Since the first successful derivations of human embryonic stem cells (hESCs) from isolated inner cell mass cells of surplus in vitro fertilization embryos, rapid progress has been achieved in the characterization and manipulation of undifferentiated hESCs in vitro [1, 2]. However, the state of the art for working with hESCs is far from ideal. Much of the current knowledge regarding conditions for deriving and culturing the undifferentiated cells was based on work performed on murine and nonhuman primate ESCs, cells derived from embryonal carcinomas (ECs), and more recently, on the work accomplished on a relatively limited number of isolated hESC lines.
The availability of hESC lines in the rapidly advancing field of stem cell biology is currently limited in the United States. Restricted hESC access is primarily due to high cost, limited government funding, and complicated intellectual property issues. A temporary and imperfect solution to these conditions has been the use of EC cells in place of hESC lines for certain applications. As the number of potential therapies and different research applications increases, so does the demand for the production and characterization of additional hESC lines grown in conditions more suitable for clinical use. Moreover, the availability of additional cell lines for comparative studies will enable the establishment of standardized methods of propagation and better refinement of characterization criteria, a prerequisite to testing for therapeutic efficacy. Although karyotypically stable relative to other cell lines, hESCs occasionally undergo spontaneous trisomies during prolonged periods in culture . Here, we propose the establishment and use of such variant hESC lines (e.g., BG01V) to better address the imposed restrictions on normal hESC lines. Data presented here suggest that stable karyotypic variant hESC lines may more closely parallel normal hESC lines than EC lines.
The BG01 hESC line was established and characterized by BresaGen, Inc. (Athens, GA, http://www.bresagen.com.au) in 2001 and is listed on the National Institutes of Health (NIH) Stem Cell Registry (http://stemcells.nih.gov/research/registry/index.asp) [4, 5]. A stable variant hESC line, designated BG01V, was derived from the karyotypically normal BG01 hESC through routine enzymatic passaging [6, 7]. This cell line contains known and stable chromosomal aberrations (XXY, +12, +17) yet possesses characteristics similar to its normal parental line. Although karyotypically abnormal, the cell line proliferates in culture, expresses markers of pluripotency, and retains the ability to differentiate into cell types representative of the three germ layers. Variant hESC lines, relative to their normal parental lines, are easier to manipulate in vitro and recover more rapidly after passaging and cryopreservation [3, 6].
Historically, pluripotent human testicular EC cells, such as the NTERA-2 cl.D1 cell line, have provided a convenient means and a reasonable model for investigation of the molecular mechanisms by which stem cells commit to specific lineages during embryonic development. The NTERA-2 cl.D1 cells were derived from the NTERA-2 cell line, which was established from a nude mouse xenograft of TERA2. The TERA2 cell line was originally derived from a metastasis of a human testicular carcinoma [8, 9]. The NTERA-2 line has been shown to differentiate to neuroectodermal lineages after exposure to retinoic acid or hexamethylene bisacetamide [10, 11]. Although EC cells can provide a convenient and robust experimental system, their differentiation capacity is often limited, unlike ESC differentiation .
To gain a better understanding of the differentiative potential of the BG01V cell, several genetic and protein-based assays were used to fully characterize the line. The resulting data were compared with NTERA-2 cl.D1 (NTERA-2) cells. We propose that BG01V cells can be useful as an additional or alternative hESC resource for selected investigations in stem cell biology.
Materials and Methods
In Vitro Propagation of BG01V and NTERA-2 Cells
BG01V hESCs (P10) were obtained from BresaGen, Inc., as a frozen stock. Approximately 1 × 106 BG01V cells were plated into each of two 9.5-cm2 wells of a six-well culture dish (Corning Life Sciences, Acton, MA, http://www.corning.com/lifesciences) containing a feeder layer of Mitomycin C-treated CF-1 mouse embryonic fibroblasts (MEFs) (American Type Culture Collection [ATCC] SCRC-1040.2; Manassas, VA, http://www.atcc.org). Cells were cultured at 37°C, 5% CO2. The growth medium used was Dulbecco's modified Eagle's medium (DMEM)/F12 (ATCC 30-2006) supplemented with ESC-Qualified fetal bovine serum (FBS) (15%) (ATCC SCRC-30-2020), knockout serum replacement (KSR) (5%) (Invitrogen Corporation), l-alanyl-l-glutamine (2.0 mM) (ATCC 30-2115), MEM nonessential amino acids (1X) (ATCC 30-2116), β-mercaptoethanol (0.1 mM) (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), penicillin (100 IU/ml)/streptomycin (100 μg/ml) (ATCC 30-2300), and basic fibroblast growth factor (bFGF) (4 ng/ml) (R&D Systems, Inc., Minneapolis, http://www.rndsystems.com). Daily medium changes began after the first 48 hours in culture. Colony formation was visible within 2 to 3 days. Cells were passaged every 4 to 5 days using collagenase IV (200 Units/ml) (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) for 15 passages (P25).
The hESC line BG01 was obtained from BresaGen, Inc., and cultured as described previously . Briefly, hESCs were maintained on MEF in DMEM/F12 supplemented with 15% FBS, 5% KSR, 2 mM nonessential amino acids, 2 mM l-glutamine, 50 μg/ml Penn-Strep (all from Invitrogen), 0.1 mM β-mercaptoethanol (Specialty Media, Phillipsburg, NJ, http://www.specialtymedia.com), and 4 ng/ml bFGF (Sigma-Aldrich). Cells were passaged by incubation in Cell Dissociation Buffer (Invitrogen), dissociated, and then seeded at approximately 20,000 cells per cm2. Under such culture condition, the ESCs were passaged every 4 to 5 days.
NTERA-2 cells were obtained from the ATCC (ATCC CRL-1973). Cells were thawed and plated into six-well culture dishes (Corning Life Sciences) without a feeder layer. The growth medium used was DMEM (ATCC 30-2002) supplemented as described above. No bFGF was used to suppress NTERA-2 differentiation. Cells were cultured at 37°C, 5% CO2, and the medium was changed every 24 hours. Cells were passaged every 4 days by mechanical scraping.
The karyotype analysis was performed using a standard G-banding technique. Cells cultured in a T75 culture flask were treated with 0.05 μg/ml Colcemid (Invitrogen Corporation 15210-040) for 1 hour, followed by dissociation using 0.25% trypsin/0.53 mM EDTA in Hanks' Balanced Salt Solution (HBSS) without calcium or magnesium (ATCC 30-2101). The cells were then collected by centrifugation (5 minutes at 240 g) and gently resuspended in a 0.06 M KCl hypotonic solution and placed in an incubator at 37°C for 25 minutes. The hypotonic effect was halted by the addition of 3:1 Carnoy's Fixative (methanol/glacial acetic acid). The cells were collected by centrifugation and resuspended by gentle mixing and run through a series of fixes prior to slide preparation. Metaphase spreads were prepared on glass microscope slides, exposed briefly to a 2% Enzar-T trypsin 40X (7000-65; Mediatech, Inc., Herndon, VA, http://www.cellgro.com) in HBSS (20-021-CV; Mediatech, Inc.) solution, and stained using a 2:1 Gurr/Giemsa stain. A total of 43 metaphase spreads were analyzed by microscopy.
Short Tandem Repeat Analysis
Frozen BG01V cells grown on MEF feeders were resuspended in phosphate-buffered saline (PBS) (ATCC SCRR-2201). A 20-μl aliquot was spotted on a labeled FTA card (Whatman, Brentford, Middlesex, U.K., http://www.whatman.com) and allowed to dry. The FTA card lyses the cells on contact and binds the DNA to the paper surface. Prior to polymerase chain reaction (PCR), a portion of the dried spot was removed with a Harris punch, washed three times with Purification Reagent (Whatman), washed once with TE Buffer (Tris-EDTA, pH 8.0), and allowed to dry. Short tandem repeat (STR) analysis was conducted using a multiplex PCR-based PowerPlex 1.2 kit (Promega, Madison, WI, http://www.promega.com). Loci analyzed include D5S818, D13S317, D7S820, D16S539, vWA, TH01, Amelogenin, TP0X, and CSF1P0. Electropherogram data were collected on an ABI 310 Genetic Analyzer. Data were analyzed using Genescan 3.1 and Genotyper 2.0 (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com). The resulting profiles were imported into an in-house database and screened against all other baseline profiles of all samples tested by the ATCC. STR analysis of the NTERA-2 cell line was performed similarly using isolated genomic DNA.
Genomic DNA was isolated from BG01V cells grown on MEFs using the GenElute Mammalian Genomic DNA Mini-prep Kit (Sigma-Aldrich). HLA DNA typing was performed by using hybridization of PCR amplified DNA with sequence specific oligonucleotide probes (Tepnel Lifecodes Corporation, Stamford, CT, http://www.tepnel.com). The target DNA is amplified by PCR and then allowed to denature and rehybridize to complementary DNA probes conjugated to fluorescently coded microspheres. A flow analyzer identifies the fluorescent intensity on each microsphere, and the determined HLA type is based on the reaction pattern compared with patterns associated with public HLA gene sequences. Assays were performed to determine the HLA-A, HLA-B, HLA-C, HLA-DRB, and HLA-DQB loci.
Mitochondrial DNA Analysis
The entire coding sequence of the human mitochondrial genome of BG01V was sequenced using an automated sequencing microarray (Human MitoChip; Affymetrix, Santa Clara, CA, http://www.affymetrix.com) as previously described . Briefly, the MitoChip can sequence 29,366 base pairs (bp) of double stranded DNA, which includes 980 bp of plasmid DNA sequence as a control for chip hybridization. Both strands of the entire mitochondrial coding region (nucleotides 573–16,024; 15,451 bp) are tiled once on the array. The forward and reverse strands of an additional 12,935 bp of the mitochondrial DNA (the coding region minus 12S and 16S RNA sequences) are tiled on the remaining features and thus provide an inbuilt duplication of sequence data for approximately 84% of the mitochondrial coding region. Automated sequence analysis of hybridized microarrays was performed using Affymetrix GeneChip DNA Analysis Software (GDAS) version 3.0, using a modification of a previously described ABACUS (adaptive background genotype-calling scheme) . Once a batch analysis is completed, the GDAS software generates a report containing individual, total number, and percentages of base calls within the batch and a detailed case-by-case list of genotype variations vis-à-vis the reference human mitochondrial DNA sequence (revised Cambridge reference sequence [RCRS]). The MitoChip has previously been demonstrated  to have more than 99.99% reproducibility of base calls in replicate experiments and is thus a highly automated and sensitive microarray tool for human mitochondrial DNA sequencing and mutation detection. To compare the mitochondrial sequence of BG01V to its parent cell line (BG01), we also sequenced the entire coding sequence from the latter, using DNA obtained from an early cell passage.
Methylation-specific PCR (MSP) was used to determine the methylation status of a select sample of specific developmental and imprinted genes in BG01V and NTERA-2 bisulfite-treated genomic DNA. MSP exploits sequence differences existing between methylated and unmethylated alleles after treatment with sodium bisulfite. The frequencies of the CpG dinucleotides within the target sequences help emphasize the sequence differences. Oligonucleotide primers for a given locus are designed to directly discriminate methylated from unmethylated sequences in bisulfite-modified DNA. The results are obtained immediately after PCR amplification and gel electrophoresis without the need for further restriction or sequencing analysis. Selected imprinted gene targets examined included the PWS/AS (snrpn) locus , the H19 promoter region (GenBank AF125183) [16, 17], and the DLK1/MEG3 differentially methylated region (DMR) . Developmentally important genes included the Xist promoter region (GenBank U50908 [gi:1575006]) [19, 20], the Oct-3/4 promoter region (GenBank AJ297527) , and the Notch1 promoter region (GenBank NT_024000 [gi:29793214]) .
Genomic DNA from the BG01V hESCs grown on feeder layers and cultured NTERA-2 cells was isolated and purified using the GenElute Mammalian Genomic DNA Miniprep Kit and protocol (Sigma-Aldrich) as provided by the manufacturer. Bisulfite conversion of the genomic DNA was accomplished using the EZ DNA Methylation Kit (Zymo Research, Orange, CA, http://www.zymoresearch.com) and protocol as provided. If necessary, promoter regions of genes were verified using PromoterInspector software (Genomatix Software GmbH, München, Germany, http://www.genomatix.de). MSP primers were designed using MethPrimer software (http://itsa.ucsf.edu/∼urolab/methprimer/index1.html)  with the exception of those already published for PWS/AS (snrpn)  and DLK1/MEG3  (Table 1). Amplification and gel analysis were performed as described . Bisulfite-converted CpGenome Universal Methylated DNA (Chemicon, Temecula, CA, http://www.chemicon.com) and bisulfite-converted male human placental DNA (Sigma-Aldrich) were used as positive controls. Negative controls consisted of amplification reactions without template.
Telomerase activity in the cultured BG01V cells and NTERA-2 cells was examined using the TRAPeze Telomerase Detection Kit (Chemicon). The BG01V cells were separated from feeder cells either by sorting SSEA-4 expressing BG01V cells using a BD FACS Aria Cell Sorter (BD Biosciences, San Diego, http://www.bdbiosciences.com) or by mechanical scraping of the ESC colonies from the culture dish. The lysate preparation and the telomeric repeat amplification protocol (TRAP) assay were performed according to the manufacturer's instructions, except for the PCR amplification steps. The PCR amplification was performed in three steps, instead of two steps suggested by the vendor: 32 cycles of 94°C for 30 seconds, 50°C for 30 seconds, and 72°C for 90 seconds with one cycle of 72°C for 10 minutes. The TRAP products, along with appropriate size markers, were electrophoresed on a 10% nondenaturing polyacrylamide gel under 50 mM Tris-borate buffer and visualized by SYBR Green (Molecular Probes, Inc., Eugene, OR, http://probes.invitrogen.com) fluorescence using a gel documentation imaging system (UVP, Inc., Upland, CA, http://www.uvp.com).
Immunophenotyping of Undifferentiated Cells
Undifferentiated BG01V ESCs and NTERA-2 EC cells were examined for expression of pluripotent hESC markers using immunocytochemistry . Cells were fixed in 2% paraformaldehyde for at least 20 minutes at room temperature. They were washed in PBS and then incubated in 3% normal goat serum (NGS) to inhibit nonspecific binding. Saponin detergent (0.5%) was used to permeablize cell membranes on samples stained for intracellular markers. Cells were then assayed with monoclonal antibodies specific for Oct-3/4 (1:250) (BD Biosciences Transduction Laboratories, Lexington, KY, http://www.bdbiosciences.com), SSEA-1 (1:100), SSEA-4 (1:50), TRA-1-60 (1:100), and TRA-1-81 (1:100) (all four from Chemicon), washed using PBS/1% NGS to remove any unbound protein, and then incubated with an Alexa Fluor 488 conjugated goat anti-mouse IgG (H+L) antibody (1:750) (Molecular Probes, Inc.). Positively stained cells were visualized using an epifluorescence microscope.
Endogenous alkaline phosphatase activity in BG01V and NTERA-2 cells was detected using the ELF 97 Endogenous Alkaline Phosphatase Detection Kit (Molecular Probes, Inc.) according to the manufacturer's instructions. Cells cultured on 12-mm round glass cover slips in 24-well plates (Corning Life Sciences) were treated with 2% paraformaldehyde for 20 minutes at room temperature. The cells were washed with PBS, treated with 0.2% Tween-20 for 20 minutes at room temperature, and rinsed with PBS. Fixed cells were then incubated with a filtered 1:20 dilution of the phosphatase substrate in situ. The immunoassayed colonies and the reaction were monitored using an epifluorescence microscope. The reaction was terminated using a stop solution consisting of PBS, 25 mM EDTA, and 5 mM levamisole, pH 8.0. Cells were rinsed with PBS before mounting on glass microscope slides.
Gene Expression of Undifferentiated Cells Using Quantitative RT-PCR
Total RNA was isolated by lysing cells in Trizol LS (Invitrogen) according to instructions. Two micrograms of total RNA was treated with DNAseI (Promega) 25°C for 15 minutes, 65°C for 10 minutes, then reverse-transcribed using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, http://www.bio-rad.com). One fortieth of the cDNA synthesis reaction was used as template for each real-time PCR using iQ SYBR Green Supermix (Bio-Rad). For each primer set, Ta is the annealing temperature determined empirically using template cDNA from NTERA-2 cells (Table 2). Tr is the temperature at which the SYBR Green fluorescence is read, chosen by examining the melt curves of the PCR products. PCR was run in an iCycler iQ Real Time Detection System (Bio-Rad) for 50 cycles of 95°C for 15 seconds, Ta 30 seconds, 72°C for 45 seconds, and for Tr 15 seconds. The relative amounts of PCR product were quantified using the relative threshold cycle (ΔΔCt) method corrected for efficiency of each amplification . The gene quantities for each sample were normalized against the geometric mean of expression of the housekeeping genes GAPDH, β-actin, and tata-binding protein (tbp). Statistically significant (p < .05) differences in gene expression were determined by the Student's t test.
Gene Expression of Undifferentiated Cells Using Microarrays
Oligo microarray experiments for undifferentiated BG01V and NTERA-2 cells were conducted with procedures of G4110B from Agilent Technologies, Inc. (Palo Alto, CA, http://www.agilent.com). The chip contains 20,170 human probes. RNA extracted from BG01V and NTERA-2 was labeled with Cy5, and human universal RNA (huURNA) (Clontech, Palo Alto, CA, http://www.clontech.com) was labeled with Cy3. Microarray data were analyzed with the Agilent G2567AA Feature Extraction Software (version 7.5). The huURNA is a high-quality reference standard for microarray experiments which permits multiple laboratories working on different microarray platforms to compare and share data. The output data were used for further computation. UniGene Build no. 176 was used to retrieve UniGene IDs. There are two replicates of microarray data. The z score method  was used to analyze data and identify significantly elevated genes.
For Illumina BeadArray, sample amplification was performed using 100 ng of total RNA as input material by the method of Van Gelder et al.  using the Illumina RNA Amplification kit (Ambion, Inc., Austin, TX, http://www.ambion.com) following the Manufacturer's instructions; labeling was achieved by use of the incorporation of biotin-16-UTP (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com) present at a ratio of 1:1 with unlabeled UTP. Labeled, amplified material (700 ng per array) was hybridized to a pilot version of the Illumina Ref-8 BeadChip according to the manufacturer's instructions (Illumina, Inc., San Diego, http://www.illumina.com). Arrays were washed and then stained with Amersham fluorolink streptavidin-Cy3 (GE Healthcare BioSciences, Little Chalfont, U.K., http://www.gehealthcare.com) in accordance with the BeadChip manual. Arrays were scanned with an Illumina BeadArray Reader confocal scanner according to the manufacturer's instructions. Array data processing and analysis were performed using Illumina BeadStudio software.
Generation and Analysis of Teratomas in SCID Mice
BG01V hESC were maintained and harvested as described. Five 6-week-old male Nod CB17-Prkdc (severe combined immunodeficient [SCID])/J mice (Jackson Laboratory, Bar Harbor, ME, http://www.jax.org) were each injected with a cell suspension of 4 × 106 cells using a 28-gauge half-inch needle in the gastrocnemius muscle of the hind limb. Palpable tumors formed in all mice and were dissected after 8 to 12 weeks. The tumors were fixed in 10% formalin (Fisher Scientific International, Hampton, NH, http://www.fisherscientific.com), dehydrated, and embedded in paraffin. Sections were cut at 4 μM and examined histologically after hematoxylin and eosin staining or immunohistochemical staining. Sections were processed for immunohistochemical analysis using antibodies to pancytokeratin (PC, catalog no. MU181-UC; BioGenex, San Ramon, CA, http://www.biogenex.com), neuron-specific enolase (NSE, catalog no. MUO55-UC; Invitrogen), smooth muscle actin (SMA, catalog no. MO851; DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com), p63 (catalog no. 8431; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, http://www.scbt.com), Oct-3/4 (catalog no. 5279; Santa Cruz Biotechnology, Inc.), and S-100 (catalog no. Z0311; DakoCytomation). All antibodies were raised in mice against human targets and used at 1:100, except for antibodies against S-100, which were raised in rabbit and used at 1:1000. Immunohistochemistry protocols were as follows: for p63 and Oct-3/4, sections were microwaved for 15 minutes at 400 W in 10 mM sodium citrate buffer (pH 6.0) and stained using Universal DAKO LSAB + Kit, Peroxidase (K0679, DakoCytomation) according to the manufacturer's instructions. Staining for PC, S-100, SMA, and NSE was performed using a DakoCytomation Autostainer Plus according to manufacturer's instructions, with no antigen retrieval (SMA, S-100), pre-treatment with 1% proteinase K for 5 minutes (PC), or with steam for 15 minutes (NSE). All slides were counter-stained with Harris' hematoxylin. Negative controls were incubated with mouse immunoglobulin G reagent grade (Lyophilized; Sigma-Aldrich) in a 1:100 dilution.
Gene Expression of Teratomas Using Real-Time RT-PCR
Total RNA was prepared from a harvested tumor using the Trizol reagent (Invitrogen) and was treated with DNase I (Promega) prior to first-strand cDNA synthesis using 2 μg RNA and the iScript system (Bio-Rad). Assays-On-Demand (Applied BioSystems), which are prequalified TaqMan quantitative PCR sets, were used in Micro-Fluidic Cards in the ABI PRISM 7900HT Sequence Detection System (Applied BioSystems). Duplicate reaction sets, each containing cDNA generated from 0.25 μg RNA, were used to compare the relative gene expression in undifferentiated BG01V and BG01V teratoma populations. The ΔCt values were averaged and normalized with GAPDH using the ΔΔCt method. Fold changes were calculated as 2-Ct, and the expression ratio of the teratoma sample relative to the BG01V sample indicated.
Sterility and Pathogen Testing
Extensive bacterial and fungal tests were performed on the BG01V and NTERA-2 cell cultures that include incubations in the following media: HTYE, TSB aerobic, Sabouraud, DMEM with 10% FBS, Blood Agar aerobic, Blood Agar anaerobic, and THIO anaerobic. The cultures were routinely monitored and reported at 48-hour, 14-day, and 21-day post incubations. In addition, mycoplasma testing was performed using a Hoechst Assay for each culture.
Pathogen testing was performed by a Clinical Laboratory Improvement Amendments-approved laboratory (Genetics and IVF Institute, Fairfax, VA, http://www.givf.com). All pathogen assays use “nested” multiplex PCR or reverse transcription (RT)-PCR formats and have a detection threshold of 10 copies of target DNA or RNA per assay (with an analytical sensitivity of 100%), with a limit of detection (LOD) of two to five copies per assay. Nucleic acids are extracted using the Qiagen QIAmp Viral RNA Kit that yields both intact RNA and DNA. One fifth of each extract is used per 50-μl “nested” multiplex PCR (Platinum Taq DNA Polymerase; Invitrogen). Amplification products are visualized by high-resolution PAGE and Sybr Green staining. For RNA targets, 10-μl of extract is used for One Step RT-PCR (Invitrogen) followed by second-round PCR prior to the PAGE analysis. For all DNA PCR assays, a human genomic sequence is amplified as an internal positive control. For RNA viruses (such as hepatitis C virus [HCV]), an endogenous human mRNA is targeted in all assays. These sample-derived targets must be scored as “detected,” indicating proper nucleic acid extraction and amplification, prior to scoring any pathogenic target.
Karyotype and Propagation In Vitro
Although early work on hESC lines supported the conclusion that these cells are karyotypically normal and stable [4, 5], there are cases in which the culture conditions employed for their propagation apparently do not comprehensively maintain chromosomal stability [3, 6]. Parental BG01 hESCs are male (XY) and have a normal complement of 46 chromosomes (data not shown). The BG01V hESC line evolved from routine enzymatic passaging by BresaGen, Inc., and exhibits a karyotype of 49 chromosomes (XXY, +12, +17). Through 25 passages, BG01V maintained the karyotype 49XXY, +12, +17 (data not shown). In comparison, the NTERA-2 cell line is hypotriploid and contains an average of 12 marker chromosomes (Fig. 1A). Despite the abnormal karyotype, BG01V cell colonies grown on MEFs exhibited uniform flat colonies. These colonies are tightly compacted, and the cells have a high nuclear/cytoplasmic ratio, as do undifferentiated cells of the parental line BG01 . The BG01V cells exhibit prominent nucleoli, and cell edges are not phase-bright.
The cells had a predictable growth rate and were easy to maintain in culture (data not shown). After thawing and plating, the BG01V cells recovered rapidly and colony formation was visible after 2 days. Enzymatic passaging was possible every 4–5 days, enabling efficient expansion of the cell line. In our hands, the normal parental hESC line, BG01, was much more difficult to successfully maintain in culture. BG01 had a slightly slower growth rate, and colony formation was not evident until 3–4 days after passaging manually every 5 to 7 days (data not shown). BresaGen, Inc., the originator of this cell line, recommends manual passage to avoid differentiation and mutation. Manual passaging by scraping and aspirating individual colonies under a microscope greatly increases “hands on” time for in vitro maintenance of this cell line, is very labor intensive, and reduces expansion capacity. Like BG01V, NTERA-2 was also easily maintained in vitro. However, this cell line did not exhibit uniform colony formation (Fig. 1B).
To ensure that work was actually performed on the appropriate cell lines, identification assays were performed and compared with previously available data. These assays included STR analysis and HLA typing. The STR profiles of the BG01V and NTERA-2 were determined as described in Materials and Methods (Fig. 2). Loci analyzed for tetra nucleotide STR analysis included D5S818, D13S317, D7S820, D16S539, vWA, TH01, TP0X, CSF1P0, and Amelogenin for gender determination. The amplicons were separated by capillary electrophoresis and analyzed using Genescan 3.1 and Genotyper 2.0 software from Applied BioSystems. Each peak in the resulting electropherogram represented an allele that is alphanumerically scored and was then entered into a database. As expected, the number of repeats and the location and peak height in the BG01V STR profile was identical to the STR profile to BG01. The NTERA-2 STR profile was identical to original NTERA-2 material (CRL-1973) deposited with ATCC.
The HLA profiles of BG01V and NTERA-2 were generated as described (Table 3). For purposes of comparison and of interest to the user community, we also performed HLA typing on the parental hESC line, BG01. HLA typing at the allelic level included HLA-A, HLA-B, HLA-C, HLA-DRB, and HLA-DQB loci. As can be seen from Table 3, BG01V had an identical HLA profile to BG01. The NTERA-2 HLA profile was interesting in its homozygosity at each locus.
Mitochondrial DNA Sequencing
Mitochondrial DNA anomalies apparently contribute to, and are correlated to, various human diseases [13, 14]. Mitochondrial DNA sequence analysis was performed on both BG01V and BG01. MitoChip sequencing yielded an approximately 96% base call rate in both samples (28,244 bp/29,366 bp for BG01V and 28,134 bp/29,366 bp for BG01); the remaining calls were deemed by GDAS version 3.0 (Affymetrix) to be of unsatisfactory quality. There were no homoplasmic sequence variations between the two cell lines (i.e., the mitochondrial coding sequence was identical across nearly 28 kb of DNA sequence between the two cell lines). These findings suggest an extremely high degree of fidelity being retained by BG01V relative to its parental cell line (BG01), at least at the mitochondrial DNA sequence level. However, it should be noted that both cell lines demonstrated identical sequence variations compared with the RCRS, consistent with the presence of coding sequence polymorphisms in the mitochondrial DNA; coding sequence polymorphisms for BG01V and BG01 were tabulated (Table 4).
MSP was employed to monitor epigenetic stability and analyze the methylation status of six specific genes for both BG01V and NTERA-2. The results are shown in Figure 3. The maternal, methylated allele and paternal, unmethylated allele of the PWS/AS (snrpn) gene locus were detected, indicating preservation of the normal imprinting pattern in BG01V cells. In contrast, only the paternal, unmethylated allele of the PWS/AS (snrpn) gene was detected in the NTERA-2 sample (Fig. 3A). MSP analysis of DLK1/MEG3 DMR and the H19 locus demonstrated the expected maternal, unmethylated allele product and the paternal, methylated allele product, indicating a normal imprint pattern in these regions in both cell lines (Fig. 3B, 3C). Both unmethylated and methylated alleles of the Xist gene promoter region were detected in BG01V and NTERA-2 cells (Fig. 3D). Only the unmethylated homologue of the Notch1 promoter region was generated in the analysis of the BG01V and NTERA-2 DNAs (Fig. 3E). Both unmethylated and methylated alleles of the Oct-3/4 promoter region were detected in BG01V, whereas only the unmethylated allele of the Oct-3/4 promoter region was detected in the NTERA-2 sample (Fig. 3F).
Germline and stem cells are known to contain enzymatic activity associated with the maintenance of chromosomal caps that prevent terminal fusion [28, 29]. Assay of telomerase activity was employed to validate the presence of hTERT activity in BG01V and NTERA-2 cells. BG01V and NTERA-2 cells tested positive for telomerase activity (Fig. 4). Heat-inactivated, RNase-treated samples of BG01V, NTERA-2, and IMR-90 were included as negative controls. The telomerase-positive control is shown in lane 8.
Immunocytochemistry and alkaline phosphatase activity assays were used to assess the presence of stem cell markers characteristic of the undifferentiated state. BG01V and NTERA-2 cells stained positively for Oct-3/4, SSEA-4, TRA-1-60, and TRA-1-81. As originally identified in EC cells, the stage-specific embryonic antigen, SSEA-4, and the keratin sulfate-associated epitopes, TRA-1-60 and TRA-1-81, were found to be expressed in ESCs later on [30, 31]. Both cell lines also demonstrated endogenous alkaline phosphatase activity as evidenced by fluorescence microscopy (Fig. 5). Phase-contrast and fluorescent images of the cells showed the expected morphology and stem cell markers previously observed on hESCs [6, 31]. BG01V cells formed well-defined, compact colonies, whereas NTERA-2 cells formed a diffuse monolayer. Although a quantitative analysis of the immunophenotyping in situ was not performed, the majority of BG01V and NTERA-2 colonies stained positive for all the markers tested. Within the positive colonies, a large majority of the individual cells stained positive. The parent line BG01 has been described as strongly and uniformly positive for the markers SSEA-4, TRA-1-60, TRA-1-81, and Oct-3/4 , and the same phenotype was observed here for BG01V. The results from quantitative immunophenotyping using flow cytometry were in agreement with the immunocytochemical staining (data not shown).
Several hundred genes were found to be significantly expressed (greater than threefold higher than huURNA) in BG01V or NTERA-2 as assayed by gene expression analysis with microarrays. The list of significant genes is available as supplemental online Tables 1 through 3. Genes including POU5F1 (Oct-3/4), DNMT3B, TDGF1 (Cripto), DPPA4 (FLJ10713), REX-1 (NM_174900.1), and NANOG (FLJ12581) were significantly expressed and in common to both cell lines. Gene ABCG2, an ATP-binding cassette subfamily G member 2, was significantly expressed in the BG01V hESC line, and a transcription factor gene SOX2 was significantly expressed in the NTERA-2 cell line only. In addition, germ line-specific genes, including SCP3, VASA, IFITM1 and IFTM2, were not present at significant levels in either cell line.
Of 264 genes highly expressed in BG01V, 22 (8.3%) were also highly expressed in both BG02 and the parental line BG01  (data not shown). For NTERA-2, 32 (9.6%) of 334 highly expressed genes were shared with BG01 and BG02. Both of these shared sets include the known hESC markers TDGF1, POU5F1, and NANOG.
Confirmation of gene expression for select targets in undifferentiated BG01V, the undifferentiated parental line BG01, and NTERA-2 was confirmed by quantitative RT-PCR (qRT-PCR) (Fig. 6). Expression of 10 ESC-associated genes was measured in real-time PCR and normalized to three housekeeping genes. BG01V shows significant differences from NTERA-2 for the genes REX1, DPPA4, DNMT3B, TDGF1, and NANOG. Interestingly, the parental line BG01 shows significant differences with BG01V in expression of REX1, SOX2, DPPA4, TDGF1, and UTF1. Unlike the results of microarray analysis, the levels of ABCG2 and SOX2 were indistinguishable between BG01V and NTERA-2. However, the genes REX1 and TDGF1 were more highly expressed in BG01V than NTERA-2 in both qRT-PCR and microarray experiments (supplemental online data). In addition, both experiments show that NTERA-2 over express DPPA4, DNMT3B, and NANOG relative to BG01V.
Comparison of BG01V with EC and Other ESC Lines
The data overall showed that BG01V shared many similarities with BG01, and we have previously shown that hESC lines are overall similar to each other . However, to directly test the overall degree of similarity of BG01V and NTERA2 to BG01 and other hESC lines, we examined global gene expression using an Illumina BeadArray, comparing each line to three pooled undifferentiated hESC lines (H1, H7 and H9). The correlation coefficients of the samples are summarized in Figure 7. The overall gene expression of BG01 is highly similar to other undifferentiated hESC lines (r2 = .9559 for genes detected at confidence >0.99). BG01V is somewhat more divergent (r2 = .9206), and NTERA2 has an even greater difference from normal hESC lines (average r2 = .8806). Genes detected at high confidence are considered differentially expressed if the level is 2.5-fold higher or lower than in the hESC pool. BG01V shows differential expression of 247 genes, very similar to the count of 251 differential genes in BG01, whereas NTERA-2 has 904 differentially expressed genes. Detailed analysis of gene expression and statistical validity will be reported elsewhere.
In the set of genes on the BeadArray, 85 of the 92 candidate “stemness” genes shared among multiple hESC lines could be identified . Of these 85 stemness genes, 79 were expressed at greater than 99% confidence in BG01V and only four (5.1%) were differentially expressed by greater than 2.5-fold compared with pooled hESC lines. NTERA-2 expressed 84 stemness genes at greater than 99% confidence, and only five (5.9%) were differentially expressed compared with the hESC pool.
Differentiation of BG01V Cells to Ectoderm, Endoderm, and Mesoderm in Complex Teratomas
To examine the differentiation capacity of BG01V cells in vivo, undifferentiated cells were injected intramuscularly to the hind limb of SCID mice. Tumors formed in all of the five mice that were injected; these tumors were dissected for analysis after 8–12 weeks. Histological, immunocytochemical, and gene expression analyses demonstrated that the teratomas contained differentiated lineages representing ectoderm, endoderm, and mesoderm (Fig. 8). The teratomas were generally well demarcated from the surrounding muscle and exhibited organized clusters of cells and primitive tissue structures (Fig. 8A), including cartilage, mesenchyme, mineralized bone, villi, smooth muscle, putative nerve bundles, liver-like hepatoid structures, ducts, cystic epithelial-lined spaces and various types of epithelia including simple cuboidal, columnar, pseudostratified with cilia and/or goblet cells, simple or stratified with single or multiple layers of vacuolated cells, and epithelial invaginations. Other cell types that were observed included isolated keratinizing cells and pigmented melanocytes (not shown). Immunostaining analysis was used to confirm the identification of particular lineages. A few small pockets of OCT-3/4+ cells (Fig. 8B) indicated the persistence of undifferentiated cells 8–12 weeks after injection. Consistent with this, hESC-like cells that were OCT-3/4+, SSEA-4+, TRA-1-81+, and SSEA-1− could be isolated and expanded from a 12-week-old teratoma (not shown). PC was expressed by endodermal aggregates (Fig. 8C, 8D) and epithelia (Fig. 8E, 8H). The basal and superbasal cells of vacuolated presumptive ectodermal epithelia also expressed p63 (Fig. 8E, 8F), distinguishing it from p63 endodermal-derived epithelia (Fig. 8G, 8H). Chondrocytes and clusters of spindloid cells immediately adjacent to cartilage expressed S100 (Fig. 8I, 8J), and smooth muscle was identified by the expression of SMA (Fig. 8K, 8L). Aggregates of neuronal lineages expressed NSE and S100 (Fig. 8M, 8N). Individual S100+ neuronal cell bodies are indicated (arrowheads).
Quantitative real-time RT-PCR was used to compare gene expression in undifferentiated BG01V cells and teratomas (Fig. 8O). A teratoma sample exhibited reduced expression of markers of pluripotency, including CD9, FOXD3, GDF3, LEFTY2, OCT-3/4, and SOX2, and elevated expression of markers of ectodermal, endodermal, mesodermal, mesendodermal, and trophectodermal lineages, including TUBB3, NEF, SOX17, GATA4, BRACHYURY, GSC, MYF5, and EOMESODERMIN. Furthermore, expression of SOX1 (ectoderm), HNF4-α (endoderm), and MYOD (mesoderm) was detected only in the teratoma sample, with crossing points at cycle 13, 6.9, and 9.3, respectively. Therefore, these analyses demonstrated the differentiation of BG01V to multiple representatives of each of the three primary germ layers in vivo.
Pathogen testing was performed by a CLIA-approved laboratory (Genetics and IVF Institute), using “nested” multiplex PCR or RT-PCR formats and had a detection threshold of 10 copies of target DNA or RNA per assay (with an analytical sensitivity of 100%), with an LOD of two to five copies per assay. Nucleic acids were extracted from BG01V, and samples were processed for RNA and DNA viral detection as described in Materials and Methods. For all DNA PCR assays, a human genomic sequence is amplified as an internal positive control. For RNA viruses (such as HCV), an endogenous human mRNA is targeted in all assays. These sample-derived targets had to be scored as “detected,” indicating proper nucleic acid extraction and amplification, prior to scoring any pathogenic target. The list of viruses tested is summarized in supplemental online Table 4. No viral contamination was detected in the BG01 or BG01V samples that have been maintained at ATCC. Thus, detailed pathogen testing as required for any master bank preparation or large-scale use of cells can be performed from a small amount of sample using commercially available reagents or validated clinical laboratories.
Although it was shown after establishment that the BG01 hESC line is karyotypically normal [4, 5], its continued propagation in vitro has given rise to a reported stable variant, BG01V [6, 7], which has been characterized and compared with hESC lines as described herein. We also analyzed various features of the human EC cell line NTERA-2 for comparative reasons. Although BG01V and NTERA-2 have disparate origins, and both are cytogenetically abnormal, they display many common cellular and molecular traits previously found in pluripotent hESCs. Our goal was to present an initial, in-depth characterization of the BG01V cell line and propose it as a viable adjunct in stem cell biology research, as well as to propose a standard scheme for characterizing and analyzing hESC lines, thereby typing them for large-scale production.
Frequent morphological analysis, along with routine karyotypic analysis, of cell lines is critical to assess the genetic and epigenetic stability of hESCs during extended periods in culture. Although both lines examined possess chromosomal aberrations, it is reasonable that karyotypic analysis of the variant hESC line should prove less demanding than routine cytogenetic analysis of the teratocarcinoma cell line because it contains far fewer abnormalities to monitor. In addition, the morphology of BG01V more closely resembles that of normal hESC lines [4, 5], whereas the EC morphology is less defined. Furthermore, as evidenced through 25 passages, BG01V cells maintain karyotypic stability with the use of enzymatic passaging techniques rather than mechanical passaging. This also facilitates efficient expansion of the cells to generate adequate quantities of material for research purposes and comparative studies.
The BG01V and NTERA-2 cells both possess trisomy 12. Duplication of the p arm of chromosome 12 has been reported in male germ cell tumors, including ECs . This region of chromosome 12 is known to contain genes such as STELLAR, NANOG, and GDF3 that influence proliferation in pluripotent cells . In addition, BG01V cells also carry trisomy 17. Anomalies involving chromosome 17 have been reported in other hESC sublines, including H7 . STAT3 and GRB2 are located on chromosome 17; interestingly, the homologues of these genes help regulate self-renewal in mouse ESCs. Collectively, such gains in gene copy number could afford a selective advantage for propagation in vitro.
As part of the continuing efforts of the Stem Cell Center at ATCC and the general international stem cell community to fully characterize and authenticate hESC lines, we are developing a comprehensive database of DNA profiles based on STR loci. The exploitation of tandemly repeated elements in the genome has become important in several fields, including genetic mapping, linkage analysis, and human identity testing. STR loci are among the most informative polymorphic markers in the genome. The profiling process involves simultaneously amplifying eight STR loci and the amelogenin gene for sex determination in a multiplex PCR reaction (Promega PowerPlex 1.2 system). This test allows for discrimination of at least 1 in 108 individuals . STR analysis can be useful in confirming and clarifying some of the anomalies identified through cytogenetic analysis. For example, the triple-repeat element for locus D16S539 correlates with a triosomy 16 or a homozygous peak for locus D13S317 may indicate a monosomy 13 as seen in the NTERA-2 cell line. The data from this analysis were entered into our database. Such data will help ensure the quality and consistency of large-scale cell culture expansions and can be periodically queried as they are maintained. This information may become the standard for identifying human embryonic cell lines. STR analysis is vital for the correct identification and verification of cells in culture. Furthermore, STR data will ensure the quality and integrity of work performed with these cells.
HLAs are a family of cell proteins found on the surface of white blood cells and other nucleated cells in the body. These proteins vary from one person to another and are critical for the activation of immune responses. HLA typing is used for tissue analysis before organ and/or cell transplantation. Performing HLA matching minimizes the possibility of rejection because transplantation of tissues and organs between genetically unrelated people usually results in rejection of the donor graft, tissue, or cell by the recipient. HLA typing is increasingly performed using DNA techniques. Such typing analyses will be critical as stem cell-based therapies evolve. In addition, the need to collect, compare, and use the typing data over long periods of time will permit stored sequence polymorphism data to be reinterpreted in the future as new alleles are discovered and newly derived hESC lines are authenticated. The Stem Cell Center at ATCC is establishing such a database to store HLA typing information. Interestingly, our assessment of the BG01V cell line revealed a heterozygous HLA genotype identical to its parental BG01 originator, whereas the profile for the NTERA-2 cell line was homozygous for all HLA alleles examined. The HLA class I (A, B, and C) and class II (DR, DQ, and DP) genes are located on the short arm of chromosome 6. Although there were two copies of chromosome 6 identified in the NTERA-2 karyotype, it would be intriguing to determine whether they were fully identical chromosomes attributable to a partial meiotic event occurring earlier in its lineage, a loss of heterozygosity, or other complex chromosomal rearrangements originating in the germinal tumor.
The mitochondrion is the only organelle outside of the nucleus to harbor its own DNA. The role of mitochondrial DNA anomalies in a variety of human conditions, including old age and cancer, has brought to the fore the importance of a sensitive and accurate, yet high-throughput, platform for sequencing the 16.5 kb of mitochondrial DNA. The Human MitoChip was recently developed to fulfill this need, and extensive characterization of this microarray has been described previously . In the current study, microarray-based sequencing of the entire coding region of both BG01V as well as its parent hESC cell line BG01 (the NTERA-2 cell line was not sequenced) using the MitoChip led to some interesting findings. First, in nearly 28,000 bp of DNA, we did not detect any homoplasmic sequence variations between the two cell lines, suggesting an extremely high degree of mitochondrial sequence fidelity in hESC cell lines despite the presence of gross karyotypic abnormalities in the variant. This is particularly striking because the mitochondrion, as the workhorse of the cell, is exposed to a large amount of reactive free radicals and is known to have inefficient DNA repair mechanisms . The mechanism(s) underlying preservation of DNA sequence identity between BG01 and BG01V are unclear but suggest that DNA repair mechanisms and/or response to DNA damage may be different in hESCs and somatic cells. Second, we found multiple common sequence variations in the two cell lines (Table 4) compared with the published RCRS human mitochondrial sequence (obtained from MitoMap, www.mitomap.org) . This was less surprising because there are multiple known mitochondrial DNA haplotypes, and the RCRS is only one common representative haplotype. Interestingly, in a recent study, six heteroplasmic sequence alterations occurring in two of nine (22%) later passage hESC lines were identified and confirmed by conventional dideoxy sequencing. Five of these alterations occurred in the coding region, three resulting in missense mutations in ND1, ND2, and ND4, and one causing a nonsense mutation in AT-Pase6 . Thus, although there are changes in the mitochondrial genome, these are at a lower frequency than what has been reported in human cancer cell lines (80–100%) . Nevertheless, the detailed cataloging of these sequence variations in BG01 and BG01V should serve as a precedent for “mitochondrial fingerprinting” of additional hESC lines, which (like STR polymorphisms) can serve as an identity test for researchers who deal with multiple cell lines in their laboratories.
DNA methylation plays a key role in tissue- and stage-specific gene regulation, genomic imprinting, and X-chromosome inactivation. It has been demonstrated to be essential for normal mammalian development . Although much effort has focused on the mechanism of CpG hypomethylation and hypermethylation of gene-specific promoter regions in tumorigenesis , little is known about its role and patterns during the differentiation of pluripotent stem cells. Moreover, less is understood concerning the potential epigenetic changes that may result throughout extensive in vitro cell culture. Thus, we proposed assaying for methylation status via MSP to examine epigenetic stability during cell culture and to indirectly determine the expression status of certain temporally regulated developmental genes important in maintaining the pluripotent state and initiating the differentiation process.
Three imprinted genetic regions, PWS/AS (Prader-Willi/Angelman) (snrpn), H19, and DLK1/MEG3, were examined for methylation status using MSP. Differential DNA methylation exists at several sites in the PWS/AS (snrpn) critical region such that the maternal allele is methylated and the paternal allele is unmethylated and transcriptionally active. In contrast, the differentially methylated regions of the H19 promoter region and DLK1/MEG3 are heavily methylated on the paternal homologue and unmethylated on the maternal homologue. MSP analysis of these regions demonstrated this to be the case in the BG01V hESCs, thereby implying epigenetic stability of the cells in culture. Although both the unmethylated and methylated alleles of the H19 and DLK1/MEG3 locus were detected in the NTERA-2 cells, only the paternal, unmethylated allele of the snrpn promoter region was present. Absence of the methylated and inactive maternal allele in this region indicates a maternal deletion of the locus through microdeletion or a paternal uniparental disomy of chromosome 15. In either case, it is evident that the NTERA-2 cell line demonstrated an imprinting anomaly at this locus.
Both the unmethylated and methylated products for the Xist promoter region were also detected in the analysis. The Xist gene is expressed exclusively from the inactive X chromosome and is required for X-chromosome inactivation to occur in early development. Interestingly, the inactivation process is random and can occur on either the paternal or the maternal X chromosome in females. In certain males who have inherited an extra X chromosome (i.e., XXY males), one of the X chromosomes is inactivated as in normal XX females. Given the XXY karyotype of BG01V, it is plausible that unmethylated and methylated MSP products are generated from one of each of the X chromosomes. Furthermore, the presence of both MSP homologues in the hESCs may be reflective of the cell state and their origin from the blastocyst. The time during development when mammals, including humans, address dosage compensation by inactivating one of the X chromosomes begins to initiate just after the blastocyst stage. The presence of the additional X chromosome in BG01V could also promote this hESC line as an in vitro model to study random X inactivation. Occurrence of both methylation products in the teratocarcinoma cell line NTERA-2 may be attributed to its abnormal cytogenetic features, including evidence of a translocation involving the X chromosome, t(Xq1q), and its molecular-level instability in general.
In a pure population of undifferentiated, pluripotent ESCs, the Oct-3/4 gene is highly expressed. This would suggest that the promoter is unmethylated. Whereas our MSP analysis detected only the unmethylated promoter region of the Oct-3/4 promoter in the NTERA-2 cells, both the unmethylated promoter and slight evidence of the methylated Oct-3/4 promoter were visible in the BG01V cells. In a culture of ESCs which contains a small percentage of cells differentiating, or beginning to differentiate, we can argue that both hyper- and hypomethylated regions of promoter may exist transiently as the cells begin to downregulate expression in committed cell types. Interestingly, the robustness of the amplicons in this assay may reflect the overall status and level of undifferentiation in the cell population and afford a means of quantitative analysis. The analysis of the Notch1 promoter generated only the unmethylated allele in both cell lines, indicating a fully unmethylated and transcriptionally active gene, reflective of its crucial role in determining cell fate.
It is well known that the cells of the human germline, including ESCs, highly express the catalytic activity of human Telomere Reverse Transcriptase, or hTERT [24, 42]. Accordingly, telomerase activity was detected in BG01V and NTERA-2 cells with similar activity levels based on a qualitative assessment of the data. Likewise, ESCs are known to exhibit elevated levels of alkaline phosphatase activity. This was evident in both cell lines by the intense staining observed using the Enzyme-Labeled Fluorescence (ELF 97) substrate system.
Undifferentiated BG01V cells were immunoreactive for Oct-3/4, SSEA-4, TRA-1-60, and TRA-1-81. These pluripotency markers were similarly expressed in BG01 and NTERA-2 cells by immunocytochemistry. Our assessment of gene expression analysis of the undifferentiated cells using microarray and qRT-PCR revealed that the three cell lines appear similar to each other with respect to the expression of a select number of genes. Using the Agilent array, 28 genes were significantly expressed above background and in common between BG01V and NTERA-2. These included upregulated markers representative of the undifferentiated state, such as POU5F1 (Oct-3/4), Cripto/TDGF1, and NANOG (FLJ12581). Expression data for BG01V were also compared with the previously published expression profiles of the parental line, BG01 and the line BG02 ; 22 genes were expressed in common between the BG01V, BG01, and BG02 cell lines, including upregulated expression of POU5F1, NANOG, and TDGF1 (data not shown). The intersection of BG01V and NTERA-2 was also notable for elevated levels of gene expression for DNA methyltransferase DNMT3β, developmental pluripotency-associated (DPPA4) (FLJ10713), and a zinc finger protein gene REX-1 (NM_174900.1) and no significant expression of markers indicating differentiation.
NTERA-2 shows less similarity to BG01V and BG01 when global gene expression is compared with a pool of undifferentiated hESC lines by bead array. The correlation coefficient for all reliably discriminated genes is quite high when BG01 is compared with the pooled hESC, significantly lower for NTERA-2, and intermediate for BG01V. In addition, the spread of the data is similar for BG01V and BG01, which both have about 250 genes differentially expressed compared with the pool. The number of differentially expressed genes in NTERA-2 cells is significantly higher at 904. Thus, gene expression in normal hESCs is modeled better by BG01V than by the EC line.
qRT-PCR analysis was used to confirm and further analyze the expression of a subset of genes, including POU5F1, DNMT3β, ABCG2, TDGF1, and REX-1. These genes were upregulated in all cell lines examined, although not always with the same fold changes as observed using microarray. As many significant differences were detected between BG01V and BG01 as between BG01V and NTERA-2. A more in-depth analysis of the expression profiles of these cell lines using MPSS (Massively Parallel Signature Sequencing) technology (Lynx Therapeutics, Inc., Hayward, CA, http://www.lynxgen.com) is in progress.
Clusters of BG01V hESC colonies xenotransplanted to SCID mice formed teratomas consisting of cell types derived from ectoderm, endoderm, and mesoderm. Antibodies specific for human PC, NSE, SMA, p63, Oct-3/4, and S-100 were used to discriminate the human from mouse cells. Highly differentiated cells and tissues derived from all three germ layers, including identified primitive muscle, cartilage and mesenchyme (mesoderm), ciliated goblet cells (gut endoderm), and neural epithelia (ectoderm) formed from a majority of teratomas. A few residual nests of pluripotent Oct-3/4 cells remained several weeks after injection. Quantitative real-time PCR demonstrated downregulation of multiple hESC markers in the tumor (CD9, Oct-3/4, SOX2, LEFTY2). Upregulation of NEF was detected in the teratoma, indicating ectoderm differentiation. Multiple endoderm and mesodermal markers were upregulated in the teratomas (i.e., Sox17, brachyury, and GSC).
To establish a baseline of characteristics of BG01V hESCs, we examined karyotype, morphology, STR and HLA profiles, mitochondrial DNA sequence, telomerase activity, methylation status, immunocytochemistry, gene expression analysis, and teratoma formation. These characteristics have been retained through 25 passages. In all measures of cell line identity and undifferentiated quality, except karyotype, BG01V is identical to its parent line BG01. Research models of early embryogenesis employing variant cell lines such as the BG01V hESC in addition to, or in lieu of, human EC cells may prove useful and valuable. Yet despite their independent advantages and disadvantages, it is likely that both systems will coexist to benefit experimental stem cell research.