CHAP TER 2 6
Induced Pluripotent Stem Cells Keisuke Okita* and Shinya Yamanaka*,† *
Center for iPS Cell Research and Application (CiRA), Institute for Integrated Cell-Material Sciences, Kyoto University, Kyoto, Japan, †Department of Stem Cell Biology, Institute for Frontier Medical Sciences, Kyoto University, Kyoto, Yamanaka iPS Cell Special Project, Japan Science and Technology Agency, Kawaguchi, Japan, and Gladstone Institute of Cardiovascular Disease, San Francisco, CA, USA
26.1 GENERATION OF iPS CELLS 26.1.1 Reprogramming Factors iPS cells are established by the forced expression of several transgenes. The classic mixture is Oct3/4, Sox2, Klf4, and c-Myc. This mixture can reprogram somatic cells of the mouse, human, rat, monkey, and dog. All of these factors have transcriptional activity, and Oct3/4, Sox2, and Klf4 regulate many ES cellspecific genes in combination. These factors also regulate their own expression. There are families of genes for Oct3/4, Sox2, and Klf4, and some of them can induce iPS cells. For example, Sox2 can be replaced with Sox1, Sox3, Sox7, Sox15, Sox17, or Sox18, and Klf4 with Klf2. Comparing the target genes among reprogramming factors and the family genes might be useful for understanding the molecular mechanisms underlying iPS cell formation. Other combinations, such as Oct3/4, Sox2, Nanog, and Lin28, have been reported for the generation of human iPS cells. Nanog is one of the most important transcription factors for stabilizing the pluripotent state in mouse ES cells. It also makes a transcriptional circuit with Oct3/4, Sox2, and Klf4. Oct3/4, Sox2, and Nanog bind and upregulate ES cell-specific genes such as STAT3 and ZIC3 with RNA polymerase II. On the other hand, they also localize to developmental regulator genes, such as PAX6 and ATBF1, with SUZ12, where they work as suppressors. The forced expression of some core components of ES cells would induce ES cell-like transcription networks in somatic cells and change their state. c-Myc is associated with many aspects of reprogramming, but its precise function is unclear. The process of iPS induction is thought to have some stochastic events dependent on cell proliferation, such as passive DNA demethylation. The expression of c-Myc blocks cell senescence, accelerates proliferation of fibroblasts, and leads to enhancement of iPS induction. c-Myc binds to more than 4,000 sites of the R. Lanza & A. Atala (Eds): Essentials of Stem Cell Biology, Third edition. DOI: http://dx.doi.org/10.1016/B978-0-12-409503-8.00026-3 © 2014 Elsevier Inc. All rights reserved.
Chapter 26: Induced Pluripotent Stem Cells
genome; therefore, it could loosen tightly packed chromosomes in somatic cells and increase the accessibility of other transcription factors to the genome during iPS induction. Overexpression of c-Myc itself also shifts the gene expression profile of mouse embryonic fibroblasts (MEFs) towards pluripotent cells. LIN28 is an RNA-binding protein and negatively regulates Let7 microRNA (miRNA) families. Let7 promotes differentiation of breast cancer cells and inhibits their proliferation. Therefore, LIN28 seems to indirectly enhance reprogramming efficiency through Let7 families. A combination of extra factors used in the induction can improve the reprogramming efficiency and quality. The addition of transcription factors, such as ESRRB26, UTF127, and SALL428, increased the efficiency. All of these factors are expressed in ES cells, and are involved in the formation of an ES-like transcriptional network. Tbx3 significantly improves the quality and the germ-line competency of mouse iPS cells. Some variations of inducing factors have been reported for iPS generation. The factor(s) in the reprogramming cocktail can be reduced if the somatic cells have sufficient endogenous expression of either of the reprogramming factor(s). For example, neural precursor cells express endogenous SOX2, KLF4, and c-MYC, and they only need OCT3/4 transgenes for iPS cell induction. The acceleration of cell proliferation and the inhibition of senescence by the suppression of the p53 and p21 pathways can also dramatically increase the efficiency. An increase in the number of cells under induction results in high iPS colony formation because the reprogramming process includes stochastic events. Suppression of p53 increases reprogramming efficiency predominantly through acceleration of cell division. On the other hand, the addition of Nanog to the reprogramming factor upregulated the net reprogramming efficiency, in a cell-division-independent manner. However, the suppression of the p53 and p21 pathways increases the genomic instability of iPS cells. Therefore, the permanent suppression of the pathway should be avoided because it would lower the quality of iPS cells. The transient suppression of inhibitors or siRNAs could be useful for the enhancement of reprogramming. iPS cell induction takes at least one week in the mouse and two weeks in humans. On the other hand, reprogramming by fusion of ES cells occurs very rapidly. The activation of endogenous Oct3/4 promoter of somatic cell nuclei is observed within two days. Although transgene expression in iPS cells requires a few days after vector transduction, the reprogramming of iPS cells seems to take much more time than that of cell fusion. ES cells must have other factor(s) that facilitate the reprogramming. Reprogramming events occur naturally in vivo during early developmental stages. The fertilized eggs erase almost all epigenetic status except imprinting before blastocyst formation, and they rebuild it as differentiation proceeds. The eggs also have high reprogramming activity, since they can produce a cloned animal after enucleation and fusion with somatic cells. Although the mechanism remains elusive, cloning might provide
26.1 Generation of iPS Cells
helpful hints for improving the generation of iPS cells. However, cloned mice have some abnormalities, such as a large placenta and a tendency to gain excess weight. There may be some limitation in the artificial reprogramming that must be considered.
26.1.2 Transduction Methods iPS cells were originally established by the delivery of transgenes by MMLV (Moloney murine leukemia virus)-based retroviral vectors. A retrovirus can robustly infect mouse fibroblasts and introduce its RNA genome into the host genome by reverse transcriptase. Therefore, the iPS cells integrate numerous transgenes, which thereby enable constant transgene expression. The inactivation of the retroviral promoter by DNA methylation is observed in ES cells as well as in iPS cells. Therefore, the expression of retroviral transgenes is gradually suppressed during the reprogramming process, and the silencing is complete when the cells become iPS cells. This automatic silencing mechanism is thought to provide effective reprogramming in somatic cells. However, the exogenous sequences remain in the genome of iPS cells and the alteration of genomic organization could induce some abnormalities. In particular, c-Myc, one of the reprogramming factors, is a proto-oncogene, and its reactivation could give rise to transgene-derived tumor formation. There have been improvements in the transduction methods for making safe iPS cells. Elimination of the c-Myc transgene for iPS cell induction is one important approach. Human and mouse iPS cells can be established from fibroblasts with only Oct3/4, Sox2, and Klf4, although the efficiency is significantly lower. Mouse iPS cells without c-Myc do not show enhanced tumor formation during the observation period (six months) in comparison to control mice. Another approach is to reduce the number of integration sites by attaching the reprogramming factors with internal ribosome entry sequences (IRES) or 2A self-cleavage peptide and putting them into a single vector. This reprogramming cassette was used with a lentivirus system containing a loxP sequence and produced iPS cells with only single insertions. The expression of Cre recombinase successfully cuts out the cassette, although a truncated long terminal repeat (LTR) remains in the iPS genome. The elimination of transgenes from the genome avoids the leaky expression of reprogramming factors, and improves the gene expression profile and the differentiation potential of iPS cells. A transposon system has also been used for iPS induction. A plasmidbased transposon vector with a reprogramming cassette can integrate into host genome with transposase. The re-expression of the transposase after establishment of iPS cells recognizes the terminal repeat of integrated transposon vector, and excises it from the genome. The excision of the transposon does not leave a footprint in most cases, so it maintains the original endogenous sequences. Non-integration methods were also reported with viral vectors
Chapter 26: Induced Pluripotent Stem Cells
(adenovirus and sendaivirus), DNA vectors (plasmid, episomal plasmid, and minicircle vector), and direct protein delivery. Although the induction efficiency of iPS cells with these methods is still low, they could become future standard methods.
188.8.131.52 Culture Conditions and Cell Signaling Culture conditions and cell signaling have a great influence on iPS generation. iPS cells are cultured in medium optimized for ES cells. Leukemia inhibitory factor and basic fibroblast growth factor are important for mouse and human ES cell maintenance, respectively. However, the roles of these cytokines in the induction process are still unclear. Wnt signaling supports the self-renewal of ES cells. The Wnt3a signal is mediated by glycogen synthase kinase (GSK) 3-β. Without the Wnt signal, GSK3-β inactivates target genes, such as β-catenin and c-Myc, by phosphorylation and proteasome-mediated degradation. Hence, the inhibition of GSK3-β with a chemical drug, such as CHIR99021, results in activation of Wnt signaling. Addition of Wnt3a or CHIR99021 enhances the reprogramming efficiency. Kenpaullone is an inhibitor whose targets are GSK3-β as well as cyclin-dependent kinase (CDK)s and can replace Kruppellike factor 4 (Klf4) in reprogramming induction from MEF with Oct3/4, Sox2, and c-Myc. Although more specific GSK3-β inhibitors, such as CHIR99021, or CDK inhibitor, purvalanol A, were unable to generate mouse iPS cells with the same combination of transcription factors and Kenpaullone itself did not increase endogenous Klf4 expression, the function of Kenpaullone is still elusive. Importantly, Li et al. found that the combination of CHIR99021 and Parnate, an inhibitor of lysine-specific demethylase 1, can generate iPS cells from human primary keratinocytes with only Oct3/4 and Klf4. The addition of vitamin C enhances iPS cell generation from both mouse and human somatic cells. Vitamin C works at least in part by alleviating cell senescence. O2 tension is also an important factor for stem cell maintenance and differentiation. For instance, low O2 tension promotes the survival of neural crest cells and hematopoietic stem cells, and prevents differentiation of human ES cells. Up to fourfold enhancement of the reprogramming efficiency is observed when the iPS induction is performed in hypoxic conditions (5% O2), in both mouse and human fibroblasts.
26.1.3 Cell Source iPS cells were first established from primary mouse fibroblast culture. Their origin was thought to be some tissue stem cells included in the culture since the efficiency of iPS cell induction was very low (less than 0.1%). Mouse iPS cells can be established from mouse hepatocytes and stomach epithelial cells and linage tracing experiments showed that most hepatocyte-derived iPS cells were from albumin-positive cells. Mouse iPS cells were also established
26.2 Molecular Mechanisms in iPS Cell Induction
from pancreatic islet β cells. Therefore, the origin of iPS cells is not only tissue stem cells but also differentiated somatic cells. Human iPS cells have been established from various tissues, including fibroblasts (adult and embryo), adult keratinocyte, adipose tissue, peripheral blood, cord blood, amniotic fluid-derived cells, and neural precursor cells. Hence, all somatic cells are thought to have the ability to yield iPS cells, although they show differential efficiency. However, it is unclear whether iPS cells from different cell sources have the same potential. Mouse iPS cells derived by the current reprogramming method from different tissues apparently have divergent characteristics. Miura et al. compared neural differentiation potential and safety of mouse iPS cells derived from MEF, tail-tip fibroblasts (TTFs), and hepatocytes. Most iPS clones form a neural sphere under in vitro-directed differentiation conditions. The neural sphere contains neural precursor cells that can produce three neuronal cell types; neuron, astrocyte, and oligodendrocyte. The neurosphere from ES cells could contribute the neural tissues when transplanted into the mouse brain. However, the neurospheres prepared from TTF-derived iPS cells tended to form teratomas after transplantation into mouse brain. Teratoma formation has been reported in the transplantation of neurospheres formed from ES cells containing undifferentiated cells that remained after the differentiation process. The population of undifferentiated cells is rare in neurospheres from MEF-derived iPS cells and ES cells, but is obvious in those from TTF-derived iPS cells (up to 20%). The study revealed that the existence of undifferentiated cells varies depending on the cell source. Accessibility to a cell source is another important point in the selection of tissues, especially for induction of human iPS cells. Human iPS cells can be established from neural precursor cells with only OCT3/4 transgenes; however, constant acquisition of the neural tissue is difficult.
26.2 MOLECULAR MECHANISMS IN iPS CELL INDUCTION 26.2.1 Epigenetics The generation of iPS cells includes epigenetic alterations. DNA methylation status and histone modifications of promoter regions including Nanog, Oct3/4, Sox2, and Fbxo15 achieve an ES-like state after reprogramming. The addition of a histone deacetylase (HDAC) inhibitor, valproic acid (VPA), improves the reprogramming efficiency in both mouse and human fibroblasts. Other HDAC inhibitors, such as suberoylanilide hydroxamic acid and trichostatin A, also work in mouse fibroblasts. Inhibitors of DNA methyltransferase, such as 5′-azacytidine and RG108, and BIX-01294 for G9a histone methyltransferase increased reprogramming efficiency. These results supported the hypothesis that the process of iPS generation involves epigenetic changes. Some of the
Chapter 26: Induced Pluripotent Stem Cells
inhibitors could abolish the use of one or two reprogramming factor(s). For example, VPA treatment of human fibroblasts enables reprogramming with only two factors, Oct4 and Sox2, and eliminates the oncogenic c-Myc or Klf4. However, it is doubtful whether these drugs fill in the exact function of reprogramming genes; rather, they seemed to enhance the induction efficiency that allows the reduction of reprogramming factor(s). iPS induction requires the establishment of an ES-like transcription factor circuit in somatic cells. In fact, iPS cells have the same expression profile as ES cells; however, they have differences in epigenetic modifications, especially in genes not involved in pluripotency. Cell differentiation is a process of limitation of the differentiation potential by epigenetic modification. Each type of somatic cell has its specific epigenetics by which cells are able to stabilize their state. The forced expression of reprogramming factors can affect several downstream genes in somatic cells and alter their epigenetic modifications. However, it is difficult to think that the factors control all genes throughout the genome. In fact, genome-wide analysis showed similar DNA methylation patterns of iPS and ES cells, but they also detected differentially methylated regions between iPS and ES cells. The uncontrolled genes would keep their epigenetic profiles even in iPS cells. This could influence the differentiation potential of iPS cells. For example, the methylation status of the enhancer binding site in the astrocyte gene, GFAP, controls the differentiation fate of neuronal precursor by changing the binding activity for an enhancer, STAT3. Without such methy lation they instead tended to become astrocytes, whereas in the presence of methylation they tended to demonstrate neuronal differentiation.
26.2.2 MicroRNAs miRNAs are small single-stranded RNAs (around 22 nt) that directly interact with target mRNAs through complementary base-pairing and inhibit the expression of the target genes. miRNAs also work at the transcriptional level. miRNAs are generated as long RNA sequences and are digested to the short mature form by Dicer. miRNAs are involved in many features of cell properties, such as proliferation, apoptosis, and differentiation, by fine-tuning gene expression. ES cells have the characteristic expression of miRNAs, and iPS cells also showed a similar expression profile. Over 70% of mRNAs in mouse ES cells are the miR-290 cluster, which contributes to the ES cell-specific rapid cell cycle progression. The cluster includes miR-291-3p, miR-292-3p, miR-293, miR-294, and miR-295. miR-291-3p, miR-294, or miR-295 increases the reprogramming efficiency from MEF with Oct4, Sox2, and Klf4. They appear to be downstream targets of c-Myc, because the miRNAs did not enhance reprogramming efficiency in the presence of c-Myc transgene, and c-Myc binds the promoter region of the cluster. The three miRNAs share a conserved seed sequence, which mainly specifies target genes, suggesting they work through common
26.3 Recapitulation of Disease Ontology and Drug Screening
targets. LIN28 is a negative regulator of Let7 miRNA families. Lin28 induced the uridylation of immature let7 RNA by a non-canonical poly (A) polymerase, TUTase4, and this leads to degradation of the RNA. Lin28 gradually decreases during ES cell differentiation, and mature let7 family miRNAs accumulate with inverse correlation. The addition of Lin28 enhances the reprogramming efficiency from both human and mouse fibroblasts. A detailed analysis showed that Lin28 accelerates the reprogramming efficiency in a cell cycle-dependent manner. This is consistent with the concept that the targets of mature let7 include oncogenic genes, such as K-Ras and c-Myc. Lin28 facilitates the expression of Oct4 at the post-transcriptional level by direct binding to its mRNA.
26.3 RECAPITULATION OF DISEASE ONTOLOGY AND DRUG SCREENING Patient-derived iPS cells are useful in understanding disease ontology. The iPS cells have the same genomic information as the patient. Many iPS cells have been established from somatic cells obtained from patients with adenosine deaminase deficiency-related severe combined immunodeficiency, Duchenne and Becker muscular dystrophy, and amyotrophic lateral sclerosis. Ebert et al. established iPS cells from skin fibroblast of a spinal muscular atrophy (SMA) patient. SMA is an autosomal recessive genetic disorder that is characterized by degeneration of motor neurons following progressive muscular atrophy. The most common cause of SMA is a mutation of the survival motor neuron 1 (SMN1) gene, and it significantly reduces the level of protein expression. The motor neurons generated from the patient’s iPS cells can recapitulate the disease ontology, as they show reduced survival motor neuron (SMN) expression in comparison to those derived from the child’s unaffected mother. Treatment with VPA or tobramycin increases SMN expression. Importantly, the same treatment also worked in the motor neurons prepared from the patient’s iPS cells. The results indicate that iPS cells could provide a useful screening system for the identification of a specific, effective drug from thousands of candidate compounds. Such patient-derived iPS cells could also be used to find developing drugs that would be harmful to the human body. Some compounds work on target tissue, but have severe side effects. Long QT syndrome is an inborn heart defect that shows characteristic prolongation of the QT interval on electrocardiogram, increases the risk of irregular heartbeat, and threatens life. It occurs only after drug administration in some individuals. Cardiomyocytes established from patients with long QT syndrome via iPS cells could therefore be used to identify any possible toxic side effect of candidate compounds before starting clinical trials.
Chapter 26: Induced Pluripotent Stem Cells
Most diseases do not have a simple cause; they are the total sum of genetic/ epigenetic issues, environment, aging, etc., in a complicated relationship between several cell types in the body. It is therefore necessary to establish a way to recapitulate late-onset disease and environmental effects in vitro or in an animal model.
26.4 iPS CELL BANKING It will require time to establish a clinical grade of useful cells from a patient’s own somatic cells. The applicability and safety of each cell type must be assessed. The clinical applications of iPS cells must also be considered from an economic point of view. Complete tailor-made iPS cell therapy would cost too much to apply to a large number of people. Therefore, a banking system should be established for iPS cells. iPS cells having various human leukocyte antigen (HLA) haplotypes should be collected to avoid immune rejection. Experience with organ transplantation has revealed that the HLA class I molecules, HLA-A and HLA-B, and the class II molecule, HLA-DR, are the most important HLA molecules to match. Therefore, the HLA matching of these loci reduces the incidence of acute rejection and improves transplant survival. Estimations of stem cell bank size have been calculated in Japanese and UK populations. The random establishment of 170 lines of iPS cells would provide donor lines for 80% or more of patients with a single mismatch at one of three HLA loci (HLA-A, -B, and -DR) among the Japanese. A comparable bank of 150 lines could provide an acceptable or better match for 84.9% of the UK population. Importantly, a bank size of only 50 lines could provide a three-locus match in 90.7% of the Japanese population, if iPS cells are established from HLA homozygous cells. Screening an HLA-type database of 24,000 individuals would be required to identify at least one homozygote for each of 50 different HLA haplotypes. This could be possible if the iPS banks cooperate with other banks, in the same manner as do cord blood banks and bone marrow banks.
26.5 SAFETY CONCERNS FOR MEDICAL APPLICATION Safety is extremely important for the clinical application of iPS cells. Each culture of iPS cells would have different properties in terms of differentiation and safety. Human iPS cells can be generated from several cell types with different combinations of reprogramming factors by various transduction methods, as described above. As yet, no one knows the best way to obtain fully reprogrammed, safe iPS cells. Assays of chimeric mice have revealed that genomic integration of c-Myc transgene is associated with a high risk of tumor formation and should be avoided. The integration of Oct3/4, Sox2,
26.6 Medical Application
and Klf4 seems to have no/little effect on tumorigenesis. However, the overexpression of Oct3/4 and Klf4 causes tumor formation, and various human tumors express OCT3/4, SOX2, and KLF4. Furthermore, retroviral insertion into the genome may itself disturb endogenous gene structure and increase tumor risks. However, there are between one and 40 genomic integration sites of retro- and lentivirus in iPS cells, and PCR-based analysis can detect them all. Therefore, it is possible to estimate the risk beforehand. Nonintegration methods have been established, but they have low induction efficiency of iPS cells, which suggests they yield reprogrammed iPS cells of lower quality than integration methods. This might be improved by using better combinations of reprogramming factors and choosing a better cell source. The retroviral induction method might be selected after a careful risk assessment if it induces better reprogramming than other transient or non-integration methods. Residual undifferentiated cells are a common problem when using stem cells for cell transplantation therapy. As described above, most mouse iPS cells can differentiate into neurospheres; however, a small portion of cells remains in an undifferentiated state in the sphere, thereby giving rise to tumor formation when transplanted. An effective protocol to eliminate undifferentiated cells should be established, such as improvement of the differentiation protocol and sorting by flow cytometry.
26.6 MEDICAL APPLICATION Mouse iPS cells have been applied to the treatment of a humanized sickle cell anemia mouse model. Homozygous mice for mutant human β-globin genes show characteristic symptoms, including severe anemia due to erythrocyte sickling, splenic infarcts, urine concentration defects, and poor health. The iPS cells established from the mouse have the same genomic mutation. The mutation was corrected by homologous recombination with a non-mutated construct. The rescued iPS cells were differentiated into hematopoietic progenitors and transplanted into a ‘patient’ mouse. The study provided proof of principle for application of iPS cells in combination with gene repair for cell therapy. Efficient gene correction methods have been established in human pluripotent stem cells. Homologous recombination occurs in human ES cells with helper-dependent adenoviral vectors. Homologous recombination was also performed using zinc-finger nuclease-mediated genome editing in both human ES and iPS cells. Human disease-corrected iPS cells have been established from Fanconi anemia patients by lentiviral delivery of a normal gene. Duchenne muscular dystrophy is caused by defect of the Dystrophine gene, which has an extremely large size of 2.4 Mbp. iPS cells from a Duchenne muscular dystrophy (DMD) patient were transferred and corrected with the
Chapter 26: Induced Pluripotent Stem Cells
human-artificial-chromosome-encoding Dystrophine gene. These techniques could supply patient-specific but gene-corrected iPS cells.
26.7 DIRECT FATE SWITCH The establishment of iPS cells introduced a new paradigm: that forced expression of master genes can alter the cell state. This contributes to the study of direct reprogramming from one somatic cell type into another cell type without the mediation of stem cells. One group screened more than 1,100 transcription factors and chose nine candidates for β-cell induction. Combinations of these genes were inserted by adenovirus vectors into the pancreata of mice. Insulin-positive cells developed in one month with the mixture of Ngn3, Pdx1, and Mafa. The cells were derived from pancreatic exocrine cells and closely resembled normal β-cells. These cells could ameliorate hyperglycemia by remodeling the local vasculature and secreting insulin in mice rendered diabetic by streptozotocin injection. Another example is the conversion of mouse fibroblasts into neurons by induction of three transcription factors, Ascl1, Brn2, and Myt1l. The induced neuronal (iN) cells expressed several neuronal markers, generated action potentials, and formed functional synapses. The iN cells are useful for neurological disease modeling and regenerative medicine. Although further study is required, these approaches could therefore form an alternative method for making specific differentiated cells from a patient’s somatic cells or iPS cells.
26.8 CONCLUSION iPS cells have tremendous potential to supply patient-specific pluripotent stem cells for use in the study of disease pathogenesis, drug discovery, toxicology, and cell transplantation therapy. Several lines of evidence support the finding that iPS cells are very similar, but not identical, to ES cells. However, there is insufficient data to definitively determine whether or not this difference is critical. Mouse and rat iPS cells can contribute to chimeric animals after injection into blastocysts. Direct and detailed comparison between iPS cells and ES cells is required. The establishment of iPS cells would also apply not only to the medical field but also to the elucidation of the control mechanisms of stem cells and the development of efficient differentiation protocols. Studies of disease pathogenesis and drug discovery have already been launched, and the results could provide relief to countless people throughout the world. The application of iPS cells to human disease will take time. In addition, both the research and medical application of human iPS cells will also be subject to a wide range of laws and research ethical policies.
For Further Study
FOR FURTHER STUDY  Aasen T, Raya A, Barrero MJ, Garreta E, Consiglio A, Gonzalez F, et al. Efficient and rapid generation of induced pluripotent stem cells from human keratinocytes. Nat Biotechnol 2008;26(11):1276–84.  Aoi T, Yae K, Nakagawa M, Ichisaka T, Okita K, Takahashi K, et al. Generation of pluripotent stem cells from adult mouse liver and stomach cells. Science 2008;321(5889):699–702.  Boyer LA, Lee TI, Cole MF, Johnstone SE, Levine SS, Zucker JP, et al. Core transcriptional regulatory circuitry in human embryonic stem cells. Cell 2005;122(6):947–56.  Ebert AD, Yu J, Rose Jr. FF, Mattis VB, Lorson CL, Thomson JA, et al. Induced pluripotent stem cells from a spinal muscular atrophy patient. Nature 2009;457(7227):277–80.  Han J, Yuan P, Yang H, Zhang J, Soh BS, Li P, et al. Tbx3 improves the germ-line competency of induced pluripotent stem cells. Nature 2010;463(7284):1096–100.  Hong H, Takahashi K, Ichisaka T, Aoi T, Kanagawa O, Nakagawa M, et al. Suppression of induced pluripotent stem cell generation by the p53-p21 pathway. Nature 2009; 460(7259):1132–5.  Huangfu D, Maehr R, Guo W, Eijkelenboom A, Snitow M, Chen AE, et al. Induction of pluripotent stem cells by defined factors is greatly improved by small-molecule compounds. Nat Biotechnol 2008;26(7):795–7.  Judson RL, Babiarz JE, Venere M, Blelloch R. Embryonic stem cell-specific microRNAs promote induced pluripotency. Nat Biotechnol 2009;27(5):459–61.  Miura K, Okada Y, Aoi T, Okada A, Takahashi K, Okita K, et al. Variation in the safety of induced pluripotent stem cell lines. Nat Biotechnol 2009;27(8):743–5.  Rowland BD, Peeper DS. KLF4, p21 and context-dependent opposing forces in cancer. Nat Rev Cancer 2006;6(1):11–23.