|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: HH NK. Performed the experiments: HH NK-K PK NK. Analyzed the data: HH NK-K MF NK. Contributed reagents/materials/analysis tools: MF. Wrote the paper: HH NK.
A major difficulty of producing induced pluripotent stem cells (iPSCs) has been the low efficiency of reprogramming differentiated cells into pluripotent cells. We previously showed that 5% of mouse embryonic fibroblasts (MEFs) were reprogrammed into iPSCs when they were transduced with a fusion gene composed of Oct4 and the transactivation domain of MyoD (called M3O), along with Sox2, Klf4 and c-Myc (SKM). In addition, M3O facilitated chromatin remodeling of pluripotency genes in the majority of transduced MEFs, including cells that did not become iPSCs. These observations suggested the possibility that more than 5% of cells had acquired the ability to become iPSCs given more favorable culture conditions. Here, we raised the efficiency of making mouse iPSCs with M3O-SKM to 26% by culturing transduced cells at low density in serum-free culture medium. In contrast, the efficiency increased from 0.1% to only 2% with the combination of wild-type Oct4 and SKM (OSKM) under the same culture condition. For human iPSCs, M3O-SKM achieved 7% efficiency under a similar serum-free culture condition, in comparison to 1% efficiency with OSKM. This study highlights the power of combining the transactivation domain of MyoD with a favorable culture environment.
iPSC technology demonstrates that differentiated cells can be dedifferentiated to a pluripotent state by introducing only a few genes, most commonly the OSKM genes , , . However, protocols for generating iPSCs are inefficient and slow. Generally less than 1% of transduced cells are reprogrammed to form iPSCs with the standard OSKM transgenes. This is partly because exogenous Oct4 and Sox2 proteins cannot gain access to their target DNA sequences, perhaps due to closed chromatin structure . We recently showed that a fusion gene, called M3O, between Oct4 and the transactivation domain (TAD) derived from MyoD, a master regulator of skeletal myogenesis, reprograms 5% of MEFs into iPSCs when combined with polycistronic SKM , . Importantly, M3O-SKM remodeled chromatin at the Oct4 and Sox2 genes to an embryonic stem cell (ESC)-like state in the majority of MEFs, including those that failed to become iPSCs, as evaluated by binding of Oct4 and Sox2 proteins and histone modification patterns at the two genes. These findings indicated that M3O-SKM increased the number of completed iPSCs as well as the background population of partially reprogrammed MEFs. Therefore, we speculated that if we could identify an additional cue that drives the partially reprogrammed population toward completed iPSCs, the efficiency of making iPSCs would drastically increase. The current study addressed the question of whether the culture environment could serve as the additional cue by changing the density of transduced cells and culture media.
We previously used feeder-free and feeder-plus protocols to make iPSCs with M3O-SKM . With the feeder-free protocol, 3.6% of MEFs transduced with M3O-SKM were reprogrammed into iPSCs in a culture medium that contained 15% fetal bovine serum (FBS). In the present study, we replaced FBS with 20% KnockOut Serum Replacement (KSR), which is optimized for maintenance of ESCs in an undifferentiated state, although its contents have not been disclosed (Fig. 1A, Protocol A). KSR has also been reported to be more effective in making iPSCs than FBS . In addition, we added the FLAG sequence to the amino terminus of M3O to create the FM3O construct (Fig. 1B). An anti-FLAG antibody was used to detect the protein with immunofluorescence staining because commercially available Oct4 antibodies recognized M3O only if the protein was denatured. The FLAG sequence did not significantly affect the efficiency of making iPSCs. The FLAG sequence was also attached to Oct4 to prepare FO as a control. The co-transduction efficiency of FM3O with SKM and FO with SKM was 87.5±3.2% and 89.2±2.3%, respectively (Fig. 1C). iPSC formation was monitored by emergence of ESC-like colonies expressing the Oct4-driven GFP transgene as previously described . GFP-positive colonies first appeared around day 4 with FM3O-SKM, and many more colonies arose over the next two days with KSR compared to FBS (Fig. 1D, top and middle). Colony formation with FM3O-SKM was rapid and highly efficient with KSR, leading to colony fusion and making it impossible to accurately count colonies. In contrast, colony formation with FO-SKM and KSR was late (day 7) and inefficient, with less than 0.1% of MEFs becoming iPSCs by day 10 (Fig. 1D, bottom).
To obtain accurate colony numbers, MEFs were seeded at various lower densities on feeder cells in the presence of KSR (Fig. 2A, Protocol B). Colonies that emerged on day 5 with FM3O-SKM were positive for GFP and the relatively specific pluripotency markers SSEA-1 and alkaline phosphatase (Fig. 2B). This study indicated that seeding transduced MEFs at 2,000 cells/well of a 12-well plate was most effective in making iPSCs (Fig. 2C). At this cell density, the number of GFP-positive colonies steadily increased over ten days, reaching a peak around day 12, when 25.8% of MEFs were reprogrammed to iPSCs with FM3O-SKM (Fig. 2D). In contrast, FO-SKM induced only 2.3% of cells to become iPSCs by day 13. Taking transduction efficiency into account, the net efficiency of making iPSCs was essentially unchanged at 29.5% (25.8/0.875) for FM3O-SKM and 2.6% (2.3/0.892) for FO-SKM (Fig. 2C). The current efficiency obtained with FM3O-SKM was substantially higher than the previous efficiency of 5.1%, which was obtained at seeding density of 50,000 cells/well in the presence of FBS . When FBS was used at cell density of 2,000 cells/well, less than 2% of MEFs were reprogrammed to iPSCs with FM3O-SKM (data not shown). Thus, KSR and low cell density were the two major factors responsible for the significant improvement of iPSC production in the current study in comparison to our previous work . Mechanistic analysis of the effects of cell density on reprogramming efficiency awaits further investigation.
The pluripotency of iPSCs prepared with FM3O-SKM using Protocol B were evaluated with a standard set of experiments. The iPSC colonies expressed nine key genes for pluripotency at similar levels to those in ESCs (Fig. 3A). Three genes highly expressed in MEFs — Thy1, Col6a2 and Fgf7 — were downregulated to ESC levels in iPSCs (Fig. 3B). When iPSCs were subcutaneously injected into NOD/SCID mice, they formed teratomas that contained tissues derived from all three germ layers (Fig. 3C). Furthermore, iPSCs formed chimeric mice upon aggregation with 8-cell-stage ICR mouse embryos (Fig. 3D). When one of the chimeric mice was mated with an ICR female mouse, all six pups had a brown coat color, proving that the derivatives of iPSCs contributed to germ cells (Fig. 3E).
To understand if the highly efficient generation of iPSCs using FM3O-SKM and Protocol B was specific to MEFs, we isolated myoblasts from adult Oct4-GFP mice (Fig. 4A) and prepared iPSCs from these cells (Fig. 4B). The co-transduction efficiency was 92.8±3.4% for FM3O-SKM and 99.0±1.5% for FO-SKM (Fig. 4C). M3O-SKM was slower and less efficient at inducing myoblasts into iPSCs (13.2%, net efficiency was 14.2%, Fig. 4D) as compared with MEFs; nonetheless, it was substantially more efficient than our previous result obtained with OSKM in the presence of KSR and feeder cells (0.0025%) .
Because a fusion gene between Oct4 and a fragment of the VP16 TAD (amino acids 446–490) also increased the efficiency of making iPSCs , we fused the same fragment and the full-length VP16 to Oct4 (called VP16SO and VP16LO, respectively, Fig. 5A). While these two VP16 fusion genes were indeed much more efficient than wild-type Oct4, they were still less efficient than M3O (Fig. 5B). This difference was not due to a difference in viral titers (Fig. 5C). We also compared the Oct4-fusion genes without c-Myc because it causes tumor formation in vivo. Although the efficiency significantly dropped for all these fusion genes, M3O-SK still induced iPSCs in 3.9% of MEFs (Fig. 5D), which was more efficient than transduction of wild-type Oct4 in the presence of c-Myc (FO-SKM, 2.3%, Fig. 2C).
Previously we showed that M3O-SKM induces iPSC formation from adult human dermal fibroblasts at an efficiency of 0.30% in feeder-free culture with KSR when seeded at 17,000 cells/well of a 12-well plate, in contrast to 0.0052% with OSKM . Following the success of Protocol B as described above, we seeded transduced human neonatal fibroblasts at densities ranging from 1,000 to 17,000 cells/well of a 12-well plate on feeder cells (Fig. 6A). The number of ESC-like colonies that expressed NANOG and TRA1-60 were counted as iPSC colonies (Fig. 6B). When transduced fibroblasts were seeded on feeder cells at 17,000 cells/well, the efficiency of iPSC formation was raised to 1.7%. However, as with mouse iPSCs, the highest efficiency was obtained at 2,000 cells/well. At this cell density, iPSC colonies emerged around day 5 with M3O-SKM and day 8 with OSKM (Fig. 6C, KSR). The colony number peaked on day 14 with 6.6% of fibroblasts becoming iPSCs with M3O-SKM, in contrast to 0.8% on day 16 with OSKM. Two additional serum-free media were less effective than KSR (Fig. 6C, Pluriton and TeSR1). Cloned iPSCs prepared with M3O-SKM and KSR also expressed other pluripotency markers, including SSEA-4, alkaline phosphatase, NANOG and endogenous OCT4 on day 27 (Fig. 7A). Pluripotency of an iPSC clone established with M3O-SKM and KSR was verified with teratoma formation (Fig. 7B). In these experiments we transduced M3O, OCT4, SOX2, KLF4 and c-MYC in separate viruses. Since KLF4 and c-MYC were endogenously expressed in fibroblasts , transduction efficiency of these genes could not be determined but transduction efficiency of SOX2 was sufficiently high (87.0% with M3O-SKM and 88.9% with O-SKM) (Fig. 7C).
In conclusion, the present study shows that the efficiency of making iPSCs can be drastically raised by changing the culture medium and the density at which transduced cells are seeded on feeder cells. A few other studies have achieved comparable efficiency of making mouse and human iPSCs. For instance, pre-B cells and myeloid progenitor cells isolated from iPSC-derived mice were converted to secondary iPSCs at an efficiency of 92% and 27.5%, respectively, after activation of the OSKM transgenes , . Another study demonstrated around a 10% efficiency of preparing mouse and human primary iPSCs (i.e. not from iPSC-derived differentiated cells) using the miR302/367 cluster . However, to the best of our knowledge, the present study achieved the highest efficiency of making primary mouse iPSCs with an OSKM variant. Furthermore, the difference in efficiency between FM3O-SKM (25.8%) and FO-SKM (2.3%) underscores the importance of the transactivation capacity of Oct4 as a major limiting factor in iPSC formation.
All animal experiments were approved by the Institutional Animal Care and Use Committee at the University of Minnesota (1002A78174).
The FLAG sequence DYKDDDDK was added to the amino terminus of M3O and Oct4 to create the FM3O and FO constructs in the pMXs-IP vector . Full-length (amino acids 411–490) and the second half (amino acids 446–490) of the VP16 TAD were fused to the amino terminus of the mouse Oct4 gene in the pMXs-IP vector to prepare the VP16LO and VP16SO constructs, respectively. Plat-E cells  were seeded at 5×105 cells/3.5 cm dish on day −4. Four micrograms of the pMXs-IP vectors encoding FO, FM3O, VP16LO, VP16SO and polycistronic SKM were separately transfected into Plat-E cells with 10 µl Lipofectamine 2000 (Invitrogen) on day −3 to prepare virus supernatant. Virus titer was measured using a Retro-X qRT-PCR titration kit (Clontech). The medium was replaced with 1.5 ml DMEM with 10% FBS on day −2. MEFs prepared from Oct4-GFP transgenic mice  were seeded at 5×104 cells/well of a 12-well plate on day −2 in DMEM with 10% FBS. Virus supernatant was harvested on day −1 and day 0 and filtered through a 0.45 µm syringe filter. Two hundred and fifty microliters of each virus supernatant was added to each well of MEFs with 10 µg/ml polybrene. In Protocol A, culture medium was changed to iPSC medium (DMEM, 20% KSR (Invitrogen), 100 µM MEM non-essential amino acids, 100 µM 2-mercaptoethanol, 2 mM L-glutamine and 1000 u/ml leukemia inhibitory factor) on day 1, and cells were maintained without subculture. KSR was replaced with 15% FBS only in the experiment shown in the middle set of Fig. 1D. Protocol B was used to measure the efficiency of making iPSCs. For this protocol, transduced MEFs were subcultured onto irradiated MEFs at 1,000–4,000 cells/well of a 12-well plate in iPSC medium containing KSR on day 1. The efficiency of making iPSCs was calculated each day thereafter by dividing the number of GFP-positive colonies by the seeded cell number. Net efficiency was obtained by dividing the above-mentioned efficiency of making iPSCs by the efficiency of viral co-transduction obtained with immunofluorescence staining of transduced cells.
Satellite cell-derived myoblasts were isolated from the hind limb skeletal muscle of one-month-old adult Oct4-GFP transgenic mice as previously described . Myoblasts were cultured on collagen-coated dishes in myoblast growth medium (HAM's F-10 medium supplemented with 20% FBS and 5 ng/ml basic fibroblast growth factor (R&D Systems)) and used to make iPSCs with Protocol B.
The human M3O fusion gene was composed of the M3 domain of human MYOD fused to the amino terminus of full-length human OCT4 cDNA . Human M3O, OCT4, SOX2, KLF4 and c-MYC (Addgene) were separately inserted into the pMXs-IP vector and transfected into Plat-A cells (Cell Biolabs) with Lipofectamin 2000 (Invitrogen) on day −3. On day −2, 2.7×104 foreskin fibroblasts obtained from a newborn boy (Lonza) were plated in a well of a 12-well plate in DMEM with 10% FBS. Virus supernatant was harvested from Plat-A cells on day −1 and day 0, filtered through a 0.45 µm syringe filter, and transduced each day into fibroblasts with polybrane. Transduced cells were harvested on day 3 with trypsin and subcultured separately in three different media at 2×103 cells/well onto irradiated MEFs in 12-well plates. The first medium (KSR medium) was composed of KnockOut DMEM/F-12 (Invitrogen), 20% KSR, 100 µM MEM non-essential amino acids, 1% insulin-transferrin-selenium (Invitrogen), 0.1 mM 2-mercaptoethanol, 1× GlutaMAX (Invitrogen) and 4 ng/ml basic FGF. Pluriton™ Reprogramming Medium (Stemgent) and mTeSR®1 Medium (STEMCELL Technologies) were prepared following the instructions provided by the manufacturers and used without additional components. Culture media were changed every other day.
iPSCs were fixed with 4% formaldehyde in phosphate buffered saline for 10 min and treated with 0.5% Triton X-100 for 3 min to permeabilize the plasma membrane. Cells were then incubated with primary and secondary antibodies for 1 hr each at 25°C using the antibodies listed in Table S1. Hoechst 33342 was used to counterstain DNA. Fluorescence photographs were taken with an AxioCam charge coupled device camera attached to an Axiovert 200 M fluorescence microscope (all from Carl Zeiss) equipped with a 10× A-Plan Ph1 Var1 objective (numerical aperture 0.25). Images were processed with Photoshop 7.0 (Adobe Systems). iPSCs were also stained with an Alkaline Phosphatase Detection Kit (Millipore SCR004) following the manufacturer's instructions.
iPSC colonies were individually harvested, and cDNA was directly prepared from cells with a Cells-to-cDNA II kit (Ambion). qRT-PCR was performed on a Realplex 2S system (Eppendorf) with a GoTaq qPCR Master mix (Promega). Primers are listed in Table S2. The mRNA level of each gene was normalized using the mRNA encoding glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The feeder-free ES cell line CGR8 (European Collection of Cell Cultures) was used as positive control.
Eight to 10 iPSCs were briefly exposed to acidic Tyrode's solution (Millipore) and transferred into a microdrop of KSOMaa solution (Millipore) that contained a zona-free 8-cell-stage mouse embryo of the ICR strain (albino). On the next day morula stage embryos that contained GFP-positive iPSCs were transferred into the uteri of 2.5 days post-coitum pseudopregnant recipient mice to obtain chimeric mice. Teratomas were induced by subcutaneously injecting one million iPSCs into NOD/SCID mice. After four weeks teratomas were isolated, fixed with 10% formalin, and embedded in paraffin. Histological sections (5 µm thick) were stained with haematoxylin and eosin to identify tissue types.
Antibodies used for immunofluorescence staining.
Primers used for quantitative RT-PCR.
We thank Michael Franklin for critical reading of the manuscript; Kimberly Stelzig and Kristin Voltzke for technical support with iPSC culture; and Toshio Kitamura for Plat-E cells and the pMXs-IP retroviral vector. We also thank the University of Minnesota Comparative Pathology Shared Resource for histological preparation of teratomas.
Competing Interests: The authors have declared that no competing interests exist.
Funding: Histological preparation was supported by the comprehensive Masonic Cancer Center NIH Grant (P30 CA077598-09). M.F. and N.K. were supported by Richard M. Schulze Family Foundation. N.K. was supported by a grant from the National Institutes of Health (R01 DK082430). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.