Ependymomas often show characteristics similar to those of neural stem cells in vivo and in vitro. However, few ependymoma cell lines that exhibit neural stem cell-like properties have been reported. In this study, we have characterized a novel cell line, designated Vn19, established from ependymoma that arose in mice inoculated intracerebrally with human BK polyomavirus. Transplanted Vn19 cells in nude mice ubiquitously expressed viral large T antigen in the nucleus and coexpressed neuronal and glial marker proteins in vivo. Remarkably, individual Vn19 cells in dispersed cultures simultaneously expressed marker proteins of neural stem cells (nestin, Bmi1, CD133), neurons (βIII tubulin, neurofilament-M) and glial cells (glial fibrillary acidic protein, A2B5, S100β, O4). Ubiquitous and homogenous expression of these multilineage marker proteins was also observed in cloned Vn19 cells. The Vn19 cells formed neurosphere-like aggregates when cultured in the presence of growth factors. Quantitative RT-PCR analysis revealed that expression of mRNA for nestin, neurofilament-H and glial fibrillary acidic protein significantly increased in Vn19 cells cultured under growth factor-deprived conditions. Among MAGE (melanoma antigen) family genes, MAGE-A (A1-8), MAGE-B (B1-3), MAGE-D1, MAGE-E1, MAGE-G1 (necdin-like 2) and MAGE-H1 were expressed in the Vn19 cells, in which neither necdin nor MAGEL2 was detectable. These results suggest that this murine ependymoma cell line recapitulates the gene expression profile in ependymal cells undergoing malignant transformation. (Cancer Sci 2011; 102: 122–129)
A growing body of evidence suggests that cancer cells originate from the transformation of normal stem cells and include their own stem cells with long-term self-renewal capability.(1) Major malignant brain tumors such as medulloblastoma, glioblastoma and ependymoma often contain cancer stem cells.(2,3) Among these brain tumors, ependymomas occur in early childhood and show characteristics similar to those of neural stem cells (NSC) in vivo and in vitro.(4,5)
Ependymal cells have recently attracted considerable attention in postnatal neurogenesis.(6) Earlier studies have demonstrated that ependymal cells generate new neurons that migrate to the olfactory bulb and serve as a quiescent stem cell population in the adult forebrain.(7,8) In contrast, another study has shown that ependymal cells are postmitotic and have no capability of self-renewal after differentiation at postnatal stages.(9) Although the identity of periventricular NSC remains elusive, ependymal cells might potentially exhibit NSC-like characteristics under specific conditions such as malignant transformation to ependymoma. Therefore, ependymoma cell lines might be useful for studying the molecular mechanisms underlying the malignant transformation of ependymal cells. Murine cell lines are especially valuable because a great deal of molecular and genetic information has been accumulated on mice. However, few mouse ependymoma cell lines that exhibit NSC-like characteristics have been reported to date.
Human BK virus (BKV), a polyomavirus that expresses its oncoprotein large T antigen, induces ependymomas with high efficiency when injected into newborn rodents.(10,11) In this study, we established a cell line from BKV-induced mouse ependymoma. This cell line coexpresses neural multilineage markers of neurons, glia cells and NSC. We also report the expression profile of MAGE (melanoma antigen) family genes, which include the tumor rejection antigen genes and the imprinted genes necdin and MAGEL2. Because expression of several MAGE family genes is controlled through epigenetic modulations, this ependymoma cell line will be useful for studying epigenetic mechanisms underlying the malignant transformation of ependymal cells.
Materials and Methods
Antibodies. Primary antibodies used for immunohistochemistry and immunocytochemistry are as follows: rabbit polyclonal antibodies against large T antigen(12) (1:500), 68 kDa neurofilament (NF-L) (1:50; Oncogene, Cambridge, MA, USA), glial fibrillary acidic protein (GFAP) (1:500; Progen, Heidelberg, Germany) and nestin (ST-1; 1:1000); mouse monoclonal antibodies against βIII tubulin (1:1000; Promega, Madison, WI, USA), Ki-67 (B56; 1:200; BD Bioscience, San Jose, CA, USA), 160 kDa neurofilament (NF-M) (NN18; 1:40; Sigma-Aldrich, St Louis, MO, USA), O4 (MAB345; 1:500; Millipore, Billerica, MA, USA), Bmi1 (F6; 1:300; Millipore), A2B5 (hybridoma supernatant 1:40; American Type Culture Collection, Manassas, VA, USA) and S-100 β-subunit (S100β)(SH-B1; 1:1000; Sigma-Aldrich); rat monoclonal antibody against CD133 (13A4; 1:200; eBioscience, San Diego, CA, USA); and rat polyclonal antibody against nestin (STR1; 1:500). Secondary antibodies are cyanine 2-conjugated anti-rabbit and anti-mouse IgG (both 1:500; Jackson ImmunoResearch, West Grove, PA, USA), and cyanine 3-conjugated anti-mouse, anti-rat and anti-rabbit IgG (all 1:500; Jackson ImmunoResearch), Alexa Fluor 350-conjugated anti-rabbit IgG (1:500; Invitrogen, San Diego, CA, USA). Chromosomal DNA was detected with 3.2 μM Hoechst 33342 (Sigma-Aldrich). Anti-nestin antibodies were raised against a synthetic peptide of carboxyl-terminal 15 amino acid residues of mouse nestin in rabbit (ST-1) and rat (STR1).
Ependymoma cell line. A 20 μL solution containing 3 × 1011 viral particles of the oncogenic BKV strain pm522(13) was injected into the right cerebral hemisphere of 13 newborn Balb/c mice within 24 h after birth. Fifteen months later, one mouse developed ependymoma (original name, Vn1919; diagnosed pathologically by Dr. T. Muto). A piece (50 mg) of the ependymoma tissue was then transplanted subcutaneously into 3–4-week-old Balb/c mice. Two weeks later the tumors were removed, cut into small pieces, treated with 0.25% trypsin and incubated in RPMI-1640 culture medium containing 20% FCS. Cultured cells were again transplanted subcutaneously into mice. The transplantable cells were maintained in DMEM supplemented with10% FCS at 37°C in humidified 5% CO2 atmosphere, and the cells with at least 30 passages were used as a stable ependymoma cell line, designated Vn19. To confirm the homogeneity of the Vn19 cell line, the Vn19 cells were cloned by limited dilution and cultured in growth-factor-reduced Matrigel matrix (Becton Dickinson, San Diego, CA, USA). To examine the effects of growth factors, the Vn19 cells were cultured in the presence of epidermal growth factor (EGF) (20 ng/mL; PeproTech, Rocky Hill, NJ, USA) and basic-fibroblast growth factor (bFGF) (20 ng/mL; PeproTech) or in the absence of the growth factors in a defined medium (DMEM/F12 + N2) (Invitrogen) for 5 days. The P19 embryonal carcinoma cells (a gift from Dr. M. McBurney) were cultured and differentiated into neurons that were enriched by cytosine arabinoside treatment.(14) Experiments using mice were approved by the Animal Experiment Committees of the Institute for Protein Research, Osaka University, and performed in accordance with the institutional guidelines and regulations.
Immunohistochemistry and immunocytochemistry. The Vn19 cells (5 × 106 cells) were injected into the back of BALB/c Slc-nu/nu mice (Japan SLC, Hamamatsu, Shizuoka, Japan). For immunohistochemistry, subcutaneous tumors were removed 2 weeks after injection, and the tumor tissues were fixed with 4% paraformaldehyde solution in 0.1 M phosphate buffer (pH 7.4) overnight and cryoprotected by immersion in 20% sucrose overnight. Frozen 12-μm-thick tissue sections were immersed in a target retrieval solution (S1699; Dako, Glostrup, Denmark) at 92°C for 25 min and then allowed to cool at room temperature for 25 min. The sections were incubated at 4°C with primary antibodies overnight and with fluorescence dye-conjugated secondary antibodies at room temperature for 45 min. For immunocytochemistry, the Vn19 cells were cultured on Matrigel-precoated dishes in 10% FCS-supplemented DMEM, fixed with methanol-acetone solution at −20°C for 20 min, and incubated with primary antibodies (overnight at 4°C) and fluorescence dye-conjugated secondary antibodies (for 90 min at room temperature). For CD133 and A2B5 immunostaining, cells were incubated with primary antibodies for 1 h at room temperature before fixing. Images were observed with a fluorescence microscope (BX-50-34-FLAD1; Olympus Optical, Tokyo, Japan), taken by a charge-coupled device camera system (DP70; Olympus Optical), and processed using Photoshop CS2 software (Adobe, San Jose, CA, USA).
RT-PCR. Total RNA was extracted with Trizol reagent (Invitrogen), and cDNA was synthesized from total RNA (5 μg) using Transcriptor First Strand cDNA Synthesis kit (Roche, Basel, Switzerland). Aliquots (2%) of cDNA were used as templates for PCR with primers listed in Table 1. The RT-PCR products were quantified using a real-time cycler (LightCycler with FastStart Program; Roche) and a reagent kit (FastStart DNA MasterPLUS SYBR Green I).(15) Statistical significance of the quantified data was tested using the Student’s t test. For semi-quantification, PCR was carried out in a 20 μL reaction mixture containing 5 μL cDNA, 2 × GC buffer I, 2.5 mM of each dNTP, 1 unit Taq DNA polymerase and 20 pmol each primer using a thermal cycler (GeneAmp PCR System 2700; Applied Biosystems, Foster City, CA, USA). The PCR conditions used were: 94°C for 1 min of initial denaturation followed by 32 cycles of denaturation (94°C for 30 s), primer annealing (54°C, 30 s for necdin, MAGEL2, MAGE-B, MAGE-D1, MAGE-E1, MAGE-G1, MAGE-H1 and Dlx2; 51°C, 30 s for MAGE-A and Bmi1; 60°C, 30 s for Sox2, CD133 and GAPDH) and primer extension (72°C for 1 min) followed by final extension at 72°C for 10 min. The PCR products were analyzed by 1.5% agarose gel electrophoresis.
Table 1. RT-PCR primers
Primer sequences used for quantitative and semi-quantitative RT-PCR analyses of mouse mRNA (cDNA) are listed. NCBI accession numbers are shown in the Reference column. GFAP, glial fibrillary acidic protein; MAGE, melanoma antigen; NF-L, neurofilament-L; NF-H, neurofilament-H.
Expression of neuronal and glial marker proteins in Vn19 cell transplants in vivo. Original tumor cells were prepared from the pathologically diagnosed ependymoma Vn1919, transplanted subcutaneously and cultured as dispersed cells. After repeating the cell transplantation procedure seven times, the cells were subcultured to passage numbers greater than 30 and used in the following experiments as an established cell line (designated Vn19). To examine whether Vn19 cells retain the tumorigenicity, we injected them subcutaneously into nude (nu/nu) mice and analyzed the expression of cell marker proteins in tumor cells in vivo (Fig. 1). Perivascular rosettes, a histopathological hallmark of ependymomas, were found in the transplanted tumor tissues by hematoxylin–eosin staining (Fig. 1A). Immunohistochemistry revealed that most of the cells expressed BKV-derived large T antigen and the early differentiated neuron marker βIII tubulin (Fig. 1B,C). Remarkably, almost all tumor cells expressed the neuronal marker NF-L and the astrocyte marker GFAP (Fig. 1D,G). The cells immunopositive for NF-L and GFAP also expressed the proliferative cell nuclear marker Ki-67 (Fig. 1E,F,H,I), suggesting that the cells expressing neuronal and astrocyte markers are proliferative. Furthermore, the transplanted Vn19 cells coexpressed GFAP and NF-M (Fig. 1J–L). In addition, part (approximately 30%) of the NF-L-positive cells were also positive for the oligodendrocyte marker O4 (Fig. 1M–O). These results suggest that individual Vn19 cells express neural multilineage markers in vivo in a simultaneous manner.
Expression of neural multilineage marker proteins in dispersed Vn19 cultures. We then used the dispersed cultures to examine whether these markers are expressed at a single cell level (Fig. 2). The Vn19 cells were positive for large T antigen and the stem cell marker Bmi1 in the nucleus (Fig. 2A,B). In addition, all Vn19 cells (100%) were positive for CD133 (NSC and ependymal cell marker), nestin (NSC marker), NF-M, GFAP, A2B5 (glial-restricted precursor marker), S100β (astrocyte and ependymal cell marker) and O4 (Fig. 2C–I). Remarkably, the Vn19 cells formed neurosphere-like aggregates when cultured in the presence of EGF and bFGF. Triple-staining analysis revealed that cells in the spheres were immunopositive for nestin, βIII tubulin and GFAP (Fig. 2J–L). Although expression levels of each marker protein differed slightly in individual cells of the spheres, all Vn19 cells (100%) expressed these marker proteins (Fig. 2M–O). These expression patterns of the cell markers suggest that this cell line consists of a homogenous cell population.
Expression of neural multilineage marker proteins in cloned Vn19 cells. In an attempt to further analyze the homogeneity of the Vn19 cell line, we cloned it by the limited dilution method and examined the expression levels of neural multilineage markers in each clone by immunocytochemistry (Fig. 3). All cloned cells showed a similar morphology (mostly spindle-shaped bipolar) and expressed nestin (Fig. 3A–D), NF-M (Fig. 3E–H), GFAP (Fig. 3I–L) and O4 (Fig. 3M–P) to similar extents. There were no appreciable differences in the cell morphology and expression levels of the marker proteins between these clones. The ubiquitous expression of these marker proteins in the cloned cells suggests that the Vn19 cells consist of a homogenous cell population. Thus, we used original (non-clonal) Vn19 cells for the following analyses.
Expression of neural multilineage genes in growth-factor-deprived cultures. We then examined whether the expression levels of neural multilineage marker genes were altered in the presence and absence of growth factors (Fig. 4). The Vn19 cells treated with EGF and bFGF formed neurosphere-like aggregates (Fig. 4A, middle panel). In the FCS-free growth-factor-deprived medium, most of the Vn19 cells lost their adhesiveness and some appeared dead (Fig. 4A, right panel). Under these culture conditions, we quantified mRNA for nestin (Fig. 4B), NF-L (Fig. 4C), NF-H (Fig. 4D) and GFAP (Fig. 4E) by real-time RT-PCR. There were no appreciable changes in the mRNA levels of nestin, NF-H and GFAP between the control cells (cultured in FCS-supplemented medium) and the cells cultured in the presence of EGF and bFGF. Only the NF-L mRNA level was markedly reduced in the FCS-free growth-factor-supplemented cultures, suggesting that FCS contains NF-L transcription stimulating factors. In contrast, deprivation of growth factors significantly increased the mRNA levels of nestin, NF-H and GFAP, whereas the NF-L mRNA level was significantly reduced by this treatment. These results suggest that expression of specific cell marker genes is suppressed in FCS-supplemented Vn19 cultures and increased under serum-deprived conditions. These responses to the growth factors, which are different from those of normal NSC, may be unique to this cell line.
Gene expression of stem cell markers and MAGE family in Vn19 cells. We determined the expression levels of the NSC marker genes by RT-PCR analysis (Fig. 5A). Sox2, an undifferentiated stem cell marker, was consistently expressed in Vn19 cells (Fig. 5A, top panel). CD133 mRNA was detected in Vn19 cells and its level was slightly increased in the presence and absence of EGF and bFGF (Fig. 5A, second panel). Bmi1 mRNA, which encodes a Polycomb complex protein that promotes proliferation of stem cells, was also detected and its levels were almost consistent in the presence and absence of the growth factors (Fig. 5A, third panel). On the other hand, Dlx2 mRNA, which encodes a homeodomain protein expressed in neuronal progenitors (transit amplifying cells) in the subependymal zone of adult mice, was undetectable in the Vn19 cells (Fig. 5A, fourth panel). As a control cell line, we used murine embryonal carcinoma P19 cells, which differentiate into neural multilineage cells in response to retinoic acid.(16) The mRNA levels of Sox2 and CD133 expressed in the P19 cells were reduced upon neuronal differentiation, whereas those of Bmi1 and Dlx2 were undetectable under undifferentiated conditions but slightly increased in the P19 cell-derived neurons.
We next examined whether the MAGE family genes are expressed in the Vn19 cells (Fig. 5B). Necdin mRNA was undetectable in the Vn19 cells in the presence or absence of the growth factors, whereas it was highly expressed in the P19 cell-derived postmitotic neurons and embryonic brain, but low in undifferentiated P19 cells and testis (Fig. 5B, top panel). MAGEL2 mRNA was also absent from the Vn19 cells, but was detectable in the P19 cells and embryonic brain (Fig. 5B, second panel). The mRNAs for MAGE-A (A1-8) and MAGE-B (B1-3) were weakly expressed in Vn19 cells, and their levels were slightly reduced in growth-factor-treated cultures (Fig. 5B, third, fourth panels). In contrast, neither the P19 cells nor embryonic brain expressed these MAGE-A and MAGE-B subfamily mRNAs. On the other hand, the mRNAs for MAGE-D1, MAGE-E1, MAGE-G1 (also known as necdin-like 2) and MAGE-H1 were expressed in the Vn19 cells, P19 cell-derived postmitotic neurons, embryonic brain and testis to varying extents (Fig. 5B, fifth to eighth panels).
The mouse ependymoma cell line Vn19 displays characteristic expression patterns of neural multilineage-specific genes. Pediatric brain tumors such as anaplastic astrocytoma, medulloblastoma and glioblastoma multiforme contain a significant fraction of dual-fate cells that coexpress βIII tubulin and GFAP(2) In contrast to these brain-tumor-derived NSC-like cells, the pluripotent embryonal carcinoma P19 cells, like normal NSC, differentiate initially into postmitotic neurons and then into astrocytes when treated with high concentrations of retinoic acid.(16,17) Furthermore, neurally differentiated P19 cells exhibit gene expression patterns similar to those seen in normal neural development and aging.(18,19) This may be because embryonal carcinoma cells are transformed from germ cells and retain characteristics seen at early stages of normal development. On the other hand, ependymal cells are born at late embryonic stages and undergo maturation during the first postnatal week.(9) Because the present cell line has been established from mice inoculated with BKV during the neonatal period, it is possible that BKV infected differentiating ependymal cells at the latest stages of neural lineage differentiation. Thus, brain tumors including ependymoma might be transformed after differentiation of neural multilineage and exhibit the gene expression patterns in dual-fate (i.e. neuronal and glial lineages) cells.
MAGE family proteins are the precursors of anti-tumor antigens that have been identified in melanomas.(20) Expression of MAGE family genes in human cancers has been extensively studied, but information about the expression of murine MAGE family genes in cancers is limited. MAGE family members are divided into two types based on their similarities in the MAGE homology domain:(21) Type I MAGE subfamily members such as MAGE-A and B groups (plus C group in humans) are expressed exclusively in cancers and testes, whereas type II MAGE subfamily members including necdin and MAGEL2 are expressed in somatic cells including neurons. Necdin, which was originally isolated from neurally differentiating P19 cells, is expressed predominantly in postmitotic cells such as neurons and skeletal muscle cells.(14,22) Necdin has strong growth suppressive activities(23,24) and is absent from transformed cell lines originating from the nervous system such as neuroblastomas and gliomas.(25) Downregulation of endogenous necdin expression is associated with the transformation of normal melanocytes to melanomas.(26) These findings suggest that necdin plays a tumor suppressor role in carcinogenesis.(27) It is noteworthy that necdin, like retinoblastoma family proteins, interacts with viral oncoproteins including SV40 large T antigen and cellular E2F family proteins.(24,28) Thus, BKV large T antigen may target necdin and retinoblastoma family proteins in ependymal cells to trigger the initial step for malignant transformation. Although endogenous necdin expression in differentiated ependymal cells remains unclear, it is tempting to speculate that the initial inactivation of endogenous necdin by viral large T antigen followed by the permanent silencing of necdin expression contributes to the transformation of ependymal cells to ependymomas.
Accumulating evidence suggests that expression of MAGE family genes is controlled by epigenetic modifications. Expression of MAGE-A subfamily genes in cancer cells is controlled by epigenetic mechanisms such as DNA methylation and histone acetylation.(29) Expression of the necdin gene is controlled by genomic imprinting, a parent-specific DNA methylation-dependent silencing in placental mammals.(30,31) In normal cells, the maternal necdin allele is completely silenced by DNA methylation, and necdin is expressed only from the paternal allele. We have recently found that CpG islands near the proximal promoter region of the necdin gene are highly methylated in Vn19 cells but low in P19 cells (Tsuyoshi Ohkumo and Kazuaki Yoshikawa, unpublished data, 2010). This cell line may provide a valuable tool for studying the epigenetic mechanisms underlying the malignant transformation of ependymal cells.
This work was supported by a Grant-in-Aid for Scientific Research B2 (21300138) from the Japan Society for the Promotion of Science (to KY). We deeply thank the late Dr. Seijiro Uchida, the late Dr. Takeshi Muto, Dr. Setsuko Shioda and Dr. Akemi Nozawa for development of the original Vn1919 cells, Dr. Yoshiaki Yogo and Dr. Hiroshi Ikegaya for valuable advice, and Ms Kazumi Imada for technical assistance.