Author contributions: N.P. and N.L.: conception and design, collection and assembly of data, data analysis and interpretation, and manuscript writing; M.A.: collection and assembly of data and data analysis and interpretation; S.K.: collection and assembly of data; C.B., M.N.A., M.G., and A.S.: provision of study material; T. Maetzig: collection and assembly of data and provision of study material; H.N.: manuscript writing; T.C. and T. Moritz: conception and design and manuscript writing.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS January 2013, 2013.
Methylation-induced gene silencing represents a major obstacle to efficient transgene expression in pluripotent cells and thereof derived tissues. As ubiquitous chromatin opening elements (UCOE) have been shown to prevent transgene silencing in cell lines and primary hematopoietic cells, we hypothesized a similar activity in pluripotent cells. This concept was investigated in the context of cytidine deaminase (CDD) gene transfer, an approach to render hematopoietic cells resistant to the chemotherapeutic agent Ara-C. When murine induced pluripotent stem cells (iPSC)/embryonic stem cells (ESCs) were transduced with self-inactivating lentiviral vectors using housekeeping (truncated elongation factor 1α; EFS) or viral (spleen focus-forming virus; SFFV) promoters, incorporation of an heterogeneous nuclear ribonucleoproteins A2 B1/chromobox protein homolog 3 locus-derived UCOE (A2UCOE) significantly increased transgene expression and Ara-C resistance and effectively prevented silencing of the SFFV-promoter. The EFS promoter showed relatively stable transgene expression in naïve iPSCs, but rapid transgene silencing was observed upon hematopoietic differentiation. When combined with the A2UCOE, however, the EFS promoter yielded stable transgene expression in 73% ± 6% of CD41+ hematopoietic progeny, markedly increased CDD expression levels, and significantly enhanced Ara-C resistance in clonogenic cells. Bisulfite sequencing revealed protection from differentiation-induced promoter CpG methylation to be associated with these effects. Similar transgene promoting activities of the A2UCOE were observed during murine neurogenic differentiation, in naïve human pluripotent cells, and during nondirected multilineage differentiation of these cells. Thus, our data provide strong evidence that UCOEs can efficiently prevent transgene silencing in iPS/ESCs and their differentiated progeny and thereby introduce a generalized concept to circumvent differentiation-induced transgene silencing during the generation of advanced iPSC/ESC-based gene and cell therapy products. STEM CELLS2013;31:488–499
Genetic modification of pluripotent stem cells such as embryonic stem cells (ESCs) or induced pluripotent stem cells (iPSCs) and their differentiation into specialized tissue prior to or after transplantation into human hosts holds substantial promise for the future of gene therapy and regenerative medicine [1, 2]. However, the implementation of such strategies has been slowed down by the silencing of transgenic promoter/enhancer elements in iPSC/ESCs and their differentiated progeny [3–8]. This problem has become particularly obvious for transgenes expressed from γ-retroviral or lentiviral vector constructs, which may be silenced by a number of epigenetic mechanisms such as DNA methylation or histone modification [9, 10]. Importantly, silencing of lentiviral transgenes has been observed not only in association with viral (spleen focus-forming virus, SFFV [6, 11]) but also with housekeeping (truncated elongation factor 1α, EFS ) and even tissue-specific (transthyretin ) promoter elements. Ubiquitous chromatin opening elements (UCOEs) have been identified as regulators of heterochromatin structure [13, 14] and specifically the A2UCOE derived from the heterogeneous nuclear ribonucleoproteins A2 B1/chromobox protein homolog 3 (HNRPA2B1-CBX3) housekeeping gene locus [13, 15] has been demonstrated to effectively negate CpG methylation-mediated transgene silencing of γ-retro- and lentiviral gene therapy vectors in chinese hamster ovary (CHO) cells or the hypermethylating teratocarcinoma cell line P19 [16, 17]. Furthermore, the A2UCOE has been used alone or in combination with suitable promoters to successfully promote transgene expression in primary hematopoietic cells following hematopoietic stem cell (HSC) gene transfer [15–18]. Therefore, we hypothesized that this element also might be used to prevent transgene silencing in pluripotent stem cells and thereof derived differentiated progeny.
We have investigated this hypothesis specifically during hematopoietic differentiation of ES/iPS cells and in the context of gene transfer of the drug resistance (CTX-R) gene cytidine deaminase (CDD). Transfer of CTX-R genes has been advocated to protect the hematopoietic system from the toxicity of anticancer chemotherapy, and a number of CTX-R genes suited for this purpose have been identified [19, 20]. CDD represents a clinically highly relevant CTX-R gene, as overexpression of this enzyme protects lymphohematopoietic cells from deoxycytidine analogs such as cytosine-arabinoside (1-β-D-arabinofuranosylcytosine, Ara-C), the most effective single-agent in the treatment of acute leukemias. Myeloprotection as well as chemoselection by hCDD gene transfer has been established in murine and human clonogenic progenitor cells [21–23] as well as in murine in vivo bone marrow transplant models [24–26]. However, in leukemias, HSC-based myeloprotective gene therapy approaches always carry the risk of inadvertent transduction of malignant cells. Hematopoietic differentiation of genetically modified iPSCs cells derived from nonleukemic sources such as fibroblasts has the potential to overcome this problem.
Thus, we here have evaluated the suitability of A2UCOE-containing self-inactivating (SIN) lentiviral vector constructs to stabilize transgenic hCDD expression in murine iPSCs and ESCs prior to and following hematopoietic differentiation. In contrast to primary HSCs, in which long-term transgene expression has been achieved with internal viral or housekeeping promoters, such as the SFFV  or EFS  promoter, in our studies, these elements on their own failed to promote stable transgene expression in iPSC/ESCs or their hematopoietically differentiated progeny. In contrast, incorporation of the A2UCOE sequence into the respective vector constructs allowed for robust and sustained hCDD expression in naïve pluripotent cells and their hematopoietic progeny and effectively protected these cells from Ara-C toxicity. Furthermore, a similar transgene promoting activity of the A2UCOE was observed during murine neurogenic differentiation, in naïve human pluripotent cells, and during the nondirected multilineage differentiation of these cells.
MATERIALS AND METHODS
The hCDD.IRES.dTomato cassette was inserted into third-generation SIN lentiviral vectors , by AgeI and SalI digestion. Next, the core 1.5 kb A2UCOE sequence  was inserted 5′ of the promoter element by ClaI digestion. Viral vector production was performed by four plasmid transfection as previously described . Briefly, 3.5 × 106 HEK293T cells were seeded 24 hours prior to transfection in 10 cm dishes and cultured in high-glucose Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Karlsruhe, Germany, www.invitrogen.com) with 10% heat-inactivated fetal calf serum (FCS), penicillin-streptomycin, and 1 mM L-glutamine. Cells were transfected with 5 μg lentiviral vector, 8 μg pcDNA3.GP.4xCTE (expressing HIV-1 gag/pol), 5 μg pRSV-Rev, and 2 μg pMD.G (encoding the vesicular stomatitis virus glycoprotein) by calcium phosphate precipitation in the presence of HEPES buffer and chloroquine. Supernatants were harvested 48 hours and 72 hours after transfection, filtered, and subsequently concentrated 50-fold by ultracentrifugation. Viral titers as determined by flow cytometry were in the range of 1–5 × 107 viral particles per milliliter.
Murine Pluripotent Cell Generation, Cultivation, and Transduction
The murine iPSC line iPSC-7 was generated from lineage negative bone marrow cells using monocistronic lentiviral vectors overexpressing Oct4, Sox2, Klf4, and c-Myc as previously described . Murine ESCs and iPSCs were cultured on mitomycin C-treated murine embryonic fibroblast feeder cells in ESC medium (knockout DMEM [Invitrogen], 15% ES-tested FCS, 1 mM L-glutamine, 0.1 mM nonessential amino acids [Invitrogen], 100 μM β-mercaptoethanol [Sigma-Aldrich, St. Louis, MO, www.sigmaaldrich.com], penicillin-streptomycin, and 103 units/ml leukemia inhibitory factor [LIF; kindly provided by the Institute of Technical Chemistry, Hannover Medical School, Hannover, Germany]). Murine ESCs and iPSCs were passaged every 2–3 days. For transduction, 5 × 104 iPSCs were seeded onto gelatinized 25 cm2 cell culture flasks and transduced with lentiviral vectors in the presence of 10 μg/ml protamine sulfate. To achieve comparable integration events for the different vector constructs, cells were transduced with equal multiplicities of infection (MOI, usually 10–20).
Hematopoietic Differentiation of Murine iPSC/ESC
Hematopoietic differentiation was carried out as previously described . In brief, iPSCs and ESCs were harvested and plated on gelatinized dishes in Iscove's Modified Dulbecco's Medium (IMDM) with 15% FCS, penicillin-streptomycin, 1 mM L-glutamine, 0.1 mM non-essential amino acids (NEAA), 150 μM monothioglycerol (MTG), and 103 units/ml LIF. After 48 hours, cells were harvested and seeded for embryoid body (EB) formation in suspension cultures containing IMDM (PAA, Pasching, Austria), 15% pretested FCS (ES-Cult, Stem Cell Technologies, Vancouver, Canada, www.stemcell.com), penicillin-streptomycin, 1 mM L-glutamine, 50 ng/ml ascorbic acid, and 150 μM MTG (all Sigma-Aldrich). On day 5 of differentiation, medium was changed and supplemented with 30 ng/ml murine stem cell factor (mSCF) and 10 ng/ml murine Interleukin-3 (mIl-3, Peprotech, Hamburg, Germany, www.peprotech.com). EBs were harvested on the desired day and dissociated using 250 U/ml collagenase in DMEM with 10% FCS. Flow cytometry analysis was carried out as described below. For clonogenic assays, 1 × 105 cells were seeded into standard methylcellulose cultures supplemented with insulin, transferrin, SCF, Il-3, Il-6, and erythropoietin (HSC007, R&D Systems, Minneapolis, MN, www.rndsystems.com). Drug resistance of colonies was assessed monitoring colony growth in the presence of increasing concentrations of Ara-C.
Neurogenic Differentiation of Murine iPSC
Neurogenic differentiation was carried out as described elsewhere , In brief, iPSCs underwent EB formation using the hanging-drop method for 3 days. Then, EBs were collected and transferred to DMEM F-12 with N2 and B27 supplements (Invitrogen), 1 mM glutamine, 1% NEAA, and 0.1 mM β-mercaptoethanol for 3 days. Thereafter, medium was additionally supplemented with 1 μM retinoic acid (Sigma-Aldrich) for 8–10 days. Subsequently, EBs were transferred onto Matrigel-coated chamber slides and immunocytochemistry for Tuj-1 (neuron-specific β-III-tubulin) was carried out.
Cultivation and Transduction of Human ESCs
Human ESCs (hES3) were cultured on irradiated human foreskin fibroblast feeder cells (kindly provided by S. Hartung, Leibniz Research Laboratories for Biotechnology and Artificial Organs [LEBAO], Hannover Medical School) in ESC medium (knockout DMEM, 20% knockout serum replacement, 1 mM L-glutamine, 0.1 mM nonessential amino acids (all Invitrogen), 100 μM β-mercaptoethanol (Sigma-Aldrich), penicillin-streptomycin, and 40 ng/ml basic fibroblast growth factor (kindly provided by the Institute of Technical Chemistry, Hannover Medical School). Cells were passaged using collagenase V (Invitrogen) once a week. For transduction, 1.5 × 105 ESCs were seeded as single cells onto Matrigel-coated 12-well plates in hESC-medium containing 10 μM ROCK inhibitor (Y-27632, Tocris, Bristol, U.K., www.tocris.com). The next day, cells were transduced with lentiviral vectors in the presence of 10 μg/ml protamine sulfate. To achieve comparable integration events for the different vector constructs, cells were transduced with equal MOI (usually 10–20).
Differentiation of Human ESCs
ESCs were harvested as single cells using Tryp-LE (Invitrogen) and transferred to suspension culture in IMDM with 20% FCS, penicillin-streptomycin, 1 mM L-glutamine, and 0.1 mM NEAA (all Invitrogen) for EB formation. EBs were harvested on the desired day and dissociated using collagenase V. Flow cytometry analysis was carried out as described below.
Alkaline Phosphatase Staining
Staining was performed using the Alkaline Phosphatase (AP) Detection Kit (Chemicon, Schwalbach/Ts., Germany, www.millipore.com) according to the manufacturer's instructions: cells were fixed with 4% paraformaldehyde, washed with Tris-buffered saline (with 0.1% Tween 20, TBST), and stained with AP staining solution.
Immunocytochemistry was performed using standard protocols. In brief, cells were fixed with methanol/acetone (50:50) and blocked with appropriate serum. Primary antibodies were incubated overnight (1:450); secondary antibodies were incubated at room temperature for 1 hour (1:1,000). Primary antibodies used: α-Tuj1 (Millipore, Billerica, MA), α-Sox2 (Invitrogen), and α-Oct4 (Santa Cruz Biotechnology, Santa Cruz, CA). Secondary antibodies used: goat α-mouse-Alexa-488 and goat α-mouse-Alexa 647 (all Jackson Immuno-Research Europe, Newmarket, U.K., www.jacksonimmuno.com).
Flow cytometric analysis was carried out as described before . In brief, cells were harvested and resuspended in fluorescence-activated cell sorting (FACS) buffer (phosphate-buffered saline (PBS) with 2% FCS and 2 mM EDTA). Cells were rinsed with FACS buffer and analyzed with a FACScalibur machine (Beckton & Dickinson, Heidelberg, Germany, www.bd.com). Antibodies used: α-SSEA-1-PE (R&D Systems), α-CD41-APC, α-Ter119-PeCy7, and α-Tra-1-60-Biotin. Streptavidin-APC was used as secondary reagent (all eBioscience, San Diego, CA, www.ebioscience.com).
Western Blot Analysis
Protein isolation was performed by cell lysis in RIPA buffer (Sigma-Aldrich) in the presence of protease inhibitors (Complete Mini, Roche, Grenzach-Wyhlen, Germany). Proteins were resolved by electrophoresis on 12% SDS-polyacrylamide gels. Proteins were transferred onto polyvinyliden flouride membranes (Hybond-P; GE-Healthcare, Waukesha, WI). After blocking in TBST containing 5% dried skimmed milk powder (TBST-milk) for 1 hour, membranes were incubated with primary antibodies (anti-Oct4: sc-5279; anti-Vinculin: V9131, Sigma-Aldrich; anti hCDD: Ab56053-100, AbCam, Cambridge, U.K.) in TBST-milk at 4°C overnight. Membranes were rinsed with TBST and incubated with the appropriate horseradish-peroxidase-coupled secondary antibodies (Jackson ImmunoResearch) at room temperature for 1 hour. Protein detection was performed with enhanced chemoluminescence reagents (SuperSignal West Femto/Pico, Thermo Scientific, Waltham, MA) and the ChemiDoc XRS+ station (BioRad, Hercules, CA).
Quantitative Reverse-Transcriptase PCR
For quantitative reverse-transcriptase PCR (qRT-PCR), total RNA was isolated using the GeneElute Mammalian Total RNA Kit (Sigma-Aldrich) followed by DNaseI digestion (Invitrogen) according to the manufacturer's instructions. RNA was reverse transcribed using avian myeloblastosis virus reverse transcriptase and oligo(dT) or random hexamer primers (Thermo Scientific). Subsequent quantification by qRT-PCR used 30–50 ng cDNA as input. TaqMan-based qRT-PCR was performed using TaqMan Universal Master Mix (Applied Biosystems, Carlsbad, CA, www.appliedbiosystems.com). The following predesigned assays were obtained from Applied Biosystems: mOct4 (Mm00658129_gH), mSox2 (Mm00488369), mNanog (Mm02384862_g1), mcMyc (Mm00487804_m1), mKlf4 (Mm01351375_g1), mActin beta (Mm00607939_s1), mGata2 (Mm00492301_m1), mRunx1 (Mm01213404_m1), and mScl/Tal1 (Mm01187033_m1). SYBR green qRT-PCR was performed using SYBR Green PCR Master Mix (Applied Biosystems) with the following predesigned Quantitect Primer Assays (Qiagen, , Hilden, Germany, www.qiagen.com): Brachyury (T, QT00094430), Kdr (Flk1, QT00069818), Gata4 (QT00031997), Sox17 (QT00204099), and Pax6 (QT00071169). For detection of hCDD, the following primers were used: Forward: GAA GCG TCC TGC CTG CAC CCT GA, reverse: GGT CCG TTC AGC ACA GAT G. Real-time PCR was performed on a StepOnePlus thermocycler (Applied Biosystems) with 40 cycles of 95°C for 15 seconds and 60° for 1 minute.
DNA Bisulfite Conversion and Sequencing
For promoter methylation analysis, genomic DNA was isolated using the GenElute Mammalian gDNA Kit (Sigma-Aldrich) and treated with bisulfite using the EpiTect Kit (Qiagen). Primers amplifying CpG islands within the respective promoter/enhancer region were designed using the MethPrimer software (http://www.urogene.org/methprimer/index1.html). Nested PCR amplified products were subcloned into pCR2.1 vectors using the TopoTA cloning Kit (Invitrogen). Individual colonies were picked, plasmid DNA was purified, and DNA was sequenced using M13 primers. Degree of methylated DNA was analyzed with the BiQ Analyzer Software . Primer sequences in 5′–3′ orientation: EFS_fwd_outer (GGG TAA ATT GGG AAA GTG ATG T); EFS_rev_outer: (TCA AAC TTC AAA ATA CAA ACA AAA C); EFS_fwd_inner (GGT AAA TTG GGA AAG TGA TGT); EFS_rev_inner (CAA ACT TCA AAA TAC AAA CAA AAC).
Quantitation of Vector Copy Numbers
Genomic DNA from UCOE.EFS.CDD and EFS.CDD transduced -7 iPSC was isolated at different time points using GenElute Mammalian Genomic DNA Miniprep Kit (Sigma-Aldrich). Quantitative real-time PCR was performed on a StepOnePlus light cycler (Applied Biosystems) using Fast SybrGreen reagent (Qiagen) and primers detecting dTomato as well as the polypyrimidine tract binding protein 2 (PTBP2) sequence, respectively. Normalization was performed using a plasmid standard harboring PTBP2 and dTomato sequence elements. Copy number calculations were performed by the Pfaffl method .
GraphPad Prism was applied to perform unpaired Student's t test or analysis of variance. Unless otherwise stated, SD is indicated. Asterisks mean: *, p < .05; **, p < .01; ***, p < .001.
In our initial experiments, we assessed the suitability of the A2UCOE to promote transgenic expression of hCDD in nondifferentiated pluripotent cells using the well-characterized murine iPSC line -7 (iPSC-7) established from bone marrow-derived hematopoietic progenitor cells as a representative model systems . Third-generation lentiviral vectors were used expressing the hCDD transgene from established housekeeping (EFS) or viral (SFFV) promoter constructs used either alone or in combination with the A2UCOE (Fig. 1A). To facilitate detection of transduced cells, the hCDD-cDNA was coupled to a dTomato marker sequence via an internal ribosomal entry site (IRES). While all constructs showed robust dTomato expression early after transduction (Fig. 1B), transgene levels varied considerably and were significantly enhanced in the presence of the A2UCOE (Fig. 1C). These differences in transgene expression were not explained by variations in transduction efficacy as similar numbers of integration events (on average 3.5 and 3.0 per cell for UCOE.EFS.CDD- and EFS.CDD-transduced iPSC-7 cells, respectively) were determined for the EFS-vectors by qRT-PCR using dTomato-specific primer sequences. Even more pronounced differences were observed when dTomato transgene expression was followed over time (Fig. 1D). Here, the EFS promoter on its own yielded relatively stable transgene expression with >60% of transgene-positive cells remaining after a 70 days (∼25 passages) time period, which were improved to >80% by combination with the A2UCOE. In contrast, dTomato expression was rapidly lost for the SFFV-construct when used alone, while prolonged and stable gene expression was reconstituted upon incorporation of the A2UCOE into the SFFV-vector.
A positive effect of the A2UCOE was also observed on hCDD transgene expression and markedly higher transgene levels were observed in UCOE.EFS.CDD compared to EFS.CDD transduced cells by qRT-PCR analysis (3.7 ± 2.2-fold, n = 3). These results were confirmed and extended by functional assays. While both constructs conferred significant Ara-C resistance to iPSC-7 cells (supporting information Fig. S1A) and allowed for efficient in vitro selection of transduced cells by Ara-C exposure (supporting information Fig. S1B), both, drug resistance as well as selection capability were significantly enhanced in UCOE.EFS.CDD transduced cells. This difference was also evident from the increase in hCDD transgene expression during the 24-day period of selection (supporting information Fig. S1C). Expression of the hCDD transgene did not interfere with the pluripotent character of the parental iPSC-7 cell line. For both constructs, UCOE.EFS.CDD and EFS.CDD, no influence on expression of the pluripotency-associated markers alkaline phosphatase and SSEA-1 or the pluripotency-associated transcription factors Oct4, Sox2, Nanog, or Klf4 was noted (Fig. 2A-2D).
Next, we investigated the effect of the A2UCOE on transgene expression during hematopoietic differentiation by comparing the EFS.CDD and UCOE.EFS.CDD vector constructs. Whereas expression from the EFS.CDD construct was rapidly lost with only 6.3% ± 0.7% and 1.4% ± 0.4% transgene-positive cells detectable at days 4 and 8, respectively, this was in sharp contrast to transduction of iPSC-7 cells with the UCOE.EFS.CDD construct. Here, only minimal loss of dTomato+ cells was observed during the initial days of the differentiation process, and 73% ± 6% of CD41+ cells still exhibited dTomato expression on day 8 (Fig. 3A, 3B). Of note, this effect was also observed in cells with a specific erythroid lineage commitment (Ter119+). Here, at day 8 of differentiation, only 4.4% of Ter119+ EFS.CDD transduced cells still expressed the dTomato reporter, whereas 38.3% of UCOE.EFS.CDD transduced cells were Ter119+/dTomato+ (supporting information Fig. S2A).
Protection from transgene silencing by the A2UCOE also was demonstrated for the hCDD transgene using Western blot and qRT-PCR analysis (Fig. 3C, 3D). Again, no or minimal hCDD expression was observed for the EFS.CDD construct from day 4 onward, whereas expression from the UCOE.EFS.CDD construct remained fairly constant throughout hematopoietic differentiation. Rapid loss of primitive Oct4-expressing cells during hematopoietic differentiation was confirmed by Western blot analysis (Fig. 3E). Differential expression of hCDD from the UCOE.EFS.CDD compared to the EFS.CDD construct was confirmed on a functional level, when CD41+ cells obtained at day 8 of the differentiation culture were seeded for clonogenic analysis in the presence of Ara-C. As depicted in Figure 3F, protection of UCOE.EFS.CDD transduced colonies from Ara-C-induced toxicity was significantly increased over EFS.CDD transduced controls. Of note, these experiments also demonstrated a significant degree of Ara-C resistance in EFS.CDD transduced colonies, most likely reflecting the residual hCDD expression levels observed by qRT-PCR analysis on days 8–16 of hematopoietic differentiation (above, Fig. 3D). To further confirm proper hematopoietic lineage specification, crucial hematopoietic transcription factors were analyzed in EBs at different time points during differentiation. Indeed, qRT-PCR revealed profound upregulation of Gata2, Runx1, and Scl/Tal1 throughout the entire differentiation process (supporting information Fig. S2B). Of note, transgenic expression of the CDD-IRES-dTomato cassette did not interfere with the hematopoietic differentiation potential of iPSC-7 cells (Fig. 3G, 3H). Thus, our studies gave no evidence for a negative impact of hCDD transgene (over)expression on pluripotent cells or their hematopoietically differentiated progeny.
Induction of resistance to the methylation of CpG residues within promoter/enhancer elements has been identified as a potential mode of action of UCOEs [15, 17]. To investigate, whether also in our hematopoietic differentiation model the inhibitory effect of the A2UCOE on transgene silencing was associated with a lack of promoter methylation, DNA bisulfite sequencing was applied to study the methylation pattern of the EFS promoter in pluripotent as well as hematopoietically differentiated iPSC-7 cells. A total of 14 CpG residues within a CpG island in the 3′ region of the EFS promoter and one CpG in the region encoding the hCDD transgene were interrogated (Fig. 4A). While only 4%–11% of the CpG residues was found to be methylated in the undifferentiated stage, upon hematopoietic differentiation nearly complete CpG-methylation was observed for the EFS.CDD construct as early as on day 4. In contrast, even at day 8 of differentiation, only moderate CpG-methylation was observed in the presence of the A2UCOE and 67% of CpGs were unmethylated at this time point (Fig. 4B).
The promoting effect of the A2UCOE on transgene expression in naïve and hematopoietically differentiated iPSC/ESCs was confirmed in CCE ESCs . In pluripotent CCE cells, transgene expression from the EFS.CDD as well as the UCOE.EFS.CDD construct was stable over a period of 30 days (supporting information Fig. S3A). Again higher expression of dTomato (MFIEFS.CDD = 98 vs. MFIUCOE.EFS.CDD = 117, mean of n = 2) as well as hCDD (supporting information Fig. S3B) was observed following transduction with the A2UCOE construct. Even more important, also significantly improved transgene expression from the UCOE.EFS.CDD versus the EFS.CDD construct was observed during hematopoietic differentiation, when cells expressing the dTomato transgene were compared at day 8 (supporting information Fig. S2C). A similar effect was observed for hCDD expression; while a 3.4-fold reduction of hCDD expression (relative to vinculin levels) was observed for EFS.CDD transduced cells at day 8, this was reduced to 1.8-fold by incorporation of the A2UCOE (supporting information Fig. S2B). Again, vector copy numbers were comparable for EFS.CDD and UCOE.EFS.CDD transduced cells with 4.8 and 3.1 integrations per genome, respectively. As for iPSC-7 cells, transgenic expression of hCDD had no influence on the hematopoietic differentiation of CCE ESCs to CD41+ cells (supporting information Fig. S2D) and rapid loss of primitive Oct4-expressing cells was observed during the differentiation process (supporting information Fig. S2E).
To investigate whether the transgene promoting activity of the A2UCOE during the differentiation of pluripotent cells represents a generalized concept and also applies to differentiation into other than the hematopoietic lineage, we subjected our murine iPSCs to a neurogenic differentiation protocol, in which neuronal fate specification is triggered by administration of retinoic acid . In line with the data from hematopoietic differentiation, transgene expression in EFS.CDD transduced cells was rapidly silenced within the first days of neuronal differentiation. After 6 days of differentiation, less than 5% of EFS.CDD transduced cells still expressed the dTomato reporter. In contrast, transgene expression was detected in >60% of UCOE.EFS.CDD transduced cells at this time point, further substantiating the concept that UCOEs can be used to effectively prevent differentiation-induced transgene silencing (Fig. 5A-5C). To ensure neurogenic differentiation, EFS.CDD and UCOE.EFS.CDD transduced iPSCs were stained for neuronal βIII-tubulin (Tuj1) at the end of differentiation, and both EFS.CDD and UCOE.EFS.CDD transduced cells unambiguously gave rise to Tuj1+ cells (supporting information Fig. S4).
In a clinical scenario, which involves transplantation of gene-corrected human iPSC/ESC progeny, the therapeutic transgene has to be expressed in the differentiated cell exerting the therapeutic effect. Thus, we next sought to investigate transgene expression in both naïve and differentiating human pluripotent stem cells. To this end, we transduced the human ESC line hES3  with our lentiviral EFS.CDD and UCOE.EFS.CDD constructs (Fig. 6A). In undifferentiated hES3 cells, a relatively moderate but consistent loss of transgene expression was observed in EFS.CDD transduced cells with less than 20% of cells remaining dTomato+ after 26 days, whereas >80% of cells expressed the dTomato marker in the UCOE.EFS.CDD transduced population at this time point (Fig. 6B). When hES3 cells (sorted for dTomato expression) were subjected to an EB-based multilineage differentiation, rapid loss of EFS-driven transgene expression was observed. After 13 days of differentiation, only ∼20% of EFS.CDD transduced cells still expressed the dTomato reporter, whereas the A2UCOE effectively prevented differentiation-induced silencing with >80% of the UCOE.EFS.CDD transduced cells remaining dTomato+ (Fig. 6C). To prove differentiation into all three germ layers and to exclude a potential impact of the transgene on differentiation propensities, we harvested EBs on day 8 of differentiation and carried out a qRT-PCR for markers of the three germ layers. As depicted in Figure 6D, untransduced, EFS.CDD, and UCOE.EFS.CDD transduced cells equally gave rise to mesodermal (T and Flk1), endodermal (Gata4 and Sox17), and ectodermal (Pax6) markers.
Using a defined 1.5 kb genetic element derived from the human HNRPA2B1/CBX3 locus, we here describe markedly improved transgene expression from SIN lentiviral vector constructs using internal housekeeping and viral promoters in undifferentiated murine and human pluripotent cells. Even more importantly, we demonstrate almost complete abrogation of silencing for transgenes expressed from the EFS promoter during the hematopoietic differentiation of murine iPSCs and ESCs and during undirected, multilineage differentiation of human ESCs. This transgene promoting activity was associated with protection from methylation at distinct CpG residues within the EFS promoter during the early phase of iPSC/ESC differentiation. In this respect, our data reflect the properties demonstrated for the A2UCOE when applied in combination with viral promoters such as the SFFV or cytomegalovirus promoter in the context of gene transfer into CHO or P19 teratocarcinoma cells [16, 17]. Also in these cells, protection from transgene silencing by the A2UCOE was at least in part dependent on the protection from promoter CpG methylation.
While the exact molecular mechanisms by which UCOEs promote an open chromatin structure and prevent transgene silencing are only partially understood, a two component model has been suggested to explain UCOE function. According to this model, UCOEs combine a central region (approx. 1 kb) of methylation-free CpG-islands associated with active histone modification marks and two closely spaced divergent promoters resulting in bidirectional transcription . For the native genomic A2UCOE, a further extended (>5 kb) region of methylation-free DNA associated with a pattern of active histone modification markers has been demonstrated . The ability to bestow this epigenetic signature onto transgenic promoter sequences at the site of vector integration probably plays a major role in the expression promoting properties of UCOEs and has been described for the A2UCOE when used in combination with the cytomegalovirus or SFFV-promoter [16, 17]. While our data certainly support the two component model described above as an explanation for the transgene-stabilizing effect of the A2UCOE during hematopoietic differentiation, it has to be noted that UCOE function to a substantial degree is context dependent and reflects the viral or cellular elements used for transgene expression as well as the specific target cell investigated. This is evident from recent work performed in the context of HSC gene therapy for chronic granulomatous disease using the myeloid-specific myeloid related protein 8 (MRP8) promoter . Here “read through” transgene expression from the A2UCOE-inherent CBX3 promoter in the presence of CpG-methylation of the MRP8 promoter is introduced as an alternative mechanism by which UCOEs stabilize transgene expression. However, these mechanisms are not mutually exclusive and both are associated with highly reliable transgene expression from individual vector integration and reduced position effect variegation [17, 18].
In our differentiation model promoter methylation-associated transgene silencing was of rapid onset and more or less completed within 3–4 days, prior to the specific induction of neurogenic or hematopoietic differentiation by cytokine administration. Therefore, transgene silencing appears to be associated with escape from the pluripotent state rather than with definitive lineage commitment, a notion that is supported by our data linking the transgene promoting effect of the UCOE to progressive loss of Oct4 expression. Given this “early” onset of the A2UCOE-mediated antisilencing effect during hematopoietic as well as neurogenic fate specification, it appears likely that similar properties of the A2UCOE may also be observed for differentiation along other lineages. Even more important, differentiation-induced transgene silencing and its prevention by the A2UCOE is not a phenomenon specific for murine pluripotent cells as similar properties were observed in human ESCs and differentiated progeny thereof. Thus, our observation bears considerable relevance to the clinical translation of pluripotent stem cell research, as it offers a generalized concept to circumvent differentiation-induced transgene silencing during the generation of advanced ESC/iPSC-based gene and cell therapy products.
With regard to safety considerations, it has to be stressed that the A2UCOE lacks classic enhancer function . Thus, it should also be devoid of the ability to promote long-range transcriptional activation from heterologous promoters, which has been identified as the basis of insertion-mediated, genotoxic events in multiple gene therapy trials [40, 41]. Nevertheless, the A2UCOE has the potential to disturb neighboring gene function by divergent transcription and aberrant splicing extending from within the viral vector into surrounding gene-containing DNA sequences. While this read-through transcription may result in activating or inhibiting effects on endogenous gene function, it may be abrogated by the introduction of mutations into the respective splice donor sites and inclusion of appropriate miRNA target sites at the ends of the A2UCOE . Furthermore, as with other insertional effects, this should not constitute serious problems in the context of ESC/iPSC-based gene therapy approaches, as here, due to the clonal nature of the gene modified cells, extensive safety studies including sequencing of the viral integration sites, and exclusion of clones with potentially risky integrations prior to final (hematopoietic) differentiation should be feasible. Therefore, this approach should match the safety of zinc finger or transcription activator-like effector nuclease nuclease-based genome editing [43–45] but at substantially reduced effort and costs. Nevertheless, differences between individual cells or cell clones with regard to the UCOE effect on transgene expression and silencing were observed in our studies. Thus, site-specific integration systems like TALE or zinc finger nucleases and cre- or flip-recombinase-mediated cassette exchange technology [46, 47] might be exploited to investigate UCOE-mediated effects on epigenetic modifications at a given locus.
We here introduce hematopoietic differentiation of genetically modified iPSCs or ESCs as a novel approach to render the hematopoietic system resistant to chemotherapeutic agents and to allow for efficient selection of genetically modified cells. The efficacy of this strategy was evaluated in the context of hCDD gene transfer, which renders hematopoietic cells resistant to Ara-C, the most important single-agent in the treatment of acute leukemias and other hematological malignancies. To this point, we demonstrated significantly increased Ara-C resistance in clonogenic hematopoietic progenitor cells derived from hCDD-overexpressing iPSCs as well as highly efficient selection of hCDD-expressing iPSCs. This novel strategy appears particularly suited for malignancies manifested in blood or bone marrow compartments such as leukemias, lymphomas, myelodysplasias, or myeloproliferative diseases as in these diseases Ara-C represents the most effective single-agent for treatment, while HSC-based myeloprotective gene therapy approaches in these diseases are hampered by the risk of inadvertent transduction of malignant cells. Hematopoietic differentiation of genetically modified iPSCs cells has the potential to overcome this problem as these cells can be generated in a patient-specific manner from nonmalignant cell sources such as fibroblasts.
As with other pluripotent stem cell-based gene therapy strategies targeting the hematopoietic system, the myeloprotective approach investigated here ultimately will require the generation of HSCs with in vivo long-term reconstituting potential (LT-HSCs), an aim that despite promising early data  has so far not been established as a reproducible procedure. Nevertheless, technological progress in recent years has considerably facilitated the in vitro generation of hematopoietic progenitor cells of various lineage commitments [49–51], and with foreseeable further advancement in this field, the generation of bona fide LT-HSCs from pluripotent cell sources appears as an achievable goal.
Taken together, we have for the first time established the transgene promoting properties of the A2UCOE in murine and human pluripotent cell populations and their differentiated progeny. This finding appears of high relevance for future iPSC/ESC-based cell or gene therapy approaches depending on stable transgene expression. In addition, promoter methylation-associated transgene silencing in our studies was of rapid onset and completed prior to the specific induction of hematopoietic or neurogenic differentiation. Thus, the transgene-stabilizing properties of the A2UCOE most likely also can be exploited for differentiation into other tissues. In this respect, our model may serve as a generalized concept to stabilize transgene expression in ESC/iPSC-derived transgenic cell therapy products.
We thank Doreen Lüttge (Hannover Medical School) for excellent technical assistance. We are grateful to Susann Hartung and Robert Zweigerdt from the LEBAO (Hannover Medical School) for providing us with the hES3 cells used in this study. This work was supported by grants from the Deutsche Forschungsgemeinschaft: Cluster of Excellence REBIRTH (Exc 62/1) and SPP1230: Grant MO 886/4–1 (T.M.) and GR 898/5-2 (M.G.).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
MNA is inventor on patents covering the biotechnological applications of UCOEs. All other authors declare no conflict of interest.