Research Article

A neurosphere-derived factor, cystatin C, supports differentiation of ES cells into neural stem cells

Takeo Kato, Toshio Heike, Katsuya Okawa, Munetada Haruyama, Kazuhiro Shiraishi, Momoko Yoshimoto, Masako Nagato, Minoru Shibata, Tomohiro Kumada, Yasunari Yamanaka, Haruo Hattori, and Tatsutoshi Nakahata
PNAS April 11, 2006 103 (15) 6019-6024; https://doi.org/10.1073/pnas.0509789103

  1. Communicated by Tasuku Honjo, Kyoto University, Kyoto, Japan, November 15, 2005 (received for review July 21, 2005)

Abstract

Although embryonic stem (ES) cells are capable of unlimited proliferation and pluripotent differentiation, effective preparation of neural stem cells from ES cells are not achieved. Here, we have directly generated under the coculture with dissociated primary neurosphere cells in serum-free medium and the same effect was observed when ES cells were cultured with conditioned medium of primary neurosphere culture (CMPNC). ES-neural stem cells (NSCs) could proliferate for more than seven times and differentiate into neurons, astrocytes, and oligodendrocytes in vitro and in vivo. The responsible molecule in CMPNC was confirmed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, which turned out to be cystatin C. Purified cystatin C in place of the CMPNC could generate ES-NSCs efficiently with self-renewal and multidifferentiation potentials. These results reveal the validity of cystatin C for generating NSCs from ES cells.

Neurogenesis is considered to be the most complex event of organogenesis during embryonic development and involves a precise signaling, along with cellular interaction cascade, to generate the functional cellular networks. The embryonic organizer allows cells in its vicinity to execute their default neural program by emitting bone morphogenetic protein antagonists (1, 2). However, other works suggest a more complex mechanism (3–5).

Neural stem cells (NSCs) are the self-renewal, multipotent cells that generate neurons, astrocytes, and oligodendrocytes (6, 7). They have great potential as a therapeutic tool for the repair of a number of central nervous system (CNS) disorders. Several in vitro systems allowing derivation of neuronal progeny from embryonic stem (ES) cells, which differentiate into all of the cell fates in a developing embryo, have been described. However, attempts to exclusively generate NSCs or neural progenitor cells from ES cells are restricted. It has been shown that neural fates emerge from ES cells in the serum-free conditions (8, 9). Although these procedures are noteworthy, both the quality and the quantity of ES-derived neural cells are not sufficient for further examination or clinical applications. Higher levels of neural differentiation are achieved by treatment of embryoid bodies with retinoic acid (RA) in the presence of FCS (10, 11) or by coculture with particular stromal cell lines (12, 13). The action of RA is pleiotropic (14, 15), whereas the effect of several stromal cell lines is attributed to an undefined neural inducing activity. These factors severely restrict the ability to use cells cultured with FCS or with particular stromal feeder cells for therapeutic treatments.

Here, we developed an efficient system for the generation of ES cell-derived NSCs (ES-NSCs) during coculture with dissociated neurosphere cells without a need for FCS or feeder cells, which creates a bottleneck for therapeutic methods. We demonstrated that neurosphere-derived factor (NDF), which induces the generation of ES cell-derived neurospheres, exists in the conditioned medium of primary neurosphere culture (CMPNC). We characterized this NDF by using chromatography and mass spectrometric identification, revealing that this activity derives mainly from cystatin C. Our results have led to the discovery of a previously uncharacterized biological activity of cystatin C, which induces NSCs from ES cells exclusively.

Results

ES Cells Differentiate into Neural Stem Cells by Coculture with Dissociated Primary Neurosphere Cells.

To determine whether NSCs can regulate the differentiation of ES cells into neural lineages, D3 ES cells, which constitutively express GFP, were cocultured in suspension with dissociated primary neurosphere cells for 21 days, and phenotypical changes in the GFP-positive ES cells were evaluated. After coculture for 21 days, round spheres measuring >100 μm in diameter could be generated, which were expressed GFP (Fig. 1 A). The number of ES cell-derived GFP positive spheres generated from 105 ES cells was 476.67 ± 32.87, by coculture with dissociated primary neurosphere cells (Fig. 1 B). When ES cells were cultured alone, the majority of ES cells died and only a few small cell clusters were generated (21.25 ± 9.50 generated from 105 ES cells). Thus, ES cells could generate spheres efficiently during coculture with dissociated primary neurosphere cells in the presence of FGF2 and EGF. Selectively FACS-sorted ES cell-derived GFP-positive cells could proliferate and again formed spheres in a defined serum-free medium with FGF2 and EGF. To exclude the possibility of cell fusion (16) between ES cells and primary neurosphere cells, we carried out FACS analysis for DNA contents of ES cell-derived cells. All ES-derived GFP-positive cells were diploid and, thus, suggested that no cell fusion occurred in this coculture system (data not shown).

Fig. 1.
Fig. 1.

ES cells differentiate into NSCs by coculture with dissociated primary neurosphere cells. (A) ES cells, which constitutively express GFP, cocultured with dissociated primary neurosphere cells form spheres in the presence of FGF2 (10 ng/ml) and EGF (20 ng/ml) after 21 days (Scale bars: 100 μm.) (B) ES cells could generate spheres efficiently (filled bar). However, when ES cells were cultured alone, the majority of ES cells died and only a few small cell clusters were generated (open bar). Data are means ± SD of triplicate determinations from two or three independent experiments. (C–H) Differentiated ES-derived spheres contain neurons (C, MAP2), astrocytes (D, GFAP), and oligodendrocytes (E, galactocerebroside). ES-derived spheres passaged five times retained also multilineage potential (F, Tuj; G, MAP2; H, MBP). (Scale bar: 50 μm.) (I) ES-derived spheres could be served to the repeated propagation. (J) The expression of Nestin, Tuj, GFAP, and Gal C genes in primary, twice-passaged, and four times-passaged ES-derived spheres was determined by RT-PCR. Lanes: 1, negative control; 2, primary ES-derived spheres; 3, P2, twice-passaged ES-derived spheres; 4, P4, four times-passaged ES-derived spheres.

When individual spheres were encouraged to fully differentiate, cells positive for the neuronal markers microtubule-associated protein 2 (MAP2) and β-tubulin type III (Tuj) appeared 3 days after differentiation (Fig. 1 C). Subsequently, cells positive for the astrocyte marker glial fibrillary acidic protein (GFAP) (Fig. 1 D) and oligodendrocyte marker Gal C (Fig. 1 E) could be detected after 7 days. The ES cell-derived spheres could be propagated repeatedly, at least 7 times (Fig. 1 I). Notably, the spheres retained multilineage potential (Fig. 1 F–H). To examine whether the characters of ES-derived neurospheres changed with passaging, gene expression profiles of ES cell-derived spheres were examined by RT-PCR. Nestin, which is expressed in NSC in vivo, the neuronal marker Tuj, the astrocytic marker GFAP, and the oligodendrocyte marker Gal C were expressed consistently during passaging (Fig. 1 J). These results suggested that ES cells differentiated into NSCs, which had both self-renewal and multilineage differentiation potentials during cocluture with dissociated neurosphere cells.

Generation of ES Cell-Derived Neurospheres by Using a Secreted Factor from Primary Neurosphere Culture.

The biological activity that induced the differentiation of ES cells into multipotential NSCs during coculture with dissociated primary neurosphere cells could derive from two possible sources: soluble factors secreted from primary neurospheres or direct cell-to-cell contact molecules. To evaluate the contribution of soluble factor secreted from primary neurospheres, we examined whether ES cell-derived neurospheres could be induced in the presence of CMPNC. Addition of the CMPNC to the ES culture significantly increased the total numbers of both cells and spheres (Fig. 2). Serial dilution of CMPNC at least up to 100 times could preserve this activity.

Fig. 2.
Fig. 2.

Generation of ES cell-derived neurospheres by using CMPNC. The addition of CMPNC increased the number of ES cell-derived spheres (A) and total cells (B). Data are means ± SD of triplicate determinations from two or three independent experiments.

To evaluate the gene expression profiles of ES cell-derived neurospheres, expression of genes restricted to neural or nonneural lineages were examined by RT-PCR (Fig. 6, which is published as supporting information on the PNAS web site). In ES-derived neurospheres formed by CMPNC treatment, expression of Rex1, which is highly expressed by inner cell mass of blastcytes and by ES cells (17), was not detected. Simultaneously, Oct4, which is expressed in ES cells and primitive ectoderm cells (17), was down-regulated. In contrast, the expression of nestin, OTX2, and Mash1 genes, all of which are expressed in NSCs in vivo, were confirmed. Furthermore, multilineage neural markers, Tuj, GFAP, and Gal C, were expressed in ES-derived neurospheres. Nonneural lineage genes, including the endodermal markers GATA4 and HNF4a (18, 19), the mesoderm marker brachyury (20), and epidermal marker CK17 (21), were not detected. These gene expression profiles observed in ES-derived neurosphere cells were comparable to those observed in primary neurospheres. Moreover, immunocytochemical analysis revealed that the majority of cells derived from CMPNC-treated ES cell-derived spheres were nestin-positive (Fig. 3 A). When individual spheres were encouraged to fully differentiate, Tuj- or MAP2-positive cells were detected at day 3 (Fig. 3 B and C). Subsequently, cells positive for GFAP, galactocerebroside, and myelin basic protein (MBP) appeared at day 7 (Fig. 3 D–F). When we evaluated the characteristics of Tuj-positive cells differentiated from ES cell-derived neurospheres, four neuronal subtype markers were detected: the dopaminergic neuron markers tyrosine hydroxyrase (TH) and anti-dopamine transporter, cholinergic neuron marker anti-acetylcholine transferase (ChAT), serotonergic neuron marker serotonin, or norepinephrine/epinephirine neuron marker dopamine-β-dehydroxyrase (DBH) were detected (Fig. 3 G–K). Among them, 75–90.9% of ES cell-derived neurospheres contained TH-positive cells during the differentiation for 7 to 14 days (Fig. 3P). This value was much higher than that for cholinergic or serotonergic neurons (ChAT: 45%, 7 days; 55%, 14 days; and serotonin: 12%, 7 days; 25%, 14 days, respectively). To further confirm the preferential generation of dopaminergic neurons, we examined the expression of the mesencephalic dopaminergic neuron markers Nurr1 and En1. These markers were presented in ES cell-derived neurospheres induced by CMPNC (Fig. 6). These data suggested that ES cell-derived neurospheres could differentiate into neurons with various kinds of neurotransmitters, predominantly into dopaminergic neurons.

Fig. 3.
Fig. 3.

CMPNC-treated ES-derived spheres had multilineage differentiation potential and self-renewal potential. Differentiated ES-derived spheres contain nestin-positive cells (A, Cy3-nestin). ES-derived spheres could differentiate into neurons (B and D, Cy3-Tuj; C, MAP2), astrocytes (D and E, FITC-GFAP), oligodendrocytes (E, Cy3-MBP; F, Cy3-Gal C). They could differentiated into various matured neurons (G, Cy3-TH; H, Cy3-DAT/FITC-Tuj; I, Cy3-ChAT; J, Cys-3-DBH; K, Cy3-Serotonin). Transplantation of ES cell-derived neurospheres into the neonatal mouse brain. All grafts were easily detected by the GFP expression as ES-derived cells (M) as compared with control (L). In serial sections, TH-positive cells (N, Cy3-TH) and GFAP-positive cells (O, Cy3-GFAP) with GFP expression could be detected around the subventricular zone. (P) A time course study of the appearance of ES-derived neurospheres with Tuj, TH, DBH, and serotonin-like immunoreactivities. Positive sphere was defined as a sphere containing 10 or more cells with immunoreactivity of each antibody. (Q) ES-derived spheres could be served to the repeated propagation in the presence of FGF2 and EGF. (R) Single-cell suspension culture in serum-free medium containing 0.3% agar were prepared. Sequential images of a single CMPNC-treated ES-derived cell at 1.5, 4, and 12 days after culturing. (Scale bars: 50 μm.)

The ES cell-derived neurospheres induced by CMPNC could be propagated repeatedly (Fig. 3 Q), at least nine times, without losing either multilineage differentiation activity or their preference toward dopaminergic lineage (see Fig. 7, which is published as supporting information on the PNAS web site). To confirm whether a single ES-NSC can form a sphere in response to CMPNC treatment, single-cell suspension cultures were prepared. Spheres could be detected at day 12 with the increasing cells in number during culture (Fig. 3 R). However the overall sphere formation value of single ES-NSC was less than that of NSC from embryos, because the majority of ES cells failed to generate spheres and simply died (Table 1, which is published as supporting information on the PNAS web site). Nevertheless, these spheres that did form retained multilineage potential. These results imply that NDF, which induces the generation of ES cell-derived neurospheres, exists in the CMPNC.

We next explored the differentiation capability of ES cell-derived neurospheres in vivo. Dissociated ES cell-derived neurosphere cells expressing the GFP marker were injected into the hemiventricular area of neonatal mice. ES cell-derived cells could be detected by GFP expression in six of eight mice at 4 weeks and seven of eight mice at 8 weeks by visual observation (Fig. 3 M). In serial sections, TH-positive cells with GFP expression could be detected around the subventricular zone (Fig. 3 N). GFP-positive cells with GFAP expression could also be detected (Fig. 3 O). On the other hand, teratoma-like structures or reduced lifespans were not observed. These results imply that ES cell-derived neurosphere cells could differentiate in vivo without leading to tumor formation.

Purification and Characterization of NDF.

Next, we evaluated the physiological characteristics of NDF in CMPNC from the viewpoint of heat instability and molecular size. Heat treatment of the CMPNC at 60°C for 30 min or 100°C for 10 min abolished its activity. This biochemical characteristic strongly suggested that NDF is protein-based. Further investigation to determine the molecular mass by using an ultrafiltration procedure demonstrated that NDF activity was recovered in the 3,000–20,000 Da molecular mass fraction. To purify the NDF, large quantities of CMPNC were generated (2 liters). After concentrating conditioned medium 200-fold by ultrafiltration, the NDF was semipurified through phenyl hydrophobic column by step gradient elution with decreasing concentration of ammonium sulfate from 1 M to 0 M. The NDF activity was detected in the fraction eluted at the 0 M concentration of ammonium sulfate. When this fraction was analyzed by SDS/PAGE, seven distinct bands were detected at the range from 3,000 to 20,000 Da molecular mass range (Fig. 8A, which is published as supporting information on the PNAS web site). After trypsin treating excised gel bands, proteins were identified by MALDI-TOF/MS peptide mapping. Bands 1–3, 4, and 7 matched similar to peptidylprolyl isomerase A, mouse cystatin C (Fig. 8 B and C) and mouse profilin2, and insulin, respectively.

Cystatin C Confers the NDF Activity.

Recently, it was demonstrated that the proliferation of NSCs in vitro and neurogenesis in vivo are stimulated by the cooperation between FGF2 and cystatin C (21). Considering the close correlation of cystatin C with neurogenesis, we evaluated whether cystatin C could replace CMPNC in inducing ES cells to differentiate into NSCs. To evaluate the generation of ES-NSCs, ES cells were cultured in suspension in the presence of FGF2 and EGF for 21 days with or without cystatin C. Addition of recombinant mouse cystatin C (R & D Systems, 1238-PI) at various concentrations (20 pg/ml to 200 ng/ml) significantly increased the number of total spheres generated from ES cell cultures (Fig. 4 A). Few ES cell-derived spheres were generated in the absence of cystatin C. The efficiency of sphere formation by using recombinant cystatin C was ≈80% compared to that using conditioned medium. In the presence of cystatin C, addition of either FGF2 or EGF significantly increased the number of ES cell-derived spheres (Fig. 4 B). Interestingly, the addition of cystatin C alone increased number of spheres, as compared with the number of spheres generated in the presence of EGF and FGF2 (P < 0.05). Furthermore, even in the presence of either FGF2 or EGF, or both at a high concentration, the number of spheres was not increased (data not shown). The expression of cystatin C was detected in primary neurospheres by RT-PCR, but not in undifferentiated ES cells. Furthermore, Western blotting analysis revealed cystatin C was contained in the CMNPC (data not shown). Therefore, we speculated that the main source of NDF activity in CMPNC is derived from cystatin C.

Fig. 4.
Fig. 4.

Cystatin C exercised the NDF activity, which promote to differentiate into NSCs from ES cells. (A) The addition of the recombinant mouse cystatin C increased the number of ES cell-derived spheres. The filled bars indicated the number of spheres, and the open bars indicated the number of small cell clusters. (B) Growth factor dependency of forming ES-derived spheres. In the presence of cystatin C, addition of either FGF2 or EGF significantly increased the number of ES cell-derived spheres. Filled bars indicate the number of spheres, and the open bars indicate the number of small cell clusters. (C) ES-derived spheres could be served to the repeated propagation over seven times and had multilineage differentiation activity (D, Cy3-Tuj/FITC-GFAP; E, Cy3-MBP). They also could differentiated into various matured neurons (F, Cy3-TH; G, Cy3-ChAT; H, Cy3-Serotonin; I, Cys-3-DBH). (Scale bars: 50 μm.) (J) Example of an HPLC chromatogram showing high levels of dopamine (DA) in the medium of ES-derived neurons (Upper, green line, medium conditioned for 24 h). Relatively low basal DA release was detected (blue line, exposure to buffer for 15 min), as compared to the high levels of DA after 15 min of KCL-evoked depolarization (Lower, red line). Data are means ± SD of triplicate determinations from two or three independent experiments.

When individual cystatin C-treated ES cell-derived spheres were cultured in the differentiation culture conditions, each of the differentiated spheres contained MAP2-positive or Tuj-positive cells at day 3 (Fig. 4 D). GFAP-positive cells and MBP-positive cells were detected at day 7 (Fig. 4 D and E). Furthermore, immunocytochemical staining for subtype-specific neuronal markers revealed that four neuronal subsets, dopaminergic, cholinergic, serotonergic, and adrenergic, were present in the Tuj-positive cell (Fig. 4 F–I). Moreover, cystatin C-treated ES cell-derived spheres could propagate at least seven times (Fig. 4 C) without losing either multilineage differentiation activity or their preferred differentiation toward dopaminergic neurons (see Fig. 9, which is published as supporting information on the PNAS web site). To further confirm the presence of dopaminergic neurons, we examined dopamine production in the induced neurons by reverse-phase HPLC. In response to a depolarizing stimulus (56 mM K+), ES cell-derived neurons released a significant amount of dopamine into the medium (Fig. 4 J). These results indicate that functional neurons producing dopamine were generated with this method, suggesting that the ES cell-derived spheres generated by cystatin C are indeed neurospheres.

Cystatin C Regulates the Commitment of ES Cell Differentiation into Neural Lineage.

Cystatin C could regulate neurosphere generation either by direct induction of ES cells into NSCs or by expansion of cells that had spontaneously differentiated into NSCs. As shown in Fig. 4 B, a small number of small cell clusters could be generated even in the absence of any growth factors. To address whether cystatin C induced NSC differentiation de novo or merely expanded cells that were already committed to a neural lineage, we compared the characteristics of spheres and small cell clusters generated in the presence or absence of cystatin C and in the absence of EGF and FGF2. Immunocytochemical analysis revealed that the majority of cells derived from cystatin C-treated ES cell-derived spheres were nestin-positive and that many more Tuj-positive cells were present in spheres that developed in response to cystatin C. In contrast, only a few nestin-positive cells were detected within the small cell clusters generated in the absence of cystatin C, and neither GFAP- nor MBP-positive cells were detected after differentiation (Fig. 5 A–E). Furthermore, small cell clusters could not reform spheres, and they simply died in the single cell culture system. Thus, they had neither self-renewal nor multilineage differentiation potentials.

Fig. 5.
Fig. 5.

Cystatin C supported to differentiate into NSCs from ES cells. (A and B) Immunocytochemical analysis of nestin-positive cells constituted small cell clusters and spheres, which developed either in the absence or presence of cystatin C. (C and D) Significantly more Tuj-positive cells differentiated from spheres developed in the presence of cystatin C than small cell clusters developed in the absence of cystatin C. (Scale bars: 100 μm.) (E) Immunocytochemical analysis of Tuj-, GFAP-, and MBP-positive cells differentiated from spheres developed in the presence of cystatin C and small cell clusters developed in the absence of cystatin C at day 3 and day 7. Data are means ± SD of triplicate determinations from two or three independent experiments. (F) Comparisons were made about the expression of Rex1, nestin, Nurr1, and En1 in the cystatin C-treated and cystatin C-untreated ES-derived sphere by RT-PCR. Lanes: 1, negative control; 2, ES cells; 3, primary neurospheres; 4, cystatin C-treated ES cell-derived spheres; 5, cystatin C untreated ES cell-derived small cell clusters.

To define the characteristics of cystatin C-treated ES cell-derived spheres and spontaneously arising small spheres, we compared the expression of several transcription factors by RT-PCR (Fig. 5 F). Although nestin was expressed in the both spheres, high levels of En1 and Nurr1, which control differentiation of dopaminergic and serotonergic neurons in the midbrain and hindbrain (22, 23), were expressed only in cell populations treated with cystatin C. Rex1 was expressed in untreated spheres, but was down-regulated in cystatin C-treated spheres. Brachyury, which is abundant in EBs, was not expressed in both (data not shown). These results demonstrate that spheres formed under these two different culture conditions are phenotypically distinct and that NSCs, which have both self-renewal and multilineage differentiation potentials, could be generated in the presence of cystatin C. This result suggests that cystatin C regulates the commitment of ES cell differentiation into NSCs.

Discussion

In this paper, we reported an efficient system for in vitro NSC induction from mouse ES cells in the presence of cystatin C. First, we showed that NSCs could be generated from ES cells by coculturing them with dissociated primary neurosphere cells. We also showed that when CMPNC was added into the serum-free culture of ES cells, they generated the neurospheres. These results suggested that soluble NDF had an active role in the generation of ES-NSCs. Finally, we then have purified cystatin C in CMPNC as the NDF activity.

Cystatin C, a cysteine protease inhibitor (24), is a molecule with pleiotropic functions (25, 26). It is a 14-kDa protein (120 amino acids) and is secreted by different cell types in vitro (27). Several proteinase inhibitors, such as thrombin inhibitors, calpain inhibitors, or cysteine protease inhibitors, have been suggested to be part of a regulatory system associated with neuronal differentiation. We showed that ES cell derived-neurospheres induced by cystatin C were nestin-positive and could differentiate into neurons and glia both in vitro and in vivo and had self-renewal potential. The expression profiles of several genes, including Rex1, Oct4, Otx2, nestin, and mash1, in cystatin C-treated ES cell-derived neurospheres were comparable to those in primary neurospheres. Furthermore, nonneural genes, including GATA4, HNF4, brachyury, and CK17, were not expressed in either the CMPNC or in cystatin C-treated ES-NSCs. Previous studies have reported the differentiation of NSCs from ES cells. Tropepe et al. (9) reported the induction of a small number of primitive neural stem cells from ES cells in the presence of leukemia inhibitory factor and FGF2, although these cells expressed an endodermal marker GATA4 in vitro and had neural and nonneural lineage potentials in vivo. Furthermore, in our laboratory, primitive NSCs could not be propagated repeated more than three times. We conclude that the characteristics of cystatin C-treated ES-NSCs in this culture system are analogous to those of NSCs. However, the efficiency of sphere formation by using recombinant cystatin C was only ≈80% of that using CMPNC. This result suggests that NDFs other than cystatin C may play a role for neural induction of ES cells.

Recently, it was demonstrated that the proliferation of NSCs in vitro and neurogenesis in vivo are both stimulated cooperatively by FGF2 and cystatin C (28). This report infers that cystatin C plays a role in supporting the propagation of cells that already have differentiated into NSCs. Here, we demonstrated that the characteristics of spheres generated in the absence or presence of cystatin C are qualitatively different, and NSCs, which had both self-renewal and multilineage neural differentiation potentials, could be generated only in the presence of cystatin C (Fig. 5 A–E). Therefore, we concluded that cystatin C had a previously uncharacterized function of regulating the commitment of ES cells differentiation into NSCs, either directly or indirectly, in addition to its ability to stimulate propagation of the cells that have already differentiated into NSCs.

How does cystatin C induce neural differentiation of ES cells? We demonstrated that addition of cystatin C alone increased the number of spheres in serum-free conditions but that high concentrations of cytokines, such as FGF2, did not stimulate to generate spheres (Fig. 4 B). However, we also observed that SU5402 (a pharmacological inhibitor of FGF receptor tyrosine kinases) and anti-FGF2 antibodies blocked ES-derived neurosphere formation by cystatin C (see Fig. 10, which is published as supporting information on the PNAS web site). These results suggest that neurosphere formation by ES cells is stimulated by the cooperation between endogenous FGFs and cystatin C. The importance of endogenous FGF2 produced from ES cells for neural lineage differentiation is in agreement with the study of Tropepe et al. (9). Taken together, this hypothesis speculates that cystatin C secreted from NSCs cooperates with endogenous FGFs, which have an active role in the promotion of early neural differentiation of ES cells. Taupin et al. (28) reported that cystatin C can be N-glycosylated and that this complex is required for its activity to induce FGF2-dependent proliferation of rat neuronal progenitor cells (NPCs) in vitro. Muotri et al. reported that L1 transcripts were enriched in N-glycosylated form of cystatin C-responsive NPCs, and L1 could retrotranspose during early neuronal differentiation, affecting the expression of neuronal genes in vitro (29). This report opens the possibility that N-glycosylation of cystatin C is required for its ability to support differentiation of ES cells into NSCs. However, the mechanism by which endogenous and exogenous FGFs effect the differentiation of ES cells into NSCs remains to be elucidated.

We demonstrated that cystatin C-treated ES-NSCs differentiate into all four neuronal subtypes, including dopaminergic, serotonergic, adrenergic, and cholinergic neurons (Fig. 4 E–H). Recently, there have been several reports on methods to produce predominantly TH-positive neurons. Stromal cell-derived inducing activity produced TH-positive neurons at an efficiency of 30% of Tuj-positive neurons (12), which is comparable to our methods. In the current system, the activity of cystatin C was compared with several previously suspected neural-induction factors (see Fig. 11, which is published as supporting information on the PNAS web site). The results obtained suggested that cystatin C is superior to previously reported factors in generating neurospheres from ES cells. One of the notable features of our system is the continuous expansion of cystatin C-treated ES-NSCs that maintain a high differentiation activity during continuous culture, making this system superior to previously reported systems (11, 12, 30). Our system, cystatin C-treated ES-NSCs, prove a good source for cell transplantation therapy, because cell replacement therapy needs large numbers of graft cells. In preliminary transplantation experiments, we demonstrated that graft cells settled in and near the subventricular zone and differentiated into dopamine neurons. Moreover, no teratoma-like structures were observed in any of the 16 mice that received grafts. Tumor formation is a major problem associated with ES cell-derived cell grafting in the treatment of neurological diseases.

Here, we have developed an efficient system for the generation of ES-NSCs by using the soluble factor, cystatin C, without a need for FCS or feeder cells, which creates a bottleneck for therapeutic methods. Furthermore, cystatin C-treated ES cell-derived cells could proliferate continuously while maintaining a high differentiation activity during continuous culture. Thus, we believe that cystatin C-treated ES-NSCs provide a good and safe cell source for CNS transplantation therapy, and this system provides a good model to investigate the mechanism of CNS development and to realize many of potential applications in neuroscience and regenerative medicine in the CNS.

Materials and Methods

ES Cell Lines.

The ES cell lines used in this study were the CCE cell line and the D3 cell line with transfected green fluorescent protein (GFP) gene driven by the ubiquitous CAG promoter. Undifferentiated ES cells were maintained on 1% gelatin-coated dishes in DMEM supplemented with 15% FCS/10 mM nonessential amino acids/0.1 mM 2-mercaptoethanol/1,000 units/ml leukemia inhibitory factor.

Primary Neurosphere Culture.

The striatums of embryonic C57BL/6 mice at day 12.5–14.0 were mechanically dissociated and were cultured as suspension in DMEM/F12 supplemented with 5 mM Hepes buffer/0.3% glucose/0.025 mg/ml insulin/0.1 mg/ml transferrin/20 nM progesterone/0.06 mM putrescine/30 nM sodium selenium (“neurosphere medium”) in the presence of 20 ng/ml epidermal growth factor (EGF) (Sigma)/10 ng/ml fibroblast growth factor 2 (FGF2) (Sigma) for 7 days (31). The concentration was set at 2.0–5.0 × 105 cells per ml.

Coculture System.

ES cells were washed three times and cocultured in suspension with dissociated primary neurosphere cells derived from the striatums of embryonic mice at day 12.5–14.0 in neurosphere medium. These mixed populations were cultured for 21 days in the presence of EGF (20 ng/ml) and FGF2 (10 ng/ml) individually, in combination with, or in the absence of any exogenous growth factors. Total concentration was set at 1.0 × 106 cells per ml. During coculture, cells were dissociated and resuspended at day 7 and 14. The number of round spheres or dissociated cells, which were discriminated by GFP expression, were counted under the microscope and spheres were defined as that measured >100 μm in diameter. We defined “small cell clusters” as cell aggregations measured <100 μm in diameter.

For differentiation, single spheres were transferred to wells coated with l-ornithine in 48-well culture plates and cultured in the neurosphere medium containing 1% FCS in the absence of any growth factors.

Cell Sorting.

Fluorescence-activated cell sorter (FACS) sorting of GFP-positive cells was performed on a FACS vantage flow cytometer/cell sorter (Becton Dickinson). Cells (1–2 × 106 per ml) were analyzed for forward scatter, side scatter, propidium iodide (PI) fluorescence, and GFP fluorescence with an argon laser (488 nm, 100 mW). Dead cells were excluded by gating on forward and side scatter and eliminating PI-positive events. D3 WT clones were used to set the background fluorescence. Viable and fluorescent cells were sorted into DMEM/F12 medium at a speed of 1,000 cells per second.

ES Cells Suspension Culture System.

ES cells were three times washed and cultured for 21 days at 1.0 × 106 cells per ml in the presence or absence of EGF (20 ng/ml), FGF2 (10 ng/ml) and CMPNC. During culture, cells were dissociated and resuspended at days 7 and 14.

Self-renewal of ES-cell NSCs was determined by single-cell suspension culture as described in ref. 32. Dissociated ES cell-derived neurosphere cells were harvested and pelleted. The supernatant was aspirated, and the cell pellet was resuspended in neurosphere medium containing 0.3% agar with a cell density of 200,000 cells per ml in the presence of FGF2 (10 ng/ml) and EGF (20 ng/ml) for 12 days. The number of round spheres was counted under the microscope at day 12.

Immunocytochemistry and Immnunohistochemistry.

Immunocytochemistory was carried out by using standard protocols. Primary antibodies are listed in Table 2, which is published as supporting information on PNAS web site. The individual specificities of these primary antibodies were tested by using appropriate tissues or cells as controls.

For immunocytochemical staining of the sphere and brain tissues in serial section, they were fixed for 20 min at room temperature in 4% paraformaldehyde. After three rinses with PBS, the spheres were equilibrated with 20% and 30% sucrose and sectioned (5–10 μm) on a cryostat. Sections were incubated overnight at 4°C with primary antibodies. After rinsing three times with PBS, sections were incubated for 1 h at room temperature with secondary antibodies. Appropriate cysnin-3-labeled (Jackson ImmunoResearch) and Alexa-488-labeled (Molecular Probes) secondary antibodies were used for visualization.

RNA Extraction and RT-PCR Analysis.

Total cellular RNA was isolated by using the RNAeasy total RNA purification kit (Qiagen) followed by treatment with RNase-free RQ DNase. For cDNA synthesis, random hexamer primers (GIBCO/BRL) were used to prime reverse transcriptase reactions. Using this method, it was possible to use the same reverse transcriptase reaction (cDNA) for PCR amplification with different sets of gene-specific primers. The cDNA synthesis was carried out by using M-MLV SuperReverse transcriptase (SuperScript II) at 42°C for 1 h. Primer sequences (forward and reverse) and the length of the amplified products are listed in Table 2. Amplified products were electrophoresed in 2% agarose gels containing ethidium bromide (25 μg/ml), and band were visualized with UV light. At least three replicates were performed.

Transplantation Procedure.

ES cells were cultured for 21 days in chemically defined medium by adding the CMPNC in the presence of FGF2 and EGF. At the end of culture, ES-derived spheres were dissociated, washed twice, and then suspended in DMEM/F12 at the equivalents 2–4 × 105 cells per μl. By using a blunt-ended 26G Hamilton syringe, 1 μl of suspension medium was injected into the hemiventricular area of neonate mice.

Protein Purification.

The CMPNC (2 liters) was generated (see Supporting Methods, which is published as supporting information on the PNAS web site). After concentrating 200-fold by ultrafiltration, CMPNC was semipurified through a phenyl hydrophobic column (HiTrap, Amersham Pharmacia Biotech) by step-gradient elution with decreasing ammonium sulfate from 1 M to 0 M. For activity determination, each fractions eluted at the various concentration of ammonium sulfate were filtered (0.22 μm) and added in the medium of ES cells suspension culture system.

Molecular Mass Spectrometric Analyses.

Mass spectrometric identification of proteins was performed as described in ref. 33. Briefly, after SDS/PAGE, proteins were visualized by silver staining and excised separately from the gels, followed by the in-gel digestions with trypsin (Promega) in a buffer containing 50 mM ammonium bicarbonate (pH 8.0) and 2% acetonitrile overnight at 37°C. Mass analyses of tryptic peptides were performed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF/MS) with a Voyager-DE/STR (Applied Biosystems). Proteins were identified by comparison between the molecular masses determined by MALDI-TOF/MS and theoretical peptide masses from the proteins registered in NCBInr.

HPLC Analysis.

Dopamine release was measured by reverse phase HPLC, as described in ref. 34. Briefly, samples were collected at the day of differentiation-conditioned medium (24 h), basal release (15 min in HBSS) and evoked release (15 min in HBSS + 56 mM KCl). Samples were stabilized and extracted by aluminum adsorption (Chromosystem). Separation of injected samples (Autosampler 540; ESA, Bedford, MA) was achieved by isocratic elution in the MD-TM mobile phase (ESA) at 0.5 ml/min. The oxidative potential of the analytical cell (ESA Model 5011; Coulochem II, Bedford, MA) was set at 350 mV. Results were validated by coelution with catecholamine standards under varying buffer conditions and detector setting.

Acknowledgments

We thank Drs. S. Masuda and K. Inui for advice on HPLC analysis and Drs. T. Yasumi and H. Hiramatsu for critical reading of the manuscript. This work was supported by grants from the Science Reseach on Priority Areas; the Creative Science Research; the Japan Society for the Promotion of Science; and the Ministry of Education, Culture, Sports, Science, and Technology.

Footnotes

  • †To whom correspondence should be addressed. E-mail: heike{at}kuhp.kyoto-u.ac.jp

  • Author contributions: T.H. designed research; T. Kato, T.H., and K.S. performed research; K.O., M.H., and T.N. contributed new reagents/analytic tools; T. Kato, T.H., K.O., M.H., K.S., M.Y., M.N., M.S., T. Kumada, Y.Y., H.H., and T.N. analyzed data; and T. Kato wrote the paper.

  • Conflict of interest statement: No conflicts declared.

  • Abbreviations:

    Abbreviations:

    ChAT,
    cholinergic neuron marker anti-acetylcholine transferase;
    CMPNC,
    conditioned medium of primary neurosphere culture;
    DBH,
    dopamine-β-dehydroxyrase;
    ES-NSC,
    ES cell-derived neural stem cell;
    GFAP,
    glial fibrillary acidic protein;
    MAP2,
    microtubule-associated protein 2;
    MBP,
    myelin basic protein;
    NDF,
    neurosphere-derived factor;
    NSC,
    neural stem cell;
    TH,
    tyrosine hydroxyrase;
    Tuj,
    β-tubulin type III.
View Abstract
Back to top
Article Alerts
Email Article
Citation Tools
Request Permissions
A neurosphere-derived factor, cystatin C, supports differentiation of ES cells into neural stem cells
Takeo Kato, Toshio Heike, Katsuya Okawa, Munetada Haruyama, Kazuhiro Shiraishi, Momoko Yoshimoto, Masako Nagato, Minoru Shibata, Tomohiro Kumada, Yasunari Yamanaka, Haruo Hattori, Tatsutoshi Nakahata
Proceedings of the National Academy of Sciences Apr 2006, 103 (15) 6019-6024; DOI: 10.1073/pnas.0509789103
Digg logo Reddit logo Twitter logo Facebook logo Google logo Mendeley logo
Proceedings of the National Academy of Sciences: 103 (15)
Table of Contents

Submit

Jump to section

You May Also be Interested in

Similar Articles