B2 cells selected from undifferentiated cell clusters present in oligodendrocyte-enriched cultures were used as parent cells, and were prepared from whole brains of 12-week-old BALB/c mice. The detailed methods for selecting B2 cells and their characterization have been described previously (11). Briefly, during the first week of culture, about 90% of the cells exhibit surface immunolabeling of galactocerebroside (GalC), a cell type-specific marker for oligodendrocytes. After 4 weeks in vitro, the proportion of oligodendrocytes in oligodendroyte-enriched cultures is reduced to 40-60%. Non-oligodendroglial cells, including 10-20% glial fibrillary acidic protein (GFAP)-positive astrocytes, and 30-40% small bipolar cells morphologically distinct from oligodendrocytes or astrocytes, populate the cultures. These small bipolar cells aggregate each other, and form a cell cluster or sphere, and have 8 µm cell bodies and 20-30 µm cell processes. The cells are harvested by a brief treatment for 10 min with phosphate buffered saline (PBS) containing 0.1% trypsin and 1 mM ethylenediaminetetraacetic acid (EDTA). They are then suspended in culture medium, and replated into T75 culture flasks. After 3 months with 6 serial passages, the culture consisted of a morphologically homogeneous cell population, designated as B2 cells. They were further maintained for 6 months with 25 serial passages in a culture medium, containing 10% horse serum (HS). The identity of mouse-derived cells was verified by karyotype analysis with G-banding technique (12). The total chromosome number of B2 cells was 40, which is identical to normal somatic cells of BALB/c mice.

B2 cell cultures were dissociated into single cells, suspended in 10% HS-containing medium, and plated in the absence of feeder cells in a 96-well microtiter plate at a density of one cell/well. The medium was renewed by half once per week. After three weeks, well-separated colonies were picked for further expansion. This procedure was repeated once more. Only one out of the 28 clones, designated as B2A1, successfully generated and expanded. Stock cultures of B2A1 cells were grown in 75 cm² tissue culture flasks in the serum-free chemically defined medium DM4 supplemented with 10% fetal bovine serum (FBS) (13). The DM4 medium consisted of Dulbecco's modified Eagle's medium (DMEM) supplemented with 10 µg/mL insulin, 10 µg/mL of transferrin, 0.3 nM of triiodothyronine, 30 nM of sodium selenite and 50 nM of hydrocortisone. All chemicals described above, except for insulin (Novo Laboratories, Willowdale, ON, Canada), were obtained from Sigma (St. Louis, MO, USA). Cultures were maintained at 37℃ in a humidified incubator containing 5% CO₂, and the cells were passed once a week.

For immunocytochemical studies, B2A1 cells were plated on poly-L-lysine (PL)-coated 9 mm round Aclar coverslips at a density of 1×10⁴ cells/coverslip. The cells were also cultured in the medium for five days, fixed in cold acid alcohol (5% acetic acid in 95% ethanol) for 15 min at -20℃, and incubated with normal goat serum for 20 min at room temperature (then processed for immunocytochemistry). To identify neural cell phenotypes, cell cultures were processed for immunostaining using cell type-specific markers: nestin for neural stem cells; low molecular weight neurofilament protein (NF-L) and β tubulin III for neurons; GFAP for astrocytes; GalC and myelin basic protein (MBP) for oligodendrocytes; and, Ricinus communis agglutinin-1 lectin (RCA-1) for microglia. The cultures were incubated with primary antibodies, followed by biotinylated secondary antibodies and avidin-biotin complex (ABC, Vector, Burlingame, CA, USA) and visualized with 3-amino-9-ethyl carbazole chromogen (AEC, Sigma). The antibodies utilized for immunochemical characterization of neural cell types are listed in Table 1.

To investigate effects upon neural cell differentiation by various growth factors and pharmacological agents in B2A1 cells, the cells were plated on PL-coated coverslips at a density of 1×10⁴ cells/coverslip. They were incubated for 72 hr in DMEM (DM4), with triiodothyronine and hydrocortisone supplement, in the presence of one of the following: 10% FBS, with various mixtures of 10 µg/mL of insulin, 10 µg/mL of transferrin, 30 nM selenite (ITS) (Sigma), 1mM dibutyryl cyclic AMP (dbcAMP) (Sigma), 10 µM retinoic acid (RA) (Sigma), 100 ng/mL of recombinant human basic fibroblast growth factor (bFGF) (GIBCO BRL, Gaithersburg, MD, USA), 100 ng/mL of recombinant human platelet-derived growth factor (PDGF) AA (Upstate Biotechnology, Lake Placid, NY, USA), 10 ng/mL of leukocyte inhibitory factor (LIF, Sigma) or 100 nM phorbol 12-myristate 13-acetate (PMA, Sigma). The immunocytochemical stains for nestin, β tubulin III, GFAP and GalC were processed from the cultures as described above.

For reverase transcriptase-polymerase chain reaction (RT-PCR) analysis, B2A1 cells were plated on 6 well plates at a density of 5×10⁵ cells/well and incubated for 5 days in DM4 with addition of one of the above cytokines. RT-PCR was performed in the cells harvested from each well using the oligonucleotide primers listed in Table 2 (14, 15). The expression of nestin, Notch1, and platelet-derived growth factor receptor-α (PDGFR-α) for neural stem cell markers were determined. To confirm the neuronal property of B2A1 cells, complementary DNA (cDNA) was amplified 35 PCR cycles for NF-L. In addition, we looked for other CNS cell type markers, GFAP for astrocytes, and MBP for oligodendrocytes. Glyceraldehyde-3-phosphate dehydrogenase (G3PDH) was used as a reaction standard. Total RNA was extracted from each sample using TRIzol reagent (GIBCO-BRL). cDNA templates from each sample were prepared from 2 µg of total RNA primed with oligo dT primers (Pharmacia, Gaithersberg, MD, USA) using 400 units of Moloney murine leukemia virus (MMLV) reverse transcriptase (GIBCO-BRL) followed by 35-40 PCR amplification cycles (94℃ for 30 sec, annealing at 55℃ for 60 sec and extension at 72℃ for 90 sec). Ten microliters of each PCR product were analyzed by 1.5% agarose gel electrophoresis. Authenticity of bands was determined by selective enzyme digestion.

For toxicity studies, B2A1 cells were seeded on both Aclar coverslips and 96-well plates at a density of 1×10⁴ cells per coverslip or well in DMEM containing 10% fetal bovine serum, 5 mg/mL glucose, 25 µg/mL of gentamicin, and 2.5 µg/mL of amphotericin B, and cultured for 7 days. The cultures were changed to a medium containing no fetal bovine serum 24 hr prior to cytotoxic treatment, and cells were exposed to the chemicals for an additional 24 hr. All chemicals were purchased from Sigma and the concentrations used for treatment were: glutamate (20 µM-2 mM), kainate (20 µM-2 mM), hydrogen peroxide (H₂O₂, 100 µM-1 mM), and buthionine sulfoximine (BSO) (1-100 µM).

Cytotoxicity was evaluated by cytopathologic changes and cell viability of B2A1 cells after treatment with exogenous ROS and excitotoxic chemicals. Cytopathologic features of B2A1 cells on Aclar coverslips were examined using a phase-contrast inverted microscope.

In order to measure viability of the B2A1 cells, the 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) colorimetric assay was performed (16). This method is based on the cellular reduction of tetrazolium salt MTT (Sigma) by the mitochondrial dehydrogenase of viable cells, which results in formation of a blue formazan product which can then be measured with a spectrophotometer. In brief, MTT was dissolved in DMEM at a concentration of 5 mg/mL, filtered through a membrane filter (45 µm in diameter), and stored at 4℃ in an aluminum foil-sealed tube. Following addition of MTT stock solution into the cultures in 96 well plates (10 µL per 100 µL medium), the cultures were incubated for 4 hr at 37℃. A portion of the MTT was converted to an insoluble bluish purple formazan by cleavage of the tetrazolium ring by cellular dehydrogenase enzymes. The medium was removed, and the formazan product was dissolved with 100 µL of acid-isopropanol (0.04N hydrochloride in isopropanol) per well. Optical density was measured using an ELISA plate-reader at a wave length of 570 nm. The test was repeated five times. Statistical analysis was done using a one way ANOVA, followed by Dunnett's multiple comparison test. The level of significance was set at P<0.05. All statistical analyses were conducted by the SPSS 15.0 statistical software program (SPSS 15.0, Chicago, IL, USA).

In serum-free DM4 medium, B2A1 cells were expressed as either bipolar or tripolar cells with a 7-8 µm cell body size. They had a tendency to grow individually or to aggregate into small clumps (Fig. 1A). Nestin immunopositivity was 97.5% in the cells (Fig. 1B).

A limited number of cells expressed cell type specific markers for neurons, such as NF-L (2.1%) and β tubulin III (1.9%) (Fig. 1C). In addition, other CNS cell markers, including GFAP (astrocyte marker), and GalC and MBP (oligodendrocyte markers) were positive using immunocytochemistry in 2.9%, 2.3%, and 1.7% of the cells, respectively (Fig. 1D). RCA-1 as a microglial marker had no immunoreactivity in the cells. The above immunoreactivities for CNS cell type specific markers are summarized in Table 3.

Cells reacting positively for nestin, β tubulin III, GFAP and GalC per total B2A1 cells following incubation in DM4 serum-free medium, treated with serum or various growth factors, are presented in Table 3. When incubated in 10% FBS, ITS, dbcAMP, bFGF, PDGF, LIF or PMA-containing culture medium, immunoreactivity for nestin was positive in 85.4-95.8% of B2A1 cells (Fig. 2A). The proportion of nestin positive cells decreased in RA-containing culture medium (84.7%), while β tubulin III positive cells were significantly increased (15.6%) (P<0.05). When treated with PDGF, 10% FBS, dbcAMP, RA or bFGF, β tubulin III was expressed in fewer cells (2.1-2.8%). Only a few cells were positive for β tubulin III (less than 0.5%) in ITS, LIF, or PMA treated culture medium (Fig. 2B).

When incubated in dbcAMP or PDGF-containing culture medium, GFAP immunolabeling was seen in 14.3-16.0% of B2A1 cells, while 10% FBS, ITS, RA, bFGF or LIF-containing medium was expressed in just 2.2-7.1% (P<0.01). Immunopositivity for GFAP was seen in 2.2% of B2A1 cells cultured in PMA-containing medium (Fig. 2C). When cultured in 10% cAMP or PDGF-containing medium, GalC positive cells were seen in 7.3-8.9% of B2A1 cells, while only 2.9-5.8% of the cells were positive in 10% FBS, RA, ITS, bFGF, or PMA-containing medium (P<0.05) (Fig. 2D).

To determine whether B2A1 cells can express markers of neural stem cells, as well as of early differentiating neurons, astrocytes and oligodendrocytes, RT-PCR was performed in cultured cells treated with various cytokines. Expression levels were compared with the control expression of the housekeeping gene G3PDH, and the data are summarized in Fig. 3. Stem cell or early precursor cell markers of CNS origin (nestin, Notch1, PDGFR-α), neuronal markers for NF-L, astrocytic markers for GFAP, and oligodendroglial markers for MBP, were consistently expressed. Even though the intensity of expression demonstrated some variation using treated serum or growth factors, the expression of genotypic markers correlated relatively well with immunocytochemical phenotypes.

Compared with control cells (Fig. 4A), swelling of cell bodies was the initial common cytopathologic hallmark in B2A1 cells treated with the cytotoxic chemicals for 4-6 hr. This was followed by fragmentation and loss of neurites, loss of cell to cell contact. Nuclear homogenization was then seen during the longer incubation time of 24 hr. Cells treated with H₂O₂ (500 µM) or BSO (1-100 µM) developed swelling of cell bodies with focally expanding parts of cytoplasmic process (axon hillocks), vacuolar changes with increased granularity of cytoplasm and cytoplasmic processes (neuritis) (Fig. 4B), and progressive disintegration of the nuclear chromatin (Fig. 4C). Glutamate or kainate-treated cells (20-2,000 µM) also showed swelling of cell bodies at 4-6 hr; however, swelling was limited to the cytoplasm of the cells and was more homogenous when compared with H₂O₂ or BSO-treated cells. As time increased, the nuclear chromatin and cytoplasm of the cells also disintegrated, and dead cells formed homogenous round bodies (Fig. 4D).

In order to investigate the relationship of H₂O₂ exposure to B2A1 cell death, the concentration of H₂O₂ which induces cell death was measured using the MTT test (Fig. 5). There was no significant cell death seen 24 hr after treatment with 1-100 µM H₂O₂ (Fig. 5A). However, viable cells were significantly reduced 24 hr after the addition of 500 µM H₂O₂ (67.7±2.6%) and 1 µM BSO (65.6±2.9%) (P<0.01 compared to control cultures). Viable cells were also reduced to 61.3-65.6% 24 hr after treatment with 1-100 µM BSO (Fig. 5B). In 20-2,000 µM glutamate or kainate group treated for 24 hr, viable cells were decreased to 66.6-74.4% (Fig. 5C, D) (P<0.01 compared to control cultures). A dose-response pattern of cell death was not present in the BSO, glutamate or kainate treated groups in this experiment.