SEARCH

SEARCH BY CITATION

Keywords:

  • Induced pluripotent stem cells;
  • Pluripotency;
  • Pluripotent stem cells;
  • Embryonic stem cells;
  • Fluorescence activated cell sorting;
  • Reprogramming;
  • Tumor cell purging

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Human induced pluripotent stem cells (hiPSC) have the potential to generate healthy cells and tissues for the study and medical treatment of a large number of diseases. The utility of putative hiPSC-based therapies is constrained by a lack of robust quality-control assays that address the stability of the cells or their capacity to form teratomas after differentiation. Here we report that virally derived hiPSC, but not human embryonic stem cells (hESC) or hiPSC derived using episomal nonintegrating vectors, exhibit a propensity to revert to a pluripotent phenotype following differentiation. This instability was revealed using our published method to identify pluripotent cells undergoing very early-stage differentiation in standard hESC cultures, by fluorescence activated cell sorting (FACS) based on expression of the cell surface markers TG30 (CD9) and GCTM-2. Differentiated cells cultured post-FACS fractionation from virally derived hiPSC lines reacquired immunoreactivity to TG30 (CD9) and GCTM-2, formed stem cell-like colonies, and re-expressed canonical pluripotency markers. Furthermore, differentiated cells from pluripotency-reverting hiPSC lines generated teratomas in immunocompromised mice, raising concerns about their safety in downstream applications. In contrast, differentiated cell populations from hESC and episomally derived hiPSC did not show any of these abnormalities. Our assays may be used to identify “unsafe” hiPSC cell lines and this information should be considered when selecting hiPSC lines for clinical use and indicate that experiments using these “unsafe” hiPSC lines should be interpreted carefully. STEM Cells 2013;31:1498–1510


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Induced pluripotent stem cell (iPSC)-based technologies have been widely adopted for the study of developmental and disease processes. However, the field lacks robust and rapid phenotypic screens to assess the safety and stability of iPSC lines or their differentiated progeny. This deficiency is a critical barrier to the clinical uptake of human iPSC (hiPSC)-derived cell types and potentially limits the interpretation of experimental studies carried out with hiPSC lines. There is an urgent need for reproducible assays that enable simple phenotypic examination of hiPSC lines. These assays should detect meaningful differences between pluripotent cell lines and provide pertinent information regarding their suitability for applications.

The discovery of mouse iPSC (miPSC) and hiPSC [1–3] has revolutionized the field of regenerative medicine. Initial characterization and comparisons between human embryonic stem cell (hESC) and hiPSC revealed strong similarities [2, 4-6] and suggested that the reprogramming process faithfully recapitulated the pluripotent state found in hESC. Despite the findings that global gene expression patterns are remarkably similar between hESC and hiPSC, the latter retain a residual transcriptional memory of the somatic cells that they were derived from [7, 8]. It appears that epigenetic mechanisms that control gene expression, such as DNA methylation, are responsible for the persistent expression of parental somatic cell genes in hiPSC [7, 8], a phenomenon also found in miPSC [9, 10]. It is remarkable that some of the differentially methylated loci in hiPSC lines studied constitute recurrent targets of epigenetic remodeling in specific genomic regions [11, 12], implying some selection process during reprogramming.

In addition to epigenetic modifications, some subtle genetic abnormalities revealed by copy number variations [13, 14] or point mutations in protein coding regions [15] have suggested that the reprogramming process might be inherently mutagenic, raising doubts regarding safety for clinical applications relying on hiPSC-derived cell types. However, a recent report demonstrates that a high proportion of abnormalities in hiPSC lines may actually reflect pre-existing genetic mosaicism in the donor cells, rather than the acquisition of mutations through the reprogramming process [16], and a further report indicates that at least for murine cells, the current reprogramming methods can indeed produce germ line competent pluripotent miPSC lines that lack identifiable genomic alterations [17].

It is clear that differences between hESC and hiPSC, collectively termed human pluripotent stem cells (hPSC), can be identified using epigenetic and genetic high-throughput technologies. An additional concern is that tumorigenesis or teratoma formation remains systematically untested for hPSC-based transplantation therapies. This is a major barrier for applied hESC and hiPSC research. It is unlikely that any hPSC differentiation protocol will ever be 100% efficient, consequently yielding a mixture of differentiated target and off-target cell types. Any residual undifferentiated cells present a significant risk of teratoma formation after transplantation, and there is an urgent need for methods that can identify and remove these unwanted cells (for review see [18]).

We have previously reported on the heterogeneity of hESC cultures and developed a screen which used TG30 (CD9) and GCTM-2 (which recognizes an epitope on the protein core of a high molecular weight pericellular matrix keratin sulfate/chondroitin sulfate proteoglycan) antibodies reactive with the cell surface epitopes on hESC, allowing characterization and quantitative comparisons between different hESC lines [19–21]. Importantly, the immunofluorescent and flow cytometric combined detection of the cell surface antigens TG30 (CD9) and GCTM-2 strongly correlated with the expression of canonical pluripotency markers such as OCT4 and can be used to identify and subdivide cell populations with and without teratoma forming ability in hPSC cultures [19–21]. Using this approach, we have shown a continuous gradient of expression of genes associated with pluripotency in hESC cultures; a gradient that can be discriminated into subsets of cells at various stages of early differentiation, as indicated by their expression of stem cell markers and lineage specific transcription factors [19–21]. Therefore, in standard hESC cultures, cells with varying levels of pluripotency markers coexist with cells that have undertaken early stages of differentiation.

In this study, we used this fluorescence activated cell sorting (FACS)-based approach coupled with colony forming, transcriptomic, and teratoma assays to phenotype and compare hiPSC lines with hESC lines to test the hypothesis that these two cell types are extremely similar. We report for the first time that virally derived hiPSC exhibit a propensity to revert to a pluripotent phenotype following early-stage differentiation.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Standard hPSC Culture

Experiments were performed with hESC line MEL1 [22] and hiPSC lines produced using a retroviral [R] strategy [R]hiPS-PDL-D1C6 [23], a lentiviral [L] strategy [L]hiPS-IMR90C2 [3], [L]hiPS-ForeskinC1 [3], and an episomal [E] vector-based strategy [E]hiPS-NHF1.3 [24], provided by the StemCore facility at Monash University (Melbourne). All pluripotent cell lines were cultured on a monolayer (1.4 × 104 cells per centimeter square) of mitotically inactivated mouse embryonic fibroblast (MEF) feeder cells, in hESC/knockout serum replacement (KOSR) media composed of Dulbecco's modified Eagle's medium: nutrient mixture F-12 (DMEM/F-12, Invitrogen, Carlsbad, CA, http://www.invitrogen.com; 11320082) supplemented with 20% KOSR (Invitrogen 10828028), 2 mM Glutamax (Invitrogen 35050061), 1% modified Eagle's medium (MEM) nonessential amino acids (Invitrogen 11140050), 0.1 mM 2-mercaptoethanol (Invitrogen 21985023), 10 ng/mL human fibroblast growth factor 2 (FGF-2, Millipore, Billerica, MA, http:/www.millipore.com; GF003AF).

FACS

Cell cultures having reached approximately 80% confluence were washed and dissociated to obtain single cell suspensions using TrypLE Express (Invitrogen 12604039). Cells were washed again and passed through a 70-νm cell strainer (Becton Dickinson, Franklin Lakes, NJ, http://www.bd.com, 352350). Wash steps were followed by centrifugation for 2 minutes, and cells were resuspended in DMEM/F-12 plus 20% v/v fetal bovine serum (FBS). Cells were incubated with the monoclonal antibodies TG30 (CD9), (used at 1.4 μg/mL) and GCTM-2 hybridoma supernatant (used at 1:5 dilution) for 30 minutes on ice, on a rocking platform. Cells were next washed and incubated with Alexa Fluor 488 goat anti-mouse IgG2a (1:500 dilution, Invitrogen A21131) and Alexa Fluor 647 goat anti-mouse IgM (1:500 dilution, Invitrogen A21238) for 30 minutes on ice. Cells were also immunostained with CD90.2-PE (Becton Dickinson, 553014) to exclude MEF cells [25]. A Becton Dickinson FACS Diva instrument was calibrated before each experiment with Rainbow Calibration Particles (Becton Dickinson 559123), and the negative gate (TG30Neg-GCTM-2Neg, arbitrarily named P4 where P denotes “population”) was set using isotype controls IgG2a (1:500, Becton Dickinson 554126), IgM (1:500, Becton Dickinson 553472), and R-phycoerythrin goat anti-mouse IgG2a (1:1,000, Invitrogen P21139). Gates for low (gate P5), medium (gate P6), and high (TG30Hi-GCTM-2Hi, gate P7) coexpression of TG30 (CD9) and GCTM-2 were set as previously described [20, 26, 27]. Nonviable cells were excluded using 0.1% v/v propidium iodide (Sigma Aldrich, St. Louis, MO, http://www.sigmaaldrich.com P4864).

Colony Forming Assays

Cells were stained and FACS-fractionated as described above. Cells from the different TG30/GCTM-2 subfractions of hESC and hiPSC were plated per triplicate at a density of 30,000 cells per well in 12-well plates (Becton Dickinson, 353503), preplated with MEF cells (1.4 × 104 cells per centimeter square). Fractionated hPSCs were cultured in MEF-CM containing 10 ng/mL FGF-2. conditioned medium (MEF-CM) was generated by harvesting supernatant from inactivated MEF cells cultured in hESC/KOSR media (+10 ng/mL FGF-2) for up to 1 week. MEF-CM was collected and filtered (0.22 μm, Millipore) daily. Fractionated cultures were refreshed daily with MEF-CM + 10 ng/mL FGF-2 for approximately 2 weeks, after which hESC-like colonies were visualized under an inverted microscope (Olympus Mount Waverly, Vic. Australia, http://www.olympusaustralia.com.au IX71) and counted manually [27].

FACS-Mediated Passaging of TG30/GCTM-2 Populations

Cultures were TG30/GCTM-2 FACS fractionated to collect 400,000 cells from the differentiated P4 and pluripotent P7 cell populations. These P4 and P7 cell collections were cultured post-FACS in 75 cm2 flasks on MEFs and with MEF-CM. For some experiments, the FACS was repeated every 2 weeks to serially reseed P4 and P7 cells for passage.

Teratoma Formation Assay

P4 (TG30Neg-GCTM-2Neg) and P7 (TG30Hi-GCTM-2Hi) cultures were harvested after 2 weeks for injection into immunocompromised mice. Briefly, cells were harvested in clumps using 5 mg/mL Collagenase Type IV (Invitrogen 17104019) in DMEM/F-12, washed, and resuspended gently in 1 mL DMEM/F-12 plus 20% FBS with 500g × 2-minute centrifugations. Cell suspensions were then centrifuged at 500g × 5 minutes, supernatant discarded, and pelleted cell clumps were resuspended to a concentration of approximately 2 × 104 cells per microliter in DMEM/F-12 containing 33% v/v Matrigel (Becton Dickinson 354234). Cell suspensions were kept on ice prior to injecting approximately 1 × 106 cells into each testis of immune-deprived mice (non-obese diabetic severe combined immunodeficient (NOD-SCID) or non-obese diabetic severe combined immunodeficient interleukin 2 receptor gamma chain (NOD/SCID IL2Rγ−/−) using a 25-gauge needle. Mice were monitored daily post-transplantations and euthanized at 12 weeks, or earlier if growing tumor(s) reached 750 mm3 in volume, based on measurements with digital calipers and calculations with the ellipsoid volume formula V = (width2 × length)/2 [28]. The teratoma and surrounding tissue were removed, fixed with 4% v/v paraformaldehyde in Sorenson's buffer, and then paraffin sections were prepared for hematoxylin and eosin (H & E) staining and histological examination under an Olympus BX51 microscope for assessment of the human tissue types generated. Protocols and use of animals in this project were undertaken with approval of the Monash University Animal Welfare Committee following the 2004 Australian Code of Practice for the Care and Use of Animals for Scientific Purposes and the Victorian Prevention of Cruelty to Animals Act and Regulations legislation.

Immunocytostaining

hPSC cell-like colonies in 12-well plates were fixed with ice-cold absolute ethanol for 5 minutes. Colonies were immunostained for 30 minutes at room temperature with antibodies diluted in DPBS supplemented with 10% v/v goat serum. Primary antibodies were against OCT-4 (1:50, Santa Cruz SC-5279, Santa Cruz, CA, http://www.scbt.com), NANOG (1:50, eBioscience, San Diego, CA, http://www.ebioscience.com 14-5769), and TRA1-60 (1:100, Chemicon MAB4360, Temecula, CA, http://www.chemicon.com). Secondary antibodies were goat anti-mouse IgG2b-AF568 (1:500, Invitrogen A21151), IgG1-AF488 (1:500, Invitrogen A21121), and IgM-AF488 (1:500, Invitrogen A21042). Excess unbound antibody in each reaction was removed with Dulbecco's phosphate buffered saline (DPBS) washes. Colonies were counterstained with 4′,6-diamidino-2-phenylindole (Molecular Probes, Eugene, OR, http://probes.invitrogen.com, 2 ng/mL in DPBS) to visualize nuclear DNA. Immunofluorescence was analyzed using a fluorescent inverted microscope (Olympus IX71).

RNA Extraction and Reverse Transcriptase Polymerase Chain Reaction Analysis

Total RNA from hESC and hiPSC was extracted with a QIAGEN RNeasy Plus Micro kit according to the manufacturer's instructions. RNA concentration and purity were determined using a NanoDrop-1000 spectrophotometer (Thermo Scientific, http://www.thermoscientific.com). Total RNA (200–500 ng) was reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com; 4374966) according to the manufacturer's instructions. Equal volumes of cDNA reaction were added to the real-time polymerase chain reaction (PCR) reactions. Target cDNA levels were analyzed by the comparative cycle time (Ct) method of real time quantitative reverse transcriptase PCR (RT-PCR) with 20 μL reactions performed with TaqMan Gene Expression Master Mix (Applied Biosystems 4369016) and 1× concentrations of selected TaqMan Gene Expression Assays (Applied Biosystems, Supporting Information Table S1). Triplicate assays were carried out on an ABI Prism 7500 Sequence Detector System (Applied Biosystems), and the mean relative level of expression and associated SDs were calculated.

TaqMan Protein Quantification of OCT4

Quantification of OCT4 protein expression in cells collected from the P4 (TG30Neg-GCTM-2Neg) subfraction was performed. Briefly, hESC and hiPSC cultures were subjected to TG30/GCTM-2 cell surface staining for FACS analysis as described above, and 40,000 cells from P4 were collected by FACS. The TaqMan Protein Expression Core Reagents Kit with Master Mix (Applied Biosystems 4405501) was used to perform relative quantitation of OCT4 protein using a specific assay for human OCT4 (Applied Biosystems 4405489) according to the manufacturer's instructions. Triplicate assays were carried out on an ABI Prism 7500 Sequence Detector System (Applied Biosystems) and the exported Ct data were analyzed using the ProteinAssist Software (Applied Biosystems).

Statistical Analysis

In pertinent experiments, differences in the levels of expression were evaluated for statistical significance according to p values gained from a two-tailed Student's t test with a 95% confidence interval using GraphPad Prism version 5.02 for Windows (GraphPad Software Inc., La Jolla, CA, USA, http://www.graphpad.com).

Microarray Analysis

Briefly, isolated RNA (200 ng), from FACS-separated populations (as above) of triplicate biological samples from discrete passages of hiPSC lines hIPS-IMR90C2 [3] and hIPS-ForeskinC1 [3], was reverse transcribed into cDNA prior to hybridization to Illumina Sentrix Human 6 V2 or V3 BeadChip arrays (Illumina Inc, San Diego, CA, http://www.illumina.com). Signal intensity for each probe was calculated using Illumina BeadStudio V2.3.41, and Bioconductor packages (http://bioconductor.org) were used for downstream normalization and statistical analysis. Specifically, all samples were normalized using the R/Lumi quantile normalization package before submission to www.Stemformatics.org [29]. Principal component analysis (PCA) was generated on quantile normalized data using the AffyCoreTools R/Bioconductor package (http://www.bioconductor.org/packages/2.11/bioc/html/affycoretools.html). Data used for the PCA included published data for the hESC lines H9 and MEL1 [21]. Our previously published pluripotency associated gene lists [21] were saved to Stemformatics.org as a public gene set entitled “Pluripotency genes from Kolle et al. 2009” and the following analyses were conducted: Differential expression was assessed using the GenePattern Comparative Marker Selection tool, which provides a pairwise t test between all conditions, and returns an adjusted p value (Bonferonni, p < .01). Hierarchical clustering of samples and gene lists was performed using the GenePattern Hierarchical Cluster tool, with a Pearson correlation > = 0.8. Microarray data has been submitted to the Gene Expression Omnibus (GEO) repository under accession numbers GSE15283 and GSE13877, and may be viewed at Stemformatics.org.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

hiPSC and hESC Lines Exhibit Similar Immunophenotypic Profiles

Comparative analyses of hESC and hiPSC have previously shown similar key features, including expression of pluripotency markers, transcriptomes, cell and colony morphology, potential for teratoma formation, and the ability to differentiate into all three germ layers [2, 4-6]. For this study, we investigated whether hiPSC lines analyzed by FACS, based on the combined expression of the cell surface markers TG30 (CD9) and GCTM-2, could generate similar FACS plots and subpopulations of cells as those seen previously for hESC lines. We found that the hiPSC lines evaluated (Fig. 1C–1F) exhibited TG30/GCTM-2 FACS profiles similar to our hESC control line (Fig. 1B). We compared the transcriptome of the four sorted populations generated from hiPSC lines with array data previously published from multiple hESC lines [21, 29]. PCA showed that the TG30/GCTM-2 fractionation drove clustering of the cell lines along component 2. Other component comparisons are shown in Supporting Information Figure S1. Examination of components 2 and 3 demonstrates a high degree of similarity between hESC and hiPSC lines, with greatest clustering driven by the FACS separation into the four fractions (Fig. 1G). Using a list of 725 genes previously found to be highly correlated with pluripotency in hESC cultures [21], we demonstrated that the hiPSC subfractions with medium (P6) and high (P7) level of expression for TG30/GCTM-2 also exhibited the highest level of expression of pluripotency markers (Fig. 1H). The subfractions with negative (P4) expression for TG30/GCTM-2 showed a decrease in the expression of pluripotency markers in hiPSC. These results support strong similarities between hiPSC and hESC lines, and consistent with our previously published findings, indicate that in both human pluripotent cell types the P4 subfraction contains mostly spontaneously differentiated cell types and that the P6/P7 subfractions consist of cells with a pluripotent phenotype [20, 21].

thumbnail image

Figure 1. Strong similarities exist between hiPSC and hESC lines. (A): Schematic representation of the combined detection of cell surface markers TG30 (CD9) and GCTM-2 using FACS. Figure shows FACS populations (P) containing pluripotent cells (P6 and P7) and the population where differentiated cells are found (P4). (B): Typical FACS plot obtained with the hESC control line MEL1: P4 (TG30Neg-GCTM-2Neg), P5 (TG30Low-GCTM-2Low), P6 (TG30Mid-GCTM-2Mid), and P7 (TG30Hi-GCTM-2Hi). Negative P4 gate was set relative to isotype controls. (C–F): hiPSC lines showed strong similarities in the TG30 and GCTM-2 coexpression profile. ([C] retroviral [R]hiPS-PDL-D1C6 [23], [D] lentiviral [L]hiPS-IMR90C2 and [E] [L]hiPS-ForeskinC1 [3] and [F] episomally-derived [E]hiPS-NHF1.3 [24]). (G): Principal component analysis (PCA) generated using quantile normalized data from the microarrays conducted in this study on the hiPSC lines [L]IMR90C2 and [L]ForeskinC1, and on the hESC lines H9 and MEL1 from Kolle et al. [21]. Figure shows that two distinct and clearly separated clusters (P6/P7 and P4) include both hESC and hiPSC lines. PC2 stand for principal component 2, etc. (H): Heat map showing the expression of 725 genes that were preferentially expressed in human pluripotent stem cells and constitute a signature for pluripotency [21, 29]. Comparisons reveal that the P6 and P7 populations largely express the pluripotency markers analyzed, whereas the P4 cells are negative for the expression of the majority of these pluripotency-associated genes. Red cells indicate that the gene for a given sample is highly expressed. Blue cells indicate low expression. The range of intermediate colors gives an idea as to relative expression levels in the mid-ranges. Abbreviations: FACS, fluorescence activated cell sorting; hESC, human embryonic stem cell; hiPSC, human induced pluripotent stem cell; PC, principal component.

Download figure to PowerPoint

Consecutive Passaging of Cells from the P4 (TG30Neg-GCTM-2Neg) Subfraction Reveals Strong Differences Between hiPSC and hESC

We then analyzed whether cells from the P7 (TG30Hi-GCTM-2Hi) and P4 (TG30Neg-GCTM-2Neg) subfractions could be serially passaged post-FACS while maintaining their respective phenotypes. P7 cells retrieved following FACS from both hiPSC and hESC cultures were able to recapitulate the characteristic profile for the cell surface markers TG30 and GCTM-2 (Fig. 2). We observed that the majority of cells (ranging from 61% to 86%) were located in the subfractions associated with pluripotency (P6 and P7) and that a small percentage (ranging from 0.3-4.4%) of P4 (TG30Neg-GCTM-2Neg) cells were also present, demonstrating spontaneous differentiation (Fig. 2). Furthermore, cultures of P7 cells displayed visually obvious pluripotent-like colonies (data not shown).

thumbnail image

Figure 2. P7 (TG30Hi-GCTM-2Hi) cells from both hiPSC and hESC can be consecutively passaged by FACS without losing their intrinsic hPSC-like properties. Representative TG30/GCTM-2 FACS immunoprofiling of consecutive FACS-based passages (A-D): A starting population of 400,000 P7 (TG30Hi-GCTM-2Hi) cells was sorted from representative hiPSC lines; ([A] retroviral [R]hiPS-PDL-D1C6 [23], [B] lentiviral [L]hiPS-IMR90C2 [3] and [C] and episomally derived [E]hiPS-NHF1.3 [24]) and one hESC (MEL1) control line [D]). Then, P7 (TG30Hi-GCTM-2Hi) cells were consecutively passaged post-FACS until reaching approximately 80% confluence (9–11 days). TG30/GCTM-2 FACS immunoprofiling was performed at each passage and only P7 (TG30Hi-GCTM-2Hi) cells were recultured. Results show that P7 (TG30Hi-GCTM-2Hi) cells continually re-established the profile seen in the initial FACS and maintained the ability to differentiate, inferred from the continued appearance of differentiated P4 (TG30Neg-GCTM-2Neg) cells. (E): Percentages of cells present in each TG30/GCTM-2 subfraction during consecutive passages of P7 (TG30Hi-GCTM-2Hi) cells (n = 3, mean ± SEM). Abbreviations: FACS, fluorescence-activated cell sorting; hESC, human embryonic stem cell; hiPSC, human induced pluripotent stem cell.

Download figure to PowerPoint

However, when P4 (TG30Neg-GCTM-2Neg) cells were cultured in a similar manner post-FACS, strong differences were observed between hiPSC and hESC lines. We used a defined number of P4 (TG30Neg-GCTM-2Neg) cells that were cultured post-FACS (Fig. 3D) for 2 weeks and reanalyzed using TG30/GCTM-2 FACS before each passage. We expected that spontaneously differentiated cells from the P4 subfraction would enrich only for TG30Neg-GCTM-2Neg differentiated cell types after consecutive passaging. This prediction held true for hESC, which showed a minor presence (3.0% ± 1.7%) of P7 (TG30Hi-GCTM-2Hi) cells at the initial passage (Fig. 3D, Passage 1), eventually disappearing after further passaging (Fig. 3D, Passage 3). In agreement with this initial minor presence of P7 (TG30Hi-GCTM-2Hi) cells, a few pluripotent-like colonies were observed (data not shown). Intriguingly, with consecutive enrichment of TG30Neg-GCTM-2Neg differentiated cell types, these cells remained GCTM-2 negative after passaging but showed a trend to become positive for TG30 (Fig. 3D), most likely indicating a subset of cells undertaking differentiation to a lineage in which TG30 (CD9) is also present [30–32].

thumbnail image

Figure 3. Consecutive cultures of spontaneously differentiated P4 (TG30Neg-GCTM-2Neg) cells from retroviral and lentiviral hiPSC lines show persistence of undifferentiated pluripotent cells. Representative FACS plots showing populations of cells (P4, P5, P6, and P7) discriminated by the level of expression of TG30 and GCTM-2. (A-D): An initial population of 400,000 P4 (TG30Neg-GCTM-2Neg) cells was collected from hiPSC lines; (A) retroviral [R]hiPS-PDL-D1C6 [23], (B) lentiviral [L]hiPS-IMR90C2 [3], and (C) integration-free [E]hiPS-NHF1.3 [24]; compared with a hESC (MEL1) control line (D). These P4 (TG30Neg-GCTM-2Neg) cells were consecutively passaged post-FACS until reaching confluence (11–13 days). Prior to each passage, TG30/GCTM-2 FACS was carried out to collect and reculture only P4 (TG30Neg-GCTM-2Neg) cells. Strong differences between hiPSC and hESC were revealed by comparison of the FACS populations in consecutive cultures of differentiated P4 (TG30Neg-GCTM-2Neg) cells. Retroviral (A) and lentivirally derived (B) hiPSC showed low and high levels, respectively, of persistent P7 (TG30Hi-GCTM-2Hi) and P6 (TG30Mid-GCTM-2Mid) cells (red arrows indicate persistent cells). Conversely, the hESC control line (D), always showed absence of P7 cells at comparative passages (Passage 3). Interestingly, the episomally derived [E]hiPS-NHF1.3 (C) also did not show persistent pluripotent cells at passage 3. (E): Percentages of cells detected in each TG30/GCTM-2 subfraction during the consecutive passages of differentiated P4 (TG30Neg-GCTM-2Neg) cells analyzed (n = 3, mean + SEM). Abbreviations: FACS, fluorescence-activated cell sorting; hiPSC, human induced pluripotent stem cells; hESC, human embryonic stem cell.

Download figure to PowerPoint

A different result was obtained when P4 (TG30Neg-GCTM-2Neg) cells from hiPSC lines were tested in a similar fashion. Two virally derived hiPSC lines (Fig. 3A, 3B) showed persistence of P7 (TG30Hi-GCTM-2Hi) cells at passage numbers (e.g., Passage 3, 0.3% ± 0.1% for [R]hiPS-PDL-D1C6 or 16.9% ± 4.2% for [L]hiPSIMR90C2) in which no P7 cells or hPSC-like colonies were observed in the hESC control or the hiPSC line [E] hiPSNHF1.3 (Fig. 3D, Passage 3). We also noticed that the proportion of persistent P7 (TG30Hi-GCTM-2Hi) cells was much higher (16.9% ± 4.2% vs. 0.3% ± 0.1%) in cultures from [L]hiPS-IMR90C2 (Fig. 3B, Passage 3) than for the retrovirally derived [R]hiPS-PDL-D1C6 (Fig. 3A, Passage 3). Interestingly, we also observed that the episomally derived and integration-free [E]hiPS-NHF1.3 demonstrated similar behavior to the hESC control, in which P4 (TG30Neg-GCTM-2Neg) cultures became negative for P7 (TG30Hi-GCTM-2Hi) cells after two or three passages (Fig. 3C). These results suggested that the persistence of P7 cells might be related to the use of genome integrative technologies to generate hiPSC lines.

Late-Passage Lentiviral hiPSC Maintain Presence of Unexpected P7 (TG30Hi-GCTM-2Hi) Cells

Previous studies have shown that the transcriptional profile of hiPSC becomes more similar to that of hESC after several passages [33], and in miPSC the epigenetic memory of the cell of origin was eliminated by continued culturing [10]. To address whether the reappearance of cells with a pluripotent (P7) phenotype seen with virally derived hiPSC may also disappear with long-term passage, we analyzed late-passage (>50 passages) [L]hiPS-ForeskinC1 for the elevated presence of pluripotent hPSC-like colonies in cultured cells taken from the P4 (TG30Neg-GCTM-2Neg) population. Our colony forming assays (CFAs) with both early (below 20 passages) and late-passage [L]hiPS-ForeskinC1 showed an elevated number of colonies derived from P4 (TG30Neg-GCTM-2Neg) cells compared with the hESC-MEL1 control (Supporting Information Fig. S2A). Representative early and late passage colonies stained positively for pluripotency markers (Supporting Information Fig. S2B). We conclude that reversion to pluripotency in differentiated cells from the virally generated hiPSC line studied does not cease with continued passaging in vitro.

Presence of OCT4 Protein in P4 Cultures

We further investigated whether a higher expression of OCT4 protein in P4 (TG30Neg-GCTM-2Neg) cells from hiPSC lines could be the underlying mechanism for the elevated number of hPSC-like colonies observed in our CFAs. OCT4 protein levels in P4 cells were quantified using TaqMan Protein Assays. OCT4 protein expression analyses for [L]hiPS-IMR90C2 showed a significant (∼50%) increase compared with hESC cells (Supporting Information Fig. S2C), indicating a possible link between elevated OCT4 protein expression and the higher level of reversion to pluripotency that [L]hiPS-IMR90C2 showed in our experiments (Fig. 3B). However, the same analysis performed with [R]hiPS-PDL-D1C6 did not show a significantly higher OCT4 protein expression; rather, it was significantly lower than the hESC controls (Supporting Information Fig. S2C). Therefore, elevated OCT4 expression in [R]hiPS-PDL-D1C6 could not account for the persistence of pluripotent stem cells in our studies.

A Possible Causative Link Between Transgene Expression and Persistence of Pluripotency in Differentiated Cells from Virally Derived hiPSC lines

Our investigations described above showed that hiPSC lines exhibited persistence of P7 (TG30Hi-GCTM-2Hi) cells at passage numbers in which the control hESC line did not. Given that the nonintegrating episomally derived [E]hiPS-NHF1.3 did not show this reversion phenomenon (Fig. 3C), and behaved identically to our hESC control line, we hypothesized that transgene expression might be the underlying cause of the reversion seen. To test this notion, we tracked the expression of the transgenes in the TG30/GCTM-2 FACS subfractions (P4, P5, P6, and P7) using specifically designed Taqman assays to detect transgene expression by quantitative RT-PCR. We detected expression of the exogenous transcription factors used for the viral hiPSC generation in all the TG30/GCTM-2 FACS subfractions from [R]hiPS-PDL-D1C6 (Fig. 4A–4D) and [L]hiPS-IMR90C2 (Fig. 4E–4H). However, we also observed significant differences in the expression of the transgenes between TG30/GCTM-2 FACS subfractions and between the types of viral vectors used for iPSC generation. In general, retroviral transgenes were expressed in P4 cells and showed only a minimal expression in pluripotent P7 cells (Fig. 4A–4D). Lentiviral transgenes, however, showed more variable expression patterns between the TG30/GCTM-2 FACS subfractions (Fig. 4E–4H), and those transgenes detected exhibited expression in all subfractions. In the case of OCT4, lentiviral transgene expression was significantly increased in P7 cells compared to all other fractions. It is tempting to speculate that differences in transgene expression between retroviral and lentiviral vectors may be underpinning the differences in levels of reversion to pluripotency between retroviral and lentiviral hiPSC lines (Fig. 3A, 3B). This would imply that transgene expression may be driving a secondary reprogramming to pluripotency and causing a fraction of TG30Neg-GCTM-2Neg cells to reacquire immunoreactivity to TG30 and GCTM-2 becoming pluripotent P7 (TG30Hi-GCTM-2Hi) cells.

thumbnail image

Figure 4. Pluripotency-reverting hiPSC lines show viral transgene expression. Transgene expression was analyzed by TaqMan quantitative reverse transcriptase polymerase chain reaction using fluorescence-activated cell sorting (FACS) fractionated P4 (TG30Neg-GCTM-2Neg), P5 (TG30Low-GCTM-2Low), P6 (TG30Mid-GCTM-2Mid), and P7 (TG30Hi-GCTM-2Hi) cells from cultures of retrovirally derived [R]hiPS-PDL-D1C6 and lentivirally derived [L]hiPSC-IMR90C2. All the TG30/GCTM-2 FACS subfractions (P4, P5, P6, and P7) showed expression of the exogenous transcription factors but with strong differences between subfractions and between viral vectors. (A–D): Expression from retroviral transgenes (R-Tg) from [R]hiPS-PDL-D1C6 are depicted (OCT4 [A], KLF4 [B], SOX2 [C], and cMYC [D]). (E-H): Expression from lentiviral transgenes (L-Tg) are shown from [L]hiPSC-IMR90C2 (OCT4 [E], NANOG [F], and SOX2 [G]), but transgenic LIN28 was not detectable [H]. Average mRNA expression levels for P4 (TG30Neg-GCTM-2Neg) cells was set to 100% in each panel for comparisons (first column). Expression levels were normalized using 18S. Error bars represent SEM from n = 3. Asterisks highlight pertinent comparisons of the P4 control sample with the other populations (P5, P6, and P7), and indicate the level of statistical significance. *, p < .05; **, p < .01; ***, p < .001; ns, p > .05, with a 95% confidence interval. Abbreviation: hiPSC, human induced pluripotent stem cells.

Download figure to PowerPoint

We have shown that differentiated P4 (TG30Neg-GCTM-2Neg) cells derived from viral hiPSC lines were able to generate presumably pluripotent cells (P6 and P7 subfractions) based on the gain of coimmunoreactivity to TG30 and GCTM-2. We investigated whether P7 (TG30Hi-GCTM-2Hi) and P6 (TG30Mid-GCTM-2Mid) cells coming from pluripotency-reverting TG30Neg-GCTM-2Neg cells could exhibit, in addition to TG30 and GCTM-2 immunoreactivity, other pluripotency-associated characteristics. To address this question, seven consecutive passages of P4 (TG30Neg-GCTM-2Neg) cells from [L]hiPS-IMR90C2 and [R]hiPS-PDL-D1C6 lines were performed. Cells were collected from combined reverted P7 (TG30Hi-GCTM-2Hi) and P6 (TG30Mid-GCTM-2Mid) subfractions (Fig. 5A, 5B) in order to show meaningful comparisons as [R]hiPS-PDL-D1C6 cells (Fig. 5B) do not show high percentages of reverted P7/P6 cells. Given that P4 (TG30Neg-GCTM-2Neg) cells from hESC lines never showed pluripotent cells after three consecutive passages, we used P7/P6 cells retrieved from standard culture of hESC-MEL1 as a comparative positive high expression control (Fig. 5C) and hESC-MEL1 P4 cells as a low expression control. After performing colony-forming assays using P7/P6 cells, we carried out quantitative RT-PCR analysis. Our results showed that cultures of control P7/P6 cells from hESC-MEL1 and reverted P7/P6 cells from [L]hiPS-IMR90C2 and [R]hiPS-PDL-D1C6 expressed the pluripotency markers OCT4 (Fig. 5D), NANOG (Fig. 5E), and LIN28 (Fig. 5F), whereas P4 cells from hESC-MEL1 exhibited comparatively lower levels. This supports our observations that P7 (TG30Hi-GCTM-2Hi) and P6 (TG30Mid-GCTM-2Mid) cells derived from pluripotency-reverting P4 (TG30Neg-GCTM-2Neg) cells exhibit several characteristics of genuinely pluripotent stem cells in vitro.

thumbnail image

Figure 5. P7 (TG30Hi-GCTM-2Hi) and P6 (TG30Mid-GCTM-2Mid) populations of cells derived from reverting differentiated cells after seven passages contain cells with a pluripotent phenotype. P4 cells from pluripotency-reverting [L]hiPS-IMR90C2 and [R]hiPS-PDL-D1C6 lines were serially passaged by TG30/GCTM-2 fluorescence-activated cell sorting (FACS) seven times (A, B): Representative FACS plots of the seventh passage of P4 cells for each cell line [L]hIPS-IMR90C2 (A), [R]hIPS-PDL-D1C6 (B) compared with a control standard TG30/GCTM-2 FACS assay for the hESC-MEL1 line (C). Red arrows indicate pluripotent cells. (D-F): A defined number of reverted cells from combined P7 (TG30Hi-GCTM-2Hi) and P6 (TG30Mid-GCTM-2Mid) subfractions was collected for colony forming assays (CFA). P7 (TG30Hi-GCTM-2Hi) and P4 (TG30Neg-GCTM-2Neg) cells derived from the hESC-MEL1 assay were also collected as high and low expression controls. CFA cultures were subjected to quantitative reverse transcriptase polymerase chain reaction analysis for expression of canonical pluripotency markers. Relative expression of OCT4 (D), NANOG (E), and LIN28 (F). mRNA expression levels from hESC-MEL1 in cultures of P6/P7 cells were set to 100% for comparison (first column in each panel). Expression levels were normalized using 18S. Error bars represent SD from n = 3. Abbreviations: hiPSC, human induced pluripotent stem cell; hESC, human embryonic stem cell.

Download figure to PowerPoint

Differentiated Cells Expressing High Levels of TG30 and GCTM-2 from Virally Derived hiPSC Lines Generate Teratomas in Immune-Deprived Mice

Teratoma formation is a major safety concern for the use of hESC and hiPSC derivatives in clinical applications. Residual pluripotent stem cells subsequent to a specified differentiation assay could potentially be tumorigenic, and in the case of [R]hiPS-PDL-D1C6 and [L]hiPS-IMR90C2 lines, “potential” pluripotent cells remained unpurged from our cultures. We demonstrated that P7/P6 cells derived from differentiated P4 cells exhibited properties of pluripotency in vitro (Fig. 5). We therefore decided to investigate whether cultures of (TG30Neg-GCTM-2Neg) pluripotency-reverting cells could induce formation of teratomas in vivo in immune-deficient mice.

We used cultures of P4 (TG30Neg-GCTM-2Neg) cells, as described in Figure 3, which were consecutively passaged at least four times prior to injections into testes of immune-deprived mice. We found that differentiated P4 cells derived from [R]hiPS-PDL-D1C6 (Fig. 6D–6F) and [L]hiPS-IMR90C2 cell lines could generate teratomas (Fig. 6J–6L). As expected, for all cell lines, cultures of P7 (TG30Hi-GCTM-2Hi) cells generated teratomas (Fig. 6). We quantified the frequency with which cultures of P4 and P7 cells generated teratomas (Table 1). We found that cultures of P4 cells did not generate teratomas when derived from integration-free [E]hiPS-NHF1.3 and hESC-MEL1 line (Table 1). In distinct contrast, cultures of P4 cells from [R]hiPS-PDL-D1C6 and [L]hiPS-IMR90C2 lines generated teratomas with 44% and 83% efficiency, respectively. The higher efficiency of teratoma formation using [L]hiPS-IMR90C2 differentiated derivatives correlates with the higher level of reversion to pluripotency observed in this lentiviral hiPSC line (Fig. 3B). We conclude that the presence of P7 (TG30Hi-GCTM-2Hi) cells is a major safety concern for the in vivo applications of virally derived hiPSC lines such as [L]hiPS-IMR90C2 and [R]hiPS-PDL-D1C6.

thumbnail image

Figure 6. Cultures of differentiated cells from virally derived hiPSC lines develop teratomas after transplantation into mice. Immuno-deprived mice were injected with cultures of either P7 (TG30Hi-GCTM-2Hi) cells as described for Figure 2, or P4 (TG30Neg-GCTM-2Neg) cells (as described for Fig. 3); derived from retroviral [R]hiPS-PDL-D1C6, lentiviral [L]hiPS-IMR90C2, episomal [E]hiPS-NHF1.3 and hESC-MEL1 lines. Injections were performed after a minimum of four consecutive TG30/GCTM-2 fluorescence activated cell sorting-mediated in vitro passages. (A-R): Representative hematoxylin and eosin histology analysis of teratomas obtained from different cultures. Human tissues of the different germ layers are shown (ectoderm, mesoderm, and endoderm). Results from P7 (TG30Hi-GCTM-2Hi) cultures of [R]hiPS-PDL-D1C6 (A-C), [L]hiPS-IMR90C2 (G-I), [E]hiPS-NHF1.3 (M-O), and hESC-MEL1 (P-R). Teratomas derived from P4 (TG30Neg-GCTM-2Neg) cultures of [R]hiPS-PDL-D1C6 (D-F) and [L]hiPS-IMR90C2 (J-L). Scale bar = 200 μm. Abbreviations: hESC, human embryonic stem cell; hiPSC, human induced pluripotent stem cell.

Download figure to PowerPoint

Table 1. Teratoma formation efficiency in the testis capsules
Type of sample injectedTeratoma formation efficiency
Cell lineCulture of TG30/GCTM-2 FACS subfractionNOD/SCID IL2Rγ−/− miceNOD-SCID miceTotal % efficiency
  1. Immune-deprived mice (NOD-SCID and NOD/SCID IL2Rγ−/−) were injected with TG30/GCTM-2 FACS-mediated consecutive cultures of either pluripotent P7 (TG30Hi-GCTM-2Hi) or differentiated P4 (TG30Neg-GCTM-2Neg) cells after a minimum of four consecutive passages, as described for Figures 2 and 3, respectively.

  2. Abbreviations: FACS, fluorescence-activated cell sorting; hESC, human embryonic stem cell; hiPSC, human induced pluripotent stem cell; IL, interleukin; NOD-SCID, non-obese diabetic severe combined immunodeficient.

[R]hIPS-PDL-D1C6P7 (TG30Hi-GCTM-2Hi)5/75/583
 P4 (TG30Neg-GCTM-2Neg)2/42/544
[L]hIPS-IMR90C2P7 (TG30Hi-GCTM-2Hi)6/67/7100
 P4 (TG30Neg-GCTM-2Neg)5/75/583
[E]hIPS-NHF1.3P7 (TG30Hi-GCTM-2Hi)4/44/4100
 P4 (TG30Neg-GCTM-2Neg)0/60/70
hESC-MEL1P7 (TG30Hi-GCTM-2Hi)4/44/4100
 P4 (TG30Neg-GCTM-2Neg)0/60/50
hESC-MEL1Unsorted, standard culture4/58/986

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

This study demonstrates that the cell-surface FACS immunoprofiling assay that we previously developed to study hESC heterogeneity is equally informative for hiPSC pluripotency phenotypes. We further confirm that the immunoprofile of hiPSC and hESC correlate with transcriptional signatures of pluripotency. Significantly, these assays reveal strong differences between hESC and hiPSC stability in cell subsets that have undergone early-stage differentiation; differences that suggest that virally derived hiPSC are unlikely to generate clinical grade material suitable for transplantation therapeutics.

Pluripotent stem cells are highly proliferative and can readily form teratomas in vivo. This property presents a serious concern when considering using differentiated derivatives of these cells for transplantation into patients. Optimized differentiation protocols using pluripotent stem cells as a starting point are unlikely to be 100% efficient and any residual self-renewing pluripotent stem cells will likely form teratomas once transplanted. Therefore, it is reasonable to expect that any cell population aimed at human therapeutics should be characterized as being free of undifferentiated pluripotent stem cells [18]. Toward this end, we have demonstrated the utility of our FACS assays based on the combined expression of surface markers TG30 and GCTM-2 in the analysis of hESC and hiPSC lines. We first demonstrated that the P6 (TG30Mid-GCTM-2Mid) and P7 (TG30Hi-GCTM-2Hi) cell subfractions constitute the pluripotent stem cells in both hESC and hiPSC lines. Cells within these subfractions were able to generate hPSC colonies that were positive for pluripotency markers, possessed a transcriptome indicative of a pluripotent state, and cultures of P7 (TG30Hi-GCTM-2Hi) cells from both hESC and hiPSC lines efficiently formed teratomas in vivo. Second, we have shown that our FACS immunoprofiling can be used to identify early-stage spontaneously differentiated cells in cultures of hiPSC (this study) and hESC lines [21]. Third, using our FACS methods we were able to identify and retrieve viable pluripotent stem cells that were intermingled in cell cultures where the majority of the population were differentiated P4 (TG30Neg-GCTM-2Neg) cells. Therefore, we propose that TG30/GCTM-2 FACS analysis could be applied to all nonreverting hPSC cell lines to clear differentiated cultures of unwanted teratoma-forming hESC or hiPSC.

Similarly, we consider that the experimental protocols outlined in this manuscript could form the basis of a diagnostic kit to definitively recognize “unsafe” hiPSC lines. We have demonstrated that in vitro, our assays are useful in the diagnosis of persistent hPSCs in cultures of differentiated cell types. We also showed that persistence of pluripotent cells constitutes a serious hazard when transplanted in vivo (at least in an immune compromised mouse model), given that the presence of P7 (TG30Hi-GCTM-2Hi) cells correlated strongly with the formation of teratomas. Remarkably, even cultures of differentiated derivatives of the [R]hiPS-PDL-D1C6 line showing P7 (TG30Hi-GCTM-2Hi) cells below 0.5% were able to generate teratomas when injected into immune-deprived mice. Consequently, it was not surprising that differentiated cell cultures showing higher numbers of P7 (TG30Hi-GCTM-2Hi) cells were also able to generate teratomas in mice. However, teratomas generated from cell preparations with low numbers of pluripotent cells raises the question of what is an acceptable level of pluripotent stem cell contamination in a sample that could potentially be used in a patient that may be immune compromised? Answers to this question are not straightforward. Animal models have shown that the generation of teratomas from pluripotent stem cells depends on the immune system, as lower cell numbers are more tumorigenic in immunodeficient nude mice compared with immunocompetent mice [34]. As well, the site of injection may impact outcomes, since teratoma formation efficiency has been reported as 25%-100% for subcutaneous injections, 60% for intratesticular, 12.5% for intramuscular, and approximately 100% following injection under the kidney capsule [35]. Similarly, as few as two pluripotent stem cells appear to able to induce teratomas in nude mice [34], which indicates that for possible clinical applications of hESC and hiPSC, we should begin to consider that the only safe cell samples aimed for transplantations into patients are those that are stringently determined as being free of pluripotent stem cells. In summary, it is possible to confidently say that cultures of differentiated cells showing a presence of P7 (TG30Hi-GCTM-2Hi) cells as diagnosed by our FACS-based methods (i.e., [L]hiPS-IMR90C2 and [R]hiPS-PDL-D1C6 lines in this study) can be considered unsafe for clinical applications. Importantly, our results suggest that it will be essential to monitor whether pluripotent cells remain following in vitro differentiation protocols, prior to the transplantation of target cell populations. Moreover, our results also suggest strongly that any residual transgene expression in differentiated cell types derived from virally produced hiPSC lines should also be treated with extreme caution.

This leads us to the question of whether genome integrative technologies can produce clinical grade iPSC lines for therapeutics. In 2007, two independent groups achieved the generation of hiPSC lines using retrovirus [2] and lentivirus [3], two genome integrative technologies. However, genome integrative technologies have significant negative consequences for the potential use of hiPSC in the clinic. First, the site of integration itself constitutes a mutation of the genome that could affect genes or DNA sequences important for the correct function of the cells harboring the DNA modification at unpredictable times during cell life span. Second, there is always a latent risk of leakage in the expression of the exogenous transcription factors, or reactivation even when the transgene has been silenced. This is even more concerning when oncogenes such as c-MYC are used for the induction of reprogramming to pluripotency, given that its reactivation is clearly responsible for the development of tumors in mouse iPSC-generated chimeras [36]. Our experiments support the notion that continued expression of the viral transgenes is causing the phenomenon of reversion to pluripotency in spontaneously differentiated cells and might be driving the persistence of pluripotent P7 (TG30Hi-GCTM-2Hi) cells in [R]hiPS-PDL-D1C6 and [L]hiPS-IMR90C2 lines. In turn, persistent P7 (TG30Hi-GCTM-2Hi) cells are likely responsible for the teratoma formation in our animal experiments using differentiated P4 (TG30Neg-GCTM-2Neg) cells. Importantly, our experiments also show that transgene expression is variable and can change during differentiation.

We further considered that the expression of the transgenes might also affect or decrease the differentiation potential of hiPSC lines. However, studies have shown, using one of the same cell lines used in this study ([L]hiPS-IMR90), that directed differentiation into cardiomyocytes or neuroepithelial cells generates functional cells exhibiting properties highly similar to human ES cell-derivatives or even integration-free hiPSC lines [37–39], despite the clearly detectable lentiviral transgene expression in the [L]hiPS-IMR90 clones [38]. A possible explanation of these results is that the levels of expression of the exogenous transcription factors might not be high enough to interfere with the differentiation potential of hiPSC line [L]hiPS-IMR90 in vitro and/or that reversion to pluripotency may only be triggered under specific and yet to be fully understood conditions.

The experimental evidence in our study demonstrates that the integration-free hiPSC line [E]hiPS-NHF1.3, generated using episomal vector methodology, and the hESC line MEL1 did not show persistence of the formation of P7 (TG30Hi-GCTM-2Hi) cells from the differentiated P4 cells. In addition, cultures of P4 (TG30Neg-GCTM-2Neg) cells from hiPS-NHF1.3 and from hESC-MEL1 did not induce teratomas when assayed in immune-deprived mice. Based on these results, it could be argued that nonintegrative technologies offer a more suitable hiPSC source for any potential human therapeutics. However, recent studies on genetic [13] and epigenetic [11] abnormalities induced during the process of reprogramming included some hiPSC that were produced using nonintegrative methodological approaches. The authors reported that the same epigenetic memory and copy number variations are exhibited by hiPSC lines devoid of viral integrations [11, 13], and consequently it is not a phenomenon that depends on the transgene expression. However, hESC lines have also been shown to display considerable genetic and epigenetic heterogeneity and have differing propensities for differentiation, much like those found in hiPSC lines [40, 41]. Furthermore, no functional assays have been reported to demonstrate that hiPSC subchromosomal changes are functionally affecting any particular gene, and therefore the effects of these novel abnormalities on the growth, differentiation, tumorigenicity, and general functionality of hiPSC remain to be fully elucidated. It is therefore our view that presently it is premature to advance any claim regarding the clinical safety of hiPSC generated by nonintegrative technologies.

Recently, an independent group working with miPSC lines reported that retrovirally derived miPSC showed residual undifferentiated cells that could not be eliminated after long-term differentiation protocols and could form teratomas in vivo [42]. The results from Fu et al. [42] are evidence of the phenomenon of reversion to pluripotency in mouse iPSC lines. Unfortunately, the transgene expression of Nanog, Oct4, Klf4, and c-Myc was not examined and as we found for the virally produced hiPS cell lines in our study, this might be the underlying cause of the reported phenomenon for murine lines. Interestingly, it is possible to obtain mice via tetraploid complementation using one of these miPSC lines [42], which suggests that despite the underlying reversion to pluripotency, these miPSC lines can fully undergo differentiation and complete developmental potential. However, it would be interesting to study the whole life span of these mice to ascertain whether tumorigenesis is elevated.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

In summary, our investigations have demonstrated that using a FACS strategy based on the coexpression of the cell surface markers TG30 and GCTM-2, it is possible to identify unsafe hiPSC lines. The ability to detect low levels of pluripotent stem cells in cultures where the vast majority are differentiated cell types also provides a sensitive assay, which constitutes a tool for the stringent elimination of teratoma-forming hESC and hiPSC from cultures of target differentiated cells prior to clinical applications.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

We thank the StemCore facility at Monash University for provision of hPSC lines, the SRC microarray facility at the University of Queensland for processing Illumina microarrays, the FlowCore facility at Monash University for FACS services. Human iPS cell lines (hiPS-IMR90 and hiPS-Foreskin) were kindly provided by Dr. James A. Thomson (University of Wisconsin). NOD/SCID IL2Rγ-/- mice were generously provided by Dr. Susie Nilsson at CSIRO. We also thank Chad Heazlewood and Daniela Cardozo at CSIRO for technical assistance with histology and mouse work and the following for critical appraisals of draft versions of this manuscript; David Haylock and Tung-Liang Chung (CSIRO), Martin Pera (University of Melbourne) and Jose Polo (Monash University). GCTM-2 and TG30 were kindly supplied by Prof. Martin Pera (University of Melbourne). This work was supported by research grants from the Australian Stem Cell Centre and the New South Wales and Victorian Government Stem Cell Research Grant Program to A.L.L. A.L.L. is a Partner Investigator on the Australian Research Council (ARC) Special Research Initiative in Stem Cell Science, Stem Cells Australia.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
STEM_1425_sm_SuppFigure1.pdf265KSupplementary Figure S1. Principal Component Analysis (PCA). PCA generated using quantile normalised data from the arrays conducted in this study on the hiPSC lines [L]IMR90C2 and [L]ForeskinC1 and from the hESC lines H9 and MEL1 from Kolle et al., 2009. [A] Separation of hiPSC and hESC samples on component 1 is likely explained by differences in the array version between the published hESC data (V2) and newer hiPSC experiments (V3) and component 2 appears to be driven by TG30/GCTM-2 fractionation. [B] Absolute measure of the percentage variance of each principal component indicating that the data can be explained by the first three components.
STEM_1425_sm_SuppFigure2.pdf216KSupplementary Figure S2. Persistence of P7 cells does not cease with long term in-vitro passaging and presence of OCT4 in P4 cultures. (A): Comparative analysis of Colony Forming Assays (CFA) between late-passage (<50 passages) lentivirally-derived [L]hIPSForeskinC1 and hESC-MEL1 control line. A defined number of P4 (TG30Neg-GCTM-2Neg) cells were cultured post-FACS in 12-well plates, generating an elevated number of hPSC-like colonies only for the lentiviral hiPSC samples. Data collected from three independent CFA (n=3), each assayed in triplicate (mean ± SEM). (B): Immunofluorescence for pluripotency markers. Colonies generated by P4 (TG30Neg-GCTM-2Neg) cells from late-passage lentiviral hiPSC samples, exhibited immunoreactivity to the pluripotency markers OCT4 (red) and NANOG (green). (C): Quantification of OCT4 protein in the P4 (TG30Neg-GCTM-2Neg) population from both hiPSC and hESC lines. P4 exhibits a low level of OCT4 protein and requires the highly sensitive TaqManR Protein Assays for quantification. A defined number of P4 (TG30Neg-GCTM-2Neg) cells was collected by FACS from standard cultures of hiPSC lines (lentiviral [L]hIPS-IMR90C2 and [L]hIPS-ForeskinC1, retroviral [R]hIPS-PDL-D1C6) and hESC control (MEL1), followed by OCT4 protein quantification. Expression levels are shown relative to hESC-MEL1 set to 100%. Data from three independent experiments (n=3), assayed in triplicates, error bars represent SEM. Asterisks highlight statistical significance with respect to the control sample hESC-MEL1. *P<0.05, **P<0.01, ***P<0.001; ns, P<0.05, with a 95% confidence interval.
STEM_1425_sm_SuppTable1.pdf12KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.