Address correspondence and reprint requests to Makoto Hamanoue, Department of Physiology, Toho University School of Medicine, 5-21-16 Ohmori-nishi, Ohta-ku, Tokyo 143-8540, Japan. E-mail: firstname.lastname@example.org
The isolation of neural stem cells (NSCs) from the brain has been hampered by the lack of valid cell surface markers and the requirement for long-term in vitro cultivation that may lead to phenotype deterioration. However, few suitable specific cell surface antigens are available on NSCs that could be used for their prospective isolation. The present study demonstrated that the expression of complex type asparagine-linked oligosaccharide (N-glycans) was detected on brain cells dissociated from embryonic and adult brain using Phaseolus vulgaris erythroagglutinating lectin (E-PHA) which binds to biantennary complex type N-glycans, and demonstrated that E-PHA bound preferentially to purified NSCs, but not to neurons, microglia, or oligodendrocyte precursor cells. The labeling of dissociated mouse embryonic brain cells or adult brain cells with E-PHA enabled the enrichment of NSCs by 25-fold or 9-fold of the number of neurosphere-forming cells in comparison to that of unsorted cells, respectively. Furthermore, a lectin blot analysis revealed the presence of several glycoproteins which were recognized by E-PHA in the membrane fraction of the proliferating NSCs, but not in the differentiated cells. These results indicate that complex type N-glycans is a valuable cell surface marker for living mouse NSCs from both the embryonic and adult brain.
Dulbecco’s Modified Eagle’s Medium:Nutrient Mixture F-12 Ham
epidermal growth factor
Phaseolus vulgaris erythroagglutinating lectin
glial fibrillary acidic protein
neural stem cell
Neural stem cells (NSCs) are an appealing resource for the development of replacement therapies for neurodegenerative diseases as they can both proliferate and differentiate into neurons, astrocytes and oligodendrocytes. Characterizing NSCs is critical, but isolating them from the brain has so far proven to be challenging. NSCs can be isolated from free floating spherical aggregates termed ‘neurospheres’ that result from proliferation of NSCs in response to both epidermal growth factor (EGF) and fibroblast growth factor-2 (for reviews see, Gage et al. 1995; McKay 1997; Okano 2002). However, the formation of neurospheres requires long-term in vitro cultivation and may lead to phenotypic deterioration. To circumvent this potential hazard of long-term culture in vitro, other methods for isolating NSCs from brain tissue with specific markers have been investigated.
Oligosaccharide chains that bind to asparagine (N-glycan) or serine/threonine (O-glycan) are major cell surface substrates. Among these two types of glycans, N-glycan expression is preferentially detected in the nervous system (Krusius and Finne 1977; Chen et al. 1998; Zamze et al. 1998). N-glycans are classified into three types according to their branched oligosaccharide chains: high mannose type (mannose), complex-type (N-acetylglucosamine) and hybrid type (mannose and N-acetylglucosamine). Specific enzymes convert the mannose type N-glycans into the complex-type and hybrid-type N-glycans, including N-acetylglucosaminyltransferase. Mice deficient for N-acetylglucosaminyltransferase are unable to convert the mannose type N-glycans to the complex-type N-glycans and exhibit a disproportionate size reduction of the cephalic area and neural tube closure defects at embryonic day 10.5 (E10.5) brain (Ioffe and Stanley 1994; Metzler et al. 1994). Biantennary complex-type N-glycans are preferentially expressed at the lateral ganglionic eminence where NSCs emerge at E11 (Flaris et al. 1995), N-glycans on NSCs may play some roles in neural development. There are two plant lectins, Phaseolus vulgaris erythroagglutinating lectin (E-PHA) and concanavalin A (ConA), which bind to biantennary complex-type N-glycans and to mannose type N-glycans, respectively (see review, Zanetta 1998). A recent study demonstrated that NSCs are recognized by E-PHA, and that mouse embryonic NSCs firmly attach to E-PHA-coated wells through cell surface N-glycan, and reported the isolation of embryonic NSCs using E-PHA-lectin coated wells (lectin panning method) (Hamanoue et al. 2008). However, the target molecules for E-PHA on NSCs remains unknown, and the isolation of adult NSCs by lectin panning method is considerably more difficult because of the existence of myelin debris (unpublished observation by Hamanoue et al.).
The present study used E-PHA and a flow cytometry system to investigate whether E-PHA could be a useful marker for isolating NSC from both the embryonic and adult brain, while also demonstrating the existence of the several glycoproteins in the membrane fractions of the embryonic brain.
Materials and methods
Pregnant and adult ICR mice were purchased from Sankyo Laboratory Service (Tokyo, Japan). The experimental protocols were approved by the research ethics committee of Toho University.
All chemical reagents were purchased from Sigma Chemical Co. (St. Louis, MO, USA) and Wako Pure Chemicals Industries (Osaka, Japan) unless otherwise noted. All culture materials were purchased from Nalge Nunc International (Rochester, NY, USA) unless otherwise noted.
See supporting information for the on-line version (Appendix S1).
Flow cytometric analysis of dissociated brain cells and primary culture cells
The dissociated cells (106 cells) were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 10 min at 25°C and then stained with Alexa Fluor 680-conjugated-E-PHA (A680-E-PHA) and 680-conjugated-ConA (A680-ConA) for 30 min at 4°C (1 ng lectin per 105 cells or per 0.1 mL of PBS). A680-conjugated lectins were prepared with an Alexa Fluor 680 Protein Labeling Kit according to the manufacturer’s instructions (Molecular Probes, Eugene, OR, USA). After washing, the cells were filtered through a 70-μm cup filter and analyzed on a FACSCalibur using CELLQuest software (BD Transduction Laboratories, Lexington, KY, USA).
The cells were treated with peptide-N-glycosidase F (PNGase F; 25 U/mL), Endoglycosidase H (Endo H; 100 mU/mL) and α-mannosidase (5 U/mL; Seikagaku Corp., Tokyo, Japan), in PBS for 16 h at 37°C in order to remove the cell surface N-glycans before lectin staining. Digestion with PNGase F was performed in PBS containing 0.75% Triton X-100.
Cell sorting by flow cytometry
The dissociated brain cells were stained with FITC-conjugated lectins [E-PHA and ConA, 1 ng lectin per 105 cells or per mL Dulbecco’s Modified Eagle’s Medium:Nutrient Mixture F-12 Ham (DMEM/F-12)] and/or the following phycoerythrin-conjugated antibodies for 30 min on ice; CD15 (Lewis X, mouse IgM), CD24 (rat IgG2c), CD31, or CD133 (prominin-I, 13A4, rat IgG1; all purchased from eBioscience Inc., San Diego, CA, USA). After washing, the cells were stained with propidium iodide (PI, 1 μg/mL) and filtered through a 70-μm cup filter. Cell sorting was performed on a MoFlo High-Performance Cell Sorter (Dako Denmark A/S, Glostrup, Denmark). Erythrocytes, dead cells and cell debris were excluded by gating on forward and side scatter and by eliminating PI positive events. The viable cells were then sorted into the growth-promoting medium and after sorting, the cells were washed with DMEM/F-12 medium.
Capacity for self-renewal and multipotency of neurosphere forming cells
To detect the capacity for self-renewal, 103 cells were plated in each well of a 96-well plate in growth-promoting medium. After 5–7 days, the resultant neurospheres were dissociated into a single cell suspension and replated into the wells of a 96-well plate. Their capacity for self-renewal was determined by the resultant secondary neurosphere formation 5–7 days after replating. To detect the capacity for multipotency, secondary neurospheres were transferred to poly-l-lysine and laminin-coated dishes and cultured in DMEM/F-12 medium containing B27 supplement without growth factors for 5–7 days. The differentiated cells were fixed and stained with antibodies for βIII-tubulin (1 : 100; mouse IgG, Chemicon, Temecula, CA, USA), glial fibrillary acidic protein (GFAP; 1 : 100) and O4 (1 : 2000; mouse IgM, Chemicon). After washing, the cells were stained with secondary antibodies conjugated with Alexa Fluor 488, Cy3 (GE Healthcare, BioScience AB, Uppsala, Sweden) and biotin (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) or with Alexa Fluor 680-conjugated streptavidin (Molecular Probes) for 1 h at 25°C. The nuclei were visualized with 13 ng/mL 4′6′-diamidino-2-phenylindole (DAPI). The fluorescence images were observed with a LSM510 META laser-scanning microscope (Carl Zeiss, Inc., Oberkochen, Germany).
Cell viability assay
The viability of the E14 brain cells was assessed by measuring the concentration of intracellular ATP (Crouch et al. 1993) with a CellTiter-Glo Luminescent Cell Viability Assay (Promega Corp., Madison, WI, USA) according to the manufacturer’s instructions.
Sample preparation of brain homogenates and the lysates of membrane fractions
See supporting information for the on-line version (Appendix S1).
Immunoprecipitation, western blot and lectin blot analysis
Crude membrane fractions were lysed in a lysis buffer, and the proteins (500 μg) were immunoprecipitated with an antibody for CD133 (2 μg, eBioscience) or for EGF receptor (2 μg, rat monoclonal IgG2a, Santa Cruz Biotechnology). The immunocomplex was collected with Immunopure Protein A/G (Pierce, Rockford, IL, USA) and digested with or without PNGase F (1 U) in a lysis buffer for 4 h at 37°C. The immunocomplex samples were boiled in Laemmli’s sample buffer.
The samples were separated on a 4–10% sodium dodecyl sulfate–polyacrylamide gels. The proteins were transferred to polyvinylidene difluoride (PVDF) membrane. PVDF membrane was incubated in a Tris-buffered saline containing 5% bovine serum albumin and 0.1% Tween 20 and then incubated with A680-E-PHA (2 μg/mL), a biotin-conjugated anti-CD133 antibody or with anti-EGF receptor antibody (sheep polyclonal IgG, upstate biotech, 1 μg/mL) for 16 h at 4°C, followed by incubation with A680-conjugated streptavidin, or with biotin-conjugated goat anti-sheep IgG (1 : 1000, Santa Cruz Biotechnology) and IRDye800-conjugated anti-mouse IgG (1 : 5000, Rockland Inc., Gilbertsville, PA, USA). Fluorescence signals were visualized with an Infrared Imaging System (LI-COR Corp., Lincoln, NE, USA).
Statistical evaluations were performed using Student’s t-test. p <0.05 was considered to be statistically significant.
Expression of complex-type N-glycans recognized by E-PHA during mouse development
Although the expression of the complex-type N-glycans is necessary for CNS development (Ioffe and Stanley 1994; Metzler et al. 1994), the proportion of N-glycans during development remains unclear. In the present study, cells dissociated from the telencephalon and thalamus were prepared at different developmental stages (E12–E16) as well as from the adult (P10w) brain. The proportion of N-glycan was evaluated using the plant lectins E-PHA and ConA as probes for oligosaccharide chains on cell surfaces. The dissociated brain cells were fixed and labeled with Alexa Fluor 680-conjugated-lectins, and then analyzed using a flow cytometer (Table 1). During the developmental stages, a very small percentage of cells consistently bound E-PHA. This percentage did not change throughout the development (3.8–4.5%), but it did slightly increase in the P10 week brain (11.0%). To ascertain whether this E-PHA binding required oligosaccharides on the cell surface, the fixed dissociated brain cells were treated with several types of glycosidases in order to remove cell surface oligosaccharides before probing with each lectin. The digestion of E14 dissociated brain cells with PNGase F, an enzyme that cleaves all types of N-glycans from glycoproteins, including mannose, complex and hybrid types, significantly decreased the binding of both E-PHA and ConA to cells by 39.5% and 47.8%, respectively; in comparison to that with no enzyme (Table 1, E14 + PNGase F). Therefore, a large fraction of lectin binding is mediated by cell surface N-glycans. Next, to ascertain whether E-PHA binding requires complex-type N-glycans, cells were treated with mannosidase or Endoglycosidase H (+Endo H), enzymes that cleave mannose type N-glycans or mannose and hybrid type N-glycans, respectively. Treatment with mannosidase significantly decreased the binding of ConA by 21.7%, but rather increased E-PHA binding by 118.4% (+Mannosidase). In comparison, treatment with Endo H significantly decreased the binding of ConA by 47.8%, however, it did not alter the binding of E-PHA (+Endo H). Therefore, a small population of brain cells that retained complex-type N-glycans was present from embryogenesis to adulthood.
Table 1. Percentage of E-PHA-positive cells during the development
Dissociated brain cells from the cerebral hemispheres stained with A680-conjugated lectins were analyzed using a FACSCalibur flow cytometer. The cells were treated with several types of glycosidases before labeling to remove cell surface N-glycans. The mean (n = 3) from three independent experiments is shown. *p <0.05 and **p <0.01 in comparison to the cells with no glycosidases. A very small percentage of the brain cells consistently exhibited E-PHA binding during development. Endo H, Endoglycosidase H; N.D., not done.
4.5 ± 0.6
3.8 ± 0.2
1.5 ± 0.1**
4.5 ± 0.1*
3.8 ± 0.4
4.5 ± 0.8
11.0 ± 3.7
2.3 ± 0
1.1 ± 0.1**
0.5 ± 0.1**
1.1 ± 0.2*
E-PHA binds to purified NSCs and astrocytes, but not neurons
To determine which type of brain cells are recognized by E-PHA, Nestin-positive NSCs, neurofilament-positive neurons, GFAP-positive type I astrocytes, A2B5-positive oligodendrocyte precursor cells (OPCs) and microglia were purified from embryonic mouse telencephalon to > 90% purity as estimated by cell-type-specific antibodies. The purified cells were fixed with 4% paraformaldehyde and stained with A680-E-PHA. According to a flow cytometric analysis, E-PHA displayed greater binding to cultured NSCs and astrocytes (39.8 ± 1.2% and 26.2 ± 0.7%, n = 3, respectively) than to neurons (2.1 ± 0.3%), oligodendrocyte precursor cells (0.2 ± 0.04%) and microglia (0.8 ± 0.04%; Fig. 1). Therefore, E-PHA binding is a characteristic of both NSCs and type I astrocytes.
E-PHAmid cells isolated from embryonic brain by flow cytometry exhibit an NSC phenotype
The binding of E-PHA to cultured NSCs prompted us to examine whether E-PHA binding could be used to isolate NSCs from mouse cerebral hemispheres. The dissociated E14 brain cells were labeled with FITC-E-PHA and then sorted by fluorescence intensity with a flow cytometer. Erythrocytes and dead cells were excluded by gating on forward and side scatter and by eliminating PI positive events. The cells were classified into three parts with E-PHA, E-PHA negative (E-PHAlow), E-PHA weakly positive (E-PHAmid), E-PHA strongly positive (E-PHAhigh) according to the fluorescence intensity (Fig. 2a). The percentage of these classified cells were 87.8 ± 1.1%, 8.2 ± 1.0% or 4.0 ± 0.8%, respectively (n = 8). To ascertain whether living NSCs could be isolated from mouse cerebral hemispheres with E-PHA, sorted cells were washed and cultured at low density (104 cells/mL) in serum-free growth-promoting medium containing fibroblast growth factor-2 and EGF. After 7 days in vitro, the number of neurospheres per thousand cells with a diameter > 50 μm was counted. Approximately 40 neurospheres were formed per thousand E-PHAmid cells, whereas only a few neurospheres were observed from either the E-PHAlow cells, E-PHAhigh cells or unsorted cells. The neurosphere frequency was calculated by dividing the number of neurospheres in the sorted cells by the number of plating cells (Table 2). The neurosphere frequency of E-PHAmid cells was 25-fold higher than that of unsorted cells (Table 2, n = 35). In comparison, less than 10 neurospheres were formed by ConAlow cells, and the neurosphere frequency of ConAlow cells was five-fold lower than that of E-PHAmid cells (n = 12). Furthermore, the neurosphere frequency of E-PHAmid was compared with that of the cell surface markers previously used for NSC preparation: CD15 (LeX), CD24 and CD133 (prominin I). The neurosphere frequency of CD15high or CD24low was 5 or 15-fold lower than that of E-PHAmid cells, respectively. The neurosphere frequency of CD133high was 1.4-fold lower than that of the E-PHAmid cells (n = 6), but the combination of CD133high with E-PHApositive achieved an even greater enrichment in comparison to that of either single stained cells (1.6-fold or 1.1-fold) or unsorted cells (28-fold). Therefore, neurosphere forming cells can be significantly enriched from the E14 brain cells through the E-PHA recognition of the cell surface.
Table 2. Neurosphere frequency of the sorted cells from embryonic brain
Dissociated brain cells from the cerebral hemispheres stained with FITC-conjugated E-PHA were sorted with the flow cytometer. The neurosphere frequency was calculated by dividing the number of neurospheres in the sorted cells by the number of plating cells. The mean plus the standard error of the mean (SEM; n = 6–36) from at least three independent experiments is shown for each condition. Purification with E-PHA permits a significant enrichment of neurosphere forming cells from the E14 brain.
0.16 ± 0.02
0.11 ± 0.03
3.96 ± 0.29
0.38 ± 0.09
0.86 ± 0.21
0.36 ± 0.15
0.20 ± 0.02
0.76 ± 0.08
0.27 ± 0.10
0.26 ± 0.02
0.62 ± 0.12
2.82 ± 0.30
4.50 ± 0.26
To verify that these neurosphere-forming cells from E-PHAmid cells possess the NSC phenotype, primary neurospheres were dissociated into single cells and cultured for 7 days. Formation of secondary neurospheres by individual cells is a hallmark of NSCs (Fig. 2b). Next, secondary neurospheres were transferred to poly-l-lysine/laminin-coated coverslips and cultured in medium free of serum and growth factors for 5 days to confirm the multipotency of NSCs. Adherent cells displayed distinct morphologies consistent with their expression βIII-tubulin, GFAP, or O4 (Fig. 2c). Therefore, a fraction of E-PHAmid cells contains NSCs.
Adult neural stem cells can be isolated by flow cytometry with E-PHA
Phaseolus vulgaris erythroagglutinating lectin binding to the cells from adult mouse brain was observed by flow cytometry (Table 1). To ascertain whether living NSCs could be isolated from adult mouse brain with E-PHA, the dissociated cells were sorted with E-PHA by flow cytometry. Erythrocytes and cell debris were excluded by gating on forward and side scatter (Fig. 3a, R1). The cells were classified into two parts, E-PHAlow and E-PHAhigh (Fig. 3b) according to the fluorescence intensity and the sorted cells were cultured for 7 days. A few neurospheres formed per 1000 E-PHAhigh cells, whereas very few neurospheres formed from E-PHAlow cells or unsorted cells. The neurosphere frequency of E-PHAhigh was nine-fold higher than that of E-PHAlow or unsorted cells (Table 3). The NSC phenotype of these neurosphere-forming cells from E-PHAhigh cells were verified by formation of secondary neurospheres (Fig. 3c) as well as by the multipotency of these NSCs to express βIII-tubulin, GFAP, or O4 under differentiating conditions (Fig. 3d). Therefore, living NSCs can be significantly enriched from the adult brain through the E-PHA recognition of cell surface.
Table 3. Neurosphere frequency of the sorted cells from the adult brain
Dissociated brain cells from the cerebral hemispheres stained with FITC-conjugated E-PHA were sorted using the flow cytometer. The neurosphere frequency was calculated by dividing the number of neurospheres in the sorted cells by the number of plating cells. The mean plus the standard error of the mean (SEM; n = 10–12) from at least three independent experiments is shown for each condition. Purification with E-PHA permits significant enrichment of neurosphere forming cells from the P10w brain.
0.03 ± 0.02
0.02 ± 0.01
0.26 ± 0.06
E-PHA labeling does not increase the number of embryonic NSCs
Phaseolus vulgaris erythroagglutinating lectin has been reported to be one of the prototypes of mitogenic lectins for T-lymphocytes (Nowell 1960) and microglial cells (Liuzzi et al. 1999). One potential explanation for the enrichment of NSCs in populations of E-PHAmid cells is that the labeling of NSCs with E-PHA promotes proliferation, thus resulting in the production of more neurospheres. To test this possibility, dissected E14 brain cells were labeled with FITC-E-PHA at the indicated concentration for 1 h at 37°C and then were cultured for 7 days after extensive washing. The viable cell number was determined by measuring the amount of intracellular ATP. Treatment with E-PHA did not increase the number of living brain cells and rather decreased the number at higher concentrations above 0.3 μg/mL in comparison to the cells without E-PHA (Fig. S1). Therefore, the enrichment of NSCs by E-PHA is a result of isolating NSCs rather than because of the promotion of proliferation.
E-PHA recognizes cell surface glycoproteins
Phaseolus vulgaris erythroagglutinating lectin is a plant lectin which binds to biantennary complex-type N-glycans (Zanetta 1998). To investigate whether E-PHA binds to glycoproteins on NSCs, brain homogenates were prepared from E14 mouse brain specimens and the glycoproteins were analyzed. A lectin blot analysis revealed that fluorescent E-PHA recognized various molecules with a wide range of molecular weight in the brain homogenate (Fig. 4a, LB, left lane). After the treatment of the brain homogenate with PNGase F, which removes all the N-glycan oligosaccharides from the asparagine residues, these E-PHA binding to the homogenate were clearly diminished (Fig. 4a, LB, 1U, 5U). The migration pattern of the proteins was not significantly altered by PNGase F-treatment on Coomassie Brilliant Blue (CBB)-stained PVDF membrane (Fig. 4a, CBB stain). Therefore, E-PHA binds to several glycoproteins in the mouse embryonic brain through the recognition of N-glycan oligosaccharides.
To investigate whether these glycoproteins were expressed specifically on the NSC cell surface, proliferating NSCs and the differentiated cells prepared from the proliferating NSCs were cultured. Each membrane fraction of the cultured cells was then analyzed by a lectin blot analysis. The glycoproteins, especially, higher molecular weight proteins recognized by E-PHA were specifically expressed in the crude membrane fraction of the proliferating NSCs, and then these proteins disappeared in the differentiated cells (Fig. 4b, LB, arrowheads). Therefore, E-PHA recognizes several membrane glycoproteins in the mouse NSCs.
E-PHA binds to the N-glycans of CD133 or EGF receptors in the embryonic brain
Considering that E-PHA is as effective as an antibody directed against CD133 for isolating NSCs (Table 2) and a previous report that E-PHA binds to the EGF receptor (Rebbaa et al. 1996), it was possible that E-PHA could recognize the carbohydrate epitopes of both CD133 and the EGF receptor. To verify these interactions, CD133 or EGF receptor was immunoprecipitated from a crude membrane fraction of E14 cerebral hemispheres and the immunocomplex was digested with or without PNGase F. The anti-CD133 antibody recognized both the glycosylated form (Fig. 5a, left lane, arrow, 130 kDa) and deglycosylated form (arrowhead, 87 kDa) of the immunoprecipitates. After incubation with PNGase F, only the deglycosylated form of CD133 remained (Fig. 5a, right lane, arrowhead). The anti-EGF receptor antibody recognized a band at the predicted molecular weight of the EGF receptor in samples not treated with PNGase F (Fig. 5c, left lane, arrow, 190 kDa). However, in samples incubated with PNGase F, the electrophoretic mobility of this band increased (Fig. 5c, right lane, arrowhead, 170 kDa), thus indicating PNGase F sensitivity. To determine whether E-PHA could bind to the glycosylated form of immunoprecipitates for the two antibodies on same membrane, the blots were stripped and reprobed with A680-E-PHA. E-PHA recognized the glycosylated form of CD133 (Fig. 5b, left lane, arrow) as well as the EGF receptor (Fig. 5d, left lane, arrow) and such binding decreased in the lanes treated with PNGase F (Fig. 5b and d, right lane, arrow). Therefore, E-PHA binds to complex type N-glycans of both CD133 and the EGF receptor in the mouse embryonic brain.
The brain cells recognized by E-PHA
Large amounts of high mannose type N-glycans and a relatively small amount of complex and hybrid type N-glycans have been detected in the adult brain with chemical carbohydrate analysis including gas chromatography and matrix-assisted laser desorption/ionization-mass spectrometry (Krusius and Finne 1977; Chen et al. 1998; Zamze et al. 1998). The present study demonstrated that the existence of a small number of complex type N-glycan-positive cells during mouse brain development. The experiment using the purified cultured cells revealed that complex type N-glycan-positive cells included Nestin-positive NSCs and GFAP-positive astrocytes. Most of astrocytes in these experiments were purified from embryonic mouse brain through multiple passages and exhibited epitheloid morphology typical of type I astrocytes. Type 1 astrocytes proliferate in response to EGF, whereas type 2 astrocytes do not (Seidman et al. 1997), and GFAP-expressing cells can generate neurons and glial cells throughout the CNS during development (Casper and McCarthy 2006). Furthermore, E-PHA binding was observed in a human pediatric high grade astrocytoma (Rebbaa et al. 1999). These results suggested that GFAP-positive type I astrocytes may have a similar N-glycan expression pattern to that of NSCs.
The current flow cytometry-based experiment demonstrated that living NSCs, which have a self-renewal activity and multipotency, can be isolated from the embryonic and adult mouse brain as E-PHA-positive cells. Among these E-PHA-positive cells in the embryonic brain, approximately 30% of the cells which showed a high fluorescence intensity of E-PHA did not exhibit a higher neurosphere frequency than that of the unsorted cells. A subset of these E-PHAhigh cells express platelet/endothelial cell adhesion molecule-1 (Newman 1994) by fluorescence-activated cell sorter (FACS) analysis using anti-platelet/endothelial cell adhesion molecule-1 antibody (unpublished observation by Hamanoue et al.). This result indicates that collecting the E-PHAmid cells among E-PHA-positive cells thus enables a more effective NSCs isolation from embryonic brain.
On the other hand, in the adult brain, we could not determine the precise E-PHAmid fraction that exhibits a high neurosphere frequency (data not shown). One plausible explanation is that the relative affinity of NSCs for E-PHA in the adult brain cells may be slightly different from that in the embryonic brain. Our results which demonstrated the E-PHA-binding in the adult brain to increase in comparison to that in the embryonic brain could also support this hypothesis. Further characterization of E-PHA-positive cells is therefore necessary for the improvement of NSC isolation.
Effects of E-PHA on cell function
As E-PHA binds to neurosphere forming cells, the neurosphere forming cells could be enriched by sorting E-PHA binding cells using flow cytometry. E-PHA not only binds to complex-type N-glycans (see review, Zanetta 1998), but it has also been reported to be mitogenic for T-lymphocytes (Nowell 1960) and microglial cells (Liuzzi et al. 1999) at a concentration higher than 1–10 μg/mL, thus raising the possibility that the observed enrichment of NSCs in E-PHA binding cells may be a consequence of the mitogenic activity of E-PHA. As the E-PHA concentration used in the current study was significantly lower (10 ng/mL) and the short-term E-PHA treatment did not promote proliferation of NSCs even at higher concentrations, NSC enrichment was thus a consequence of isolating NSCs cells from mouse brain rather than an effect of proliferation. However, at a higher concentration (0.3 μg/mL), E-PHA exhibits cell growth inhibition through the competitive inhibition of the EGF receptor with the ligand EGF (Rebbaa et al. 1996), therefore a low concentration of E-PHA of below 0.1 μg/mL is strictly required to achieve optimal NSCs isolation.
Targets of E-PHA
Of the many cell surface markers used previously as markers for NSCs, glycoproteins are the likely binding partners for E-PHA rather than gangliosides, proteoglycans or LeX, because the oligosaccharides of these molecules differ from the biantennary complex type N-glycans (Gooi et al. 1981; Zanetta 1998). The present study demonstrated the existence of several glycoproteins recognized by E-PHA in the membrane fractions from the cerebral hemisphere. Among these membrane glycoproteins, E-PHA was confirmed to bind to the carbohydrate epitope of the EGF receptor using a lectin blot analysis. Therefore, E-PHA binding to the EGF receptor likely contributes to NSC isolation. Ciccolini et al. used a fluorescent EGF conjugate for NSCs isolation from embryonic and adult brain instead of antibodies specific for EGF receptor as most antibodies specific for EGF receptor recognize an intracellular epitope (Ciccolini et al. 2005). However, the separation process with EGF should be performed quickly and kept at 4°C to minimize the risk that receptor/ligand turnover could decrease cell isolation (Chen et al. 1989; Ciccolini et al. 2005). In contrast, the incubation time and temperature during cell separation does not appear to be a critical factor for use of E-PHA. Because short-term E-PHA treatment could bind to the multiple cell surface molecules including EGF receptor on NSCs and this binding efficiently inhibits the growth of NSCs for 7 days after the administration of this treatment.
The membrane glycoprotein CD133 has both Endo-H-sensitive mannose type N-glycans and Endo-H-resistant and PNGase F-sensitive complex type N-glycans (Florek et al. 2005) and E-PHA bound to the carbohydrate epitope of CD133. Although CD133 is an efficient marker of stem cells in human or mouse fetal tissue (Uchida et al. 2000; Corti et al. 2007; Coskun et al. 2008), NSCs cannot been isolated from adult brain by monitoring CD133 expression (Pfenninger et al. 2007). In contrast, NSCs can be isolated from both embryonic and adult brain by monitoring E-PHA, thus indicating that E-PHA is useful cell surface marker available for all ages. E-PHA could recognize glycoproteins other than CD133, including the EGF receptor, in the adult brain.
Previous reports demonstrated that several glycoproteins, other than CD133 and EGF receptor, have the complex N-glycans, including; CD3, T-cell receptor (Chilson and Kelly-Chilson 1989), annexin V (Gao-Uozumi et al. 2000), glycoprotein on human neuroblastoma cells, LA-N-1 (Ross et al. 1995), arginine receptor (Finger et al. 1996) and P-glycoprotein (Rebbaa et al. 1999). However, the present study demonstrated that several molecules exist in the membrane fractions of the proliferating NSCs that are larger than the glycoproteins previously reported. Furthermore, these glycoproteins on the proliferating NSCs disappears during differentiation into matured cells, thus indicating the specific expression of complex type N-glycans on the NSC membrane. In addition, a recent study suggested that a kind of cell adhesion molecules with complex type N-glycan may express on NSCs (Hamanoue et al. 2008). A future analysis of these target molecules recognized by E-PHA on NSC is necessary not only for further improvement of NSC isolation but also to better understand the function of complex type N-glycan on NSCs.
In conclusion, the present study demonstrated that NSCs can be efficiently isolated with E-PHA by flow cytometry from both the embryonic and adult brain. The current flow cytometry-based experiment has two advantages in comparison to the lectin panning method with E-PHA that enable the purification of embryonic NSC (Hamanoue et al. 2008). First is that performing flow cytometry with E-PHA makes the purification of adult living NSCs possible. Although the lectin panning method is quite a rapid method without disturbing the NSC viability during separation and is also suitable for analyzing the intact NSCs signaling from the embryonic brain, the lectin panning method is not suitable for adult NSC purification because of the existence of myelin (unpublished observation by Hamanoue et al.). Second is the fact that flow cytometry with E-PHA makes it possible to achieve a higher enrichment of NSCs. Especially, it is important to achieve further improvements in NSC purification by combining E-PHA and the other cell surface markers. This technique avoids the potential complications associated with long-term in vitro cultivation, and helps with the precise understanding of the function of native NSCs for cell replacement therapies addressing a variety of neurodegenerative diseases.
We thank Takashi Yamada for support with flow cytometry. This work was supported by a grant from the Ministry of Education, Culture, Sports, Science and Technology (MEXT; to M.H., Y. M., K.T. and H. O.), a project research grant from Toho University School of Medicine (to M.H.), a grant from the Solution-Oriented Research for Science and Technology (SORST) program of the Japan Science and Technology Agency (JST) and also supported in part by a grant-in-aid from the 21st Century COE program of the Ministry of Education, Culture, Sports, Science and Technology of Japan to Keio University. The mouse monoclonal antibody against Nestin, developed by Susan Hockfield, was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the National Institute of Child Health and Human Development (NICHD) and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242, USA.