Phosphoserine phosphatase (PSP) metabolizes the conversion of l-phosphoserine to l-serine, classically known as an amino acid necessary for protein and nucleotide synthesis and more recently suggested to be involved in cell-to-cell signaling. Previously, we identified PSP as being enriched in proliferating neural progenitors and highly expressed by embryonic and hematopoietic stem cells, suggesting a general role in stem cells. Here we demonstrate that PSP is highly expressed in periventricular neural progenitors in the embryonic brain. In the adult brain, PSP expression was observed in slowly dividing or quiescent glial fibrillary acidic protein (GFAP)-positive cells and CD24-positive ependymal cells in the forebrain germinal zone adjacent to the lateral ventricle and within GFAP-positive cells of the hippocampal subgranular zone, consistent with expression in adult neural stem cells. In vitro, PSP overexpression promoted proliferation, whereas small interfering RNA-induced knockdown inhibited proliferation of neural stem cells derived from embryonic cortex and adult striatal subventricular zone. The effects of PSP knockdown were partially rescued by exogenous l-serine. These data support a role for PSP in neural stem cell proliferation and suggest that in the adult periventricular germinal zones, PSP may regulate signaling between neural stem cells and other cells within the stem cell niche.
Disclosure of potential conflicts of interest is found at the end of this article.
Neural stem cells are multipotent cells of the embryonic and adult brain that have the potential to produce multiple lineages of mature neural cells. Neural stem cells have traditionally been identified by retrospective assessment of their self-renewal and multipotency using techniques such as the neurosphere assay [1, 2]. Neural stem cells are found in the embryonic brain, as well as in the ventricular zone and in the subventricular zone of the lateral ventricle and in the subgranular zone of the dentate gyrus in the adult brain [3, 4].
Recent studies have described various spatial and temporal differences in neural stem cells [5, –7]. For example, in the embryonic brain, stem cells are highly proliferative, giving rise to neuronal, astrocytic, and oligodendroglial precursors at different stages of development. In contrast, the current model of adult neurogenesis suggests that slowly cycling neural stem cells in the subventricular zone (SVZ) give rise to rapidly cycling transient amplifying progenitor cells, which then produce immature neuroblasts for integration into the nervous system [8, , , , –13]. Thus far, it seems clear that, at least in the adult brain, the slowly cycling stem cells are glial fibrillary acidic protein (GFAP)-positive [14, 15], although GFAP-positive cells do not exist in the early embryonic brains . Furthermore, even in the adult brain, clearly not all GFAP-positive cells, or even the majority of them, are stem cells.
Both cell-extrinsic and cell-intrinsic factors have been shown to influence the maintenance and regulation of the neurogenic system in vivo [17, 18]. To identify further candidate factors, we used a strategy of stepwise screening using stem cell-containing culture systems in vitro [19, –21]. We performed gene expression profiling on cultures of neural and non-neural stem cells to identify genes enriched in neural stem cells. We further stratified these in vitro candidates by examining the regions in vivo of their expression to confirm their presence in stem cell-containing regions. A few genes, including phosphoserine phosphatase (PSP), were remarkably enriched in germinal zones of neurogenesis.
PSP has been characterized in various organisms for over 20 years as the enzyme that metabolizes the conversion of l-phosphoserine to l-serine, and it is the rate-limiting enzyme in the primary serine synthesis pathway [22, , –25]. No protein phosphatase activity has ever been reported for PSP. l-Serine is an amino acid necessary for protein and nucleotide synthesis, as well as potentially playing a role directly in regulating neuronal differentiation. d-Serine, glycine, and l-phosphoserine have all been shown to bind as cofactors or ligands to various extracellular receptors [22, , –25]. All three molecules are within one step metabolically of PSP.
The purposes of this study were to determine the PSP-expressing cell types in brain throughout development and to determine the function of PSP using in vitro neural progenitor cultures. In the brain, we identified high levels of PSP expression in neural stem or progenitor cells throughout brain development: it is highly enriched in cells in the germinal zone lining the ventricles during embryonic stages and in the GFAP-positive cells in neurogenic regions in the adult brain. Functional studies with overexpression and knockdown of PSP demonstrated both intrinsic and extrinsic roles for PSP in proliferation of neural stem cells from both embryonic and adult brains. These data suggest that PSP is expressed in the neural stem cell niche and may regulate proliferation of neural stem cells.
Materials and Methods
All experiments were performed using embryonic and adult CD1 mice obtained from the Jackson Laboratory (Bar Harbor, ME, http://www.jax.org) and with the approval of UCLA's Animal Research Committee following NIH guidelines.
In Situ Hybridization
Hybridization was performed essentially as described . Probes of two different fragments of PSP (accessions) or full-length PSP yielded the same results. Sense controls showed no labeling in all cases.
Development of Antibody
PSP antibody was produced by immunizing two New Zealand White rabbits with synthetic peptide to the C terminus of the protein (WYITDFVELLGELEE) conjugated to KLH. Western blot identified specific antisera that recognized an appropriately sized 25-kDa band that was absent in preimmune serum and peptide-blocked antisera. Antibody was then purified by affinity column (Sigma-Genosys, St. Louis, http://www.sigmaaldrich.com).
Postnatal CD1 mice were perfused transcardially with ice-cold phosphate-buffered saline (PBS) followed by ice-cold 4% paraformaldehyde in PBS, pH 7.4. Brains were removed, fixed in 4% paraformaldehyde overnight, sunk in 20% sucrose PBS, frozen in 4-methylbutane, and stored at −80°C until use. Sections (40 μm) were cut on a cryostat and stored in PBS 1% azide at 4°C until use.
Embryonic mice, 17 days postconception, were dissected from timed pregnant CD1 dams, and fixed in 4% paraformaldehyde for 10 days. They were then embedded in 2% low melting point agarose and cut on a Vibratome (Vibratome, St. Louis, http://www.vibratome.com) instrument into 100-μm sections and stored in PBS 1% azide until further use.
Free-floating sections were incubated overnight in 24-well plates on a rotator at room temperature in the presence of 0.1% azide, 0.25% Triton, and 5% normal goat serum in 500 μl of PBS and primary antibody at the following concentrations: anti-βIII Tubulin (TuJ1), 1:1,000 (MMS-435P; Covance, Princeton, NJ, http://www.covance.com); anti-mammalian achaete-scute homolog 1 (anti-MASH1), 1:20 (556604; BD Pharmingen, San Diego, http://www.bdbiosciences.com/pharmingen); anti-proliferating cell nuclear antigen (PCNA), 1:10,000 (M 0879; DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com); anti-5-bromo-2′-deoxyuridine (BrdU), 1:5,000 (PAB105P; Maine Biotechnology, Portland, ME, http://www.mainebiotechnology.com); anti-PDZ-binding kinase (anti-PBK) monoclonal, 1:50 (612,170; BD Transduction, San Jose, CA, http://www.bdbiosciences.com); anti-GFAP, 1:1,000 (AB1540; Chemicon, Temecula, CA, http://www.chemicon.com); anti-PSP, 1:500; anti-CD24, 1:1,000, with no Triton (sc-19651). 3-Phosphoglycerate dehydrogenase (3PGDH), 1:1,000, was a gift of Dr. Watanabe ).
For BrdU and PCNA, antigens were retrieved by incubating sections for 1 hour at 65°C in 50% formamide, 2× SSC, and 30 minutes in 2.0 N HCl at 37°C. Secondary antibodies were diluted 1:1,000 and included cy2-, cy3-, and cy5-conjugated antibodies (Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com) and Alexa 350-, 488-, 568-, and 594-conjugated antibodies (Molecular Probes Inc., Eugene, OR, http://www.probes.invitrogen.com). In all cases, no primary controls yielded no labeling.
For double labeling with 3pgdh and PSP, PSP labeling was performed as described above with Alexa 568-conjugated anti-rabbit secondary antibody, and then tissue was blocked for 30 minutes in PBS with 5% normal goat serum and 5% normal rabbit serum. 3pdgh was then incubated for 4 hours at 1:1,000 in the same buffer, followed by 30 minutes in Alexa 488 anti-rabbit secondary. No 488 signal was seen when the second primary was omitted, and PSP-positive puncta in hypothalamus were always negative for Alexa 488, demonstrating the efficacy of the blocking steps.
Nuclei were counterstained with 4,6-diamidino-2-phenylindole-containing mounting medium (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) or with Topro-3-iodide (Molecular Probes), a nuclear stain fluorescing in the far red range (650 nm), by exposing tissue sections for 5 minutes to a 20 μM solution in PBS.
All fluorescent images were acquired on either a Leica TCS-SP MP confocal and multiphoton inverted microscope (Heidelberg, Germany, http://www.leica.com) and a two-photon laser setup (Spectra-Physics, Irvine, CA, http://www.newport.com/spectralanding) or a Zeiss LSM 510 META confocal microscope (Carl Zeiss, Jena, Germany, http://www.zeiss.com), using lasers and filters appropriate for the fluorophores, and pseudocolored images were overlaid with Zeiss software or Adobe Photoshop (Adobe Systems Inc., San Jose, CA, http://www.adobe.com). Infrared wavelengths were most often pseudocolored blue. To confirm colocalization of antibodies, stacks of <1-μm-thick high-power confocal images were acquired and examined both as three-dimensional reconstructions and as animations moving through the z-plane with the Zeiss LSM image browser.
Immunocytochemistry of adherent progenitors was performed as described previously [20, 27], using the following antibodies: TuJ1 (1:500; Berkeley Antibodies, Covance, http://www.crpinc.com) and O4 (1:50; Chemicon). Primary antibodies were visualized with Alexa 568-conjugated (red) and Alexa 488-conjugated (green) secondary antibodies (Molecular Probes). Hoechst 33342 (blue) was used as a fluorescent nuclear counterstain.
Neural Progenitor Cultures
Neurosphere cultures were prepared as described previously . Briefly, cortical telencephalon was removed from E12 CD1 mice, and cerebral cortex was isolated from older animals (Charles River Laboratories, Wilmington, MA, http://www.criver.com). Cells were dissociated with a fire-polished glass pipette and resuspended at 50,000 cells per milliliter in Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 medium (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with B27 (Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com), 20 ng/ml basic fibroblast growth factor (bFGF) (Peprotech, Rocky Hill, NJ, http://www.peprotech.com), penicillin/streptomycin (Gemini Bioproducts, West Sacramento, CA, http://www.gembio.com), and heparin (Sigma-Aldrich). Growth factors were added every 3 days. For differentiation, culture medium was replaced into Neurobasal medium (Invitrogen) supplemented with B27, 2% fetal bovine serum (FBS), and 10-6 M of all-trans-retinoic acid (Sigma-Aldrich), on poly-l-lysine (PLL)-coated dishes, and maintained for up to 5 days. For secondary sphere formation assay, the primary spheres were dissociated and plated into 96-well microwell plates in a 0.2-ml volume of growth medium at 40,000 cells per milliliter, and the resultant sphere numbers were counted at 7 days.
To assay the influence of gene knockdown or overexpression, the neurosphere culture system was modified. Neurospheres were propagated for 1 week and then dissociated with trypsin (0.05%) followed by trituration with a fire-polished pipette. The cells were then placed in DMEM/Ham's F-12 medium with 2% FBS (Gibco-BRL) and plated onto polyornithine/fibronectin-coated glass coverslips . After 6 hours, the serum-containing medium was removed and the cells were placed back in the neurosphere growth medium without heparin and supplemented with bFGF (20 ng/ml). Transfection was then performed as described below. To assay the sphere-forming potential of the transfected cells, they were lifted off the plate with trypsin (0.05%), incubated briefly in medium containing 10% FBS to inactivate trypsin, spun, and then placed into neurobasal medium supplemented with B27, bFGF, and heparin. To assay the function of cells expressing enhanced green fluorescence protein (EGFP) driven by the maternal embryonic leucine-zipper kinase (MELK) promoter, neurospheres at 7 days in vitro (DIV) were plated onto coverslips as described above and transfected. Some cultures were then placed into neurosphere conditions to assay sphere-forming potential, whereas others were propagated and differentiated on the coated coverslips after transfection.
GFAP-Positive Astrocyte-Enriched Cultures
Primary astrocyte cultures were prepared from P1 mouse cortices as described previously . Briefly, as cells became confluent (12–14 DIV), they were shaken at 200 rpm overnight to remove nonadherent cells and obtain pure astrocytes and passaged onto PLL-coated coverslips for RNA collection or fibroblast growth factor stimulation.
Reverse Transcription Polymerase Chain Reaction
Total RNA was isolated from each sample using TRIzol (Gibco-BRL), and 1 μg of RNA was converted to cDNA by reverse transcription following the manufacturer's protocol (Impron, Promega, Madison, WI, http://www.promega.com). For semiquantitative reverse transcription-polymerase chain reaction (RT-PCR), the amount of cDNA was examined by RT-PCR using primers for the glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) gene as an internal control from 20–40 cycles. After correction for GAPDH signal for each set, the resultant cDNA was subjected to PCR analysis using gene-specific primers listed in supplemental online Table 1. The protocol for the thermal cycler was as follows: denaturation at 94°C for 3 minutes, followed by cycles of 94°C (30 seconds), 60°C (1 minute), and 72°C (1 minutes), with the reaction terminated by a final 10-minute incubation at 72°C. Control experiments were done either without reverse transcriptase and/or without template cDNA to ensure that the results were not due to amplification of genomic or contaminating DNA. Each reaction was visualized after 2% agarose gel electrophoresis for 30 minutes, and expression levels were compared between the cDNA samples on the same gel. PSP primers were as follows: sense, 5′-ccaggaaccgcggaggaaaactt-3′; antisense, 5′-cggctgtcggctgcatctcatc-3′. Sequences of other genes will be provided upon request.
Flow Cytometry and Sorting
Flow cytometry and sorting with LeX antibody  and propidium iodide were performed with a FACSVantage flow cytometer (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) using a purification-mode algorithm. Gating parameters were set by side and forward scatter to eliminate dead and aggregated cells. Background signals were determined by incubation of the same set of progenitors without primary antibody. Isolated LeX+, isolated LeX−, dividing (4n) cells, and nondividing (2n) cells were dissolved in TRIzol for RNA purification. E12 progenitors were labeled with LeX antibody (Invitrogen) for 30 minutes, and Alexa 530 was used for flow cytometry and sorting.
Cells were transfected using Lipofectamine 2000 (Invitrogen) following the manufacturer's protocol. For transfection of plasmid vectors, the cells were incubated with reagents for 6 hours with the primary progenitor cells. For transfection of the double-stranded siRNA complex, serial dilutions of siRNA from 5 to 200 nM were tested to obtain specific knockdown of the gene of interest, and 100 nM was chosen as the concentration for functional study. Incubation with siRNA complex was 6 hours with primary cells. For negative control of plasmid transfections, EGFP expression vector driven by the cytomegalovirus promoter was used. For negative control of siRNA transfections, either mock, negative control siRNA (catalog no. AM4611; Ambion, Austin, TX, http://www.ambion.com) or luciferase siRNA (catalog no. 4627; Ambion) transfection was used.
Vector and siRNA Synthesis
The full-length coding region of mouse PSP was amplified by PCR using mouse embryonic neurospheres as a template and subcloned into pGEM-T Easy vector (Promega). After sequence verification, the PSP fragment was subcloned into pCMV-Tag vector (Stratagene, La Jolla, CA, http://www.stratagene.com) at the NotI site. siRNA was synthesized using the Silencer siRNA Construction Kit following the manufacturer's instructions (Ambion). Sequences for double-strand RNA synthesis of PSP were as follows: antisense, 5′-aaattctgtggtgtggaggcctgtctg-3′; sense, 5′-aacctccacaccacagaatcctgtctg-3′.
PSP Expression Is Highly Enriched in Neural Stem and Progenitor Cells Cultured from Embryonic and Postnatal Cortices
During early development, neural stem cells in the ventricular zone proliferate rapidly and give rise to neuronal progenitors. We cultured these progenitor cells using the neurosphere method  and compared the expression of PSP mRNA and protein between proliferating neurospheres and sister cultures differentiated by withdrawal of growth factors (Fig. 1A–1D). For the analysis of the PSP protein, we raised antisera against a peptide corresponding to the C-terminus of PSP. The anti-PSP antibody was first tested by immunocytochemistry. When the PSP-Flag fusion protein was expressed in cultured neural progenitor cells, the signals of the PSP antibody colocalized with those of anti-Flag antibody (supplemental online Fig. 1). Western blot of protein from cultured neurospheres recognized a 25-kDa band that was blocked by preincubation with PSP peptide (Fig. 1B). The PSP protein was abundant in proliferating neurospheres, whereas differentiated cells expressed less PSP protein. Protein samples from muscle and liver in mice at postnatal day 0 were used as negative and positive controls for the Western blot, respectively . These results were consistent with mRNA expression in neural progenitor cells (Fig. 1B).
To determine whether PSP was universally expressed within the cultures or was restricted to a more progenitor-like population within neurosphere cultures, we used a cell surface antibody, LeX, to enrich neural progenitors from these cultures as described previously [28, 31] (Fig. 1C). Neurosphere-initiating neural stem cells in the embryonic cortices are exclusively found in the LeX-positive fractions . Strikingly, PSP was highly restricted to the LeX-positive neural progenitor population, similar to other stem cell-enriched genes, such as MELK and SOX2 [28, 32, –34] (Fig. 1C). Next, we separated embryonic neural progenitors into dividing (4n) and nondividing (2n) populations and examined gene expression (Fig. 1D). Similar to the results with the LeX antibody, SOX2 and MELK were enriched in the dividing population within neural progenitor culture, whereas Msi1 was detected in nondividing population. PSP expression was restricted to the dividing population of embryonic neural progenitors. These data suggest that PSP mRNA and protein are highly enriched and virtually restricted to proliferating neural progenitors in the developing central nervous system (CNS).
Postnatal neurogenesis is different from embryonic neurogenesis, and the neural stem cells in the adult SVZ are a subset of the GFAP-positive cells . These slowly cycling neural stem cells in the SVZ give rise to rapidly cycling transient amplifying progenitor cells, which then produce immature neuroblasts for integration into the nervous system [5, 6, 36]. To test whether PSP could be involved in adult neurogenesis, we used a cell culture system that simulates the transition of those cell types in vitro (Fig. 1E–1G). GFAP-positive astrocytes from P0 cortices (>95%) were cultured and then were stimulated with bFGF to produce GFAP-negative progenitors, followed by TuJ1-positive neuroblasts  (Fig. 1E). Previously, we demonstrated that this culture system is an in vitro model to recapitulate adult neurogenesis in the SVZ . As was expected, RT-PCR demonstrated that GFAP expression wanes and MASH1 expression waxes throughout this induction (Fig. 1G). PSP expression was strongly downregulated following the addition of bFGF. These data suggest that PSP is expressed by GFAP-positive progenitors rather than their daughter cells, such as rapidly amplifying progenitors or committed neuroblasts.
PSP Is Expressed in the Germinal Zones Throughout Development
To validate our in vitro studies, we investigated PSP expression during neural development using in situ hybridization and immunohistochemistry (Figs. 2, 5). In the embryonic brain, PSP transcripts were strongly expressed along the ventricular wall from E13 to P0, as well as the developing dentate gyrus of hippocampus at E17 and P0 (Fig. 2A). Strong signal was also observed in the embryonic liver (Fig. 2A). This PSP expression in the developing liver was also detected at the protein level by Western blot with the PSP antibody (Fig. 1B) . No specific signal was detected using the PSP sense probe (data not shown). Enriched expression of PSP in the developing germinal zones was also observed using immunohistochemistry (Fig. 2B). The PSP protein was detected primarily in the ventricular zone region rather than in the TuJ1-positive cortical plate region (Fig. 2B). Collectively, these data demonstrate that PSP mRNA and protein are expressed in the germinal zones of mouse embryonic brain.
Function of PSP in Neural Stem (or Multipotent Progenitor) Cells in the Embryonic Brain
Only neural stem cells (or other self-renewing, multipotent progenitors), not lineage-committed progenitors, are capable of forming new tripotent neurospheres under clonal conditions [1, 2, 37]. Therefore, the number of new neurospheres depends on the neural stem cell population within neurospheres. By taking advantage of this phenomenon, we directly tested PSP function in neural stem cells in culture. Overexpression of PSP enhanced the clonal neurosphere forming capacity of neural stem cells derived from E12 telencephalon (Fig. 3). A similar tendency was obtained by experiments with P0-derived cells, although the effect was not statistically significant (Fig. 3B). The effect of knockdown was more prominent (Fig. 3C), as treatment with PSP siRNA resulted in a reduction of the neurosphere forming capacity of neural stem cells both in E12 and P0 cells in a dose-dependent manner. For example, at P0, the reduction of neurosphere formation was approximately 75% compared with control siRNA-treated progenitors (Fig. 3C).
PSP Does Not Significantly Affect Differentiation Potential of Neural Progenitors
Next, we assessed whether alteration of PSP expression affects neural progenitor differentiation potential (Fig. 4). PSP siRNA treatment did not clearly influence morphology or fate of both E12 and P0 progenitors (data not shown). After transfection of either control siRNA or PSP siRNA, P0 progenitor cells were differentiated for 3 days and stained with three major cell type markers: TuJ1 for neurons, GFAP for astrocytes, and O4 for oligodendrocytes. As the majority of cells were stained with GFAP, we focused our evaluation on potential changes in the numbers of neurons and oligodendrocytes. Withdrawal of growth factors resulted in the appearance of both TuJ1-positive neurons and O4-positive oligodendrocytes (Fig. 4A). PSP downregulation did not result in any significant shift of potential for immature neural progenitors to give rise to cells in all three lineages (Fig. 4B, 4C). These data suggest that PSP regulates neural stem cell proliferation without significant effect on their differentiation potential.
PSP Is Expressed in Ependymal Cells and SVZ Astrocytes in the Adult Brain
We then examined whether PSP was expressed in adult neural stem cells in vivo by immunohistochemistry. Like PSP mRNA, the strongest PSP protein expression in the adult brain was detected specifically in cells lining the entire ventricular system (Fig. 5). However, by examining the PSP-expressing cells with confocal microscopy, we determined that PSP was expressed in dual layers (Fig. 5B). First, PSP was expressed in a population of cells lining the ventricle but also in a second layer of cells approximately 20 μm more lateral. This dual-layer pattern was unique to the lateral ventricles and was not seen in the other ventricles. In addition to this labeling along the lateral ventricles, PSP was also expressed in glia-like cells in the hippocampal subgranular layer (supplemental online Fig. 2). In addition to these neurogenic regions with strong PSP-labeled cells, faint labeling was also occasionally detectable in glia in other regions, including the corpus callosum, cortical white matter, and strings of beaded puncta, suggesting axonal varicosities, in the hypothalamus (data not shown).
Heterogeneous cell populations exist in the neurogenic region along the lateral ventricles, including GFAP-positive neural stem cells, PBK/T-Lak cell originating protein kinase (TOPK)-positive rapidly amplifying progenitors , and TuJ1-positive neuroblasts, as well as a lining of ependymal cells along the ventricular surface. To determine whether the PSP-positive cells lining the ventricles were ependymal cells, we examined colabeling with CD24 using confocal microscopy . Indeed, the PSP-positive cells lining all of the ventricles were CD24-positive (Fig. 5C), whereas cells of the second, more lateral layer found exclusively in the lateral ventricle were not CD24-positive. In addition to ependymal cells, the subventricular zone of the lateral ventricle contains GFAP-positive cells, at least some of which are neural stem cells . To determine whether PSP was expressed in GFAP-positive cells, we examined colocalization of GFAP and PSP. We found that the PSP-positive cells in the lateral layer did contain GFAP-positive fibers, as did the PSP-positive ependymal cells, albeit more dimly (Fig. 5G). We also found that the PSP-positive cells in the subgranular zone of hippocampus were GFAP-positive (supplemental online Fig. 2), whereas GFAP-positive cells in other parts of the hippocampus were not PSP-immunoreactive (not shown).
Other populations of cells within the SVZ include transient amplifying progenitors that arise from the slowly proliferating GFAP-positive cell and immature migrating neurons that arise from the transient amplifying progenitors [5, 6]. To determine whether PSP is expressed in the transient amplifying progenitor, we double-labeled the PSP-positive cells with PCNA, MASH1, and PBK/TOPK. We did not detect any PSP-positive cells double-labeled with PCNA-positive rapidly cycling progenitors (Fig. 5E). Likewise, MASH1 and PBK/TOPK, which are expressed in proliferating neuronal and glial progenitors of the SVZ [40, 41], were not double-labeled with PSP (Fig. 5F; supplemental online Fig. 3). Next, we tested double labeling of PSP with TuJ1, which is expressed strongly in immature neuroblasts as well as more mature neurons. Interestingly, although PSP was not expressed by the TuJ1-positive cells, the dual layers of PSP near the lateral ventricle seemed to envelop the bright TuJ1-positive immature neurons (Fig. 5D). Likewise, in the dentate gyrus, PSP-positive cells were not TuJ1-positive but adjacent to those cells (supplemental online Fig. 2). These results are consistent with PSP being expressed in the GFAP-positive, slowly cycling stem cells adjacent to immature neurons. Collectively, our expression data provide strong evidence that PSP is highly expressed in the rapidly dividing neural stem cells in the embryonic brain and that in the adult, PSP is expressed in slowly cycling, GFAP-positive SVZ cells, which may be stem cells. PSP is also strongly expressed in ependymal cells lining all the ventricles, including the lateral ventricles. However, our data do not indicate whether or not these cells are also stem cells.
PSP Regulates Proliferation of Adult SVZ Progenitors
The data described above indicated that PSP is strongly enriched in the GFAP-positive SVZ cells. Recent studies, including ours, have demonstrated that GFAP-positive cells but not other cell types in this region are neurosphere-initiating neural stem cells [11, 15, 30, 14]. We tested PSP function in adult neural stem or progenitor cells using neurosphere cultures from adult SVZ (Fig. 6). Similar to the experiments with embryonic progenitors (Figs. 3, 4), either PSP or control siRNA was transfected into progenitors from adult SVZ, and neurosphere forming capacity was examined under clonal conditions. As was the case in embryonic cells, treatment of adult neural progenitors with PSP siRNA resulted in downregulation of PSP compared with control siRNA-treated cells (Fig. 6A). Subsequently, treatment of adult neural stem/progenitor cells with PSP siRNA resulted in a diminished number of secondary neurospheres formed under clonal conditions (Fig. 6B). These findings suggest that PSP regulates not only embryonic but also adult neural stem/progenitor cell proliferation. Although the data described in Figure 4 indicate a lack of effect of PSP downregulation on postnatal neural progenitor differentiation, we do not yet know whether there is any role for PSP in the differentiation of adult progenitors.
l-Serine Production Is a Partial Mechanism of Action of PSP in the SVZ Neural Stem/Progenitor Cells
The specific expression of PSP in putative neural stem cells in vivo, as well as in the dual layer of expression suggest that serine metabolism may play important roles cell intrinsically and/or by providing an important component of the neurogenic niche in the brain. We reasoned that if PSP regulates neural stem cells through its activity as an enzyme in the l-serine synthesis pathway, then it should be coexpressed with other members of that pathway. Another member of the same pathway is the enzyme 3PGDH, which is two steps upstream of PSP (http://www.genome.jp/dbget-bin/www_bget?path:ot00260). Expression of 3PGDH has previously been examined in the mouse brain [26, 42]. We examined coexpression of PSP and 3PGDH (Fig. 7A). All PSP-positive cells were also positive for 3PGDH. In contrast, 3PGDH expression seemed to extend well beyond PSP-positive regions, and consistent with previous studies [26, 42], 3PGDH appeared to be robustly expressed in astrocytes throughout the brain. These data suggest that l-serine metabolism in general and PSP in particular may be important regulators of stem cells and their niche.
To some extent, it is not surprising that decreasing the expression of an enzyme involved in l-serine metabolism decreases proliferation in a very actively proliferating population of cells. l-Serine plays a significant role in the synthesis of new nucleotides and new proteins, both of which are required for cell cycle progression . We next asked, does exogenous l-serine rescue the decreased neural stem cell proliferation induced by the PSP knockdown? To answer this question, siRNA-treated adult neural stem/progenitor cells were stimulated to form clonal neurospheres with or without l-serine. Addition of l-serine in the medium showed a partial rescue in the number of clonal neurospheres, suggesting that the effect of downregulation of PSP was partially due to an autocrine lack of l-serine. However, several lines of evidence suggest that this simple explanation for the impact of PSP siRNA is incomplete. First, the rescue by l-serine was only partial, suggesting other effects of siRNA knockdown. Second, in vivo PSP was highly expressed not by the highly proliferative transient amplifying cell but by the relatively slowly cycling GFAP-positive cells. This suggests that PSP may also have a paracrine role in the regulation of the neurogenic niche.
GFAP-positive SVZ cells express both 3PGDH and PSP. To examine potential paracrine effects, conditioned medium from PSP siRNA-treated astrocytes was used. Astrocytes from P0 cortices were transfected with PSP or control siRNA, and their conditioned medium was added to neural progenitors in E12 telencephalon following dissociation. The number of secondary neurospheres was reduced by the medium from PSP siRNA-treated astrocytes, although the effect was not dramatic compared with the direct knockdown in E12 progenitors (Figs. 3C, 7C). However, although the impact on the number of spheres was small, the effect of conditioned medium derived from PSP siRNA-treated GFAP-positive cells on the size of spheres was much more dramatic.
PSP Expression in the CNS Germinal Zones Throughout Development
Previously, we performed extensive gene expression profiling and candidate stratification for neural stem cell and stem cell-associated genes [20, 21]. Here, we undertook studies to determine the potential role of PSP in the brain. Analysis of expression indicated that PSP is highly expressed in the germinal zones of embryonic mice, whereas in the adult, PSP is expressed by slowly dividing or quiescent cells that are situated adjacent to more rapidly cycling cells.
In vitro data with PSP siRNA and overexpression support a direct role for PSP in the proliferation of neural progenitors, in a cell autonomous or autocrine fashion. Serine not only is important as an amino acid but also can play a role in DNA synthesis by serving as a substrate for purine biosynthesis . Thus, it would be reasonable to presume that adequate expression of PSP would be required for optimal proliferation of neural progenitors, especially under conditions where this proliferation is being driven to its maximal extent, such as in the neurosphere formation assay in vitro.
However, serine and phosphoserine may also play a role in paracrine signaling. In the adult brain, PSP is not expressed by rapidly proliferative cells, yet knockdown of PSP inhibits adult neurosphere formation, suggesting that PSP expressed by relatively quiescent cells influences either their own entry into the cell cycle or the proliferation of more rapidly proliferative progenitors. Conditioned medium derived from PSP siRNA-treated astrocytes (GFAP-positive cells) inhibited neurosphere formation and growth as compared with control conditioned medium, suggesting that PSP regulates the production of a substance that regulates proliferation. The target of this substance could be stem cells themselves, as suggested by the effects on neurosphere number, and/or other proliferative progenitors, as indicated by the effects on neurosphere size.
The latter observation—that conditioned medium derived from PSP siRNA-treated cells has a greater effect on sphere size, whereas direct siRNA treatment has a more profound effect on sphere number, may seem somewhat contradictory. However, this points to a limitation of the methods used. Given our estimated rates of transfection efficiency using this system (65%; ), we reason that it is likely that PSP siRNA abolishes sphere formation in virtually every sphere-forming progenitor that is transfected. Thus, there would be no opportunity to determine direct or indirect effects on other progenitors that would have existed within the sphere and that would likely influence sphere size.
The identity of the secreted substance regulated by PSP could be serine itself. This is consistent with our observation that addition of serine to the medium at least partially rescues effects of PSP knockdown on neurosphere formation. The fact that the rescue is partial could be related to methodological issues, such as degradation of added serine or lack of access to all cells, or it could indicate that PSP plays other important roles. An intriguing possibility is that the function of PSP could be to actually remove a substance that inhibits progenitor proliferation. A recent screen of small molecule libraries identified l-phosphoserine as one candidate that inhibited proliferation and enhanced neuronal differentiation in embryonic progenitor cells . Phosphoserine can bind on the N-methyl-d-aspartate site of the NDMA receptor or act as an agonist for metabotropic glutamate receptors [45, 46]. High expression of PSP in neurogenic regions may serve to eliminate this signaling molecule by catabolizing phosphoserine to l-serine, regulating the level of proliferation and differentiation in neurogenic regions of adult brain.
In Vivo Function of PSP
The present study demonstrated that PSP plays a role in proliferation of embryonic and adult neural progenitors, at least partly through regulation of l-serine production. However, it is yet to be determined whether PSP has the same role in vivo. The absence of 3PGDH, the enzyme upstream of PSP, leads to reduction of l-serine resulting in hypoplasia of the telencephalon, diencephalon, and mesencephalon . Furthermore, both patients with PSP deficiency and patients with 3PGDH deficiency have impaired brain development, resulting in microcephaly—an expected result if these enzymes regulate stem or progenitor cell proliferation. In one patient with PSP deficiency, treatment with oral serine led to normalization of serine levels and some improvement in head growth [48, 49], a finding consistent with a role for PSP in supplying l-serine. It will be intriguing to examine the in vivo function of PSP for brain morphogenesis during murine development and adulthood.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
This work was supported by National Institute of Mental Health Grant MH065756 and the Miriam and Sheldon Adelson Program in Neural Repair Research. I.N. was supported by a Institute for Stem Cell Biology and Medicine (ISCBM)-California Institute for Regenerative Medicine (CIRM) fellowship. J.D.D. was supported by a Howard Hughes Medical Institute (HHMI) predoctoral fellowship. The 3pgdh antibody was a gift of Dr. Masahiko Watanabe (). I.N. and J.D.D. contributed equally to this work.