Pluripotentiality and Conditional Transgene Regulation in Human Embryonic Stem Cells Expressing Insulated Tetracycline-ON Transactivator


  • Diego S. Vieyra,

    1. Stem Cells and Regenerative Medicine Center and Center for Cell and Gene Therapy, Department of Pediatrics, Baylor College of Medicine, Houston, Texas, USA
    Search for more papers by this author
  • Margaret A. Goodell Ph.D.

    Corresponding author
    1. Stem Cells and Regenerative Medicine Center and Center for Cell and Gene Therapy, Department of Pediatrics, Baylor College of Medicine, Houston, Texas, USA
    • Stem Cells and Regenerative Medicine Center, Baylor College of Medicine, One Baylor Plaza, N1030, Houston, Texas 77030, USA. Telephone: 713-798-1265; Fax: 713-798-1230
    Search for more papers by this author


Conditional manipulation of gene expression by using tetracycline (TET)-ON based approaches has proven invaluable to study fundamental aspects of biology; however, the functionality of these systems in human embryonic stem cells (hESC) has not been established. Given the sensitivity of these cells to both genetic manipulation and variations of culture conditions, constitutive expression of TET transactivators might not only be toxic for hESC but might also impair their ability to self-renew or differentiate into multiple tissues. Therefore, the effect of these transactivators on the biology and pluripotentiality of hESC must first be evaluated before broad use of TET-ON methodologies is applied in these cells. Improved insulated lentivectors that display stable transgene expression and minimal insertional transactivation have been described for hESC. By using insulated lentivectors that allow simultaneous expression of TET components and fluorescent reporters, here we demonstrate that hESC constitutively expressing the TET-ON transactivator rtTA2SM2 can be derived and expanded in culture while retaining inducible transgene expression and pluripotentiality, including marker expression, a normal karyotype, and the ability to generate multiple tissues of different germ layer origin in teratomas. We also show that these cells retain the ability to control the expression of a stable integrated transgene in a doxycycline-dependent manner, which demonstrates that an insulated TET-ON lentiviral system is functional in hESC. Together, our results indicate that improved TET regulators like rtTA2SM2 in combination with insulated lentiviral-based systems offer alternative strategies for conditional gene expression in hESC.

Disclosure of potential conflicts of interest is found at the end of this article.


The main property that distinguishes human embryonic stem cells (hESC) from any other cells is their ability to self-renew and differentiate into any cell type of the human body, a property called pluripotency [1]. Because hESC can be genetically modified and expanded in culture for long periods of time as undifferentiated cells, hESC offer an invaluable model system to experimentally address aspects related to human development as well as a potential source of tissues and biomolecules for biomedical applications [2]. However, our knowledge of the molecular mechanisms controlling the biology of hESC is hampered, in part, by methodological limitations, including the efficient conditional manipulation of gene expression in these cells.

Several groups have recently reported success in the conditional manipulation of genes in hESC [3, [4]–5]. However, these studies have also shown that the particular biology of hESC imposes experimental challenges that affect the efficacy of standard conditional gene expression approaches. Some current culture methodologies require hESC and their differentiated progeny to be expanded as clumps of cells such as colonies, embryoid bodies (EBs), and teratomas, which limit the efficiency of gene delivery and ectopic control of transgene expression. Similarly, the derivation of gene regulatable hESC lines has been shown to be hindered by toxicity of stably expressed regulators, including Cre [5] and tetracycline (TET) transrepressors [4]. Therefore, alternative conditional gene approaches are still needed for hESC.

TET transactivators offer alternative and unique features for hESC models. Its inducer, doxycycline (dox), has proven to be well tolerated and to display efficient and broad tissue penetration in murine and primate models [6]. Also, improved TET-ON transactivators with decreased toxicity and increased sensitivity of response have been developed [7]. However, the functionality of TET-ON systems has not been explored in hESC models. Furthermore, the effect of TET regulators on pluripotentiality of hESC has not been investigated, which is a fundamental aspect to be evaluated before broad use of TET-based methodologies in hESC.

Lentiviral approaches have been shown to allow fast and efficient production of stably modified hESC [8, [9]–10]. Lentivectors can efficiently and stably transduce both dividing and nondividing cells and can carry inserts up to more than 15 kilobases (kb), which allows large and complex transgene arrangements. Improved insulated lentivectors that display stable transgene expression with minimum risk of insertional oncogenic transactivation have been developed by Ramezani and colleagues [11]. These lentivectors have been successfully applied to derive transgenic stem cell lines [11], including hESC [10]. Recently, Ma and colleagues [10] have reported derivation of green fluorescent protein (GFP) hESC that retained the ability to differentiate into fluorescent progeny after extended culture.

Using insulated lentivectors, we investigated whether cells constitutively expressing the improved TET-ON transactivator rtTA2SM2 [7] could be derived from hESC and, if so, whether these cells could be expanded in culture while retaining pluripotency and inducible transgene properties. We then investigated whether a TET-ON-responsive system could be customized into an insulated lentivector from which we could track and manipulate transgene expression from hESC. We demonstrate that transgenic hESC constitutively expressing rtTA2SM2 can be derived and expanded in culture as pluripotent cells that retain dox-inducible transgene regulation, multilineage differentiation potential, and expression of rtTA2SM2 in highly differentiated tissues derived from them. We also demonstrate functionality of an insulated lentiviral TET-ON system in hESC. Together, our results indicate that improved TET regulators like rtTA2SM2 in combination with insulated lentiviral-based systems offer alternative strategies for conditional gene expression in hESC models.

Materials and Methods

hESC Culture and Differentiation Assays

H9 hESC (NIH registry WA09; obtained from WiCell Research Institute, Madison, WI,, TET1, enhanced green fluorescent protein (EGFP)1, and TET1-Luc cells were grown on gamma-irradiated mouse embryonic fibroblast feeders in 80% Dulbecco's modified Eagle's medium-F12 (Gibco, Grand Island, NY,, 20% Knockout Serum Replacement (Gibco), 1 mM glutamine (Gibco), 0.1 mM βmercapto-ethanol (Sigma-Aldrich, St. Louis,, 1% nonessential amino acids (Gibco), and 4 ng/ml human recombinant basic fibroblast growth factor (bFGF) (Invitrogen, Carlsbad, CA, as described [12]. Karyotypic analysis of cells was performed by analysis of 60 metaphases of KaryoMAX (Gibco)-treated samples. Cytogenetics analysis (G-banding) of cells was carried out at the Texas Children's Hospital Cytogenetics Lab. Generation of EBs was performed as in [13]. Briefly, clumps of 200–300 undifferentiated hESC were transferred into low attachment plates and cultured in suspension in hESC medium without bFGF for 2 weeks. In vivo teratoma assays were performed as in [8, 12]. Clumps of approximately 300 transduced cells were injected in the hind limb of nonobese diabetic (NOD)/severe combined immunodeficient (SCID) mice. Eight weeks later, animals were sacrificed, and tumors were immediately removed, rinsed in phosphate-buffered saline (PBS), fixed in 4% formalin, and embedded in paraffin. Five-μm serial sections were then processed for hematoxylin-eosin staining and immunofluorescence.

Construction of Lentivectors and Production of Lentiviruses

We engineered pLTET1 as follows. An EcoRI fragment containing the cDNA of rtTA2SM2 [7] was digested from pBTE [14] (gift from F.J. Kern under approval of W. Hillen) and subcloned into NotI of pTRIDENT32 [15]. A BamHI/XhoI fragment containing rtTA2SM2 cDNA flanked by two mammalian glial-and-testis-specific Homeobox internal ribosome entry sequences was digested from this construct and subcloned into BclI/SalI of pLSIND1D2Lk1, a lentivector we derived from pLSIN-EF1α-GFP-SAR/HS [11] (gift from R. Hawley), to improve plasmid replication and facilitate further subcloning into our lentivectors. pLSIND1D2Lk1 was engineered by removing two 2-kb fragments of nonviral DNA in two digestion-filling-religation steps (HpaI/SexA1, PacI/AatII) followed by subcloning of a multiple cloning site linker into EcoI/SalI, downstream of the EF1 promoter. We engineered pLTRET-Luc as follows. An XhoI/NotI fragment containing the TET response element (TRE)-tight promoter and firefly luciferase cDNA from pTRE-Tight Luciferase (Clontech, Palo Alto, CA, was subcloned into NheI/NotI of pTRIDENT32, excised with CeuI/ppoI, and resubcloned into pLSIND1D2LK2-DsRedEx after removal of a PmeI fragment. We derived the latter from pLSIND1D2Lk1 by removing a BamHI fragment containing EGFP and subcloning the cDNA of DsRed Express (from pCMV-DsRed-Express; Clontech) into XbaI and a CeuI/ppoI linker into ClaI. Production and titration of lentiviruses were performed in 293T cells (American Type Culture Collection, Manassas, VA, as in Ramezani et al. [11]. Viral supernatants were concentrated by using centrifugation columns (Amicon; Millipore, Billerica, MA, and stored at −80°C.

Derivation of TET1, EGFP1, and TET1-Luc Cells

TET1 and EGFP1 cells were derived by transduction of passage 35 H9 hESC with pLTET1 and pLSIND1D2Lk1, respectively. TET1-Luc cells were derived by transduction of passage 10 TET1 cells with pLTET-Luc. Transduction of hESC was performed at multiplicity of infection (MOI) 10 as in [10] with minor modifications. Dissected clumps of 2-day-old plated hESC and viral particles were incubated together in low attachment plates in the presence of hESC medium and 8 μg/ml polybrene at 37°C for 4 hours, after which the clumps were rinsed with hESC medium and plated onto fresh feeder plates. Under these conditions, we observed averages of 50–200 EGFP or DsRed positive cells per 3-cm plate 4–6 days after transduction with pLTET1 or pLTET-Luc, respectively. Due to our lack of success in establishing cultures from fluorescence-activated cell sorting (FACS)-sorted cells, TET1, EGFP1, and TET1-Luc cultures were propagated from mechanically dissected clumps of transduced cultures. Enrichment for positive transductants was performed by successive rounds of manual dissection under fluorescent stereomicroscope, as described [16].

Transient Transfection and Luciferase Activity Assays

Transient transfections were performed as in [17] on passage 10 TET1 cells (Fig. 2) by using 6 μl of Lipofectamine 2000 (Invitrogen), 2.5 μg of TRE-tight-Firefly (Clontech), and 0.5 μg of CMV-Renilla (Promega, Madison, WI, luciferase plasmids in a final volume of 340 μl. Two μg/ml doxycycline (Sigma) was added 4 hours post-transfection. Thirty-six hours after addition of doxycycline, both cells from supernatants and attached cells were resuspended in lysis buffer, pooled, and stored at −20°C until analysis. For TET1-Luc experiments, protein extracts were prepared from cells grown in the presence of 2 μg/ml dox for 3 days. Fresh medium containing dox was replaced daily. Extracts containing similar numbers of EGFP+DsREDEx+ were produced from pools of trypsinized cultures processed by FACS and fluorescent microscopy. Cells were rinsed with PBS and resuspended in lysis buffer as in transient transfection assays. Analysis of samples for luciferase activity was performed on a single-tube automated injector luminometer (model TD-20/20; Turner BioSystems, Sunnyvale, CA, by using the dual luciferase reporter assay kit (Stratagene, La Jolla, CA,, as recommended by the manufacturer's instructions.

FACS and Immunofluorescence Microscopy

Preparation of hESC and EB samples for FACS and immunohistochemical analyses was carried out as described in [13, 16] with minor modifications. For FACS and immunofluorescence microscopy studies on EBs and undifferentiated cells, cells were fixed in 2% paraformaldehyde (Sigma) PBS−/− (Gibco). For Oct4 staining, fixed cells were permeabilized with 0.2% Triton X-100 (Sigma) PBS+/+ (Gibco). We used the following primary antibodies: mouse anti-tet repressor (TETO3; Boca Scientific, Boca Raton, FL,, rabbit anti-GFP (Molecular Probes, Eugene, OR,, rabbit anti-oct4 (Santa Cruz Biotechnology Inc., Santa Cruz, CA,, mouse anti-SSEA4 (Chemicon, Temecula, CA,, mouse anti-Tra-1-81 (Chemicon), and goat anti-Firefly Luciferase (Abcam, Cambridge, U.K., Secondary antibodies were: rat anti-mouse IgG1 phycoerythrin (PE) (Cedar Lane, Hornby, ON, Canada,, goat anti-rabbit IgG PE (Cedar Lane), Alexa 594 donkey anti-mouse IgG (Molecular Probes), Alexa 594 donkey anti-rabbit IgG (Molecular Probes), PE-goat anti-mouse IgM (Miltenyi Biotec, Bergisch Gladbach, Germany,, and cy5 donkey anti-goat (Jackson Immunoresearch Laboratories, West Grove, PA, TET1-Luc cells, EBs, and teratoma tissue sections were analyzed by confocal laser microscopy in a Zeiss Axiovert 100M microscope. Dual-fluorochrome images were collected by using the LSM 510 acquisition software (Carl Zeiss, Jena, Germany, All images were collected by using slow scan, and signal gain settings were kept constant for each detector to minimize differences among samples.

Reverse Transcription-Polymerase Chain Reaction Assays

RNA isolation was performed by using the RNAqueous Kit (Ambion, Austin, TX, We used Superscript III (Invitrogen) and recombinant Taq (Invitrogen) for reverse transcription (RT) and polymerase chain reaction (PCR) reactions, respectively. We designed the following primers: TGGAGGAACAGGAGCATCAAGTAG and GAGCATGTCAAGGTCAAAATCGTC for rtTA2SM2; AAGCTGACCCTGAAGTTCATCTGC and GTCTTGTAGTTGCCGTCGTCCTT for EGFP; GCCCCGTAATGCAGAAGAAGACTA and TCCAGCTTGGAGTCCACGTAGTAG for DsRedEX; and ACCTCCCGGTTTTAATGAATACGA and TCTCAGTGAGCCCATATCCTTGTC for firefly luciferase. We used the following set of previously characterized primers [18]: ACCAGGACCTGCTCAATGTC and ATCTCCACGGTCTTCACCAC for glial fibrillar acidic protein (GFAP); AGAACCTGTCACAAGCTGTG and GACAGCAAGCTGAGGATGTC for α-fetoprotein (AFP); and CCCTGCACCAGCCCCAATCAGA and CGAAGCCCAGCCCGGTCAA-CT for cardiac troponin (CTN1).

Real-Time PCR Studies

Real-time quantitative PCR was performed on the ABI Prism 7500 Sequence Detection System (Applied Biosystems, Foster City, CA, Amplification was performed using SYBR Green Master Mix (Applied Biosystems). Lentivirus copy number was estimated by the “standard curve method” using serial dilutions of linearized pLTET “spiked” with 50 ng of nontransduced human genomic DNA (Promega). Standards were set up to range from 300,000 to 30 copies of lentivector per reaction (supplemental online Fig. 5). Genomic DNA was prepared from FACS sorted cells by using the DNeasy Tissue Kit (Qiagen, Hilden, Germany, We used a previously characterized [19] set of primers raised against the long terminal repeat of human immunodeficiency virus-based lentivectors: GCCTCAATAAAGCTTGCCTTGA and TCCACACTGACTAAAAGGGTCTGA. The “Δ ΔCt method” was used to quantify dox-inducible transgene response of TET1-Luc cells. To this end, we designed the following set of primers using Primer Express Software (Applied Biosystems): GCTATGAAGAGATACGCCCTGGTT and CAACACCGGCATAAAGAATTGAAG for luciferase and CTACGGCGTGCAGTGCTTC and CCCTCGAACTTCACCTCGG for EGFP; the latter was used as an internal control to normalize relative amounts of luciferase transcripts. Ct values of dox-untreated cells were used as calibrators. The relative amount of luciferase gene expression for each sample was calculated and plotted as the average ± standard error.


Construction of pLTET1 and Derivation of rtTA2SM2-Expressing Cells

To experimentally address whether rtTA2SM2-expressing hESC could be derived and expanded in culture, we aimed at producing hESC coexpressing rtTA2SM2 and EGFP by lentiviral transgenesis [8, [9]–10], so transduced hESC and their progeny could be readily tracked by green fluorescence [8, 10, 20]. To this end, we engineered pLTET1, an insulated lentivector derived from pLSIN-EF1α-GFP-SAR/HS [11], whose functionality has been well established in hESC [10]. In pLTET1, strong mammalian internal ribosome entry sequences allow simultaneous expression of rtTA2SM2 and EGFP from the same transcript (Fig. 1A). Ubiquitous and stable transgene expression is achieved through the elongation factor 1α (EF1a) promoter and scaffold attachment region sequences, respectively. The insulator element from the globin locus prevents transcriptional activation of genes flanking the site of lentiviral integration [11] (Fig. 1A).

Figure Figure 1..

Derivation and functional characterization of TET1 cells. (A): Schematic representation of pLTET1 lentivector used to derive TET1 cells. (B–E): Phase contrast and indirect fluorescence microscopy studies on TET1 cells. TET1 cell colonies displayed typical morphology of undifferentiated hESC (B), strong expression of EGFP (C), and rtTA2SM2 (D–F). Scale bar = 100 μm. (F): Reverse transcription-polymerase chain reaction assay on RNA extracts isolated from TET1 cells. Control: rtTA2SM2-expressing 293T cell line derived by transduction of pLTET1. EGFP1: EGFP-expressing hESC. hESC: nontransduced hESC. (G): TET1 cells displayed doxycycline-inducible reporter activity when transiently transfected with a TRE-Firefly luciferase plasmid; however, EGFP1, hESC, and feeders did not show inducible luciferase activity. Values were normalized to Renilla luciferase internal transfection controls in each individual experiment. Average and standard deviations for three independent experiments. Abbreviations: EF1a, elongation factor 1α; EGFP, enhanced green fluorescent protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; hESC, human embryonic stem cell; INS, insulator; LTR, long terminal repeat; mires, mammalian internal ribosome entry sequence; ppT, polypurine track; SAR, scaffold attachment region.

Transduction of H9 hESC with TET1 lentivirus (MOI10) resulted in hundreds of EGFP-positive colonies that could be expanded in culture until EGFP homogeneity (Figs. 1, 2). TET1 colonies exhibited typical morphology of undifferentiated hESC [12], including defined borders and homogeneous cell type with prominent nuclei (Fig. 1B). TET1 cells displayed robust and homogenous expression of both EGFP and rtTA2MS2 as indicated by immunofluorescence microscopy (Fig. 1C–1E) and RT-PCR studies (Fig. 1F). The presence of functional rtTA2SM2 in these cells was evaluated by assessing dox-dependent induction of luciferase reporter activity after transfection with a TRE (TET response element)-luciferase construct. Cotransfection of TET1 cells with TRE-Firefly and CMV-Renilla luciferase plasmids allowed us to normalize levels of reporter induction to efficiency of transfection in each experiment. Under the assayed conditions, dox-treated TET1 cells displayed an average of 120-fold more firefly luciferase activity than dox untreated and nontransduced cells (Fig. 1G). However, under the same conditions, no inducible luciferase activity could be detected in nontransduced hESC, hESC expressing EGFP, or feeders (Fig. 1G), indicating that a dox-inducible luciferase activity was present only in TET1 cells. Together, these results demonstrated that cells constitutively expressing functional rtTA2SM2 could be derived from hESC.

Figure Figure 2..

Expression of markers of undifferentiated hESC and pluripotency in TET1 cells. (A–F): Fluorescence-activated cell sorting analysis of TET1 cells and untransduced hESC. TET1 cells expressed typical markers of undifferentiated hESC, including OCT-4 and Tra-1-81, at similar levels of expression to control hESC. (G–I): Teratoma assay. TET1 cells retained pluripotency. Xenotransplanted TET1 cells produced encapsulated teratomas containing derivatives of different embryological origin, including glandular (G), neural (H), and cartilage (I) tissue. Hematoxylin-eosin staining. Scale bar = 100 μm. Abbreviations: EGFP, enhanced green fluorescent protein; hESC, human embryonic stem cell.

Expression of hESC Markers and Pluripotency in rtTA2SM2-Expressing Cells

We then asked whether TET1 cells retained well-established hESC culture properties, specifically, expression of undifferentiation markers, normal karyotype, and pluripotency. To this end, we performed comparative studies between TET1 cells and nontransduced hESC. FACS analysis indicated that TET1 cells expressed typical cell surface and nuclear markers that are present in undifferentiated hESC, including TRA-1-81 and OCT-4 (Fig. 2A–2F). Except for the robust expression of EGFP in TET1 cells, levels of expression and percentage of cells expressing these hESC markers were indistinguishable between TET1 cells and control hESC (Fig. 2A–2F). Karyotypic analysis of these cells also indicated normal number and structure of chromosomes for TET1 cells (supplemental online Fig. 1). These observations led us to investigate whether TET1 cells retained pluripotency. To this end, we performed transplantation studies in NOD/SCID mice. Palpable tumors could be detected in animals transplanted with both control and TET1 cells after 8 weeks. Detailed examination of these tumors revealed that, like control hESC, TET1 cells generated typical encapsulated teratomas containing multiple tissues of endodermic, ectodermic, and mesodermic origin, including glandular, neural, and cartilage tissues, respectively (Fig. 2G–2I). Altogether, these data indicated that constitutive expression of rtTA2SM2 did not impair the ability of TET1 cells to be passaged in culture as undifferentiated pluripotent cells.

Retention of rtTA2SM2 Expression in Highly Differentiated Progeny of TET1

We further evaluated the effect of constitutive expression of rtTA2SM2 on the differentiation of TET1 cells. Specifically, we asked whether expression of rtTA2SM2 was retained in highly differentiated progeny of TET1 cells. We reasoned that if rtTA2SM2 was toxic for differentiated TET1 progeny, we would not detect rtTA2SM2-expressing cells after differentiation. A classic in vitro differentiation approach involves the generation of EBs, in which multiple and varied cell types can be produced [13]. As shown in Figure 3A–3D, 2-week-old TET1 EBs expressed strong levels of EGFP and rtTA2SM2 together with several tissue-specific markers, which were present only in differentiated hESC. These included typical markers of ectoderm, such as GFAP, mesoderm, such as CTN1, and endoderm, such as AFP (Fig. 3D). These results indicated that in vitro differentiated progeny of TET1 cells retained robust expression of rtTA2SM2 and EGFP for at least 2 weeks. However, these conditions might have favored the generation of a few, rather than multiple, differentiated cell types, and the analyzed time frame might not have been long enough to unmask a toxic effect associated with rtTA2SM2 expression during differentiation. Thus, we investigated expression of rtTA2SM2 for an extended period of differentiation using a teratoma model. Because palpable teratomas develop in a 2-month time frame, and on the other hand, these tumors are composed of multiple, semiorganized tissues, we reasoned that this in vivo differentiation assay could offer a more stringent approach to analyze retention of rtTA2SM2 expression during TET1 differentiation. Confocal immunofluorescence microscopy studies on tissue sections of 8-week-old TET1 teratomas revealed that highly differentiated tissues of different embryonic origin exhibited strong and homogenous expression of rtTA2SM2 and EGFP (Fig. 3E–3P). These tissues included glandular (Fig. 3H–3J), neural (Fig. 3K–3M), and cartilage (Fig. 3N–3P), which demonstrated that highly differentiated cell types and tissues expressing rtTA2SM2 could be derived from TET1 cells. Together, this indicated that a robust expression of rtTA2SM2 persisted for a significantly long period of time during in vitro and in vivo differentiation of TET1 cells.

Figure Figure 3..

Retention of rtTA2SM2 expression in highly differentiated progeny of TET1 cells. In vitro (A–D) and in vivo (E–P) differentiated TET1 cells. (A–C): Confocal microscopy studies on TET1 EBs. Antibody controls in supplemental online Figure 2. (D): Reverse transcription-polymerase chain reaction assay on RNA samples isolated from undifferentiated TET1 cells and TET1 EBs. Retention of EGFP and rtTA2SM2 expression and upregulation of tissue-specific markers typical of glia (GFAP), liver (AFP), and muscle (CTN1) in 2-week-old EBs derived from TET1 cells. Scale bar = 100 μm. (E–P): Ubiquitous expression of EGFP and rtTA2SM2 in TET1 teratomas. (E–G): Micrograph showing a whole section of a TET1-derived teratoma and representative tissues of different embryonic origin present on it (H–P), including glandular (H–J), neural (K–M), and cartilage (N–P). Scale bars = 100 μm (E–G) and 20 μm (H–P). Abbreviations: AFP, α-fetoprotein; CTN1, cardiac troponin; EGFP, enhanced green fluorescent protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GFAP, glial fibrillar acidic protein; Undiff., undifferentiated.

Conditional Regulation of a Stably Integrated Transgene in rtTA2SM2-Expressing Cells

To further examine the functionality of rtTA2SM2 after extended culture of TET1 cells, we then asked whether these cells were able to induce transgene expression from a stably integrated construct in a dox-dependent manner. To this end, we constructed pLTET-Luc (Fig. 4A). In this insulated lentivector, while the TRET promoter allows dox-dependent control of the luciferase gene, the EF1a promoter constitutively drives the expression of DsRedEX, a variant of red fluorescent protein that displays increased solubility and faster maturation [21]. Transduction of passage 10 TET1 cells (Fig. 2) with pLTET-Luc (MOI10) generated EGFP+/DsRedEX+ colonies (Fig. 4B–4E) that retained expression of rtTAS2M2, EGFP, and DsRedEX for extended culture (six passages), as accounted for immunofluorescence microscopy and RT-PCR analyses (Fig. 4B–4F). These data indicated stable integration of pLTET-Luc and functional regulation of DsRedEX from the EF1a promoter. The functionality of the TET-ON-responsive promoter was examined by analyzing the ability of TET1-Luc cells to induce luciferase expression upon addition of dox into the culture medium. Quantitative real-time PCR (QRT-PCR) studies indicated a dose-dependent induction of luciferase RNA transcripts that directly correlated with dox concentration (Fig. 4F). Dox-induced luciferase transcripts could be detected at a dose as low as 0.5 μg/ml, although maximum level of luciferase expression was observed at 2 μg/ml dox (Fig. 4F). Under this condition, dox-treated TET1-Luc cells displayed an average of 40-fold more luciferase RNA transcripts than untreated cells (Fig. 4F). This dox-inducible transcriptional regulation was also retained after a second cycle of exposure to dox (supplemental online Fig. 4), although the magnitude of the response was lower than that observed in single-exposure experiments, probably as a result of the continuous overexpression of luciferase upon temporally close and repeated cycles of dox. Altogether, these studies indicated conditional and reversible regulation of the luciferase transgene in TET1-Luc cells.

Figure Figure 4..

Conditional regulation of a stably integrated transgene in rtTA2SM2-expressing cells. (A): Schematic representation of pLTET-Luc lentivector used to derive TET1-Luc cells from TET1 cells. (B–E): Phase contrast and indirect immunofluorescence microscopy studies of passage 4 TET1-Luc cells. Scale bar = 50 μm. (F): Dox dose-dependent induction of luciferase RNA in TET1-Luc cells. Quantitative real-time reverse transcription-polymerase chain reaction analysis on RNA extracts isolated from these cells after 3 days of culture in presence (+) and absence (−) of dox; n = 3. (G): Luciferase assay. Protein extracts containing a similar number of cells were produced from the indicated cultures and assayed for luciferase activity; (+) and (−) indicate cultures grown in presence and absence of dox, respectively. Average and standard deviations for three independent experiments. (H–W): Confocal microscopy analysis of undifferentiated TET1-Luc cells (H–O) and EBs derived from them (P–W), grown in presence (Dox+ [H–K, P–S]) and absence (Dox− [L–O, T–W]) of dox during 3 days. Isotype controls in supplemental online Figure 3. Scale bar = 10 μm. Abbreviations: Dox, doxycycline; EF1a, elongation factor 1α; EGFP, enhanced green fluorescent protein; hESC, human embryonic stem cell; INS, insulator; LTR, long terminal repeat; ppT, polypurine track; SAR, scaffold attachment region; TRE-T, TET response element tight promoter.

In agreement with these studies, luciferase activity assays indicated dox-inducible expression of luciferase protein. Under the assayed conditions, dox-treated TET1-Luc cells displayed an average of 26-fold induction of luciferase activity compared with dox-untreated cells. Based on the QRT-PCR data, we believe that our luciferase activity study underestimated the magnitude of the response of TET1-Luc to dox. Nevertheless, in spite of the clear dox-dependent induction of RNA and luciferase activity (Fig. 4F, 4G), the luciferase values contrasted with those from transient transfection studies. This variation could be due to differences in molecular ratios for transactivator/responsive construct in transient versus stable transfection experiments, different rates of transcription, and translation of integrated versus nonintegrated transgenes, as well as the use of different luciferase reporter constructs in our transient and stable expression assays. To address these possibilities, we analyzed the number of integrated lentiviral copies in TET1-Luc cells. QRT-PCR analysis estimated average copy numbers per cell of 1.63 ± 0.3 and 1.88 ± 0.1 for LTET1 and LTRE-Luc, respectively (supplemental online Fig. 5). Thus, the dox-inducible response displayed by TET1-Luc resulted from a few integrated copies of LTET1 and LTRE-Luc, which highly contrasted with the transient transfection assay setting in which millions of transgene copies are present in the cells as episomes. Together, these results highlight the sensitivity and performance of this lentiviral system in hESC.

Lastly, the dox-inducible transgene regulation displayed by TET1-Luc was further characterized by analyzing protein expression of luciferase by immunofluorescence microscopy in undifferentiated and differentiated TET1-Luc cells. Both undifferentiated colonies and plated EBs derived from passage 6 TET1-Luc strongly induced luciferase protein in a dox-dependent manner 3 days after addition of dox (Fig. 4H–4W). Highly differentiated cells displaying flattened morphology, prominent cytoplasm, and elongated nuclei exhibited readily detectable expression of luciferase in dox-treated cultures (Fig. 4P–4S). These data correlated with RNA and luciferase activity studies and further supported functionality of the integrated pLTRE-Luc in both undifferentiated and differentiated TET1-Luc cells. These results demonstrated retention of functional rtTA2SM2 and conditional regulation of a stably integrated transgene for at least 16 passages in TET1 cells. In summary, our data indicate stable and reversible conditional transgene responses to dox for TET1-Luc. Together, our results demonstrate functionality of these insulated TET lentivectors in hESC.


Our study makes fundamental observations regarding the effect and functionality of rtTA2SM2 and an insulated lentiviral TET system in hESC. First, hESC constitutively expressing the TET-ON transactivator rtTA2SM2 can be derived and expanded in culture as undifferentiated pluripotent cells (Fig. 1) that retain dox-inducible transgene regulation (Figs. 1G, 4F–4O), normal karyotype (supplemental online Fig. 1), and multilineage differentiation potential (Figs. 2, 3). Second, expression of rtTA2SM2 is well tolerated and retained in both undifferentiated and highly differentiated cell types and tissues derived from these cells over an extended period of time (Fig. 3). Lastly, TET-ON components customized into insulted lentiviral systems display functionality in hESC (Figs. 1, Figure 2., Figure 3.4).

Successful conditional gene regulation in hESC has been recently described by other groups [3, [4]–5]. However, toxicity of stably expressed regulators has been reported for these cells. Vallier and colleagues [5] described toxicity for a stably expressed Cre-ER derivative upon its activation with hydroxytamoxifen. By tightly controlling exposure of these cell lines to this inducer, authors overcame this limitation. By following a different approach, Nolden and colleagues [3] developed a hESC model in which ectopic control of transgene expression was achieved by adenoviral delivery of Cre, preventing continued exposure of hESC to the recombinase. An all-in-one lentiviral TET-suppressor system has been characterized in a variety of cell models other than hESC [22]. However, Zhou and colleagues [4] have reported toxicity for a TET suppressor lentiviral system in hESC. Encouraging studies by Adachi and colleagues [23] demonstrated efficient performance of a nonviral TET-OFF system in both undifferentiated and differentiated nonhuman primate embryonic stem cells, suggesting that TET-OFF systems could also be applied to hESC models.

In our study, we have examined the effect of an improved TET-ON transactivator on both the derivation and propagation of undifferentiated transgenic hESC lines as well as on the pluripotency of these cells, which we consider fundamental aspects to be addressed before broad application of these systems in hESC. We then explored the functionality of a customized insulated lentiviral TET-ON system in these cells. Our results indicate that an insulated lentiviral TET-ON system is functional in hESC and offers unique features that can be exploited to study the biology of hESC. First, it does not interfere with the most distinctive biological properties of hESC. Undifferentiated transgenic hESC lines expressing rtTA2SM2 could be expanded in culture without loss of pluripotency, and rtTA2SM2-expressing tissues could be derived from them. Conditional manipulation of transgene expression was retained in undifferentiated and differentiated hESCs. Second, selection and expansion of TET regulatable hESC do not require either addition of dox to the medium [4] or ectopic induction of gene reporter expression [22, 23]. This minimizes toxic and uncontrolled differentiation effects during the isolation and expansion of transgenic cells. Third, a dual insulated system allows extra control of transgene response, in addition to dox. As shown in mouse and primate TET-ON models [6], it is possible that hESC tolerate different levels of both transgene induction and background expression (“leakiness”) in a gene- and context-dependent manner. Given the presence of insulator elements in these constructs, it could be possible to tune up/down these effects with minimum risk of insertional transactivation by increasing/decreasing the MOI of any or both TET lentivectors. In summary, our study indicates that improved TET regulators like rtTA2SM2 and TET insulated lentiviral systems are functional in hESC, and therefore they offer alternative conditional gene expression strategies for hESC models.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.


We acknowledge C. Andreu-Vieyra for insightful discussions and analysis of confocal microscopy data; A. Rosen for technical assistance; R. Hawley for pLSIN-EF1α-GFP-SAR/HS; F.J. Kern for pBTE; and T.P. Zwaka for comments on the manuscript. This work was supported by Grant 1R01DK075355 to M.A.G.; D.S.V. Received a Canadian Institutes of Health Research Postdoctoral Fellow. M.A.G was a Stohlman Scholar of the American Leukemia and Lymphoma Society.