All mice were maintained in the Hebrew University-Hadassah vivarium, following institutional IACU guidelines. Pros1^fl/fl (Burstyn-Cohen et al., 2009), Nestin-Cre (Tronche et al., 1999), and Nestin-GFP (Mignone et al., 2004) mice were previously described. In vivo experiments on adult mice were performed on mice 2–4 months of age.

Mice were deeply anesthetized by intraperitoneal Keatmine/Xylazine solution, and transcardially perfused as previously described (Burstyn-Cohen et al., 2012). The brains isolated, immersion-fixed overnight at 4°C in 4% PFA in PBS, cryopreserved for 24 h at 4°C in 30% sucrose/PBS, and embedded in OCT (Tissue-Tek). Fifty micron-thick cryosections were collected into PBS and stored in Methanol at −20°C until used. For immunostaining, sections were washed three times in PBS, blocked (5% FBS, 0.1% Triton X-100 in PBS) 1 h at room temperature, followed by primary antibody incubation overnight at 4°C. Next, after PBS washes, sections were incubated with a fluorophore-coupled secondary antibody in blocking buffer for 1 h at room temperature, Hoechst counterstained, and mounted with Fluoromount G (Southern Biotech). TUNEL staining was performed using the In situ Cell death Detection Kit (Roche) according to manufacturer's instructions. Images were taken with an Olympus BX51 fluorescent microscope mounted with a DP72 camera, or with a Nikon C2 confocal LASER scanning microscope.

For in vivo proliferation studies, intraperitoneal injections of bromodeoxyuridine (BrdU, Sigma, 50 mg/kg) were given for 3 days, and animals were sacrificed 24 h after last BrdU injection. To detect BrdU, cryosections were incubated in 50% formamide for 2 h at 65°C, rinsed in 2XSSC for 15 min/ room temperature (RT). DNA was denatured in 2N HCl for 30 min at 37°C and neutralized with 0.1 M Boric Acid, followed by three washes in 1X TBS (5M NaCl, 1M Tris in DDW) and blocked in blocking buffer (3% FBS, 0.25% Triton X-100 in TBS) for 1 h at RT, followed by incubation with primary rat anti-BrdU for 72 h at 4°C. Next day, the sections were washed three times in 1XTBS and re-blocked in blocking buffer for 15 min at RT, followed by incubation with fluorescently-conjugated secondary antibody in blocking buffer for 1 h at RT. Sections were documented after nuclei staining, as described above.

For BrdU incorporation in neurospheres, cultures were incubated with BrdU (10 μM) for 4 h, fixed in 100% ice-cold methanol for 20 min on ice, and washed three times in 1XPBS. DNA was denatured in 2N HCl for 10 min at 37°C and neutralized with 0.1M Boric Acid for 10 min at RT, followed by three washes in 1X PBS and blocked in the above blocking buffer for 1 h at RT. Finally, spheres were incubated with anti-BrdU antibody and processed as described above. All antibodies used in this study are listed in Supplementary Table 1.

Staged embryos (morning of vaginal plug was considered E0.5) were isolated, forebrains dissected, and the meningeal membranes and blood vessels removed. Telenchephalic cortices were dissected and processed individually to maintain genetic homogeneity. Tail DNA was taken for genotyping. Tissues were mechanically dissociated and Trypsin digested (0.125%, Biological Industries) for 15 min at 37°C, then neutralized for 2 min with trypsin inhibitor (1.25 mg/ml, Sigma), and DNAse I-digested 600 U/ml, Sigma) for 2 min at RT. Serum-free media (DMEM/F12; 100 U/ml penicillin; 0.1 mg/ml streptomycin; 2 mM L-Glutamine) was added and cells were centrifuged at 1,200 rpm for 10 min (acceleration = 9, deceleration = 1). Pelleted cells were resuspended in growth medium (serum free medium as above, supplemented with 1x B27 (Gibco) and hbFGF (10 ng/ml, PeproTech Asia) and seeded (300,000 cells/ml) in 10 cm cell culture plates in growth medium. Fresh hFGF was added daily. Neurospheres were grown for 2 days in vitro (2 div), then mechanically dissociated and reseeded at a clonal density of 300,000 cells/ml to generate secondary neurospheres. For self-renewal assays, isolated NSCs (embryonic or adult-generated) were plated at clonal density of 2,500–5,000 cells per well in 2 ml growth medium as above supplemented with 20 ng/ml EGF (PeproTech Asia). All cultures were maintained at 37°C in 5% CO₂. To measure self-renewal, individual neurospheres were mechanically dissociated, then replated at a clonal density of 2,500–5,000 cells per 3.5 cm diameter well. Primary and secondary neurospheres were counted after 3 div. To evaluate subpopulations, neurospheres were acutely dissociated in Accutase solution (1X, Millipore, Temecula), plated on pre-coated poly-Orinithine (0.5 mg/ml, Sigma) sterile cover slides, allowed to adhere for 2–3 h and fixed immediately in 4% PFA to prevent differentiation. Cells were further processed for immunofluorescence as described above. PROS1 was purchased from Enzyme Research Laboratories (South Bend, IN). To determine PROS1 dosage we performed preliminary experiments using different PROS1 concentrations. Thirty-five nanomoles per liters of PROS1 was found to be effective, without apparent toxicity effects on NSCs.

Plasmid constructs were transfected into NSCs by electroporation, using the NEPA21 electroporation system (Nepa Gene, Japan), according to manufacturer's instructions. Mouse Pros1 cDNA was excised from the pcDNA6-mPros1 vector (a gift from Greg Lemke, Prasad et al., 2006) using EcoR1 and Xho1, and cloned using the same restriction sites into the pCAGGS expression vector harboring an IRES-nuclear GFP reporter. shBmi-1 and control sh-irr vectors were a gift from Dinorah Friedman-Morvinski.

Eight to twelve weeks old mice were deeply anesthetized and sacrificed. The brains isolated and placed in ice-cold HBSS (Biological Industries) containing 10 mM Hepes (Biological Industries) until dissected midsagitally; the dentate gyri isolated and kept in HBSS/10 mM Hepes on ice. All solutions were used ice-cold, unless otherwise stated. The tissue was enzymatically dissociated in 5 ml of Solution 1 [5.4 mg/ml D-Glucose (Sigma), 15 mM HEPES (Biological Industries) 680 ng/ml Trypsin (Sigma) and 700 ng/ml hyaluronidase (Sigma) in 1 × HBSS] for 30 min at 37°C, with gentle trituration after 15 min. Cells were mixed with 5 ml Solution 3 (40 mg/ml BSA (Sigma), 20 mM HEPES, 1 × EBSS (Biological Industries) passed through a 70 μm strainer, and centrifuged at 1,300 rpm for 5 min. The pellet was resuspended in 10 ml Solution 2 [308 mg/ml Sucrose (Sigma) in 0.5 × HBSS], and centrifuged at 2,000 rpm for 10 min, then resuspended in 2 ml Solution 3, and carefully loaded onto 12 ml of Solution 3, and pelleted by centrifugation at 1,500 rpm for 7 min. Cellular debris was removed by centrifugation; the cells resuspended in 15 ml medium in neurosphere growth medium [DMEM:F12, 100 U/ml penicillin; 0.1 mg/ml streptomycin; 2 mM L-Glutamine, 1 × B27 w/o vitamin A (Gibco), 10 mM Hepes]. EGF (20 ng/ml) and FGF2 (10 ng/ml) was supplemented daily. After 3 div, primary neurosphere cultures were resuspended in 15 ml growth medium, cellular debris was removed by centrifugation; the cells resuspended in growth medium and replated. Growth medium was replaced every 3 div, neurospheres were taken for analysis at 10 div. For self-renewal assays, at 3 div cells were treated as above to rid of cellular debris, and resuspended at 2,500 cells/1.5 cm diameter dish (24 well) thus are considered secondary neurospheres. Four days later, cells were collected, dissociated. and replated at the same density to generate tertiary neurospheres.

Total RNA was extracted from DG (carefully dissected and cleaned from adjacent tissue under a dissecting stereoscope) or cultured neurospheres with TRIZOL (Sigma). cDNA was synthesized (qScript cDNA synthesis kit, Quanta Biosciences). Real-Time reactions were performed in triplicate using KAPA SYBR FAST qPCR Kit (2X) following manufacture's recommendation on a CFX96 Real Time PCR system (Bio-Rad). Expression levels were calculated relative to Gapdh, using the ΔΔ threshold cycle (Ct) method. Primer sequences are provided in Supplementary Table 2. Quantitative RT-qPCR determinations were done in at least three biological replicates with similar results.

Protein was extracted from DG or neurospheres in lysis buffer (50 mM Tris pH7.5, 150 mM NaCl, 1%Triron X-100, 0.5% NP-40, 0.1% SDS, 0.5 mM EDTA) supplemented with protease inhibitor cocktail (Sigma), incubated on ice for 30 min, and cellular debris removed by centrifugation. Protein concentration was determined by the BCA method (Thermo Scientific). Samples (5–20 μg) were run on 10% acrylamide gels and transferred to PVDF membranes (Millipore). ECL signal was detected using a ChemiDoc™ MP Image System (Bio-Rad), and band quantification was done using ImageJ.

Cells were prepared as previously described (Moussaud and Draheim, 2010). Briefly, adult mice were transcardially perfused with cold 1X PBS. DG were isolated and a single cell suspension was prepared by enzymatic digestion [20 units/ml papain (Worthington) in 116 mM NaCl, 5. NS, no statistical significance 4 mM KCl, 26 mM NaHCO₃, 1 mM NaH₂PO₄, 1.5 mM CaCl₂, 1 mM MgSO₄, 0.5 mM EDTA, 25 mM glucose, 1 mM cysteine] for 60–90 min at 37°C, 5% CO₂. After incubation, fetal bovine serum (FBS) was added to quench enzymatic activity (20% final concentration in HBSS, Biological Industries). Cells were collected by centrifugation (200 g,7 min, RT), resuspended and incubated for 5 min at room temperature in 0.5 mg/ml DNAse I (Sigma), passed through a 70 micron cell strainer (BD bioscience), and centrifuged (200 g, 7 min). Cells were then separated by a gradient density centrifugation using 0.9 M sucrose in HBSS and centrifuged at 200 g for 20 min with slow acceleration and no brake, resuspended in 2% FBS in PBS, and incubated with primary antibody for 30 min on ice. Cells were twice washed in 2% FBS and incubated with secondary antibody for 30 min on ice. Samples were washed in 2% FBS, filtered through 70 micron mesh and read by an LSR II FACS (BD Biosciences, San Jose, CA) and analyzed using FlowJo software (Ashland, OR).

Data represent mean ± SEM of experiments. At least three independent experiments were performed per assay. Significance was determined using a Student's t-test, considering P < 0.001 as extremely significant (^***), P < 0.01 as highly significant (^**) and P < 0.05 as statistically significant (^*).

We recently reported that the conditional deletion of Pros1 in NSPCs led to a 52 and 57% increase in SVZ and SGZ NSPC proliferation, respectively (Zelentsova et al., 2016). Similar results depicting the increased NSPC proliferation in the SGZ are presented in Figures 1A,B. Moreover, increased NSPC proliferation was observed in culture from NSPCs isolated from either E14.5 cortices or adult SGZ NSPCs, demonstrating a similar role for Pros1 in NSPC proliferation regardless of their developmental stage or topographic location. Unexpectedly, the increased NSPC proliferation did not result in depletion of the NSPC pool. Rather, NSPC proliferation continued throughout adulthood, indicating functional NSPC proliferation in aged mice (Zelentsova et al., 2016).

We set to explore the consequences of augmented NSPC proliferation on the NSPC population. For this, we first assessed the numbers of Sox2⁺ cells located in the SGZ by immunostaining and found a 50% increase in the numbers of Sox2⁺ cells confined to the SGZ of Pros1-cKO adult mice, compared to control littermates (Figures 1C–E), suggesting increased numbers of NSPCs. To verify these Sox2⁺ cells are indeed NSPCs, we crossed our mice to the nestin-GFP reporter mice (Mignone et al., 2004) allowing the visualization and quantification of bona-fide NSPC based on nestin-GFP expression, and on morphological features distinguishing radial glial-like NSCs from horizontal NSPCs. We found a significant increase in the overall numbers of both radial glial-like QNP and horizontal NSPCs in Nestin-Cre⁺; Pros1^fl/+; Nes-GFP (cHet; N-GFP) mice, also seen by nestin-GFP immunoreactivity in western blots of DG extracts (Figures 1F–I). Glial-like radial NSCs were scored based on both nestin-GFP expression and the presence of radial morphology with an extensive dendritic tree (Figures 1H,J). The highly-proliferative horizontal NSPCs also express nestin-GFP, but their accurate scoring was not possible by IHC due to over-crowdedness of cells (Figures 1F–H). We therefore took advantage of GFAP downregulation as NSCs transit from the radial to horizontal mode (Filippov et al., 2003; Seri et al., 2004), while sustaining expression of nestin-GFP (Encinas et al., 2011), and FACS- analyzed single cell suspensions of dentate gyri. Radial quiescent GFAP⁺; Nestin-GFP⁺ and horizontal, proliferating GFAP⁻; nestin-GFP⁺ cells were scored. In agreement with our IHC analysis, we observed a significantly higher percentage of GFP-positive cells in Pros1-cHet mice than in control mice (Figure 1K and Supplementary Figure 1A). Additionally, the GFAP⁻; GFP⁺ population was also increased compared to control mice (Figure 1L and Supplementary Figure 1B), indicating an overall increase in both radial/quiescent and horizontal/amplifying NSPCs. Noteably, the deletion of a single Pros1 allele was sufficient to generate this phenotype. Taken together, these data demonstrate increased numbers of NSPCs following ablation of PROS1.

Pros1-cKO adult brains contain many cells in which Pros1 expression is retained, including astrocytes, endothelial cells, and microglia (Burstyn-Cohen et al., 2009; Butovsky et al., 2014; Zelentsova et al., 2016), in which Nestin-Cre is not active, and may thus contribute to these effects. Investigating NSC self-renewal in Pros1 null mice is impossible due to embryonic lethality caused by hypercoagulability and vascular dysgenesis (Burstyn-Cohen et al., 2009). We therefore tested the effects of PROS1 deficiency in isolated NSPC-derived neurosphere cultures, thus excluding any possible PROS1 complementation by Pros1-expressing niche cells. Wild type neurospheres isolated from E14.5 embryonic NSPCs grown for 10 days expressed PROS1, which co-localized with all three NSPC markers tested: Sox2, nestin, and vimentin (Supplementary Figure 2), indicating that PROS1 expression in NSPCs is maintained under culture conditions. Neurospheres generated from Pros1-cKO NSPCs E14.5 brains had significantly more NSPCs, as indicated by the number of cells positive for the Nestin, Sox2, vimentin and nanog (P < 0.01 for nestin and nanog; P < 0.001 for Sox2 and vimentin; N = 5; Figures 2A–G and Supplementary Figure 2). Spheres generated from adult SGZ NSPCs also had significantly more cells positive for NSPC markers (P < 0.004 for nestin and Sox2; P < 0.002 for vimentin; N = 4; Figures 2H–J). Similarly, elevated nestin, decreased DCX and increased S100b immune-reactivity were observed in cultured neurospheres (Supplementary Figure 3). These results are consistent with our in vivo observations indicating elevated numbers of NSPCs following Pros1 deletion, and indicate that increased NSPC numbers is a capacity maintained when grown in culture, and is common to NSPCs generated from both embryonic and adult brains.

Another feature of Pros1-cKO NSPCs cultured from E14.5 brains is that they yielded significantly larger neurospheres than those generated from control littermates. Pros1-cKO spheres appeared with a more dense morphology, including a distinct core, whereas individual cells were still apparent in control spheres, resembling immature spheres (Figures 3A,B). Primary neurospheres were dissociated and grown to generate secondary spheres. Both primary and secondary spheres from Pros1-cKO embryos were significantly larger than control spheres (Figures 3C,D), demonstrating that increased sphere size is inherent to NSPCs, and that neither NSPC numbers nor sphere size depend upon factors provided by niche cells. To understand whether the increased neurosphere size was also a feature shared by adult hippocampal NSPCs, we isolated, cultured, and assayed neurospheres generated from adult SGZ NSCs. As for E14.5 spheres, and in line with their increased NSPC proliferation, adult-generated spheres from Pros1-cKO mice produced significantly larger secondary and tertiary spheres than adult control NSPCs (Figures 3E–H). Taken together, these results show that PROS1 functions to maintain NSPC numbers in a cell-autonomous manner, and that this feature is maintained upon passages in culture. Additionally, Pros1 functions similarly in NSPCs generated from embryonic and adult SGZ, suggesting a common role in regulating NSPC homeostasis in development and adulthood.

In the process of adult neurogenesis, only a small subset of newborn neuroblasts survive to maturity, while most newborn cells undergo apoptotic death in the first 4 days of their life (Sierra et al., 2010). The remaining neuroblasts mature within 4 weeks, during which they extend dendrites, migrate into the subgranular zone and finally synapse and integrate within the pre-existing neuronal network, contributing to hippocampal functions such as learning and memory (van Praag et al., 1999; Deng et al., 2010; Murai et al., 2014). To understand whether the elevated numbers of NSPC observed in Pros1-cKO may be a result of increased neuronal cell death, we assessed cell death by the Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay. We found control and Pros1-cKO DG had similar numbers of TUNEL-positive and cleaved caspase 3-positive (C-Casp3) cell counts (Supplementary Figures 4A–F). Similarly, comparable levels of apoptotic cells were measured in control and Pros1-cKO cultured neurospheres by TUNEL and C-Caspase 3 from embryonic and adult-derived NSPCs (Supplementary Figures 4G–M) indicating that loss of Pros1 does not enhance apoptotic cell death, or delay the clearance of newborn cells. Taken together, enlarged neurosphere size, and lack of increased cell death, all indicate the viability of Pros1-cKO newborn cells.

The generation of larger-sized neurospheres could stem from either increased NSPC proliferation, increased NSPC self-renewal, or both. The increased NSPC proliferation following Pros1 ablation was recently reported (Zelentsova et al., 2016), however the direct impact of Pros1 ablation on regulation of NSPC self-renewal was not previously investigated. Most NSPCs in the adult DG are self-renewing and multipotent (Suh et al., 2007; Bonaguidi et al., 2011). Maintaining the delicate balance between self-renewal and differentiation is critical to brain homeostasis, as it preserves a constant pool of NSPCs for future tissue renewal, while generating the brain's functional cells at the same time. Given the increased number of cells expressing NSPC markers in Pros1-cKO SGZ (Figures 1C–L) and in isolated NSPCs (Figures 2A–J) we hypothesized that loss of PROS1 would affect the self-renewal capacity of NSPCs. We next performed a direct self-renewal assay which revealed that embryonic Pros1-cKO NSPCs have a significantly higher capacity to generate primary and secondary neurospheres, as is also the case for adult-generated SGZ NSPCs (Figures 3I–L), demonstrating that elimination of Pros1 expression in NSCs increases stem cell self-renewal in both embryonic and adult neurospheres grown in vitro. Importantly, increased numbers of NSPCs does not result from elevated levels or cell death, as TUNEL and cleaved-Caspase 3 assays yielded similar values for Pros1-cKO and control cells both in vivo and in vitro (Supplementary Figure 4). Taken together with our observation of increased radial and horizontal NSPCs in vivo (Figures 1F–L), with more numerous Nestin⁺, Sox2⁺, and vimentin⁺ cells in isolated neurospheres (Figures 2A–J) we conclude that deletion of Pros1 results in increased neural stem cell self-renewal.

Bmi-1 is a negative transcriptional regulator of p16^Ink4a and p19^Arf in NSCs (Molofsky et al., 2003, 2005). Bmi-1 deficiency in NSCs leads to reduced self-renewal (Molofsky et al., 2003), while its overexpression resulted in larger neurospheres with more self-renewing divisions (He et al., 2009). Given its role in NSC self-renewal; and the increased self-renewal observed in Pros1-deficient NSCs, we hypothesized the increased self-renewal observed in Pros1-cKO NSCs may be mediated through Bmi-1. We found elevated Bmi-1 expression in Pros1-cKO neurospheres derived from embryonic NSPCs (Figure 4A). Both p16^Ink4a and p19^Arf expression were significantly inhibited in Pros1-cKO spheres (Figure 4A), consistent with Bmi-1 being a transcriptional repressor for both genes. Akin to embryonic NSPCs, Bmi-1 expression was significantly increased in adult cKO NSPCs (Figure 4B). These results indicate Pros1 similarly affects the expression of Bmi-1 pathway genes in both embryonic and adult NSPCs.

To determine whether Bmi-1 expression and self-renewal were directly related to PROS1 levels, we performed a rescue experiment by electroporating a GFP-tagged Pros1 expression vector into Pros1-cKO neural stem cells, restoring Pros1 expression in cKO NSPCs to control levels (Figure 4C). Compared to treatment-naïve and control-treated (GFP-encoding vector) cells, rescuing Pros1 expression in cKO NSPCs repressed Bmi-1 expression (Figure 4D), and elevated the transcript levels of p16^Ink4a and p19^Arf, consistent with Bmi-1 being a repressor for both genes (Figures 4E,F). Phenotypically, the rescue of Pros1 expression in cKO NSCs reversed their increased self-renewal levels (Figures 4G,H).

Next, to verify that the elevated Bmi-1 levels in Pros1-cKO NSCs were directly related to increased self-renewal in our cells, we inhibited the elevated Bmi-1 levels in Pros1-cKO NSCs by means of sh-RNA (Figure 5A). This lead to a sharp increase in p16^Ink4a and p19^Arf expression (Figures 5B,C), reversing the increased proliferation and self-renewal phenotypes (Figures 5D–F). The decreased self-renewal following Bmi-1 knockdown in cKO NSCs also corresponded to restoration of stem cell numbers, as indicated by immune-detection of Nestin, Sox2, and vimentin (Figures 5G–I). Taken together, these results demonstrate that Pros1 negatively regulates NSPC self-renewal, through inhibition of Bmi-1.

Finally, to gain insight into the mechanistic action of PROS1, we took advantage of the secretory nature of PROS1 and tested whether NSPC cultures expressing and secreting PROS1 could rescue both Bmi-1 levels and the self-renewal phenotype in Pros1-cKO cells. Conditioned medium (CM) from control Pros1^fl/fl NSPCs (C-CM) rescued the expression levels of Bmi-1, p16^Ink4a and p19^Arf in Pros1-cKO NSPCs (Figures 6A–C), bringing them to control levels. Moreover, supplementation of Pros1-cKO NSPCs with purified PROS1 protein also reversed Bmi-1, p16^Ink4a and p19^Arf expression (Figures 6A–C), highlighting the direct effect of PROS1 on Bmi-1 signaling. The addition of purified PROS1 further rescued both the self-renewal phenotype of Pros1-cKO cells bringing them to levels comparable to those of control NSPCs (Figures 6D,E), and the numbers of cells expressing the NSC markers Sox2, nestin, and vimentin (Figures 6F–H). Consistent with lower levels of Bmi-1 (and higher p16^Ink4a and p19^Arf), cells were also less proliferative, as indicated by reversal of the high incorporation levels of BrdU (Figure 6I). Taken together, our results point to PROS1 as a novel regulator of NSPC homeostasis, regulating the self-renewal capacity of NSPCs through Bmi-1 and its downstream effectors p16^Ink4a and p19^Arf.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.