nagy's laboratory. The maintenance of undifferentiated mouse embryonic stemcells, embryoid body formation and culture were carried out as previously described .
All cell cultures were incubated at 37°C in a 5% CO2 atmosphere.
RA and RARβ agonist treatment
Neurospheres were passaged as described above and single cells were seeded in 6-well plates at a density of 100 cells/mm2 with fresh culture medium containing different concentrations of all-trans-RA (10-9, 10-8 and 10-6 M; Sigma Chemical Co.) dissolved in dimethyl sulfoxide (DMSO; Sigma Chemical Co.). Control cells were cultured with the same dilutions of the RA vehicle, DMSO. Some cultures were treated with a RARβ-specific agonist  at a concentration of 10-8 M dissolved in DMSO, which we observed is effective in these cultures (R Martín-Ibáñez et al., in preparation). NPCs were allowed to grow for 3 DIV and then were pelleted for RNA extraction and RT-PCR analysis or processed for BrdU immunocytochemistry.
Mixed neuron-glial LGE primary cultures were grown for 3 DIV and RA or RARβ agonist (BMS641) dissolved in DMSO was added to the medium at the concentration of 10-8 M. Fresh RA/RARβ agonist was added every 24 hours of culture and 3 hours prior to fixation.
Embryoid bodies were formed as described above and RA was added to the culture medium at different concentrations (10-9, 10-8 and 10-6 M). They were grown for 4 DIV and then pelleted for RNA extraction and RT-PCR analysis.
To over-express NolzI, cells were transfected with the pNolz-IRES2-DsRED plasmid, that was obtained by clonation of the human NolzI contained in the pOTB plasmid (MGC full-length (IRAU) collection, clone ID 4053098) into EcoRI and SmaI restriction sites of p-IRES2-DsRED-Express plasmid (BD Biosciences) coding for DsRED fluorescent protein. As a control, we used pIRES2-DsRED-Express empty plasmid. Tranfection was made with 9 μg of NolzI-RED or RED plasmid.
To reduce NolzI expression, three different siRNAs against NolzI mRNA were used (Silencer Pre-designed siRNAs, IDs 89661, 169777, 89565, Ambion, Applied Biosystems, Foster City, CA, USA). Tranfection was made with 2 μM of each siRNA or 6 μM of negative control siRNA (Silencer Negative Control #1 siRNA, Ambion).
Low passage (four to seven) embryonic neurospheres were disaggregated and transfected by nucleofection following the manufacturer's protocol (Amaxa Biosystems, Lonza Iberica SA, Barcelona, Spain). Using the A33 Nucleofector program (Amaxa Biosystems), 5 × 106 cells were transfected. Viable cells were counted by trypan blue exclusion in a Neubauer chamber after nucleofection.
For BrdU incorporation assays, 65,000 cells/cm2 were seeded after nucleofection. Two days later cells were collected and incubated for 10 minutes in 24-well plates with Matrigel-treated coverslips and they were incubated for 10 minutes in BrdU-containing media at a final concentration of 2 μg/ml. Just after incubation, cells were fixed and processed for immunocytochemistry.
The cell cycle exit index was analyzed as described elsewhere . Briefly, we performed the same procedure as for BrdU incorporation except that neurospheres were pulsed with BrdU the same initial day of inducing hNolz over-expression and cells were fixed 3 days later. Thereafter, neurosphere cultures were processed for BrdU and Ki67 immunostaining. BrdU-positive but Ki67-negative cells were counted as the cells that left the cell cycle during the experimental period.
In order to analyze cell cycle duration, 10 neurospheres were plated 3 days after transfection in 96-well plates with complete medium and supplemented with 1 μM BrdU. Neurospheres were attached at 1, 3, 6, 12 and 24 hours after treatment in 96 well-plates pre-coated with Matrigel. Ten minutes later, cells were fixed in 4% PFA and processed for immunocytochemistry.
For the self-renewal assay, 65,000 cells/cm2 were seeded after nucleofection and the total number of neurospheres obtained 5 days later was counted (passage 0 (P0) after transfection). Cells were dissociated and 2,500 cells/cm2 were seeded and counted again on day 5 (passage 1 (P1) after transfection).
For differentiation after nucleofection, 50,000 cells/cm2 were seeded and the same protocol as for non-nucleofected cells was followed.
To over-express Nolz1 in primary cultures, we transfected the cells with the pLV-Nolz-IRES-EGFP plasmid or the pLV-IRES-EGFP plasmid, which encode hNolz and EGFP or EGFP only, respectively. The pLV-IRES-EGFP plasmid was generated using the pRRLsinPPT plasmid (pRRL) constructed by the Miami Project to Cure Paralysis Viral Vector Core Lab based on the lentiviral transducing plasmid developed by Naldini et al. . Briefly, the multiple cloning site (MCS) of the pRRL plasmid was substituted by the MCS-IRES-EGFP from the PRV-IRES-EGFP (Genetrix SL, Tres Cantos, Madrid, Spain) using the BamHI and the SalI restriction sites. To construct the pLV-Nolz-IRES-EGFP, the hNolz gene from the pNolz-IRES2-DsRED plasmid was cloned into pLV-IRES-EGFP between the MCS BamHI and XhoI sites. Primary cultures were transfected 24 hours after seeding with 0.5 μg of the corresponding plasmids per well (24-well plate). The transfection was performed using Lipofectamine LTX (Invitrogen SA), following the manufacturer's instructions. Three days after transfection cells were fixed with 4% PFA for immunocytochemistry analysis.
For DEAB (Sigma Chemical Co.) treatment, primary cultures were transfected 12 h after seeding as described above. Then, 24 h later DEAB was added to the medium at a concentration of 10-8 M. Three days after treatment cells were fixed with 4% PFA for immunocytochemistry analysis.
For RAR inverse agonist administration, primary cultures were transfected 12 h after seeding as described above. Thereafter, the RAR inverse agonist (BMS493 ) was added to the medium at a concentration of 10-8 M in DMSO, which was repeated every single day. At 3 DIV cells were fixed with 4% PFA for immunocytochemistry analysis.
Production of viral particles and cell transduction
To over-express Gsx2, the human Gsx2 gene from the pcDNA-hGsx2 plasmid, kindly provided by Dr Peter Marynen (Université de Leuven, Belgium), was PCR-cloned into the retroviral vector pRV-IRES-EGFP using the MCS BamHI and XhoI sites.
For retrovirus production, 293T cells were plated at a density of approximately 6 × 104 cells per cm2. The following day, cells were transfected by a three-plasmid system (the pRV-Gsx2-IRES-GFP plasmid, the plasmid that expresses HIV-1 gag and pol genes, and the plasmid that expresses vesicular stomatitis virus G) using the calcium phosphate/DNA co-precipitate method. The transfection mixture remained on the cells for 7 h before the transfection medium was replaced with fresh medium. The supernatant from vector-producing 293T cells was recovered every 22 h during 3 days before being harvested, passed through a 0.45-μm-pore-size filter to remove producer cells, and then subjected to two centrifugations at 4°C and 22,000 × g for 90 minutes to concentrate the virus. The virus-containing pellet was dissolved in 1% bovine serum albumin. Viral concentrate (20 μl) from pRV-Gsx2-IRES-EGFP or pRV-IRES-EGFP was used to transduce 3.5 × 106 dissociated NPCs in a 6-well plate as described previously . Transduced NPCs were growth as neurospheres as described above for 5 DIV before being pelleted for RNA extraction.
Generation of the anti-Nolz1 antibody
Anti-Nolz1 polyclonal antibodies were obtained from the serum of immunized rabbits with a keyhole limpet hemocyanin (KLH)-conjugated oligopeptide coding for amino acids 2 to 14 of the Nolz1 sequence (MSTAPSLSALRSSKH; Figure 1A). Pre-immune serum was obtained from the same rabbits before immunization.
All immunostaining was performed using the following antibodies: polyclonal anti-NolzI (1:10,000), monoclonal anti-GFAP (1:500; Sigma Chemical Co.), monoclonal anti-Tuj1 (1:500; Sigma Chemical Co.), monoclonal anti-BrdU (1:50; Dako A/S, Glostrup, Denmark), monoclonal anti-MAP2 (1:200; Sternberger Monoclonals, Lutherville, MD, USA), polyclonal anti-nestin (Rat 401; 1:40; Developmental Studies Hybridoma Bank; The University of Iowa, Iowa), polyclonal anti-Tle4 (1:200; generous gift from Dr Stefano Stifani, McGill University), polyclonal conjugated FITC-GFP (1:200; ABCAM, Cambridge, UK), polyclonal anti-Ki67 (1:200; Thermo Fisher Scientific SLU, Alcobendas, Madrid, Spain). For Nolz negative controls, pre-immunization serum was used at the same concentration as NolzI antibody-containing serum. Preparations were counterstained with DAPI to visualize the nucleus.
No signal was detected in control immunostaining assays in which the primary antibody was omitted.
In situ hybridization
We analyzed the expression of several genes by radioactive in situ hybridization as described elsewhere . The following oligonucleotide probes were used: mouse NolzI - complementary to nucleotides 3,226 to 3,266 of the NolzI sequence (GenBank accession number NM_145459); mouse Tle1 - complementary to nucleotides 1,942 to 1,983 of the Tle1 sequence (GenBank accession number NM_011599); mouse Tle2 - complementary to nucleotides 1,351 to 1,389 of the Tle2 sequence (GenBank accession number NM_019725); mouse Tle3 - complementary to nucleotides 3,434 to 3,474 of the Tle3 sequence (GenBank accession number NM_001083927); mouse Tle4 - complementary to nucleotides 1,614 to 1,652 of the Tle4 sequence (GenBank accession number NM_0011600).
Quantitative PCR assays
Expression of several genes was evaluated by Q-PCR assays performed as previously described , using the following TaqMan® gene expression assays (Applied Biosystems): 18S, Hs99999901_s1; nestin, Mm00450205_m1; β-tubulin III, Mm00727586_s1; GFAP, Mm00546086_m1; MAP2, Mm00485230_m1; Gsx2, Mm00446650_m1; mouse NolzI, Mm00520908_m1; Raldh3, Mm00474049_m1; RARα, Mm00436264_m1; RARβ, Mm01319674_m1; RARγ, Mm00441083_m1; CRBP1, Mm00441119_m1; Cyp26b1, Mm00558507_m1; Tle4, Mm01195160_m1. To specifically recognize hNolz, a customized Taqman® Assay was designed consisting of the following primers: forward, CCTCGCCCTCCTCCAAAC; reverse, GCCCGATTTGGTGTCCTTGT; reporter, TCTCCTCGGTTGCCTCC. To provide negative controls and exclude contamination by genomic DNA, the reverse transcriptase was omitted in the cDNA synthesis step, and the samples were subjected to the PCR reaction with each TaqMan® gene expression assay.
Analysis and quantification was performed with the Comparative Quantitation Analysis program of the MxPro™ Q-PCR analysis software version 3.0 (Stratagene, La Jolla, CA, USA), using the 18S gene expression as internal loading control. All Q-PCR assays were performed in duplicate and repeated for at least three independent experiments. The results were expressed as relative levels with respect to the expression of the same gene in the control condition, considered as 100%.
We analyzed the levels of Nolz1 protein in transfected neurospheres or in the striatum at different developmental stages. Samples (at least n = 3 per time point) were prepared and processed for western blotting as described elsewhere . Blots were incubated overnight at 4°C with anti-NolzI antibody (1:50,000). The secondary antibody was a horse radish peroxidase-conjugated anti-rabbit IgG (1:3,000; Promega Biotech Iberica, SL., Madrid, Spain) and the signal was developed using the ECL western blotting analysis system (GE Healthcare Europe GMBH, Cerdanyola del Vallès, Barcelona, Spain).
In order to determine the role of NolzI on the proliferation of progenitor cells in vitro, we counted the number of cells that incorporate BrdU. BrdU- and EGFP-positive (transfected) cells were detected by immunocytochemistry and the total number of cells determined by DAPI counterstaining. For over-expressing experiments the results were expressed as the percentage of proliferating cells with respect to the transfected (EGFP-positive) cells, while for siRNA experiments, the results were expressed as the percentage of proliferating cells with respect the total number of cells (n = 5). We also counted by phase contrast the total number of neurospheres 5 days after hNolz or control transfection.
Cell death was evaluated by counting the number of apoptotic nuclei stained with DAPI after hNolz over-expression. The results were expressed as the percentage of dying cells with respect to transfected (EGFP-positive) cells (n = 4). Results were normalized with respect to control-transfected NSCs (considered as 100%).
We estimated the cell cycle time as previously described [1, 29]. Briefly, BrdU was added to neurosphere cultures during the last 1, 3, 6, 12 and 24 h of culturing. The number of BrdU-positive cells and the total number of cells, determined by DAPI counterstained nuclei, in each neurosphere was counted. The percentage of proliferating cells was calculated for each time point after BrdU administration. Regression analyses of the active portion of each BrdU labeling curve were used to estimate cell cycle time assuming that all cells proliferate at the same rate and that every cell is labeled at the end of a single cycle . The r2 of this lineal correlation was used to calculate the cell cycle duration. We counted at least 30 neurospheres in each condition in 3 transfected cultures.
We analyzed the cell cycle index as the number of cells that retain BrdU but leave the cell cycle (Ki67-negative cells) after a 3-DIV pulse label. Thus, we counted the fraction of BrdU+/Ki67- cells and normalized to the total number of BrdU-positive cells in the culture. Results were expressed as absolute percentages in each condition.
To determine the effect of Nolz1 on the differentiation of LGE primary cultures, we counted the number of cells per coverslip overexpressing hNolz or EGFP that colocalized with different markers, such as nestin, Tuj1 and MAP2 3 or 5 days after the transfection. The results are expressed as the percentage of transfected cells colocalizing with the different markers with respect to the total number of transfected cells. Between 50 and 200 transfected cells per coverslip were counted per transfection (n = 3 to 5).
Brain slice electroporation
Coronal brain slices (250 μm) from E14.5 mice embryos were obtained with a vibratome. Slices were plated onto culture membranes with minimum essential medium supplemented with 10% FBS and 50 U/ml penicillin-streptomycin. After 1 hour in the incubator, media was changed to Neurobasal supplemented with B27, 1 mM HEPES, 50 U/ml penicillin-streptomycin and 2 mM L-glutamine. Two hours later, slices were electroporated with 8 μg of DsRED-Nolz or DsRED plasmid. After 48 hours in culture, electroporated slices were fixed during 2 hours with 4% PFA, dehydrated with increasing ethanol concentrations and stored until processing for immunohistochemistry.
We counted the number of Ki67-positive cells present in the electroporated zone of brain slices. First, we took a picture of the slices showing the RED fluorescence of the electroporated plasmids prior to fixation. Then, the slices were immunostained against Ki67 and the number of positive cells included in the electroporated area for the control side (DsRED plasmid) or experimental side (Nolz1-DsRED) were counted. The area to count was fixed by delineating the electroporated DsRED positive area in the pre-immunostaining image using ImageJ, and transferring this area to the Ki67 stained image. At least four different slices were counted for each condition.
Luciferase RARE reporter assay
To monitor the RARE activity we used the Cignal RARE Reporter Assay Kit (SABioscience Corporation, Frederick, MD, USA) following the manufacturer's instructions. Mixed neuron-glial LGE primary cultures were performed as described above and 24 hours after seeding cells were transfected with the inducible RARE-responsive mixture and the pLV-Nolz-IRES-EGFP or the pLV-IRES-EGFP plasmids. The transfection was performed using Lipofectamine LTX (Invitrogen SA) following the manufacturer's instructions. Two days later, cultures were processed to evaluate luciferase using the Dual-Luciferase Reporter Assay System (Promega). The firefly/Renilla luciferase ratio was calculated for each well, and results are expressed as the mean of four independent experiments and normalized with respect to control-transfected primary cultures (considered as 100%).
All results are expressed as the mean of independent experiments ± standard error of the mean. Results were analyzed using the Student's t-test or one-way ANOVA followed by the Bonferroni post-hoc test.