To analyze Patz1 expression levels across development in different cellular populations we took advantage of different publicly available datasets. Normalized fragments per kilobase of exon model per million reads mapped (FPKM) expression values were downloaded from “Mouse RNA-seq time-series of the development of seven major organs” experiment of EMBL-EBI Expression Atlas (EMBL-EBI Expression Atlas, 2020; Papatheodorou et al., 2020). The experiment comprises gene expression data (mRNAseq) from different tissues (among them heart, brain, liver, testis, ovary) across several time points from embryonic day 10.5 (E10.5) to postnatal day 63 (P63). Expression data relative to Patz1 in this dataset are represented in the heatmap (A) and in the barplot (B) in Figure 1. Two single cell RNA-seq dataset were also analyzed (©2020 10× Genomics 10K E18 Brain cells; Telley et al., 2019- GSE118953). We downloaded row counts and metadata relative to 10,000 cells from E18 mouse brain from the 10X Chromium website and data relative to the Telley dataset using GSE118953 accession number in the GEO database¹. Clustering analysis was performed with R version 4.0.3 (2020-10-10) using Seurat package version 3.2.3 (Satija et al., 2015; Butler et al., 2018; Stuart et al., 2019). The dataset was normalized using standard SCT normalization pipeline². A cell cycle score was also assigned to each single cell using the Seurat pipeline but was not used as variable to regress during normalization because considered biologically relevant. We considered the top 60 statistically significant principal components as input for UMAP dimensional reduction, using the DimPlot function. To identify cellular clusters, we used FindClusters function considering 0.2 resolution for both datasets. The robustness of the clustering was validated using an in silico downsampling approach. To annotate cell clusters, top 20 differentially expressed genes (markers) per cluster were considered and the most relevant were plotted using DotPlot function to check cell type specificity. Data relative to 10X-Chromium dataset were shown in Figures 2A–D, whereas data relative to the dataset of Telley et al. (2019) were shown in Figures 3A–C. We also analyzed bulk-RNAseq data derived from purified NS/PCs from adult murine SVZ (Belenguer et al., 2021). The normalized expression matrix of the dataset of Belenguer et al. (2021) was downloaded from GEO database using GSE138243 accession number. Expression data relative to Patz1, Cd9, Slc1a3, Tbr2, Dcxa, and Cd24a genes were reported in the heatmap (scaled data) and in the floating box plot (normalized expression) in Figures 2E,F.

Patz1-knockout mice were previously described (Valentino et al., 2013b). C57/BL6 wild type mice were used to collect mouse brains at different embryonic and post-natal stages of development. Mice were maintained under standardized non-barrier conditions in the Laboratory Animal Facility of Istituto dei Tumori di Napoli (Patz1-knockout mice and their wild type controls) and in the Animal Facility of the Institute of Genetics and Biophysics (C57/BL6 wild type mice showed in Figure 1) in Naples, Italy. All studies were conducted in accordance with the Italian regulations for experimentations on animals (prot. no. 576/10 approved by the Italian Ministry of Health on 3 February 2011).

Tail pieces from both embryos and adult mice were lysed in tail digestion buffer (50 mM Tris, pH 8.0; 100 mM EDTA, pH 8.0; 100 mM NaCl; 1% SDS) plus proteinase K (0.6 mg/ml) over-night at 55°C. Genomic DNA was extracted by phenol/chloroform/isoamyl alcohol (1:1), precipitated with 95% ethanol and dissolved in TE (Tris/EDTA) buffer, pH 7.5. A set of three primers was used to detect both normal and mutant Patz1 alleles by polymerase chain reactions (PCR), as previously described (Monaco et al., 2018). The primers used were: 5′—GCC TTC TTG ACG AGT TCT TC–3′/5′—CCA CAC CAT CAA AGT TGG–3′ for the knockout allele; 5′—AAG CAA GTG GCT TGT GAG–3′/5′—CCA CAC CAT CAA AGT TGG–3′ for the wild type allele.

For embryonic brain collection, C57/BL6 mice were crossed. Noon of the day on which the vaginal plug was detected was considered as 0.5 days post coitum (dpc) in the timing of the embryo collection. Timed pregnant females were sacrificed by cervical dislocation and embryos were dissected from decidual tissue in cold PBS. For each embryonic stage two different embryonic litters were analyzed. For each litter, brains from three embryos were dissected into the different territories. Corresponding brain territories of the same litter were pulled together to increase starting material. At each postnatal stage, two mice were sacrificed by cervical dislocation and brains were extracted and then dissected into the different territories. Each postnatal brain territory was analyzed separately.

The dissection was started separating the head from the body using scissors, then, a midline incision along the integument from the neck to the nose was made to expose the skull or its primordium. Any residual tissue was removed using scissors or tweezers. Before to proceed to surgery, all the dissection instruments were disinfected with 70% ethanol. Moreover, the instruments were sterilized with 70% ethanol before the removal of each tissue in order to avoid contaminations.

Five hundred microliter of Trizol reagent (Sigma, St. Louis, MO, United States) were added to each sample, except for adult cerebellum and mesencephalon in which 1 ml was added, for subsequent RNA extraction according to the manufacturer’s protocol. Recovered RNA was reverse transcribed using M-MuLV Reverse Transcriptase (Roche Diagnostics). qPCR was performed with SYBR Green PCR Master Mix (Life Technologies) under the following conditions: 10 min at 95°C, followed by 40 cycles (15 s at 95°C and 1 min at 60°C). Each reaction was performed in duplicate in at least three independent experiments. The 2^–ΔΔCt method (Livak and Schmittgen, 2001) to calculate the relative expression levels (fold change) was used. Amplification of two housekeeping genes, Glucose-6-Phospate Dehydrogenase (G6PD) and Ribosomal Protein S9 (RPS9) genes, was used to normalize the amount of cDNA. For analysis of brain territories in Figure 1, E11.5 or E14.5 samples have been used alternatively as calibrators to calculate the relative expression. For analysis of neurospheres, both proliferating and differentiated in Figure 6, one of control samples, randomly chosen and different for each independent experiment, was used as calibrator to calculate the ΔΔCt. Sequences for forward (Fw) and reverse (Rv) primers were as follows: Patz1 Fw = 5′-GAG CTT CCC CGA GCT CAT-3′; Patz1 Rv = 5′–CAG ATC TCG ATG ACC GAC CT–3′; Nanog Fw = 5′–GCC TCC AGC AGA TGC AAG–3′; Nanog Rv = 5′-GGT TTT GAA ACC AGG TCT TAA CC-3′; Nestin Fw = 5′—GAG AAG ACA GTG AGG CAG ATG AGT T–3′; Nestin Rv = 5′—GCC TCT GTT CTC CAG CTT GCT–3′; Tubb3 Fw = 5′—GAA TGA CCT GGT GTC CGA GT–3′; Tubb3 Rv = 5′—CCG ATT CCT CGT CAT CAT CT–3′; G6pd Fw = 5′—CAG CGG CAA CTA AAC TCA GA–3′; G6pd Rv = 5′—TTC CCT CAG GAT CCC ACA C—3′; Rps9 Fw = 5′—CTG GAC GAG GGC AAG ATG AAG c—3′; Rps9 Rv = 5′—TGA CGT TGG CGG ATG AGC ACA—3′.

Heterozygous Patz1-mutant mice were mated and vaginal plug monitored each day to establish the coitum day. At 15.5 days dpc pregnant mice received a single intra-peritoneal injection of 5′-bromo-2′-deoxyuridine (BrdU) (50 mg/kg) and, 2 h later, they were euthanized. One embryo for each genotype was extracted, post-fixed in 4% paraformaldehyde for 24 h and then kept in 70% ethanol at 4°C before to be processed for immunofluorescence analysis. Embryos were then dehydrated through ascending ethanol, washed in Toluene and Toluene-Paraffin 1:1, included in paraffin and then sectioned using a microtome to undergo histological analyses (Giorgio et al., 2014). For details see elsewhere (Q4Lab, 2016). Briefly, cutting angle and thickness have been set, respectively, at 5 and 8 μm. Paraffin sections have been dewaxed in xylene (pre-warmed at 55–65°C) twice for 5 min and then rehydrated through sequential steps in gradually less concentrated ethanol solutions and rinsed in tap water, at the end. Slides have been heated in a 10 mM sodium citrate pH 6.0 in a microwave to expose the antigens (Pisapia et al., 2020). Tissue sections have been incubated 1 h at R.T. with a solution 1% milk, 10% FBS, 1% BSA, 1X Sodium Azide, 0,5% Tween in PBS 1X and then O.N. at 4°C with a rat-monoclonal BrdU antibody (Novus Biologicals NB500-169) at dilution 1:400. Sections have been rinsed and incubated 1,5 h at R.T. with a Donkey α-rat Alexa Fluor- 488 secondary antibody (Molecular Probes A-21208) at 1:400 dilution, counterstained with Hoechst 1:5,000 and then mounted.

Sections were dewaxed in xylene, hydrated in graded series of alcohol and subjected to heat-induced antigen retrieval (Guadagno et al., 2017). After blocking endogenous peroxidase activity, the tissue was incubated for 1 h at R.T. with the blocking solution previous described and then with our previously characterized polyclonal antibody able to recognize both human and mouse Patz1 proteins (Valentino et al., 2013a) at a dilution 1:40 (V/V) O.N. Subsequently, the slices were rinsed and incubated with the biotinylated secondary antibody, at room temperature, for 30 min. The bound antibody complexes were stained for 3–5 min with diaminobenzidine, and slides were then counterstained with hematoxylin (30 s), dehydrated and mounted.

Before surgery all the dissection instruments have been disinfected with 70% ethanol. Adult mice or pregnant females were euthanized. SVZ dissection from 4 wild type, 3 Patz1^+/– and 2 Patz1^–/– adult mice or total forebrain dissection from 6 wild type, 6 Patz1^+/– and 2 Patz1^–/– 13.5 dpc embryos was carried out under stereoscopic control in ice-cold disodium phosphate buffer solution supplemented with glucose 0.6% (PBSg). The dissected samples were transferred into a new sterile cold PBSg buffer and brought under the hood where they were mechanically dissociated in single cell suspension.

For each sample, cells were counted and plated at the density of 3,000 cells/cm² in 14–16 ml of serum-free medium comprising 1:1 mixture of DMEM and F12 supplemented with 2 mM glutamax, N2 supplement (1:100), B27 minus Vitamin A (1:50) (Invitrogen), 20 ng/ml EGF (Sigma), 20 ng/ml bFGF (Sigma) and penicillin/streptomycin (1:100), into T75 neurosphere-flasks in a prewarmed culture medium, to obtain a single embryo cell culture. 1 × 10⁵ and all cells recovered from dissection were plated for embryonic and adult neurospheres, respectively. The number of neurospheres/brain were counted under microscope after 4 days in culture for embryonic cells and 7 days for adult ones, and their diameter measured by ImageJ. Only neurospheres > 40 μm in diameter were considered (Scopa et al., 2020). Primary neurospheres (P0) were then disaggregated by incubating with TripleExpress dissociation reagent (Invitrogen) for 5 min at room temperature. After the incubation, the dissociation reagent was aspirated, the neurosphere were mechanically dissociated and replated (1 × 10⁵/plate) to form secondary neurospheres (P1). Initial self-renewal activity was evaluated by counting the number of P1 neurospheres/cell seeded ^∗ 100.

For differentiation assay, neurospheres at passage 3 were dissociated to obtain single cell suspension, as described above and plated on poly-D-lysine coated multiwell plates at a density of 50,000 cells/cm². Cells were grown in absence of growth factors for 6 days in a 1:1 mixture of Neurobasal and DMEM/F12 medium supplemented with B27 plus Vitamin A (1:50), 2 mM glutamax (Invitrogen) and penicillin/streptomycin (1:100). Half of the differentiation medium was replaced every other day.

ImageJ/Fiji software was used to perform image analysis (Rueden et al., 2017). To compute the percent of BrdU⁺ cells (Figure 3F), 10 different images were analyzed per sample by defining the number of BrdU⁺ cells over Hoecsht⁺ cells on 5 different regions of interest (ROIs) manually set on VZ/SVZ using Cell Counter ImageJ Plug-in. The lateral ventricle area was computed over 5 different images per sample spanning from anterior to posterior regions, by segmenting Hoecsht signal and automatically defining ROIs over segmentation using ImageJ default threshold algorithm (Figure 3G). BrdU intensity signal per cell was computed by segmenting single positive nuclei of 3 different images per sample using ImageJ MaximalEntropy threshold function and computing the Mean Intensity arbitrary unit (a.u.) score using the Analyze Particle function. Representative quantifications of this analysis are shown in the dot plot in Figure 3H.

To compute the percent of β-III-tubulin⁺ cells (Figure 6F), 5 different images were analyzed per sample by defining the number of β-III-tubulin⁺ cells over Hoecsht⁺ cells using Cell Counter ImageJ Plug-in. NeuroJ Plung-in was used to perform neurite guided tracing and perform number (Figure 6H), length (Figure 6I), and fluorescence score (Figure 6J) analysis of 3–5 neurons per image for at least 3 images per sample.

Pearson correlation test was performed among the Patz1 expression levels detected and the age of the analyzed samples (Figure 1) and between Patz1 and Nanog expression in Patz1^+/+ and Patz1^+/– adult neurospheres (Supplementary Figure 1). Ordinary one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test was applied for comparison of all the other sets of data. The values analyzed are the average ± SD of at least two biological replicates and three independent experiments. All tests were assessed using GraphPad Prism 7 software, La Jolla (CA), United States. Statistical significance was indicated by ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001; ^****p < 0.0001 vs. wild type control when not differently specified in the figures by linking lines.

To get insight into Patz1 expression pattern during embryonic development, we took advantage of the Expression Atlas publicly available dataset containing RNA-seq data from different tissues across multiple timepoints during both embryonic and postnatal development (Figures 1A,B). As shown in Figure 1A, Patz1 expression is frequently higher during embryogenesis than in postnatal and adult stages, with the strongest expression of Patz1 in embryonic brain respect to the other tissues analyzed (heart, liver, ovary, and testis). Interestingly, during brain development, a decreasing trend of Patz1 expression with respect to time, from E10.5 to P63, is quite evident (Figure 1B).

To confirm and expand these findings, we analyzed, by means of qPCR, Patz1 transcript distribution levels in specific mouse brain territories at different time points during both embryonic (11.5, 14.5, and 17.5 dpc) and postnatal (14 days and adults) life. The results are shown in Figure 1C and confirm the existence of a temporal Patz1 expression gradient, with higher levels of expression at early embryonic stages, that decrease during embryogenesis and even more during postnatal (P14) and adult life, when Patz1 expression reaches the minimum measured level. This trend is detectable in the whole brain (Pearson correlation index = −0.6249; p = 0.0014) and also in the specific structures analyzed, as reported in Figures 1C–E. Conversely, there is no preferential spatial distribution at the different embryo and adult stages. Interestingly, looking at the relative expression of Patz1 in the adult, a higher expression level was detected in telencephalic structures (olfactory bulbs, cortex, and hippocampus), in which neurogenesis mainly occurs after birth, respect to the most posterior ones (Figure 1C).

Given the complex cellular heterogeneity and temporal dynamics of embryonic and adult brain tissue, we took advantage of publicly available single-cell and bulk RNA sequencing datasets to get insight into Patz1 expression level in neural progenitors during brain development. By analyzing the 10× Genomics 10K E18 Brain cells dataset, we found an enrichment of Patz1 expression in early apical radial glia/neural stem cells (RG/NSCs) and intermediate progenitors (IPCs) (Figures 2A–D). We also examined Patz1 expression in a dataset of bulk RNA sequencing from sorted types of progenitors from mouse adult SVZ (Belenguer et al., 2021), finding that Patz1 was expressed in the entire proliferating lineage with an enrichment in neural progenitor cells (NPCs) and neuroblasts (NBs) (Figures 2E,F). A further analysis, on single-cell RNA sequencing of sorted progenitor cells belonging to presumptive somato-sensory cortex of E12, E13, E14, and E15 mouse embryos (Telley et al., 2019), confirmed an enrichment of Patz1 expression during development in early apical radial glia/neural stem cells (RG/NSCs), mainly at the E12 stage (Figures 3A–C). Accordingly, Patz1 immunodetection at E15.5 showed a strong immunoreactivity in the cortical plate as well as in the VZ/SVZ of wild type embryo (Figure 3D). As expected, no Patz1 immunosignal was detected in corresponding brain sections of null mutants (Figure 3D). All these data point out to a possible function of Patz1 during embryonic neurogenesis. Then, to have a first indication of the impact of Patz1 depletion on proliferating cells in the periventricular area, we examined the in vivo proliferation inside the SVZ of 15.5 dpc Patz1-null and heterozygous mouse embryos, compared to wild type control, using a saturation BrdU (5′-bromo-2′-deoxyuridine) pulse-labeling method (Kuhn et al., 1996; Hayes and Nowakowski, 2002) that could label the entire pool of proliferating NS/PCs within a 12 h-period (Figure 3E). BrdU is a thymidine analog that incorporates into dividing cells during DNA synthesis. BrdU-labeling appears weaker and unevenly distributed in the SVZ of Patz1-null and heterozygous mutants compared to control brains. Quantitative analysis indicates a reduction of the number of proliferating cells (Figure 3F) and a significant decrease of BrdU incorporation level (Figure 3H) coupled to an overall reduction of lateral ventricle area (Figure 3G) in both Patz1 mutants compared with control. Interestingly, the heterozygous shows mean values lower than wild type and higher than Patz1 null mutant for all parameters analyzed.

To effectively measure the number of NS/PCs and their proliferation rate, we performed neurosphere assays on both Patz1 mutant and wild type brains. In this assay NS/PCs derived from SVZ of Patz1^–/–, Patz1^+/– and wild type adult mice (more than 1 month up to 21 months of age) as well as from 13.5 dpc embryonic telencephalon (from which SVZ arises), were allowed to proliferate in culture at very low density to form spheroid cell aggregates called neurospheres. The number of neurospheres was determined after 4 days in culture for embryonic and 7 days for adult neurospheres (Figures 4A, 5A). Under these conditions, primary neurosphere colonies are derived from single cells and can be used as a good model of the number of in vivo NS/PCs (Morshead et al., 2003; Coles-Takabe et al., 2008). In agreement with in vivo BrdU labeling data (Figures 3E–H), the number of primary neurospheres (P0) isolated from Patz1^–/– embryonic telencephalons was significantly lower than wild type counterparts (mean difference −1208; 95% CI of diff. −2377 to −39.62) (Figures 4B,C). The diameter of each neurosphere, an indirect measure of the ability of cells to grow and proliferate, was also determined, showing a significant reduction in neurospheres derived from Patz1^–/– embryos compared to the wild type (mean difference −12,36; 95% CI of diff. −24.48 to −0.2338) (Figure 4D). Furthermore, to assess the initial NS/PC self-renewal capability, primary neurospheres were dissociated into single cells after 4–7 days in culture and 10,000 cells per sample were cultured as above to form secondary neurospheres (P1) that were counted (day 8 and 14 for embryonic and adult neurospheres, respectively). As shown in Figure 4E, we observed a significant reduction (mean difference = −1.787; 95% CI of diff. −3.261 to −0.3126) in the number of P1 neurospheres from Patz1^–/– embryos compared to wild type controls, indicating the impairment of the self-renewal ability of Patz1-null NS/PCs. The reduction of the number (mean difference −332.9; 95% CI of diff−561.5 to −104.4), the diameter (mean difference −81.52; 95% CI of diff. −145.8 to −17.26) and the self-renewal capability (mean difference −3.425; 95% CI of diff. −6.053 to −0.7973) of NS/PCs lacking Patz1 was also evident in the SVZ of adult brain (Figure 5). Even though no significant difference was observed in the number and size of neurospheres from the heterozygous Patz1^+/– brains, compared to both Patz1^–/– and wild type samples, data from both embryonic and adult samples suggest for Patz1^+/– heterozygotes an intermediate trend between Patz1-null mutants and wild type controls (Figures 4C,D, 5C,D). Noteworthy, both embryonic and adult Patz1^+/– NS/PCs showed a self-renewal capability significantly higher than Patz1-null homozygotes and comparable to wild type (Figures 4E, 5E).

Overall, these data provide evidence for a role of Patz1 in the maintenance, proliferation and self-renewal capability of NS/PCs at both embryonic and adult stages.

RNA extracted from embryo-derived neurospheres, after at least three passages in culture (Figure 4A), under proliferating conditions, was used to analyze expression of genes characteristic of stemness (Nanog) and neural progenitor (Nestin) phenotype in proliferating neurospheres, or neuronal (Tubb3) cells. Nanog is a DNA binding homeobox transcription factor involved in embryonic stem cell proliferation, renewal, and pluripotency, which is a general marker of stemness; Nestin, neuroepithelial stem cell protein, is an intermediate filament protein that has generally been considered to be a marker of NS/PCs and is crucial for their self-renewal and survival; while Tubb3, coding for β-III-tubulin, correlates with neuronal differentiation (Zhang and Jiao, 2015). In agreement with the above data, both Nanog and Nestin genes were significantly downregulated in homozygous Patz1-null NS/PCs with respect to wild type (Figures 6A,B), confirming also at molecular level that Patz1 inactivation strongly affects the stem/progenitor phenotype. Moreover, Patz1-null heterozygous neurospheres also showed a significant downregulation of Nanog and Nestin expression with respect to the control ones. However, Nestin expression levels of Patz1^+/– NS/PCs settle in the middle between Patz1^–/– and wild type NS/PCs, whereas Nanog expression of Patz1^+/– NS/PCs was as low as in the homozygous Patz1-null neurospheres. These results might suggest a significant reduction of the Nestin⁺ and Nanog⁺ cells in Patz1-knockout NS/PCs compared to wild type controls. However, according to the role of Patz1 as a positive regulator of Nanog expression in mouse embryonic stem cells (Ow et al., 2014), Patz1 might activate Nanog expression also in NSCs and the reduced levels of Nanog could be partly due to the direct modulation of Nanog promoter by Patz1. Consistently, Patz1 and Nanog gene expression were highly correlated in adult neurospheres as well (Pearson correlation index = 0.9; p = 0.0069) (Supplementary Figure 1).

Conversely, the expression of Tubb3 did not change significantly in the three groups (Figure 6C), even though a slight increasing trend was visible among the different genotypes from Patz1^+/+ to Patz1^–/– proliferating neurospheres, which might suggest an accelerated differentiation into neurons of the Patz1-null neurospheres. To verify this hypothesis, neurospheres were cultured for 6 days in differentiating conditions and neuronal differentiation was confirmed by β-III-tubulin immunodetection (Figure 6D). Tubb3 transcript levels after neurospheres differentiation did not show any significant difference among the different genotypes (Figure 6E), as well as the number of β-III-tubulin⁺ cells detected (Figure 6F) and the mean fluorescence level per neuron (Figure 6G). However, both number and length of neurites were higher in Patz1^–/– than in both Patz1^+/– and Patz1^+/+ cells (Figures 6H–J), indicating that Patz1-null neurospheres can go toward an enhanced neuronal differentiation compared to control. Consistent with the genotype, Patz1 expression in heterozygous was about half the level of wild type cells, while it was not detected at all in Patz1-null samples (Figure 6K). Interestingly, its expression was lower in differentiated than in proliferating cells, once again supporting a role of Patz1 in counteracting neuronal differentiation.