Normal sulfation levels regulate spinal cord neural precursor cell proliferation and differentiation

Background Sulfated glycosaminoglycan chains are known for their regulatory functions during neural development and regeneration. However, it is still unknown whether the sulfate residues alone influence, for example, neural precursor cell behavior or whether they act in concert with the sugar backbone. Here, we provide evidence that the unique 473HD-epitope, a representative chondroitin sulfate, is expressed by spinal cord neural precursor cells in vivo and in vitro, suggesting a potential function of sulfated glycosaminoglycans for spinal cord development. Results Thus, we applied the widely used sulfation inhibitor sodium chlorate to analyze the importance of normal sulfation levels for spinal cord neural precursor cell biology in vitro. Addition of sodium chlorate to spinal cord neural precursor cell cultures affected cell cycle progression accompanied by changed extracellular signal-regulated kinase 1 or 2 activation levels. This resulted in a higher percentage of neurons already under proliferative conditions. In contrast, the relative number of glial cells was largely unaffected. Strikingly, both morphological and electrophysiological characterization of neural precursor cell-derived neurons demonstrated an attenuated neuronal maturation in the presence of sodium chlorate, including a disturbed neuronal polarization. Conclusions In summary, our data suggest that sulfation is an important regulator of both neural precursor cell proliferation and maturation of the neural precursor cell progeny in the developing mouse spinal cord.


Background
Complex carbohydrate moieties attached either to secreted extracellular matrix molecules or to membrane bound cell adhesion molecules are important players for central nervous system (CNS) development, because they can bind a multitude of cytokines or growth factors [1][2][3]. Among these carbohydrate structures, the chondroitin sulfate (CS) glycosaminoglycans (GAGs) have been proved to be particularly important for neural precursor cell (NPC) biology, because interference with CS-GAG biology using the bacterial enzyme chondroitinase ABC (ChABC) strongly affects NPC proliferation and differentiation [4][5][6][7]. CS-GAGs are highly sulfated, and the sulfation pattern results from the specific expression pattern of various CS-sulfotransferases during CNS development and in adulthood [8]. Sulfated CS-GAGs also play a role in pathophysiological situations. Up-regulation of CS proteoglycans (PGs) bearing complex CS-GAGs after CNS injury is associated with a specific sulfation pattern on CS-GAGs, mediating potentially inhibitory properties of PGs on axonal regeneration [9,10]. These inhibitory properties can be overcome using ChABC in combination with defined motor tasks [11,12]. However, different sulfation patterns appear to differentially affect axonal regeneration after CNS lesion [13].
Besides CS-GAGs, the heparan sulfate (HS)-GAGs constitute a second major class of sulfated glycans in the CNS. HS-GAGs are known for their strong impact on fibroblast growth factor (FGF) and Sonic hedgehog signaling [14]. In addition, some other glycan structures can be sulfated as well. Among these, the complex LewisX glycan, which is expressed by NPCs during development and in adulthood [15][16][17], and the human natural killer 1 antigen are the most prominent ones [18].
The transfer of sulfate on the glycosaminoglycan backbone of CS-GAGs and HS-GAGs is catalyzed by CS-or HS-specific sulfotransferases located in the Golgi apparatus [19,20]. 3′-phosphoadenosine 5′-phosphosulfate (PAPS) serves as the sulfate donor. The enzymatic generation of PAPS can be inhibited using sodium chlorate (NaClO 3 ) [21], resulting in hypo-sulfated GAG chains. Therefore, NaClO 3 is often used to pharmacologically interfere with GAG biology. Here, we used NaClO 3 to further elucidate the role of sulfation for differentiation and proliferation of spinal cord NPCs cultivated as free floating neurospheres. Neurospheres treated with NaClO 3 showed a reduction in growth accompanied by changes in neuronal differentiation and maturation. In contrast, differentiation towards glial lineages seemed to be largely unaffected. Thus, we propose that the proliferation of spinal cord NPCs and the maturation of spinal cord NPC-derived neurons depend on a specific sulfation pattern during development.

Animals
All experiments were performed according to international rules using timed-pregnant wild-type NRMI mice kept under standard housing conditions. The age of the embryos was determined following the Theiler Stages, and the day of the vaginal plug was considered as E0.5.

Neurosphere culture
Cultivation of mouse spinal cord NPCs was carried out as previously described [22]. Briefly, the spinal cords of E13.5 to E18.5 mouse embryos were dissected using small forceps. Afterwards, the isolated spinal cords were enzymatically dissociated with 30 U/mL Papain (Worthington, New Jersey, USA), resulting in a single cell suspension. After centrifugation for 5 min at 200 g the cell sediment was resuspended in neurosphere medium containing DMEM and nutrient mixture F12 (in a ratio of 1:1; both Gibco, Karlsruhe, Germany), 2 % (v/v) B27 (Invitrogen, Karlsruhe, Germany), 1 % (v/v) L-glutamine (Invitrogen) and 1 % (v/v) penicillin/streptomycin (Invitrogen). Under proliferative conditions, either 1,250 cells/mL (clonal density) or 100,000 cells/mL (high density culture) were plated, and either 10 ng/mL (clonal density) or 20 ng/mL (high density culture) EGF and FGF2 (both PreproTech, Rocky Hill, USA) were added. FGF2 containing cultures were additionally supplemented with 0.25 U/mL (clonal density) or 0.5 U/mL (high density culture) heparin (Sigma-Aldrich, Munich, Germany). After one week at 37°C and 5 % (v/v) CO 2 , neurospheres had formed. For clonal analyses, the cells were kept for one week without any agitation, in order to avoid neurosphere fusion events. As recently described, 30 mM NaClO 3 or the respective solvent were added for the entire cultivation period [23].
For differentiation analyses, primary neurospheres were dissociated using trypsin/ethylenediaminetetraacetic acid (EDTA; Invitrogen) for 4 min at 37°C, and single cells were plated at 5,000 cells/well onto poly-DL-ornithine /laminin coated four-well dishes (Greiner, Frickenhausen, Germany) for another 4 days in the presence of 1 % (v/v) FCS and either 30 mM NaClO 3 or the respective solvent. Afterwards, the cells were either subjected to immunocytochemical or electrophysiological analyses.

Immunohistochemistry
Pregnant mice were killed by cervical dislocation, and the embryos were immediately removed. The embryos' trunks were washed once with PBS and then fixed in 4 % (w/v) paraformaldehyde (PFA) at 4°C. Depending on the age of the embryo, the fixation time varied between 30 min (E9.5) and 20 h (E18.5). After fixation the embryos were transferred to 20 % (w/v) sucrose for cryoprotection. Finally, the tissue was embedded in Tissue Tec Freezing Medium (Jung, Nussloch, Germany) and cut into 16 μm thin sections on a cryostat CM3050S (Leica, Solms, Germany). The sections were mounted on superfrost slides (Thermo Scientific, Schwerte, Germany) and stored at −20°C until further use.
For immunohistochemical analyses of neurospheres, the neurospheres were transferred into a 1.5 mL Eppendorf tube, briefly washed with PBS and then fixed for 1 h in 4 % (w/v) PFA at 4°C. Subsequently, the PFA was removed, and 20 % (w/v) diethylpyrocarbonate-treated sucrose was added for cryoprotection for 4 h at 4°C. Finally, the neurospheres were embedded using Tissue Tec Freezing Medium and cut into 16 μm thin cryosections on a cryostat CM3050S. The cryosections were rehydrated and blocked for 1 h at room temperature with PBT1 (PBS + 1 % (w/v) BSA + 0.1 % (v/v) Triton X-100) and 1.7 % (w/v) NaCl-PBS (PBS + 0.9 % (w/v) NaCl) in a ratio of 1:1 plus 10 % (v/v) normal goat serum, followed by incubation with the primary antibodies diluted in PBT1 + 5 % (v/v) normal goat serum overnight at 4°C. The next day, the sections were washed three times with PBS and subsequently incubated with species-specific antibodies coupled with either Cy2 (1:250) or Cy3 (1:500) (Dianova, Hamburg, Germany) diluted in PBS/A (PBS + 0.1 % (w/v) BSA) for 3 h at room temperature. Hoechst 33528 (Sigma) was included (diluted 1:10 5 in PBS), to additionally label the nuclei. The sections were washed three times with PBS and finally mounted with ImmuMount (Invitrogen).

Immunocytochemistry
After removal of the culture medium, adherent cells were briefly washed twice with PBS/A. In case of membrane bound or extracellular epitopes (O4, 473HD), the incubation with the primary antibody diluted in PBS/A was directly carried out for 30 min at room temperature. Then, the cells were washed again three times with PBS/A and fixed with 4 % (w/v) PFA for 10 min at room temperature. To detect intracellular epitopes, the fixation was performed prior to the incubation with the primary antibodies diluted in PBT1. After incubation with the primary antibody, the cells were washed three times with PBT1 and the incubation with either Cy3-or Cy2-or HRP (1:500)-coupled species-specific secondary antibodies (Dianova) diluted in PBS/A was carried out at room temperature for 30 min. Hoechst 33528 (1:10 5 ) was additionally added to visualize the nuclei. Finally, the cells were washed twice with PBS and mounted in PBS and glycerine (2:1). For BrdU incorporation analysis, 1 μM BrdU was added to the neurospheres for 1 h. Then, the neurospheres were dissociated, and single cells were plated on a poly-DL-ornithine substrate for two hours. Finally, the BrdU-immunocytochemistry was carried out using the BrdU-labeling and detection Kit I (Roche, Mannheim, Germany) according to the manufacturer's instructions. The diaminobenzidin staining was carried out by incubating the cells with freshly prepared diaminobenzidin (Sigma-Aldrich) diluted in double distilled water for 10 min after incubation with the HRP-coupled secondary antibody. Finally, the cells were washed twice with double distilled water and mounted in PBS and glycerine (2:1).

Western blot
Neurospheres were homogenized and solubilized by mechanical agitation in 4°C cold cell lysis buffer (50 mM Tris-HCl pH 7.4; 150 mM NaCl; 5 mM EDTA; 5 mM ethyleneglycotetraacetic acid; 1 % (v/v) Triton X100; 0.1 % (v/v) Na-desoxycholate; 0.1 % (v/v) SDS) and incubated for further 30 min on ice. The lysate was then cleared by centrifugation (16,000 g) at 4°C. The protein concentration was determined using a protein quantification kit (Pierce, Rockford, USA) according to the manufacturer's instructions. 10 μg protein was separated on a 7 % (v/v) SDS-gel and transferred to a PVDF membrane (Roth, Karlsruhe, Germany). After transfer the membrane was blocked with 5 % (w/v) skim milk powder in Tris-buffered-Saline (TBS) for one hour at room temperature. The primary antibodies were diluted in 5 % (w/v) skim milk powder in TBS + 0.05 % (v/v) Tween20 (TBST) and incubated at 4°C over night. Subsequently, the membrane was washed three times with TBST for 10 min and the incubation with the HRP-coupled secondary antibodies (1:5000) diluted in 5 % (w/v) skim milk powder in TBST was carried out at room temperature for one hour. Finally, the membrane was washed again three times with TBS, and the signal was detected using enhanced chemiluminescence reagent (Pierce, Rockfort, USA).

Electrical recordings
Whole cell patch-clamp recordings of spinal cord NPCderived neurons were performed using borosilicate glass pipettes of a mean resistance of 2-6 MΩ. The glass pipettes were filled with a solution, mimicking the intracellular ion concentrations (100 mM K-gluconate, 0.1 mM CaCl 2 , 1.1 mM EDTA, 5 mM MgCl 2 , 5 mM NaCl, 10 mM HEPES, 3 mM MgCl 2 -ATP, pH 7.35, 235 mOsm). For the experiments, the cell culture medium was removed and the cells were washed twice with the bath solution (110 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl 2 , 0.8 mM MgCl 2 , 10 mM HEPES, 10 mM glucose, pH 7.35, 250 mOsm) and finally kept in the bath solution at room temperature for no longer than 1 h. The setup was equipped with an inverse microscope including phase contrast optics. The data were acquired using an EPC7 patch-clamp amplifier and processed via Pclamp 6 software. After liquid junction potential correction, the cell was depolarized from −77 mV to 53 mV in 5 mV steps, in order to analyze both the mixed sodium and potassium current density. The cell capacitance was calculated from the integral of the charging curve after application of a step depolarization of 10 mV amplitude. The sodium current was determined from the peak amplitude for a step depolarization to −12 mV. The potassium current was determined after 200 ms of depolarization to 28 mV. To account for the cell size, the calculated currents were normalized to the cell capacitance. Thus, the current densities data are displayed in the units A/F. The data were pooled from five independent preparations. In total, 29 control cells and 31 NaClO 3 treated cells were analyzed. Each cell with a leak current of more than 20 pA was excluded from the final data analysis, resulting in 12 control cells and 16 NaClO 3 treated cells.

Documentation and data analysis
Pictures were taken at an Axioplan2 microscope with the AxioCam HRc camera using the AxioVision 4.4 and 4.5 software (Zeiss, Jena, Germany). For quantitative analyses of immunocytochemical antigen detections, a minimum of 200 Bisbenzimid-positive nuclei were counted in at least three independent experiments per antibody and culture condition.
To quantify the neurosphere formation from primary spinal cord cells, the number of all neurospheres in the whole culture flask was counted after one week. In order to prevent the formation of non-clonal neurospheres, the culture flasks were not taken out of the incubator during the cultivation period. The diameters of single neurospheres as well as the neuronal morphology were analyzed using ImageJ v1.41. Unless otherwise stated, the data are expressed as mean ± SD. Statistical significance was assessed using the paired and unpaired two-tailed Student's t test and the P-values are given as *P ≤0.05, **P ≤0.01, and ***P ≤0.001.

Results
The sulfation-dependent 473HD-epitope is expressed by neural precursor cells in the embryonic mouse spinal cord Since systematic data concerning the expression of GAG chains throughout spinal cord development are missing, we started with an expression analysis of the 473HD-epitope, a representative CS-GAG present on the chondroitin sulfate proteoglycan phosphacan/RPTPβ/ζ [25,27]. Immunohistochemical detection of the 473HD-epitope on frontal spinal cord sections between E9.5 and E18.5 show its upregulation towards the end of neurogenesis at E12.5 (Additional file 1: Figure S1A, B). The expression was particularly high in the ventral spinal cord between E13.5 and E15.5 (Additional file 1: Figure S1C, D). Towards the end of embryogenesis at E18.5, the 473HD-epitope could be detected within the whole spinal cord except for the central canal region (Additional file 1: Figure S1E). Further immunohistochemical analyses on frontal E13.5 spinal cord sections ( Figure 1A) revealed that the 473HD-epitope was expressed by Nestin-positive NPCs in vivo ( Figure 1B-D). Note that most of the ventricular zone lacks immunoreactivity for the 473HD-epitope except for a distinct region within the ventral spinal cord. To investigate the cellular source, we dissociated the spinal cord from various embryonic ages and plated single cells in low density for two hours on a poly-DL-ornithine substrate. After that, we immunocytochemically characterized the cells using various cell type specific markers. We observed that many 473HD-positive cells co-expressed the NPC markers Nestin, BLBP and GLAST ( Figure 1E-G). In contrast, we never observed 473HD immunoreactivity on βIII-Tubulin-positive young neurons ( Figure 1H). We further quantified the relative number of 473HD-positive cells expressing the NPC markers Nestin, BLBP and GLAST at E13.5, E15.5 and E18.5. Our findings are summarized in Table 1. The percentage of 473HDpositive cells co-expressing one of the mentioned markers was about 5 % for each marker at E13.5 but increased within the next two days to around 10 % ( Figure 1I and Table 1). Towards the end of embryogenesis at E18.5, the percentage of Nestin-and-473HD-positive cells decreased again, while the BLBP-and-473HD populations increased and the GLAST-and-473HD populations did not change ( Figure 1I and Table 1). Finally, we determined the overall 473HDpositive cell population throughout development and found a general increase in the relative amount of 473HD-positive cells between E12.5 and E18.5, consistent with our immunohistochemical analyses (E12.5: 6.2 ± 1.9 % (n = 4); E13.5: 9.0 ± 3.3 % (n = 10); E15.5: 15.2 ± 2.7 % (n = 10); E18.5: 23.2 ± 5.3 % (n = 8); Figure 1J).

Sodium chlorate efficiently reduces the level of the sulfation-dependent 473HD-epitope in spinal cord neural precursor cell cultures
Several studies dealing with GAG biology were based on the usage of NaClO 3 , in order to interfere with the sulfation levels of the GAG chains. In this study we applied NaClO 3 and asked whether alterations in sulfation levels might regulate proliferation, survival and differentiation of spinal cord NPCs grown as free floating neurospheres. We cultured primary neurospheres from E13.5 spinal cord cells and analyzed the expression of the sulfationdependent 473HD-epitope and its carrier protein RPTPβ/ ζ after one week. Western blot analyses of neurosphere detergent extracts revealed that neurospheres expressed high levels of the 473HD-epitope under standard culture conditions. The addition of NaClO 3 strongly reduced the 473HD levels in comparison to the solvent control ( Figure 2A). However, the expression levels of its carrier protein itself appeared not to be affected (Figure 2A). In an independent experiment, neurosphere cryosections were labeled for the 473HD-epitope as well as RPTPβ/ζ. Consistent with the western blot analysis, the 473HDimmunoreactivity was strongly reduced when sulfation levels were decreased using NaClO 3 ( Figure 2B). The RPTPβ/ζ immunoreactivity, however, seemed to be enhanced, potentially reflecting an enhanced accessibility of the epitope rather than an increased expression level ( Figure 2B), since the western blot analysis did not reveal changes in the RPTPβ/ζ expression. Yet, these data demonstrate the suitability of NaClO 3 to suppress normal sulfation levels in spinal cord NPC cultures.

Sodium chlorate affects spinal cord neural precursor cell growth
We recently reported on a decreased neurosphere formation capacity from primary cortical cells upon chlorate treatment [8]. Thus, we started to analyze the influence of NaClO 3 on NPC proliferation by assaying the clonal neurosphere formation from E13.5 primary spinal cord cells. In order to investigate whether NaClO 3 might just affect NPC subpopulations, we grew neurospheres under different growth factor conditions. Figure 3A,B shows representative pictures of neurospheres grown in the presence of EGF and FGF2 under control or NaClO 3 conditions. Typical neurospheres of 100 μm to 200 μm diameter formed within one week under control conditions. In contrast, NaClO 3 -treated neurospheres were smaller. Moreover, the number of primary neurospheres was significantly reduced upon NaClO 3 treatment (without: 75.6 ± 35.2; NaClO 3 : 33.8 ± 15.1; n = 5; P = 0.04; Figure 3C). In parallel, neurospheres were grown in high density cultures in the presence of EGF and FGF2, and the total cell number, as a measurement for the overall proliferation rate, was determined after one week. In the presence of NaClO 3 we observed strongly reduced total cell numbers in the culture flasks (without: 2,240,000 ± 149,000; NaClO 3 : 622,000 ± 219,000; n = 4; P = 0.001; Figure 3D).
Finally, we further analyzed the effect of NaClO 3 on the size of neurospheres grown in the presence of EGF and FGF2 for one week. For that purpose, photomicrographs of single neurospheres were used to measure the diameter of respective colonies. Data from four independent experiments revealed that the addition of NaClO 3 resulted in a shift towards smaller neurospheres at the expense of larger neurospheres ( Figure 3F). However, since there were still some large neurospheres in the presence of NaClO 3 , NaClO 3 itself might only affect the proliferation rate of a subpopulation of NPCs.

Sodium chlorate affects Erk-signaling
To gain first insights into potential molecular mechanisms mediating the cell biological influence of NaClO 3 on spinal cord NPCs, we analyzed the activation levels of Erk1/2 and Akt, two effector molecules downstream of tyrosine kinase receptors. For that purpose, we used neurospheres from high density cultures grown in the presence of either EGF and FGF2 or FGF2 alone. Both conditions were additionally supplemented with heparin. The addition of NaClO 3 resulted in a reduced expression of the 473HD epitope ( Figure 5A), confirming the functionality of NaClO 3 in that experiment. While the level of phosphorylated Akt (pAkt) was not changed, we observed increased levels of phosphorylated Erk1/2 ( Figure 5B). Moreover, the overall expression level of the EGFR in the neurospheres was reduced upon NaClO 3 treatment ( Figure 5C). This was in line with the increased percentage of neurons in the presence of NaClO 3 .

Sodium chlorate affects morphological and functional maturation of neural precursor cell-derived neurons
The observation that neuronal differentiation was enhanced in the presence of the sulfation inhibitor NaClO 3 prompted us to further examine neuronal maturation with regard to morphology. To address that issue, we plated dissociated primary neurosphere cells under differentiating conditions and analyzed the neuronal morphology after four days in terms of neurite number, length of the longest neurite and basal soma size. Figure 6A,B shows representative photomicrographs of neurons grown in the absence and presence of NaClO 3 . The mean number of neurites was not affected by NaClO 3 (without: 1.95 ± 0.17; NaClO 3 : 1.88 ± 0.19; n = 6; Figure 6C). In contrast, the mean length of the longest neurite was significantly reduced upon NaClO 3 treatment (without: 385.5 ± 94.6; NaClO 3 : 306.8 ± 77.8; n = 6; P = 0.044; Figure 6D). The average soma size was also not changed (without: 1,426 ± 218; NaClO 3 : 1,454 ± 253; n = 6; Figure 6E). Next, we analyzed if NaClO 3 also affects electrophysiological properties of NPC-derived neurons. Thus, we performed whole cell patch-clamp recordings of NPC-derived neurons. To take potential neuronal subpopulations into consideration, we specifically analyzed neurons with a triangular cell soma ( Figure 6F). Upon supra-threshold depolarization, the neurons usually responded with an inward current (sodium current component) followed by a prolonged outward current (potassium current component). Following culture in NaClO 3 , peak sodium currents were strongly decreased in comparison to cells cultured under control conditions ( Figure 6F). To ensure that these differences were not due to differences in cell size, we first determined the cell capacitance and did not observe any changes in response to NaClO 3 treatment (without: 13.3 ± 2.7 pF; NaClO 3 ± 12.2 ± 1.3 pF; n = 5 preparations; Figure 6G). Then, we measured both the sodium inward and potassium outward current and normalized both components to the cell capacitance, yielding current densities. We found that the sodium current density was significantly reduced in the presence of NaClO 3 (without: -17.5 ± 2.9 A/F; NaClO 3 : -6.3 ± 3.6 A/F; n = 5 preparations; P <0.001; Figure 6H). In contrast, the potassium current density was not significantly changed (without: 23.4 ± 8.9 A/F; NaClO 3 : 16.8 ± 3.0 A/F; n = 5 preparations; Figure 6H).
To further analyze the influence of NaClO 3 on neuronal maturation, we extended the differentiation period for another three days. After that, we first determined the relative number of Caspase3-positive cells to investigate whether potential differences might be due to changes in the rate of apoptotic cell death. However, we did not notice any differences in the percentage of Caspase3-positive cells ( Figure 7A,B). Next, we counted the number of MAP2-positive cells and found a significant increase in the percentage of MAP2-positive cells upon NaClO 3 treatment (without: 2.3 ± 0.5%; NaClO 3 : 8.4 ± 3.6% (n = 4; P = 0.015)). Finally, we analyzed the polarization of the neurons in our differentiation assay by staining for MAP2 (dendritic compartment) in combination with Tau (axonal compartment). While under control conditions, a substantial number of MAP2positive cells also had a Tau-positive axon; the relative number of clearly polarized neurons in the presence of NaClO 3 was significantly reduced (without: 25.9 ± 6.3%; NaClO 3 8.0 ± 2.1%; (n = 4; P = 0.002)). We also observed  In order to analyze the neuronal morphology, the βIII-Tubulin immunoreactivity was visualized using diaminobenzidine. In the control situation the neurons often displayed a multipolar morphology with long neurites. Upon NaClO 3 treatment the neurites appeared to be shorter. (C-E) To assess the morphological maturation, the neurite number, the length of the longest neurite and the soma size were measured. The latter two are given in arbitrary units. While the neurite number and the soma size were not altered in the presence of NaClO 3 (n = 6), the mean length of the longest neurite was significantly reduced (n = 6; P <0.05). (F) Whole cell patch-clamp recordings of individual NPC-derived neurons were performed under voltage clamp conditions. To take neuronal subpopulations into account, only neurons exhibiting a triangular soma shape were analyzed. Upon depolarization, the neurons responded with a brief sodium inward current followed by a delayed potassium outward current. The inward component was strongly reduced in the presence of NaClO 3 . (G) The determination of the cell capacitance revealed no changes in the overall cell size upon NaClO 3 treatment (n = 5 preparations). (H) Analysis of the two current components expressed as current density clearly showed a reduced sodium current density in the presence of NaClO 3 (n = 5 preparations; P <0.001). In contrast, the potassium current density was not affected (n = 5 preparations). Scale bar: 25 μm. a similar phenomenon in two independent experiments after a ten-day differentiation period (data not shown).

Discussion
GAG chains on proteins of the extracellular matrix or on membrane bound cell adhesion molecules are complex carbohydrates that participate in many biological processes during development and in adulthood, in part via their highly specific sulfation patterns. Sulfation of GAGs is performed by sulfotransferases that have been shown to be particularly relevant in developmental processes [8,[29][30][31][32][33][34]. In this study, we initially investigated the expression of the 473HD-epitope, a unique CS-motif, during embryonic spinal cord development and found that it is expressed at late embryonic ages. Its pronounced expression within the NPC compartment, moreover, suggests a functional role of sulfated GAG chains in the regulation of spinal cord NPC biology. Along these lines it has been shown that highly sulfated complex GAG structures are expressed by cortical NPCs [35][36][37] and regulate proliferation and differentiation of embryonic and adult cortical NPCs [4][5][6][7]38]. Thus, we analyzed the role of changes in sulfation on the proliferation and differentiation potential of mouse embryonic spinal cord NPC using NaClO 3 as a potent inhibitor of eukaryotic sulfation reactions. Low concentrations of NaClO 3 (2 mM to 5 mM) preferentially inhibit sulfation of CS-GAGs, whereas higher concentrations (15 mM to 30 mM) are required to inhibit sulfation of HS-GAGs as well [39]. We used a concentration of 30 mM, to assure an almost complete sulfation blockage of both CS-GAGs and HS-GAGs. To monitor the effectiveness of NaClO 3 , we analyzed the expression level of the 473HD-epitope, and observed an almost complete reduction of the epitope after NaClO 3 treatment. This observation is in line with a former study demonstrating that the 473HDepitope depends on a distinct sulfation pattern [40]. Some residual 473HD immunoreactivity in the western blot as well as in the immunohistochemical analyses is likely due to the fact that NaClO 3 might not affect every neurosphere cell as soon as the neurospheres reach a certain size. However, since the total amount of the carrier protein RPTPβ/ζ does not change, our data prove the functionality of NaClO 3 to specifically affect sulfation events.
We next analyzed clonal neurosphere formation capacity from E13.5 spinal cord NPCs. Treatment of neurosphere cultures with 30 mM NaClO 3 resulted in significantly fewer and smaller neurospheres. This effect was independent from the growth factor used to propagate the neurospheres (EGF, FGF2 or both; Figure 4 C-E). Looking at the diameter of the spheres, it became obvious that there are more small-diameter neurospheres and fewer large-diameter neurospheres when the cultures were treated with NaClO 3 ( Figure 4A,B,F). This phenomenon has been observed previously for the embryonic telencephalic neurospheres [8]. The reasons for the altered neurosphere growth could be that NPCs within the neurosphere divide more slowly in the presence of NaClO 3; NPCs die more frequently; or NPCs have a higher differentiation capacity, and therefore NaClO 3 might affect stem or precursor cell maintenance. Our BrdU incorporation analysis did not reveal any changes in the S-phase of the cell cycle. However, while the relative number of pH3-positive cells present in the G2-phase of the cell cycle was significantly enhanced, the percentage of M-phase cells was significantly decreased, indicating a potential G2-phase arrest in the presence of NaClO 3 . Although a G2-arrest is a common phenomenon preceding apoptotic cell death, especially in cancer cells [41], we did not observe an increased percentage of Caspase3-positive cells, Thus, it is unlikely that NaClO 3 affected cell death rates in our culture system. Besides the difference with regard to the G2/M-phase of the cell cycle, we also documented a significant increase of βIII-Tubulin-positive young neurons in response to NaClO 3 treatment. Furthermore, the overall expression level of EGFR in the neurosphere cultures was reduced. This is also in line with an increased percentage of βIII-Tubulin-positive neurons, since the acquisition of the EGFR is generally considered a hallmark for the developmental switch from neurogenic NPCs towards gliogenic NPCs [42]. However, based on our data, we conclude that NaClO 3 primarily affects spinal cord NPC proliferation by changing the cell cycle kinetics. This results in a higher number of neurons already under proliferative conditions. However, it remains unsolved whether NaClO 3 truly promotes neurogenic differentiation. It might be that the enhanced percentage of neurons simply reflects the reduced proliferation capabilities of the surrounding cells.
Regarding potential molecular mechanisms, we initially reasoned that the suppression of sulfation might compromise growth factor signaling, leading to changes in the activation levels of downstream effector molecules such as Erk1/2 and Akt. Surprisingly, we observed an increased Erk1/2 activation in the presence of NaClO 3 . Yet an increased Erk1/2 activation has already been reported previously in the context of a G2 arrest in different cancer cell lines [43,44]. It is unlikely that the increased Erk1/2 activation depends on increased receptor activation, because the lack of sulfation should result in a reduced growth factor signaling strength. Along these lines, we recently showed that the enzymatic degradation of CS-GAGs compromises FGF signaling in cortical neurosphere cultures, resulting in a reduced MAP kinase activation [5]. Here, we think that NaClO 3 might reduce the activation levels of phosphatases through a yet unknown mechanism. In this context, it is noteworthy that RPTPβ/ζ bears sulfated CS-GAGs [25] and regulates Erk1/2 activation levels in human keratinocytes [45]. Therefore, the sulfation status of RPTPβ/ζ might have a critical influence on its activation level and as a consequence also on the Erk1/2 activation levels.
Our laboratory has recently shown that degradation of CS-GAGs from telencephalic NPCs using ChABC results in a pronounced glial differentiation at the expense of neuronal cell types [6]. With regard to spinal cord NPCs, we observed a higher proportion of neuronal cells already under proliferative conditions. Under differentiating conditions we documented an enhanced relative number of βIII-Tubulin-positive cells after four days and MAP2-positive cells after one week following NaClO 3 treatment. These data are in line with previous studies demonstrating an increased neuronal differentiation of mouse embryonic stem cells upon siRNA mediated knock down of either the PAPS transporters 1 and 2 or HS-specific sulfotransferases [46]. In contrast to the increased percentage of neurons, the relative number of glial cells was not affected.
To further elucidate the phenotype of the βIII-Tubulinpositive cells, we took a closer look at the morphology of these cells and valued the morphology as a level of maturation. The spinal cord NPC-derived neurons exhibited no changes with respect to soma size and total primary neurite number upon NaClO 3 treatment. But they significantly differed in the length of their longest neurite. This was surprising, since CSPGs and HSPGs are primarily known for their inhibitory influence on neurite outgrowth, particularly under pathological conditions [10,47]. Moreover, CS-GAGs have recently been shown to influence the morphology of NPC-derived neurons in terms of neurite number, length and branching [4,48]. Yet there is growing evidence that specific CS-GAGs also promote neurite outgrowth depending on both the mode of their presentation (that is, homogenous substrate or alternating substrate) and on the neuronal cell type [13,25,49]. Thus, we believe that our data hint towards an attenuated maturation of spinal cord NPC-derived neurons in the presence of NaClO 3 in vitro. In line with this hypothesis, we also documented a strongly reduced neuronal polarity after seven and ten days in culture. To check whether this delayed morphological maturation is also relevant at a functional level, we analyzed the differentiated neurons grown from NaClO 3 -treated dissociated cultures in comparison with the control group using whole cell patchclamp recordings. We determined that the overall cell capacitance of the cells was not changed. This is in line with the fact that the soma size was not changed either. However, we observed a significantly reduced sodium current density, while the potassium current density remained unaltered. Lower sodium currents suggest a delay in functional neuronal maturation, because sodium currents usually increase upon differentiation of neuronal precursors during embryogenesis [50]. Furthermore, neurite outgrowth and sodium channel-dependent excitability have previously been shown to be tightly coupled during functional neuronal maturation [51]. Interestingly, the addition of FGF2 to rat postnatal hippocampal neurons significantly enhances the sodium current density, demonstrating a growth factor signaling-dependent regulation [52]. This might be mediated by highly sulfated GAG chains. In this context, it is noteworthy that we recently demonstrated changes in spontaneous synaptic activity of primary hippocampal neurons upon ChABC-mediated CS-GAG degradation [53].