|Home | About | Journals | Submit | Contact Us | Français|
Induced pluripotent stem (iPS) cells have been derived from fibroblast, stomach, and liver cultures at extremely low frequencies by ectopic expression of the transcription factors Oct4, Sox2, c-myc, and Klf4, a process coined direct or in vitro reprogramming [1–8]. iPS cells are molecularly and functionally highly similar to embryonic stem cells (ESCs), including their ability to contribute to all tissues as well as the germline in mice. The heterogeneity of the starting cell populations and the low efficiency of reprogramming suggested that a rare cell type, such as an adult stem cell, might be the cell of origin for iPS cells and that differentiated cells are refractory to reprogramming. Here, we used inducible lentiviruses  to express Oct4, Sox2, c-myc, and Klf4 in pancreatic β cells to assess whether a defined terminally differentiated cell type remains amenable to reprogramming. Genetically marked β cells gave rise to iPS cells that expressed pluripotency markers, formed teratomas, and contributed to cell types of all germ layers in chimeric animals. Our results provide genetic proof that terminally differentiated cells can be reprogrammed into pluripotent cells, suggesting that in vitro reprogramming is not restricted to certain cell types or differentiation stages.
Pancreatic β cells are mature, fully differentiated cells, whose defining characteristic is the expression of insulin. In vivo lineage-tracing studies have demonstrated that in healthy adult mice, the β cell population is maintained by self-duplication, not an adult stem cell . Upon injury, insulin-producing β cells are also produced from facultative endocrine progenitors . Importantly, these progenitors do not express insulin. Moreover, insulin-expressing β cells do not contribute to any other cell type in vivo . Because of their easily defined identity and stable cell fate, pancreatic β cells are an ideal cell type with which to test whether iPS cells can be derived from a terminally differentiated cell type.
We first tested whether β cells can be cultured under iPS cell culture conditions. To this end, we explanted pancreatic islets from 3- to 4-month-old mice that expressed GFP under the control of the Pdx1 promoter . Pdx1 expression in the postnatal pancreas is confined to β cells, in which it regulates insulin expression . As shown in Figure 1, most islet cells were GFP+ and maintained GFP expression in culture for at least 12 days. Rare GFP− fibroblast-like cells appeared after ~1 week (Figures 1D and 1E). Most of these cells are probably derived from the pancreatic mesenchyme , whereas rare cells may also originate from β cells that have dedifferentiated in culture, as previously observed . On the basis of the expansion of islets in culture, we estimated that β cells divided once before arresting. Incubation with a lentivirus constitutively expressing tdTomato showed that roughly 50% of GFP+ islet cells (148 of 279 counted cells) became infected, compared with 80% of adult fibroblasts (209 of 261 cells), indicating that cultured islet cells can be transduced with lentivirus, albeit at a lower efficiency than fibroblasts (Figures 1F and 1G and data not shown).
To genetically mark β cells in the adult, we crossed RIP-Cre mice, in which the Cre gene is controlled by the β cell-specific rat insulin promoter , with ROSA26-lacZ reporter mice (Figure 2A). Immunostaining of pancreas sections showed that lacZ expression was restricted to insulin+ cells contained within β islets, thus confirming the specificity of the transgene and excluding the possibility that non-βcells had been labeled (Figures 2B and 2C) . On the basis of this observation, we conclude that most, if not all, cells with an active rat insulin promoter in the adult pancreas correspond to differentiated β cells.
Pancreatic islets from RIP-Cre/lacZ mice were isolated, explanted in culture for 2 days, and infected with doxycycline-inducible lentiviruses expressing Oct4, Sox2, c-myc, and Klf4 as well as a lentivirus constitutively expressing the reverse tetracycline-dependent transactivator (rtTA). Eighteen to twenty-four days after adding doxycycline to the cultures, colonies emerged that were visually different from pancreatic islets and resembled iPS colonies derived from fibroblasts (Figure 3B). The delayed kinetics with which iPS colonies appeared from pancreatic cultures compared with fibroblast cultures  may reflect the slow replication rate of islet cells. A total of 24 colonies were picked at day 24, and ~80% (19/24) of the colonies could be expanded in the absence of doxycycline, demonstrating independence of viral gene expression and activation of the endogenous pluripotency program, as has been seen previously [9, 16].
To verify the cellular origin of iPS cells, we stained the 19 expanded clones with the β-galactosidase substrate 5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside (Xgal). Four of the clones (21%) stained homogenously positive with Xgal, demonstrating that they were derived from insulin-expressing β cells (Figure 2C), whereas all other clones were entirely Xgal−. This number is significantly lower than the labeling efficiency of β cells with the RIP-Cre transgene (>80%), suggesting that Xgal− cells within explanted islets may be more easily reprogrammed or more efficiently infected by lentiviruses compared with β cells. Alternatively, the ROSA26 locus may have been silenced in some of the Xgal− iPS cell clones. iPS cells derived from β cells will be referred to as β iPSs.
We estimated the reprogramming efficiency to range between 0.1% and 0.2% (for calculation, see Experimental Procedures), which is similar to that seen in fibroblast cultures with the same lentiviral system . To rule out the possibility that the RIP-Cre transgene becomes activated during the reprogramming process, we generated iPS cells from tail fibroblasts of RIP-Cre/lacZ mice in which the insulin promoter is silenced. Out of more than 50 iPS clones analyzed, none stained positive for Xgal, thus confirming the tightness of the RIP-Cre/lacZ system (data not shown).
All four β iPS clones expressed the pluripotency markers SSEA1, Nanog, Sox2, and Oct4 at the protein level (Figures 3D–3F and data not shown), indicating faithful molecular reprogramming. Real-time quantitative PCR analysis showed that β iPSs expressed endogenous c-myc, Klf4, Oct4, and Sox2 at levels comparable to ESCs, whereas viral gene products were essentially undetectable (Figure S1 available online).
To test the differentiation potential of β iPS cells, we injected the four cell lines into the flanks of SCID mice. All β iPS lines formed teratomas, which on histological examination showed differentiated cell types representative of all three germ layers, including glandular structures, muscle fibers, keratinized epithelia, and cartilage (Figures 4A–4E). Next, we injected three of the four iPS clones into blastocysts to assess their potential to contribute to normal development. All clones gave rise to E14.5 fetal chimeras and neonatal animals that were highly chimeric on the basis of whole-mount staining for lacZ activity, and cross sections of individual organs showed widespread tissue contribution (Figures 4F–4J and data not shown). One surviving chimera showed obvious coat-color chimerism but was cannibalized by its littermates at three weeks of age (data not shown). Together, these analyses demonstrate that the β iPS cells were reprogrammed to a pluripotent state.
Our findings allow three major conclusions. First, using genetic-lineage analysis, we demonstrate that cells from a defined cellular lineage can be reprogrammed into iPS cells that broadly contribute to tissues in chimeras. Second, our data show that reprogramming is not only confined to mesodermal derivatives, which have been used previously, but is also possible with an endodermally derived cell type. This is in agreement with a recent report showing reprogramming of stomach and liver cells into iPS cells . Third, terminal differentiation, at least in the β cell lineage, does not restrict a cell’s potential to be reprogrammed by the four reprogramming factors c-myc, Klf4, Oct4, and Sox2. In a previous report, iPS cells have been derived from genetically marked, albumin-expressing liver cells . Because albumin is also expressed in hepatic progenitor and stem cells [17–19], this experiment did not unequiovocally demonstrate the reprogramming of differentiated cells.
Our data indicate that the consistently low efficiency of iPS cell derivation from somatic cells is unlikely to reflect the reprogramming of rare stem or progenitor cells present in the starting cell population, suggesting that reprogramming is a universal process that is not restricted to certain cell types or differentiation stages. This does not exclude the possibility, however, that adult stem and progenitor cells are more amenable to in vitro reprogramming than differentiated cells. In agreement with this, pro B cells have been reported to be more easily reprogrammed into iPS cells than mature B cells, which required additional genetic manipulation of the B-cell-specific transcription factor Pax5 . We believe that stochastic epigenetic-remodeling events are necessary for successful reprogramming, which results in the low overall reprogramming efficiencies. The requirement for such remodeling events or the frequency at which they occur might be cell type specific. Thus, it is conceivable that the genetic or pharmacologic manipulation of epigenetic modifiers may significantly increase reprogramming efficiencies in different cellular contexts.
Derivation and handling of Pdx1-GFP, RIP-Cre, and R26R-lacZ mice were described previously. We crossed RIP-Cre mice with R26R-lacZ mice to obtain RIP-Cre/lacZ animals. Genotyping was done by PCR with the following oligonucleotides for RIP-Cre: 5′-TAGCACCAGGCAAGTGTTTG-3′ and 5′-ATGTTTAGCTGGCCCAAATG-3′. Genotyping for ROSA26 was done as described .
The pancreas was perfused through the bile duct with digestion solution (low-glucose DMEM [GIBCO] with 10 mM HEPES [GIBCO], 0.25 mg/ml Liberase RI [Roche], and 0.1 mg/ml ovalbumin trypsin inhibitor [Roche]), dissected, and incubated at 37°C for 20 min. Cold washing solution (low-glucose DMEM with 10 mM HEPES, 10% FBS [Hyclone], and 0.1 mg/ml OTI) was added, and islets were centrifuged, washed twice, and filtered through a 500 μm diameter wire mesh. Islets were centrifuged, washed twice in washing solution, resuspended in Histopaque 1077 (Sigma), and vortexed. The islet suspension was carefully overlaid with washing solution (without serum) and centrifuged for 20 min at 10°C, separating islets from exocrine tissue. The islet layer was collected at the interface, pelleted, and washed twice. Finally, pure islets were handpicked under a dissecting scope, pelleted, washed, and cultured in ES medium on laminin coated plates.
The number of islets that attached 1 day after seeding was determined to be ~500 islets per 35 mm plate. On the basis of an average number of 100 β cells per islet (K.B. and D. Melton, unpublished data), and a labeling efficiency of 80% in RIP-Cre mice , we calculated the total number of Xgal+ cells present in the cultures to be ~40,000. Because ~50% of cells were infected by individual lentiviruses (Figure 1), we calculated the number of Xgal+ β cells infected by all four viruses to be 2500 (40,000 × 0.54). Out of 2500 cells, we obtained four Xgal+ iPS clones at an efficiency of 0.16%.
iPS cells were harvested by trypsinization, preplated onto untreated culture plates for removal of feeders as well as differentiating cells, and injected into the flanks of NOD/SCID mice, with ~5 million cells per injection. Mice were sacrificed 3 weeks later, and teratomas were isolated and processed for histological analysis.
Female BDF1 mice were superovulated with PMS and hCG and mated to BDF1 stud males. Zygotes were isolated from females with a vaginal plug 24 hr after hCG injection. After 3 days of in vitro culture in KSOM media, blastocysts were identified, injected with iPS cells, and transferred into pseudo-pregnant recipient females. Pups were delivered by Cesarean section at day 19.5 and nurtured by foster mothers.
Cultured cells were fixed with 0.2% glutaraldehyde for 15 min and incubated in X-gal staining solution (0.1 M phosphate buffer with 2 mM MgCl2, 5 mM potassium ferrocyanide, and 5 mM potassium ferricyanide containing 1 mg/ml X-gal [5-bromo-4-chloro-3-indolyl-β-D-galactosidase]) for 12–16 hr at 37°C. Teratomas and newborn mice were fixed overnight in 1.5% paraformaldehyde (PFA) and washed with PBS before incubation in X-gal staining solution. For analysis of tissue chimerism, PFA-fixed newborn mice were equilibrated in 30% sucrose solution and frozen in OCT compound. Ten micrometer cryosections were prepared and incubated in X-gal staining solution overnight.
iPS cells were cultured on pretreated coverslips, fixed with 4% PFA, and permeabilized with 0.5% Triton X-100. The cells were then stained with primary antibodies against mOct4 (Santa Cruz, sc-8628), mSox2 (Chemicon, AB5603), and mNanog (Abcam, ab21603); this was followed by staining with the respective secondary antibodies conjugated to Alexa Fluor 546 (Invitrogen). Nuclei were counterstained with DAPI (Invitrogen). Cells were imaged with a Leica DMI4000B inverted fluorescence microscope equipped with a Leica DFC350FX camera. Images were processed and analyzed with Adobe Photoshop software.
Cells were harvested and incubated with either an antibody against SSEA-1 (MC-480, Developmental Hybridoma Bank) or an IgM isotype control, subsequently incubated with APC-conjugated mouse anti-mouse IgM antibody (II/41, eBiosciences), and analyzed on LSR2 (BD Biosciences). Dead cells were excluded by staining with DAPI. Data were analyzed with FlowJo software (Tree Star).
The generation and structure of tet-inducible lentiviruses expressing c-myc, Klf4, Sox2, and Oct4 has been described in detail elsewhere . In brief, a tet-inducible lentivirus termed LV-tetO was generated by replacement of the ubiquitin promoter elements from the FUΔGW vector  with tetO sequences. cDNAs for c-MYC (T58A mutant), Klf4, Oct4, and Sox2 were isolated from pMIG or pMX vectors and cloned into LV-tetO. The lentivirus constitutively expressing tdTomato was a kind gift of Dr. Niels Geijsen, MGH. To produce infectious viral particles, we transfected 293T cells cultured on 10 cm dishes with the LV-tetO vectors (11 μg) together with the packaging plasmids VSV-G (5.5 μg) and Δ8.9 (8.25 μg) using Fugene (ROCHE) transfection reagent. Viral supernatants were harvested on 3 consecutive days starting 24 hr after transfection, yielding a total of ~30 ml of supernatant per virus. Viral supernatant was concentrated ~100-fold by ultracentrifugation at 20,000 rpm for 1.5 hr at 4°C and resuspension in 300 μl PBS. Viral concentrates were stored at −80°C. Infections were carried out in 1 ml ES medium containing 5 μg/ml polybrene with 5–10 μl of each viral concentrate per 35 mm plate. Fibroblasts were infected at passages 1 or 2 at a density of ~200,000 cells/plate. Primary islets cells were infected 2 days after seeding them at a density of ~500 islets/plate. The medium was replaced 24 hr after infection and supplemented with 1 μg/ml doxycycline another 24–48 hr later. Irradiated feeders cells were added to the pancreatic cultures ~10 hr after viral induction.
RNA was isolated from cells with the TriPure reagent (Roche), and this was followed by RNA clean up with the RNeasy Minikit (QIAGEN). cDNA was produced with the First Strand cDNA Synthesis Kit (Roche). Real-time quantitative PCR reactions were set up in triplicates with the Brilliant II SYBR Green QPCR Master Mix (Stratagene) and run on an Mx3000P QPCR System (Stratagene). Primer sequences are listed in Table S1.
We thank Doug Melton, Chad Cowan, Nabeel Bardeesy, Yuval Dor, and members of the Hochedlinger lab for critical review of the manuscript. We also thank Nimet Maherali for qPCR reagents and Rita Martinez for technical support with animal husbandry. M.S. was supported by a postdoctoral fellowship from the Schering Foundation. Support to K.B. was from the Howard Hughes Medical Institute and an award from S. Kurtzig. K.H. was supported by the NIH Director’s Innovator Award, the Harvard Stem Cell Institute, the V Foundation, and the Kimmel Foundation.
One figure and one table are available at http://www.current-biology.com/cgi/content/full/18/12/890/DC1/.