A 6-year-old male patient was admitted to the hospital for headache, nausea and vomiting, which had continued for 1 month. Magnetic resonance imaging (MRI) revealed a solid mass in the pineal region with hydrocephalus (Fig. 1A). The patient underwent a series of chemotherapy and radiation therapy following endoscopic third ventriculostomy and biopsy, which were diagnosed as PB. The tumor recurred on the left frontal lobe 10 months after the initial therapy and was removed using an occipital transtentorial approach. The final pathological diagnosis was also consistent with PB. A year later, the tumor recurred in the fronto-parietal region and the patient's condition deteriorated (Fig. 1B). The tumor was removed again and a fresh tumor specimen was obtained in the operating room through the cryostat laboratory. The patient relapsed with leptomeningeal seeding 2 months after the surgery and subsequently passed away, despite a series of additional surgery, chemotherapy and radiation therapy. The patient diagnosed with GBM was a 61-year-old male, who was presented with tingling sensation in the left hand and twitching in the left arm and face. MRI showed a 3.5-cm-sized enhanced mass in the right parietal lobe. The patient underwent total removal of the tumor. Pathology showed GBM and the patient underwent concurrent chemoradiation therapy as well as a series of chemotherapy (temozolomide). Chemotherapy was reinitiated after radiological evidence of tumor recurrence.

A surgical specimen from the rPB patient was freshly obtained from the operating room. Informed consent was provided according to the Declaration of Helsinki. Neuropathologists diagnosed each surgical specimen according to the WHO classification criteria (19). TSs were isolated from the rPB specimen using a modification of previous methods for TS isolation from human brain cancers (11,13,15,20,21). Briefly, the cell isolation procedure was performed within 60 min of PB removal using a mechanical dissociation method. The surgical specimen was minced and dissociated with a scalpel in Dulbecco's modified Eagle's medium/Nutrient Mixture F-12 (DMEM/F-12; Mediatech, Manassas, VA, USA) and then passed through a series of 100-μm nylon mesh cell strainers (BD Falcon; BD Biosciences Franklin Lakes, NJ, USA). Cell suspensions were washed twice in DMEM/F-12 and cultured in complete TS media composed of DMEM/F-12 containing B27 supplements (1X; Invitrogen, San Diego, CA, USA), 20 ng/ml of basic fibroblast growth factor (bFGF; Sigma, St. Louis, MO, USA), 20 ng/ml of epidermal growth factor (EGF; Sigma), and 50 U/ml penicillin/50 mg/ml streptomycin (11,22–25). In addition, glioblastoma (GBM) TSs (TS13-20), isolated from a primary GBM patient, were used for comparison with rPB TSs.

For investigation of surface and intracellular antigen expression profiles, rPB TSs were transferred to cover slides, fixed with 2% paraformaldehyde for 7 min, and then treated with a 3:1 ratio of methanol and acetic acid for 3 min. The cells were then washed and permeabilized by incubating with 0.1% Triton X-100 for 10 min. After blocking with 1% bovine serum albumin (BSA; Amresco, Solon, OH, USA) for 1 h, cells were incubated with primary antibodies for 2 h at room temperature. The following antibodies were used: rabbit anti-CD133 (1:250, ab19898; Abcam Dawinbio Inc., Hanam, Korea), rabbit anti-musashi (1:250, ab52865; Abcam) and rabbit anti-podoplanin (1:250, ab10274; Abcam). Primary antibody against CD133 was detected with goat anti-rabbit IgG conjugated with Alexa Fluor 555 (1:2,000; Invitrogen), which is spectrally similar to Cy3. Alexa Fluor 488-conjugated goat anti-rabbit IgG (1:2,000; Invitrogen) was used to detect antibodies against musashi and podoplanin. The cells were mounted with Vectashield H-1200 mounting media containing 4′6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA) to counterstain nuclei. Phosphate-buffered saline (PBS; Dawinbio Inc., Hanam, Korea) was used for all washing steps, and antibody diluent reagent solution (Invitrogen) was used to dilute antibodies. As a negative control, only the secondary antibody was used. A fluorescence microscope (1X71; Olympus Korea, Co., Ltd., Seoul, Korea) and DP Controller software (Olympus Korea) were used for observing and photographing the cells.

Total RNA was isolated from rPB TSs (TS13-19) and GBM TSs (TS13-20) using an RNeasy kit (Qiagen, Hilden, Germany). cDNA was synthesized from 2 μg of total RNA using the SuperScript II First-Strand Synthesis system (Invitrogen). The following oligonucleotide primer pairs were used for PCR: GAPDH (glyceraldehyde-3-phosphate dehydrogenase), 5′-AGG GGT CTA CAT GGC AAC TG-3′ and 5′-ACC CAG AAG ACT GTG GAT GG-3′; CD133, 5′-GCC AGC CTC AGA CAG AAA AC-3′ and 5′-TAC CTG GTG ATT TGC CAC AA-3′; podoplanin, 5′-CCA GCG AAG ACC GCT ATA AB-3′ and 5′-AGA GGA GCC AAG TCT GGT GA-3′; musashi-1, 5′-ACC CCC ACA TTC TCT CAC TG-3′ and 5′-GAG ACA CCG GAG GAT GGT AA-3′; GFAP (glial fibrillary acidic protein), 5′-AGA TCC ACG AGG AGG AGG TT-3′ and 5′-CGG CGT TCC ATT TAC AAT CT-3′; Olig2 (oligodendrocyte transcription factor 2), 5′-CAG AAG CGC TGA TGG TCA TA-3′ and 5′-AAGGGTGTTACACGGCAGAC-3′; TUBB3 (β-tubulin III), 5′-CAT CCA GAG CAA GAA CAG CA-3′ and 5′-GCC TGG AGC TGC AAT AAG AC-3′; β-catenin, 5′-GCT TGG TTC ACC AGT GGA TT-3′ and 5′-GAG TCC CAA GGA GAC CTT CC-3′; snail, 5′-GAG CAT ACA GCC CCA TCA CT-3′ and 5′-TTG GAG CAG TTT TTG CAC TG-3′; Zeb1 (zinc finger E-box binding homeobox 1), 5′-GAC AGG GCT GAA AGT AGT CAA GC-3′ and 5′-GGT AGT TAG CAC GGG TTG GA-3′.

Total RNA from rPB TSs was prepared using an RNeasy kit (Qiagen) and reverse transcribed using a First-Strand cDNA Synthesis kit (Invitrogen) according to the manufacturer's instructions. Changes in expression levels were quantitatively analyzed using a real-time PCR machine (StepOne Plus; Applied Biosystems, Foster City, CA, USA). Specific commercial TaqMan probes (Applied Biosystems) were used to quantify levels of the following mRNAs: CD133, podoplanin, nestin, musashi-1, Sox2, Oct4, β-catenin, snail and Zeb1. For analysis of relative gene expression, rPB TSs were tested in triplicate; transcript levels of GBM TSs were monitored as an internal control.

The multipotency of rPB TSs was tested by examining neural lineage expression by immunocytochemical staining, as previously described (11,20,26). Briefly, after being seeded onto chamber slides (Lab-Tek II; Nalge Nunc International, Rochester, NY, USA), cells were grown in neural differentiation media containing 10% fetal bovine serum (FBS; Lonza) and 1X B27 supplement (Invitrogen) for up to 14 days. Cells were then fixed with 2% paraformaldehyde for 7 min at 4°C, and permeabilized by incubating with 0.1% Triton X-100 for 10 min. After blocking with 1% BSA (Amresco) for 1 h, cells were immunostained with the following antibodies: rabbit anti-GFAP (1:200 dilution; Dako, Carpinteria, CA, USA), mouse anti-MBP (myelin basic protein, 1:200 dilution; Chemicon International, Inc., Temecula, CA, USA), mouse anti-NeuN (1:100 dilution; Chemicon) and mouse anti-TUBB3 (Tuj1, 1:200 dilution; Chemicon). The primary antibodies were detected with Cy3-conjugated anti-mouse or anti-rabbit secondary antibodies (1:200 dilution; Jackson ImmunoResearch Laboratories, West Grove, PA, USA), as appropriate. Nuclei were counterstained with DAPI (Vector Laboratories). Slides were examined and photographed using a fluorescence microscope.

For invasion assays, collagen I/Matrigel matrices were prepared from 2.4 mg/ml of high-concentration rat tail collagen type I (Corning Life Sciences, Tewksbury, MA, USA), 2.1 mg/ml of Matrigel (Corning Life Sciences), 10% NaHCO₃ and 2X TS culture medium. The solution was well mixed and kept at 4°C before use. Single rPB TSs and GBM TSs were each placed individually onto 100 μl collagen I/Matrigel matrices in a 96-well plate and plates were incubated at 37.4°C in a 5% CO₂ environment for 30 min. After full gelation of the matrix, 100 μl of TS culture medium was added on top. The dynamic morphology of rPB TSs and GBM TSs was observed by collecting images using an inverted microscope (Optinity KI 400; Intron Biotechnology, Inc., Seongnam, Korea). For quantification, the maximal area covered by migrating edges of cells was used as the parameter for defining invasiveness, calculated as the invaded area at a certain time/spheroid area at initial time x 100. Data were analyzed using ToupView image analysis software (x64 v3.7.1460; AmScope, Irvine, CA, USA).

Four-to 8-week-old male athymic nude mice (Central Lab. Animal Inc., Seoul, Korea) were used for experiments. Mice were housed in micro-isolator cages under sterile conditions and observed for at least 1 week before study initiation to ensure proper health. Lighting, temperature and humidity were controlled centrally. All experimental procedures were approved by the Yonsei University College of Medicine Institutional Animal Care and Use Committee. Mice were anesthetized with a solution of Zoletil (30 mg/kg; Virbac Korea, Co., Ltd., Seoul, Korea) and xylazine (10 mg/kg; Bayer Korea, Seoul, Korea), delivered intraperitoneally. rPB TSs were implanted into the right frontal lobe of nude mice using a guide-screw system within the skull, as previously described (22,24,25,27–29). Mice received 5×10⁵ rPB TSs via a Hamilton syringe (Dongwoo Science, Co., Seoul, Korea), inserted to a depth of 4.5 mm. rPB TSs were injected into three mice simultaneously using a multiple micro-infusion syringe pump (Harvard Apparatus, Holliston, MA, USA) at a speed of 0.5 μl/min. Body weights of mice were checked every other day. If body weight decreased by >15% compared to the original weight, mice were euthanized according to the study protocol. Formalin-fixed, paraffin-embedded tissue blocks were used for the generation of slides for histologic examination. Hematoxylin and eosin (H&E)-stained sections were reviewed by a pathologist.

DNA was extracted from paraffin embedded tissues and their stem cells using a QIAamp DNA Mini kit (Qiagen) according to the manufacturer's instructions. DNA extracted from reference tissues and stem cells was measured using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and diluted to 1 ng/μl. For specific detection of human DNA in xenografts extracted from mouse tissue, DNA was quantified using a Quantifiler Duo DNA Quantification kit (Applied Biosystems), according to the manufacturer's instructions. The DNA quantity ranged from 42 to 126 pg/μl. Extracted DNA was tested for the genotypes of amelogenin and 23 forensic STRs (D3S1358, D1S1656, D2S441, D10S1248, D13S317, D16S539, D18S51, D2S1338, CSF1PO, TH01, vWA, D21S11, D7S820, D5S818, TPOX, D8S1179, D12S391, D19S433, FGA, D22S1045, Penta E, Penta D and DYS391). PCR amplification using PowerPlex Fusion (Promega Corp., Madison, MI, USA), Kplex-16 (BioQuest, Inc., Seoul, Korea) and Euplex-13 (BioQuest), performed according to the manufacturer's instruction using 1 ng of DNA extracted from human tissue and stem cells, were used to maximize STR genotypes obtained from degraded DNA in paraffin-embedded tissue. DNA extracted from mouse tissue was amplified by PCR essentially as described by the manufacturer, except that 5 μl of the extracted DNA was used and two additional PCR cycles were performed to increase the PCR yield of each multiplex system. All PCR amplifications were performed in duplicate in two additional independent experiments. PCR products were separated by capillary electrophoresis using an ABI PRISM 3130xl Genetic Analyzer (Applied Biosystems), and the results were analyzed using GeneMapper ID Software version 3.2 (Applied Biosystems). The genotypes of amelogenin and STRs were determined based on observation of each allele at least two additional times in independent tests.

Total RNA was extracted from rPB-TSs and 4 GBM-TSs that were established using a Qiagen miRNA kit according to the manufacturer's protocol. Expression profiles of TSs from rPB and GBM (control) were obtained using an Illumina HumanHT-12 v4 Expression BeadChip (Illumina, Inc., San Diego, CA, USA). Data were variance stabilizing transformed and normalized with the quantile normalization method using R/Bioconductor lumi package. Complete linkage hierarchical clustering with distance metric by taking distance = (1-corelation)/2, was performed and depicted as heatmap. Comparative Marker Viewer version 7.13 of GenePattern module was used for differential expression of two groups. Gene set enrichment analysis was performed using gene sets of hallmark signatures and those composed of genes upregulated in metastasis and epithelial mesenchymal transition, all of them provided by the Molecular Signatures Database (v5.0 MSigDB) (30).

Data are expressed as means ± standard deviations. Comparisons between two groups were done using the Student's t-test. P-values <0.05 or false discovery rate <25% were considered statistically significant.

Cells isolated from the rPB specimen yielded spheroids when cultured in TS complete media (Fig. 2A). These spheroids (TS 13-19), termed rPB TSs, proliferated ~5- to 6-fold in 10 days. This proliferation pattern was similar to that of GBM TSs (TS 13-20), which grew ~7-fold in 10 days (Fig. 2B).

To examine the stemness of rPB TSs, we used neurosphere formation assays, immunocytochemical staining and RT-PCR and qPCR analyses. Results from these studies were compared with those for GBM TSs. Neurosphere assays confirmed the presence of cells with extensive self-renewal ability in the rPB specimen. The size of neurospheres increased over time for both rPB TSs and GBM TSs (Fig. 3A), as could be seen grossly by light microscopy and as quantified by sphere radius measurements on days 3, 7 and 10. Immunocytochemical staining of rPB TSs identified cells expressing markers associated with stem cells and brain tumor stem cells (31), including CD133, podoplanin and musashi (Fig. 3B). To further assess the stemness of rPB TSs at the gene level, we used RT-PCR and qPCR analyses to measure CD133, podoplanin, nestin, musashi, Sox2 and Oct4 expression levels. Whereas both rPB TSs and GBM TSs expressed all stem cell surface markers, expression of CD133, podoplanin, musashi and Sox2 was significantly higher in rPB TSs (Fig. 3C and D), whereas expression of nestin and Oct4 was higher in GBM TSs.

To assess the multilineage differentiation capacity of rPB TSs, we cultured them in neuro-glial differentiation media as described in Materials and methods. Immunocytochemical staining with Tuj1 demonstrated that rPB TSs expressed TUBB3, a marker for immature neurons; however, they did not express other differentiation markers such as GFAP, MBP, or NeuN, which are specific for astrocytes, oligodendrocytes and mature neurons, respectively (Fig. 3E). RT-PCR analyses showed that rPB TSs expressed higher level of TUBB3, but lower levels of GFAP and Olig2 compared with GBM TSs (Fig. 3F). These results make sense given that pineocytes are specialized neurons (32), and give rise to tumors that are distinct from tumors of glial origin, such as GBMs, which express surface markers for oligodendrocytes, astrocytes and mature and immature neurons.

rPB TSs invasion patterns were investigated using 3D invasion assays. Light microscopy showed no overt evidence for invasion by rPB TSs at 0, 72 or 144 h, whereas GBM TSs were clearly invasive at 72 and 144 h (Fig. 4A). From 0 to 144 h, the area occupied by rPB TSs increased ~1.2-fold, whereas that of GBM TSs increased by 63.7-fold. At 144 h, the areas were significantly different between the groups (P<0.05). The relative invasiveness, calculated by comparing maximal areas, was 43% for rPB TSs and 5911% for GBM TSs (P<0.005). An RT-PCR analysis of epithelial-mesenchymal transition (EMT) markers detected expression of β-catenin and Zeb1 in rPB TSs and GBM TSs, and additionally detected snail in GBM TSs (Fig. 4B). A qPCR analysis found that all three EMT genes were expressed in both rPB TSs and GBM TSs, but were expressed at significantly lower levels in rPB TSs (Fig. 4C).

To assess the difference in the expression of genes between rPB TSs and GBM TSs, we performed whole-genome expression profiling. rPB TSs showed a gene expression profile distinct from that of GBM TSs of different origin, as depicted in the heatmap shown in Fig. 4D. Differential expression analysis revealed 4809 genes with false discovery rate <25%. Highly ranked genes in the comparative marker selection analysis included ARID3B, NOP56, RPOS4Y1 and JARID2, all of which were known to be expressed in embryonal tumors or cell lines derived from them. Gene set enrichment analysis revealed the activation of myc-targeted genes in rPB TSs with statistical significance whereas gene sets upregulated in metastasis and epithelial mesenchymal transition were activated in GBM TS.

To determine the tumorigenic potential of rPB TSs, we injected them into the right cerebrum of male athymic nude mice. Mice were euthanized when their weight decreased by >15% relative to their original body weight. A cross-sectional slice showed the presence of an intracerebral xenograft tumor (Fig. 5A, right image). To determine whether the xenograft tumor replicated phenotypes of the parent tumor, we compared histological features of xenograft tumors with those of the parent tumor tissue obtained in the initial surgery (Fig. 5A, right image, corresponding to the tumor in Fig. 1A) and from surgery after recurrence (Fig. 5A, middle image, corresponding to the tumor in Fig. 1B). A pathologist confirmed that histopathological findings of all three tissues showed common characteristics of embryonal tumors, including high cellularity, increased mitotic index and high nucleus-cytoplasm ratio. A test of the genotypes of amelogenin and 23 forensic STRs (D3S1358, D1S1656, D2S441, D10S1248, D13S317, D16S539, D18S51, D2S1338, CSF1PO, TH01, vWA, D21S11, D7S820, D5S818, TPOX, D8S1179, D12S391, D19S433, FGA, D22S1045, Penta E, Penta D and DYS391) showed that all autosomal STR profiles were identical, indicating that the genomic alterations found in the parental rPB were replicated in the corresponding xenograft tumors (Fig. 5B).