Switching cell fate: the remarkable rise of induced pluripotent stem cells and lineage reprogramming technologies

Review

Switching cell fate: the remarkable rise of induced pluripotent stem cells and lineage reprogramming technologies Vimal Selvaraj, Jennifer M. Plane, Ambrose J. Williams and Wenbin Deng Department of Cell Biology and Human Anatomy, Institute for Pediatric Regenerative Medicine, School of Medicine, University of California, Davis, 2425 Stockton Boulevard, Sacramento, CA 95817, USA

Cell reprogramming, in which a differentiated cell is made to switch its fate, is an emerging field with revolutionary prospects in biotechnology and medicine. The recent discovery of induced pluripotency by means of in vitro reprogramming has made way for unprecedented approaches for regenerative medicine, understanding human disease and drug discovery. Moreover, recent studies on regeneration and repair by direct lineage reprogramming in vivo offer an attractive novel alternative to cell therapy. Although we continue to push the limits of current knowledge in the field of cell reprogramming, the mechanistic elements that underlie these processes remain largely elusive. This article reviews landmark developments in cell reprogramming, current knowledge, and technological developments now on the horizon with significant promise for biomedical applications. Introduction: the reprogramming potential of cells During mammalian development, cells differentiate by the activation and repression of specific gene networks. These events are mediated by transcription factors that ultimately govern functional specialization. This developmental program is a carefully orchestrated unidirectional process, during which a pluripotent cell becomes multipotent, and gradually loses phenotypic plasticity and becomes terminally differentiated. Deviation from this developmental program could result in abnormalities and pathology. Evidence from the cancerous transformation of differentiated cells has long implied that random reprogramming of gene networks to a primordial cell fate can occur. However, the process of regulated reprogramming of cell fate, that is, directly turning one cell type into another, by artificial means might prove to have enormous potential for medical and research advances. The potential of cell fate reprogramming has not been well explored until recently, in part, because of significant knowledge gaps in the understanding of complex gene networks and the nuclear state. After the pioneering work by John Gurdon in somatic cell nuclear transfer in Xenopus laevis [1], the first evidence for the experimental reversal of cell differentiation in mammals came from the holistic manipulation of a sheep epithelial cell nucleus by its transplantation into an enucleated oocyte, which resulted in the production of the normal adult sheep Corresponding authors: Selvaraj, V. ([email protected]); Deng, W. ([email protected]).

214

Dolly - a milestone that showed that the nuclear transfer technology could reverse the cell fate of somatic cells to stem cells [2]. It has been discovered that undefined factors present in the oocyte cytoplasm can reprogram the epigenGlossary Totipotent: The ability of a cell to give rise to all cell types of the body, including those that make up the extra-embryonic tissues such as the placenta. The zygote (fertilized oocyte) is considered to be totipotent. Pluripotent: The ability of a cell to give rise to all the different cell types of the body. Pluripotent cells can differentiate into cells of all three germ layers but do not contribute to extra-embryonic tissues such as the placenta. Pluripotent cells include ES and iPS cells. Multipotent: The ability of cells to develop into more than one cell type of the body. Multipotent cell types in the body comprise lineage-committed progenitors, including organ-specific adult stem cells. Epigenetic regulation: The process by which histone and DNA modifications regulate specific gene expression patterns (Figure 2). Epigenetic modifications of the genome are passed on to daughter cells during cell division. Epigenetic code: The combination of epigenetic features that create the characteristic phenotypes of different cell types. The overall epigenetic state of the cell is referred to as the epigenome. Heterochromatin: Refers to tightly packed DNA in which genes are repressed epigenetically to various extents. Stem cells are largely devoid of heterochromatin. Euchromatin: Refers to lightly packed DNA in which genes are often under active transcription. Loose DNA in euchromatin regions allows access to transcription factors and polymerases to facilitate expression of genes. However, not all euchromatin regions are essentially transcribed. Imprinting: Refers to an epigenetic phenomenon by which genes are expressed in a mono-allelic (parent of origin specific) manner. Genomic imprinting is inherited from either the maternal or paternal genome. Telomere: Refers to a region of repetitive DNA at the ends of the chromosomes. Telomeres serve to protect the chromosomes and maintain genomic stability. The length of the telomere is maintained by telomerase. Gradual telomere attrition is considered a normal phenomenon of aging. ES cells: Undifferentiated cells derived from a preimplantation embryo (inner cell mass of a blastocyst) that are capable of self-renewal, and can develop into cells and tissues of the three primary germ layers. IPS cells: Stem cells generated from somatic cells by the induced expression of specific reprogramming factors. Transformation of the epigenetic code of somatic cells to a primordial stem cell state has been demonstrated to be regulated by embryonic transcription networks triggered by Oct3/4, Sox2, Klf4, c-Myc, Nanog and Lin28. Dedifferentiation: A process by which a differentiated cell reverts to a less differentiated state: pluripotent or multipotent. Lineage reprogramming: Type of reprogramming that brings about a direct fate switch from one type of differentiated cell to another without the generation of a pluripotent intermediate. Experimental approaches for lineage reprogramming require the activation of cell-type-specific transcriptional networks. Synonym: transdifferentiation. Transcriptome: Refers to the library of all RNA molecules, or transcripts, produced by a particular cell type. These include mRNA, rRNA, tRNA and other non-coding RNAs. The transcriptome is directly linked to the epigenome and is unique for each cell phenotype. Proteome: Refers to the entire complement of proteins expressed by a particular cell type. The proteome is directly linked to the translation of mRNA transcripts and is unique for each cell phenotype.

0167-7799/$ – see front matter ß 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibtech.2010.01.002 Available online 9 February 2010

Review

Trends in Biotechnology

Vol.28 No.4

Figure 1. Dedifferentiation and transdifferentiation. (a) Pluripotent ES cells derived from the inner cell mass of blastocysts can differentiate into cells of the three germinal layers. (b) Certain terminally differentiated adult cells can be induced to transdifferentiate or undergo a lineage switch from one cell type to another. (c) Terminally differentiated cells can also be dedifferentiated to generate iPS cells that have very similar properties when compared with those of ES cells. (d) These iPS cells can also differentiate into cells of the three germinal layers and generate germline-transmitting chimeras when injected into blastocysts. (e) As an ultimate testament of its pluripotency, the advanced technique of tetraploid blastocyst complementation, in which the inner cell mass is formed only by a few injected iPS cells allows the direct generation of iPS-cell-derived animals.

ome of the transplanted nucleus to a totipotent state. This process, termed somatic cell nuclear transfer (SCNT) or simply cloning, subsequently has become a prototypic example for the process of cellular dedifferentiation [3,4]. However, evidence that differentiated cells can be reprogrammed came even before the popularity of SCNT. It was first reported that the transcription factor MyoD formed the nodal point of conversion of fibroblasts to cells of the myogenic lineage [5]. Genes transcribed as a result of MyoD gene induction in fibroblasts could reprogram them efficiently to myocytes. This type of reprogramming that could bring about a direct fate switch without the generation of a pluripotent intermediate is what is known today as ‘lineage conversion’ or ‘transdifferentiation’. A quantum leap in the field of cell reprogramming was defined by the recent discovery that differentiated cells can be directly reprogrammed to induced pluripotent stem (iPS) cells using defined reprogramming factors [6]. The ramifications of cellular reprogramming mediated by four transcription factors have launched the field back to the limelight, and the past 3 years have seen an explosion of scientific curiosity and industrial interest. This is mainly because iPS cells derived by cellular dedifferentiation are virtually indistinguishable from embryonic stem (ES) cells [7,8], and thereby could potentially replace ES cells for various clinical applications, circumventing crucial ethical concerns regarding destroying embryos. Notably, iPS cells also present the benefit of being patient-specific autologous cells that should avoid immune rejection if used for cell therapy in regenerative medicine. Besides applications, new knowledge gained by this seminal discovery has forced reevaluation of current models that depict the plasticity of somatic cells. It appears that mammalian cells attain

functional specializations that are different from each other during development, but they retain the potential to be transformed into other cell types when provided with the right environmental stimuli or induced with specific transcription factors (Figure 1). In this article, we examine the rapidly evolving technology that surrounds induced manipulation of cell fate, and its use, mechanisms and potential prospects in biotechnology and medicine. Induced pluripotency: generating iPS cells from differentiated cells Trials based on the hypothesis that factors responsible for maintenance of pluripotency in ES cells might induce pluripotency in somatic cells led Yamanaka and colleagues to identify four genes, Oct3/4 (also known as Pou5f1), Sox2, Klf4 and c-Myc, encoding transcription factors that could reprogram murine and human fibroblasts to iPS cells [6,9]. Almost immediately, a multitude of studies have since reproduced this result in several cell types and species examined to date (Supplemental Material, Table 1). As an alternative to the use of the above four genes, it has also been shown that Nanog and Lin28 could replace Klf4 and c-Myc to achieve pluripotency in human fibroblasts [10]. In addition, use of all six reprogramming factor genes simultaneously has resulted in a synergistic effect [11], which suggests complex regulation of this reprogramming process. However, the genomic integration of these factors limits the utility of these cells because of the risk of insertional mutations. Moreover, copies of oncogenes such as c-Myc increase the potential tumor risk when these cells are used for regenerative therapy. Therefore, subsequent studies have focused on generating safer iPS cells by alternative approaches. 215

Review

Trends in Biotechnology Vol.28 No.4

Figure 2. Epigenetic control of cell fate reprogramming. (a) Differentiated cells maintain their transcriptome with stable DNA and histone modifications for preserving celltype-specific gene expression. Major epigenetic regulation of gene expression occurs through histone acetylation that enhances target gene expression, histone methylation that inhibits target gene expression, and DNA modifications such as methylation of the promoter region that can prevent transcription factor binding. In a simplified model, reprogramming cell fate requires initiation of gene expression specific to the target cell type that would most likely be within genomic loci that are silenced in a differentiated unrelated cell. Therefore, modification of cellular gene expression patterns involves regulation of HDACs, HMTs and DNMTs that can modulate genome-wide DNA and histone modifications. The reprogramming efficiency and generation of iPS cells can be enhanced using small molecules to inhibit these enzymes. (b) In addition, factors that inhibit cell proliferation also affect epigenetic plasticity. In dividing cells, the chromatin is considered to be in an open state and thereby more amenable to chemical modification by histones and DNA. However, certain events can make a cell refractory to gene expression changes required for reprogramming to occur. The tumor suppressor gene, p53, nicknamed the ‘guardian of the genome’, presents a barrier to epigenetic reprogramming. The p53 pathway is physiologically activated in conditions of DNA damage to induce cell cycle arrest, which leads to global chromatin condensation. Suppression of p53 by siRNA or induction of undifferentiated embryonic cell transcription factor 1 can significantly enhance reprogramming efficiency. Although the link between p53 activation and the reprogramming pathway is unclear, this physiological cellular response appears to ensure iPS cell genomic integrity. Chromatin-remodeling proteins such as Chd1 promote reprogramming by ensuring an open state of chromatin. Telomerase activity is also crucial for reprogramming: senescence associated with telomere attrition is inhibitory towards epigenetic reprogramming.

Experiments have demonstrated that it is possible to reduce the number of transcription factors used to generate iPS cells under different conditions. For example, iPS cells could be generated from fibroblasts in the absence of c-Myc [12,13]. However, this was at the cost of reprogramming efficiency. It was also observed that endogenous expression levels in the starting cell type could influence the need for these factors. Oct3/4 in combination with c-Myc or Klf4 [14,15], or even Oct3/4 alone [16,17] can reprogram neural stem cells that already express Sox2 at levels similar to ES cells. In addition to genes encoding transcription factors, recent studies also have revealed that expression of ES-cellspecific microRNA (miRNA) molecules such as miR-291, miR-294 and miR-295 can enhance the efficiency of induced pluripotency by acting downstream of c-Myc [18]. Nonetheless, the mechanistic link between the above factors and epigenetic regulators remains to be elucidated. To aid in epigenetic modifications during the reprogramming process, several studies have examined the use of small-molecule compounds (Figure 2). Reprogramming efficiency was enhanced by the histone deacetylase 216

(HDAC) inhibitor valproic acid by potentially compensating for c-Myc [19]. Moreover, the inhibition of histone methyl transferase (HMT) with BIX-01294, in combination with the activation of plasma membrane L-type Ca2+ channels with BayK8644 in fibroblasts, increases reprogramming efficiency by compensating for Sox2 [20]. Inhibition of DNA methyl transferase (DNMT) activity by 5aza-cytidine also improved reprogramming efficiency [8]. Recently, another small molecule, transforming growth factor a RI kinase inhibitor II has also been reported to replace Sox2 by the induction of Nanog [21]. However, the effects of these small-molecule agents are not directed towards specific reprogramming events, and they do not completely replace the need for transcription factors for induced pluripotency. In an attempt to eliminate concerns brought about by genome integrating approaches for iPS cell generation, studies have utilized creative ways to induce these transcription factors in cells. The use of transient expression with plasmids [22], adenoviruses [23] and transposon vectors [24], which code for the different transcription factors,

Review

Trends in Biotechnology

Vol.28 No.4

Box 1. Emerging technology and challenges in cell reprogramming Generation of iPS cells Technology for the generation of iPS cells is evolving continuously to make the process more efficient and the cells safer in potential recipients. Three future strategies are being considered for generating safe iPS cells for clinical use.  The use of synthetic antigene RNAs (agRNAs) to induce the expression of endogenous genes [72]. Designed agRNAs for endogenous pluripotency promoter regions can activate these genes and bring about relatively safe reprogramming.  The method of transcriptome-induced phenotype remodeling (TIPeR) [73], by which the entire transcriptome of a stem cell can be transferred to a target somatic cell to bring about cell fate reprogramming might represent a holistic approach for iPS cell generation. This approach facilitates the freedom from overexpression artifacts.  The development of synthetic regulatory proteins that are artificial transcription factors [74], fused to protein transduction domains for the pluripotency genes. These artificial proteins may enable derivation of iPS cells without causing any genetic modifications. In vivo lineage reprogramming At this time, research on lineage reprogramming strategies is focused mainly on discerning effective reprogramming factors,

has been successful in generating iPS cells, albeit with an even lower efficiency. Another approach in which the genome-integrated factors flanked by loxP sites could be removed via a cre recombinase-mediated excision after the completion of reprogramming also has been devised [25]. Moreover, direct introduction of recombinant proteins for the different factors also has demonstrated preliminary success in inducing pluripotency [26]. However, the protein-based methods possess very low efficiency and might prove difficult to scale up, and thus are incompatible with the need to generate various individual cell lines from many different patients to study the nature and complexity of disease. The field of iPS cells continues to suffer from a crucially low efficiency of reprogramming and a high proportion of genomic or karyotypic abnormalities in the reprogrammed cells. To overcome concerns and problems of the current reprogramming technologies, testing is under way for several innovative methods for iPS cell generation (Box 1). Research should be directed towards deducing the mechanisms that underlie the dedifferentiation process in order to better manipulate them. Lineage reprogramming: fate switch of differentiated cells During development, mammalian cells undergo a progressive restriction of cell fate, which leads to functional specialization and a terminally differentiated state. These differentiated cells normally maintain this state stably throughout their lifetime. However, several in vitro studies have demonstrated that certain terminally differentiated cell types can exhibit a fate switch when placed in stimulating environments, or after specific transcription factor induction (Supplemental Material, Table 2). Could adult cells ensconced in their native environment or niche respond to ectopic induction? In principle, a direct reprogramming of abundant and easily accessible cell types in vivo to repair diseased or damaged tissues presents an attractive source of patient-specific cells for regenerative therapy. The proof of principle for therapeutic transdifferentiation has been established by the restoration of hear-

which would be contingent on starting and target cell types. It is unclear currently how much influence the cellular niche has on the steady-state of cells. Therefore, certain terminally differentiated cell types nestled in functional networks could be refractory to reprogramming. Current technology for in vivo lineage reprogramming involves the ectopic expression of powerful transcription factors that act as master regulators. At times, this may force its specific transcriptional programs in the starting cell without stable/ complete phenotypic transformation and result in a hybrid phenotype or cellular mimicry. For that reason, lineage reprogramming may be limited to developmentally related cells with the least epigenetic differences. An attractive target is the stress-induced activated cell that has a natural proclivity for regeneration at the environs of a lesion. For example, reactive astrocytes after CNS injury are more plastic and could serve as an abundant cellular substrate for reprogramming to neurons [70]. Strategies to specifically target this cell type could facilitate neurogenesis in the injured/diseased brain [71]. Despite the challenges ahead, regenerative lineage reprogramming is a practical approach and will form an important chapter in the future of regenerative medicine.

ing in deaf guinea pigs by the lineage reprogramming of non-sensory cochlear cells of the inner ear to auditory hair cells by adenoviral delivery of the transcription factor gene Atoh1 [27]. After this first milestone, the potential for therapeutic lineage reprogramming has been explored in numerous areas of regenerative medicine. Here, we focus on two target cell types that are of significant clinical interest: pancreatic b cells and neurons. Type 1 diabetes mellitus is characterized by loss of b cells in the pancreatic islets of Langerhans, which leads to insulin deficiency. In these islets, endocrine cells are present amidst the complex duct system that collects from the exocrine acinar cells that secrete pancreatic digestive enzymes. During embryonic development, pancreatic endocrine cells differentiate early from the endoderm-derived duct cells. Soon thereafter, they divide into the distinct lineages of glucagon-producing a cells and insulin-producing b cells [28]. Early studies of the regenerative capacity of the pancreas have demonstrated that injury brought about by partial pancreatectomy or pancreatic duct ligation can lead to neogenesis of b cells [29,30]. Although it was initially perceived that adult progenitors existed in the pancreas, recent studies have shown that these injury-induced b cells arise by the lineage reprogramming of pancreatic duct cells [31] (Figure 3). Although specific mechanisms are not known, it could be argued that the duct cells transition to an activated state that is more plastic in response to injury. This lineage conversion could be replicated in vitro when duct cells are transduced with a transcription factor gene Pdx1 [32]. Moreover, downstream signaling events mediated by the activation of glucagon-like peptide-1 (GLP-1) receptors enhanced Pdx1-mediated effects [33]. Trials using exendin-4, a GLP-1 receptor ligand in conjunction with Pdx1 overexpression in endoderm-derived hepatocytes, have led to the expression of pancreatic genes [34]. However, a complete phenotypic conversion of hepatocytes to functional b cells could not be dictated by Pdx1 alone, potentially because of the greater epigenetic distance between the two cell types. A recent breakthrough for b-cell regeneration has come with the discovery that in vivo 217

Review

Trends in Biotechnology Vol.28 No.4

Figure 3. Cell fate plasticity specific for two target lineages of important clinical interest. (a) Generating pancreatic b cells for the treatment of type 1 diabetes. Several cell types have demonstrated the potential for transdifferentiation into pancreatic b cells both in vitro and in vivo. Transcription factor Pdx1 acts as a master regulator in the differentiation of b cells. Forced expression of Pdx1 along with the treatment of exendin-4 in hepatocytes can lead to their transdifferentiation into insulin-producing b cells in vitro. Introduction of adenoviral vectors that express Pdx1, Ngn3 and MafA in the murine pancreas in vivo can lead to transdifferentiation of exocrine cells into b cells. Even epigenetically distant cells such as fibroblasts can be reprogrammed partially to express insulin genes when treated with an insulinoma extract. Pancreatic duct cells also have a propensity to transdifferentiate into b cells when cultured with pancreatic stromal cells. An injury-induced activated state of pancreatic duct cells brought about by a traumatic event such as pancreatic duct ligation can also lead to their in vivo spontaneous transdifferentiation into insulin-producing b cells. (b) iPS-cell-based approach for generating pancreatic b cells and neurons. Somatic cells dedifferentiated by viral induction of Oct3/4, Sox2, c-Myc and Klf4 into iPS cells can be differentiated into isogenic pancreatic b cells or neurons for cell therapy. This possibility is considered a major application for iPS cells in regenerative medicine. (c) Generating neurons for the treatment of CNS degenerative diseases. Lineage-committed oligodendrocyte precursor cells can dedifferentiate to neural stem cells when cultured sequentially with bone morphogenetic proteins and fibroblast growth factor 2. These neural stem cells can proliferate and produce neurospheres and mature neurons. Epigenetically distant mesodermal myoblasts can also transdifferentiate into neuroectodermal cells by treatment with the chemical reversine or ciliary neurotrophic factor. Treatment with retinoic acid and medium that contains insulin, transferrin and selenium, or nerve growth factor in medium that contains B27 supplement and fetal bovine serum can differentiate neuroectodermal cells to neurons. The neurogenic potential of astrocytes is of particular interest because of their abundance in the CNS and their formation of a reactive state in response to local injury. Induction of Pax6, Ngn2 or Mash1 in astrocytes can transdifferentiate them to neurons in vitro. Injury-induced activation of astrocytes in vivo can induce the expression of markers specific for neural progenitors, which suggests that a fate change can occur in astrocytes. Therefore, reprogramming targeted to reactive astrocytes in vivo represents an avenue for therapeutic intervention for regenerative transdifferentiation.

overexpression of Pdx1, along with the transcription factor genes expressed in pancreatic endocrine progenitors Ngn3 and MafA in pancreatic acinar cells, can reprogram them into endocrine b cells [35]. This approach for generating more b cells in vivo by direct lineage reprogramming offers a novel approach for treating type 1 diabetes. 218

Similar to b-cell regeneration, preliminary studies have suggested that a lineage reprogramming approach could also be relevant for neuronal regeneration. In the mammalian central nervous system (CNS), there remains only an insignificant capability for regeneration by neurogenesis. Therefore, most CNS degenerative diseases cause

Review

Trends in Biotechnology

Vol.28 No.4

Box 2. Cell reprogramming: lessons learned and key questions Can every cell type be reprogrammed? Cell types from all three germ layers have been reprogrammed successfully to iPS cells using defined transcription factors. The question remains as to whether extremely specialized differentiated cells, such as neurons, can be fully reprogrammed. Although similar stochastic models apply, the transdifferentiation/lineage reprogramming potential of differentiated cells depends primarily on epigenetic distance. Thus, cells types arising from common progenitors are likely to interchange their phenotype with relative ease. Is iPS cell reprogramming complete? During reprogramming, the genome-wide epigenetic code must be reformatted to the target cell type to avoid a partially reprogrammed cell fate. Initially, epigenetic aberrations were not detected in fibroblast-derived iPS cells [8], however, a study that has compared iPS-cell-derived chimeras has found an increased rate of perinatal mortality in liver- and stomach-cell-derived compared with fibroblast-derived iPS cells [75]. Significant variations also have been detected in the teratoma-forming capacity of iPS cells derived from different adult tissues [76]. These findings suggest that there could be some epigenetic memory retention in iPS cells, depending on their tissue of origin. The implications of these findings could be crucial and necessitate genome-wide epigenetic screening to validate cell quality before use in vital applications.

irreparable damage. During embryonic development, CNS progenitors, the so-called radial glial cells, represent a major population of neural progenitors that generate neurons and the two major lineages of glial cells: oligodendrocytes (myelin-forming cells) and astrocytes [36]. In the adult brain, neural progenitors become restricted to two sites: the subventricular zone and the dentate gyrus. In response to injury, these cells have a limited capacity of ectopic migration to the site of injury, and give rise to neurons, oligodendrocytes and astrocytes [36]. Interestingly, the astroglial characteristics displayed by these adult progenitors raises the question whether mature astrocytes in non-neurogenic regions can retain some precursor-like properties. Recent in vitro studies have demonstrated that adult astrocytes can become neurons by the expression of one of the three transcription factors: Pax6, Ngn2 or Mash1 [37,38] (Figure 3). It is well known that Pax6 is strongly expressed in neural stem cells and its downstream targets Ngn2 and Mash1 regulate their differentiation into neurons. Therefore, the transcriptional network initiated by Pax6 in astrocytes can effectively reprogram them into functional neurons. Astrocytes respond to many forms of CNS injury by proliferation and assume a reactive state termed reactive gliosis. Several studies have observed that these reactive astrocytes re-express markers, such as nestin, normally seen in neural stem cells [39,40]. This property by itself indicates a certain degree of dedifferentiation towards a multipotent lineage of neural stem cells. Analysis of injured tissue, when cultured in vitro, has demonstrated that local astrocytes gain the ability to form multipotent neurospheres that can differentiate into neurons and oligodendrocytes in addition to astrocytes [41]. Therefore, there is great potential that forced in vivo expression of Pax6 or Ngn2 in astrocytes could reprogram them to neurons. This approach for initiating neurogenesis in the injured CNS through lineage repro-

Is there a link between pluripotency and cancer? The tumor suppressor p53, nicknamed ‘the guardian of the genome,’ is known as the master regulator that helps prevent cancer. Inactivation or deletion of p53 significantly increases reprogramming efficiency [49–53]. Reprogramming factor expression appears to upregulate p53 downstream targets and induces the DNA damage response and senescence that act as a barrier to reprogramming (Figure 2). These similarities between reprogramming-induced senescence and DNA-damage-induced senescence suggest that mechanisms that prevent tumorigenesis significantly overlap with those that prevent reprogramming. Intrinsic quality control might be lost without p53, which leads to reprogrammed cells with heavy DNA damage or truncated telomeres. This common link in the genesis of cancer stem cells and iPS cells also underscores a strong connection between tumor formation and induced pluripotency. Can iPS cells be safe for clinical use? Current techniques used to reprogram somatic cells into iPS cells are unsuitable for clinical applications because they involve genomeintegrating viruses that pose a threat of insertional mutagenesis. The drawback of non-genome and protein-based approaches to iPS cell generation is their low efficiency rate. Therefore, current iPS cells are unlikely to be used in clinical regenerative medicine. With improvements in safe derivation and validation, we can be optimistic that an improved version of iPS cells will be available for future therapeutics.

gramming offers a novel way to treat several neurodegenerative diseases. It appears that a common theme is that the response of cells to injury triggers a regenerative response. Mammals lack the dramatic regenerative capacities seen in lower vertebrates. In mammals, injury-induced activated cells are not capable of physiological regeneration, but targeting these cells with specific factors could be an efficient strategy to promote in vivo lineage reprogramming and repair in degenerating lesions. This approach holds a distinct advantage over stem cell transplants and is poised to become a crucial part of regenerative medicine. However, challenges remain before significant progress can be made on this front (Box 2). Modus operandi: mechanisms of kick-starting master transcription networks Every differentiated cell of a living mammal is shaped by its characteristic epigenetic signature and interactions with the local environment. When a somatic nucleus is transferred into an enucleated oocyte, the proteome and transcriptome of the oocyte impose its pattern of gene expression and bring about dramatic epigenetic modifications to match those of the zygote. However, the success rate of this reprogramming process is merely 0.5–3% [42]. Moving away from this holistic to a more defined approach, generating iPS cells from somatic cells has an even lower efficiency of 0.01–0.1% [6]. Interestingly, even the induction of secondary iPS cells using differentiated primary iPS cells that contain an inducible form of the transcription factors still exhibits a low efficiency of 3–5% [43]. Recent studies have hypothesized and demonstrated that reprogramming by these transcription factors is a continuous stochastic process [44,45]. This concept is best explained by the understanding that gene expression is a noisy process [46], because of the stochasticity of molecular processes, even in stable steady-state cell phenotypes. Molecular 219

Review noise can give rise to significant heterogeneity at the cell and population level [46,47]. According to this stochastic model, almost all cells can eventually give rise to iPS cells on continued growth and reprogramming factor expression. Latencies for reprogramming are not normally distributed, however, increase in cell division rate directly rescales and accelerates the kinetics of reprogramming. In corroboration, cell senescence presents a significant roadblock to reprogramming [48]. Independent of cell division rate, triggering cell-intrinsic mechanisms like Nanog overexpression can also accelerate reprogramming kinetics [45]. Therefore, cell fate determination during the reprogramming process is controlled by interplay between ectopically induced factors, and stochasticity of molecular processes that are restricted by deterministic regulatory mechanisms (Figure 2). Five recent studies have reported that activation of p53, the tumor suppressor gene, blocks epigenetic reprogramming [49–53]. p53 triggers the DNA damage response by activation of its target gene p21(Cip1) that brings about cell cycle arrest. Reducing the level of p53 expression in cells markedly enhances their reprogramming efficiency and prevents chromatin condensation. Linking global chromatin programs, chromodomain helicase DNA-binding protein 1 (Chd1), which promotes euchromatin formation,

Trends in Biotechnology Vol.28 No.4

is crucial for reprogramming and pluripotency [54]. Telomerase expression also regulates the reprogramming process; telomere shortening associated with aging is reversed during iPS cell generation [55]. iPS cell telomeres acquire similar epigenetic marks to ES cells, with a low density of H3K9me3 and H3K20me3 regions. Moreover, the epigenetic landscape shifts more completely towards euchromatin regions marked by H3K4me3 and also H3K4me3– H3K27me3, as in ES cells, in contrast to genes repressed by H3K27me3 or DNA methylation. This addresses why the use of HDAC inhibitors [19], HMT inhibitors [20] and DNMT inhibitors [8] can distinctly increase reprogramming efficiency. Over and above the regulatory network of factors, crucial decisions on epigenetic landscape modifications are mediated actively by the dominant master transcription factors that can ultimately impose their cell fate specification. In ES cells, it has been deduced that the transcriptional network established by Oct3/4, Sox2 and Nanog is responsible for the global regulation of pluripotency [56]. However, the roles of c-Myc, Lin28 and Klf4 in reprogramming are not well defined. c-Myc is not an absolute requisite for pluripotency but helps the efficiency of reprogramming, perhaps by recruiting chromatin remodelers and widespread miRNA repression [57]. Similarly, recent findings

Figure 4. ‘Cellular U-turn’-based iPS cell applications. (a) One of the primary objectives for generating iPS cells is to perform a complete cellular U-turn from differentiated cells to pluripotent stem cells, and then back to the desired differentiated lineage that may be used for cell therapy in regenerative medicine. These cells are superior to EScell-based technology because they circumvent the complication of immune rejection. In animal models from which the derivation of ES cells has proven problematic, iPS cells provide means for genetic engineering applications for scientific and commercial purposes. (b) Somatic cells from individuals with genetic diseases can be reprogrammed to iPS cells and then differentiated to the afflicted cell type to study directly the progression and pathology of the disease. This can greatly aid in designing and screening target drugs, thus accelerating drug discovery. In vitro iPS cell models to aid in research and testing have been developed for Parkinson’s disease [64], Lou Gehrig’s disease [62], familial dysautonomia [66], thalassemia [65] and spinal muscular atrophy [63]. Moreover, generation of iPS cells from somatic cells with known genomic mutations will enable targeted genome repair by homologous recombination and generation of healthy cells of the desired differentiated lineage for cell therapy. Proof of concept for genetic correction in iPS cells has been established for treating diseases such as sickle cell anemia [68] and Fanconi anemia [69].

220

Review also show that the RNA binding protein Lin28 can block specific miRNAs such as those of the let7 family, thereby preventing differentiation [58]. The mechanisms that explain the role of Klf4 in the reprogramming process remain to be deciphered. As these factors show considerable synergy [11], it is also conceivable that they recruit a cascade of interacting proteins and feedback loops, which adds to the complexity of this process. In lineage reprogramming, a change in cell fate can be triggered by shifting the balance between stochasticity of molecular processes and feedback loops of deterministic regulatory mechanisms by induced cell-type specific factors. In this model, cell fate can be considered non-terminal, but held by reversible epigenetic barriers that can be overcome by appropriate stimuli. Therefore, lineage reprogramming might not present excessive prohibitive barriers because cells require much less epigenetic modification compared with that for the generation of iPS cells. The efficiency of lineage reprogramming thereby largely depends on the epigenetic distance between the starting and target cells and the regulatory power of the transcription factor. In agreement, efficiencies in lineage reprogramming are much higher than reprogramming to pluripotency. For example, fibroblasts are reprogrammed to myogenic cells with a 25–50% efficiency [5], mature B cells are reprogrammed to macrophages with an 35% efficiency [59]; even in vivo reprogramming of pancreatic acinar cells to islet cells shows a high efficiency of 20% [35]. In cases of injury, cell stress might bring about increased noise and can provide means for individual cells to dynamically explore different phenotypes, as seen in microorganisms [47,60,61], thereby facilitating lineage reprogramming in vivo. However, identifying an optimal starting cell type is the key to success for regenerative lineage reprogramming strategies. In summary, one must outcompete the current stable transcription network, as well as enact a change upon the epigenetic code, to reprogram cell phenotype. However, the fundamental processes that connect transcription factor networks to epigenetic regulators remain unclear. An understanding of these complex regulations is imperative to address several existing concerns in reprogramming applications (Box 2). Multifarious applications of cell fate reprogramming: technology yields new medicine Cell fate reprogramming is poised to open a new era in medicine and biotechnology. The ‘cellular U-turn’ approach by which patient-specific somatic cells can be dedifferentiated into iPS cells, and subsequently re-differentiated into target cells opens the door to personalized medicine (Figure 4). These cells could be used to generate ex vivo lineage-committed cells that are suitable for regenerative cell therapy, thus avoiding immune rejection. For example, iPS cells could be differentiated into oligodendrocyte progenitors and used to treat demyelinating diseases such as multiple sclerosis. In patients with genetic diseases, differentiating iPS cells to the afflicted cell type can recreate the ‘disease in a dish’, thereby allowing careful dissection of underlying causes and formulating treatments directly on the diseased cells. Several human conditions such as Lou

Trends in Biotechnology

Vol.28 No.4

Gehrig’s disease [62], spinal muscular atrophy [63], Parkinson’s disease [64], thalassemia [65] and familial dysautonomia [66] have already been modeled using patientspecific iPS cells, which allows greater understanding of live disease pathogenesis rather than postmortem pathology. The list of iPS cells derived from diseased genomes continues to grow, adding to the research toolkit for studying numerous diseases [67]. Taking personalized medicine to the next level, iPS cells from patients with known genetic disorders can be corrected by homologous recombination, differentiated into lineage-committed progenitors, and transplanted back into their affected tissues/ organs to restore function. In animal models, this has been demonstrated for sickle cell anemia [68]. In human cells, the proof of principle for correcting genetic defects in iPS cells has been shown for Fanconi anemia [69]. iPS cells could also find novel applications in drug discovery. Currently, new drug development continues to suffer from the limited ability to predict the efficacy and toxicity of drugs developed and tested in animal models. As a result, several promising treatments in rodents and nonhuman primates fail in human clinical trials. Differentiated cells and/or tissues derived from human iPS cells can address this issue by providing an unlimited source of cells to screen drug efficacy and toxicity. Moreover, specific ethnic and idiosyncratic differences in drug action and metabolism can also be evaluated with iPS cells derived from selected individuals, thereby, making possible customized treatments for individual conditions. Looking ahead, an alternative novel strategy of lineage reprogramming for tissue regeneration is also gaining substantial momentum. In this approach, abundant human cells, such as fibroblasts or adipocytes, could be collected from a patient and converted into specific therapeutically important target cells, such as pancreatic b cells, by lineage reprogramming, and used for the treatment of type I diabetes in the same patient. Lineage reprogramming can also be directed in vivo to the target cell type. For example, lineage reprogramming of non-sensory cochlear cells of the inner ear to auditory hair cells to restore hearing [27], and lineage reprogramming of pancreatic acinar cells to insulin-producing b cells to treat type I diabetes [35]. This approach brings about direct fate switch without the generation of a pluripotent intermediate, thereby circumventing potential risk factors in the use of stem cells such as teratoma formation. Studies examining the feasibility of lineage reprogramming of different starting cell types and identifying master transcription regulators important for each strategy are currently under way [34,70,71]. This approach surely has unique advantages, promise and potential as an important facet of future regenerative medicine. Full speed ahead: a future reprogrammed As a result of its tremendous potential in a wide variety of clinical and research applications, there is great interest in cellular reprogramming. Unencumbered by the ethical issues of using cloned human embryos to create ES cells, the advent of clinically safe iPS cells will launch a new era in human medicine. In vivo lineage reprogramming could ultimately empower the human body to fix itself. However, there is still much to learn about how this reprogramming 221

Review process works. However, even in its infancy, it has galvanized many in the stem cell field and is set to become one of the most exciting and rapidly evolving areas of biomedical research today. Acknowledgements We thank all our laboratory members and colleagues for stimulating discussions pertinent to this review. We express our special thanks to Dr. Daniel H. Feldman and Dr. Olga Chechneva for critical reading of this manuscript. We apologize to colleagues whose work we could not cite because of space limitations. J.M.P. is supported by an NIEHS training grant. Work in our laboratory is supported by grants from the NIH (RO1 NS059043 and RO1 ES015988), National Multiple Sclerosis Society, Feldstein Medical Foundation and Shriners Hospitals for Children. We declare no conflicts of interest that relate to this article.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.tibtech.2010. 01.002. References 1 Gurdon, J.B. et al. (1958) Sexually mature individuals of Xenopus laevis from the transplantation of single somatic nuclei. Nature 182, 64–65 2 Wilmut, I. et al. (1997) Viable offspring derived from fetal and adult mammalian cells. Nature 385, 810–813 3 Hochedlinger, K. and Jaenisch, R. (2002) Monoclonal mice generated by nuclear transfer from mature B and T donor cells. Nature 415, 1035– 1038 4 Eggan, K. et al. (2004) Mice cloned from olfactory sensory neurons. Nature 428, 44–49 5 Davis, R.L. et al. (1987) Expression of a single transfected cDNA converts fibroblasts to myoblasts. Cell 51, 987–1000 6 Takahashi, K. and Yamanaka, S. (2006) Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126, 663–676 7 Maherali, N. et al. (2007) Directly reprogrammed fibroblasts show global epigenetic remodeling and widespread tissue contribution. Cell Stem Cell 1, 55–70 8 Mikkelsen, T.S. et al. (2008) Dissecting direct reprogramming through integrative genomic analysis. Nature 454, 49–55 9 Takahashi, K. et al. (2007) Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131, 861–872 10 Yu, J. et al. (2007) Induced pluripotent stem cell lines derived from human somatic cells. Science 318, 1917–1920 11 Liao, J. et al. (2008) Enhanced efficiency of generating induced pluripotent stem (iPS) cells from human somatic cells by a combination of six transcription factors. Cell Res. 18, 600–603 12 Nakagawa, M. et al. (2008) Generation of induced pluripotent stem cells without Myc from mouse and human fibroblasts. Nat. Biotechnol. 26, 101–106 13 Wernig, M. et al. (2008) c-Myc is dispensable for direct reprogramming of mouse fibroblasts. Cell Stem Cell 2, 10–12 14 Kim, J.B. et al. (2008) Pluripotent stem cells induced from adult neural stem cells by reprogramming with two factors. Nature 454, 646–650 15 Eminli, S. et al. (2008) Reprogramming of neural progenitor cells into induced pluripotent stem cells in the absence of exogenous Sox2 expression. Stem Cells 26, 2467–2474 16 Kim, J.B. et al. (2009) Direct reprogramming of human neural stem cells by OCT4. Nature 461, 649–653 17 Kim, J.B. et al. (2009) Oct4-induced pluripotency in adult neural stem cells. Cell 136, 411–419 18 Judson, R.L. et al. (2009) Embryonic stem cell-specific microRNAs promote induced pluripotency. Nat. Biotechnol. 27, 459–461 19 Huangfu, D. et al. (2008) Induction of pluripotent stem cells by defined factors is greatly improved by small-molecule compounds. Nat. Biotechnol. 26, 795–797 20 Shi, Y. et al. (2008) Induction of pluripotent stem cells from mouse embryonic fibroblasts by Oct4 and Klf4 with small-molecule compounds. Cell Stem Cell 3, 568–574

222

Trends in Biotechnology Vol.28 No.4 21 Ichida, J.K. et al. (2009) A small-molecule inhibitor of Tgf-beta signaling replaces Sox2 in reprogramming by inducing Nanog. Cell Stem Cell 5, 491–503 22 Okita, K. et al. (2008) Generation of mouse induced pluripotent stem cells without viral vectors. Science 322, 949–953 23 Stadtfeld, M. et al. (2008) Induced pluripotent stem cells generated without viral integration. Science 322, 945–949 24 Woltjen, K. et al. (2009) piggyBac transposition reprograms fibroblasts to induced pluripotent stem cells. Nature 458, 766–770 25 Chang, C.W. et al. (2009) Polycistronic lentiviral vector for ‘‘hit and run’’ reprogramming of adult skin fibroblasts to induced pluripotent stem cells. Stem Cells 27, 1042–1049 26 Zhou, H. et al. (2009) Generation of induced pluripotent stem cells using recombinant proteins. Cell Stem Cell 4, 381–384 27 Izumikawa, M. et al. (2005) Auditory hair cell replacement and hearing improvement by Atoh1 gene therapy in deaf mammals. Nat. Med. 11, 271–276 28 Habener, J.F. et al. (2005) Minireview: transcriptional regulation in pancreatic development. Endocrinology 146, 1025–1034 29 Rosenberg, L. et al. (1988) Induction of nesidioblastosis will reverse diabetes in Syrian golden hamster. Diabetes 37, 334–341 30 Bonner-Weir, S. et al. (1993) A second pathway for regeneration of adult exocrine and endocrine pancreas. A possible recapitulation of embryonic development. Diabetes 42, 1715–1720 31 Bonner-Weir, S. et al. (2008) Transdifferentiation of pancreatic ductal cells to endocrine beta-cells. Biochem. Soc. Trans. 36, 353–356 32 Noguchi, H. et al. (2003) PDX-1 protein containing its own antennapedia-like protein transduction domain can transduce pancreatic duct and islet cells. Diabetes 52, 1732–1737 33 Zhou, J. et al. (2002) Exendin-4 differentiation of a human pancreatic duct cell line into endocrine cells: involvement of PDX-1 and HNF3beta transcription factors. J. Cell. Physiol. 192, 304–314 34 Aviv, V. et al. (2009) Exendin-4 promotes liver cell proliferation and enhances PDX-1-induced liver to pancreas transdifferentiation. J. Biol. Chem. 284, 33509–33520 35 Zhou, Q. et al. (2008) In vivo reprogramming of adult pancreatic exocrine cells to beta-cells. Nature 455, 627–632 36 Kriegstein, A. and Alvarez-Buylla, A. (2009) The glial nature of embryonic and adult neural stem cells. Annu. Rev. Neurosci. 32, 149–184 37 Heins, N. et al. (2002) Glial cells generate neurons: the role of the transcription factor Pax6. Nat. Neurosci. 5, 308–315 38 Berninger, B. et al. (2007) Functional properties of neurons derived from in vitro reprogrammed postnatal astroglia. J. Neurosci. 27, 8654– 8664 39 Duggal, N. et al. (1997) Nestin expression in reactive astrocytes following focal cerebral ischemia in rats. Brain Res. 768, 1–9 40 Itoh, T. et al. (2006) Cultured rat astrocytes give rise to neural stem cells. Neurochem. Res. 31, 1381–1387 41 Lang, B. et al. (2004) Astrocytes in injured adult rat spinal cord may acquire the potential of neural stem cells. Neuroscience 128, 775–783 42 Solter, D. (2000) Mammalian cloning: advances and limitations. Nat. Rev. Genet. 1, 199–207 43 Wernig, M. et al. (2008) A drug-inducible transgenic system for direct reprogramming of multiple somatic cell types. Nat. Biotechnol. 26, 916– 924 44 Yamanaka, S. (2009) Elite and stochastic models for induced pluripotent stem cell generation. Nature 460, 49–52 45 Hanna, J. et al. (2009) Direct cell reprogramming is a stochastic process amenable to acceleration. Nature 462, 595–601 46 McAdams, H.H. and Arkin, A. (1997) Stochastic mechanisms in gene expression. Proc. Natl. Acad. Sci. U. S. A. 94, 814–819 47 Kaern, M. et al. (2005) Stochasticity in gene expression: from theories to phenotypes. Nat. Rev. Genet. 6, 451–464 48 Banito, A. et al. (2009) Senescence impairs successful reprogramming to pluripotent stem cells. Genes Dev. 23, 2134–2139 49 Marion, R.M. et al. (2009) A p53-mediated DNA damage response limits reprogramming to ensure iPS cell genomic integrity. Nature 460, 1149–1153 50 Li, H. et al. (2009) The Ink4/Arf locus is a barrier for iPS cell reprogramming. Nature 460, 1136–1139 51 Utikal, J. et al. (2009) Immortalization eliminates a roadblock during cellular reprogramming into iPS cells. Nature 460, 1145–1148

Review 52 Kawamura, T. et al. (2009) Linking the p53 tumour suppressor pathway to somatic cell reprogramming. Nature 460, 1140–1144 53 Hong, H. et al. (2009) Suppression of induced pluripotent stem cell generation by the p53-p21 pathway. Nature 460, 1132–1135 54 Gaspar-Maia, A. et al. (2009) Chd1 regulates open chromatin and pluripotency of embryonic stem cells. Nature 460, 863–868 55 Marion, R.M. et al. (2009) Telomeres acquire embryonic stem cell characteristics in induced pluripotent stem cells. Cell Stem Cell 4, 141–154 56 Kalmar, T. et al. (2009) Regulated fluctuations in nanog expression mediate cell fate decisions in embryonic stem cells. PLoS Biol. 7, e1000149 57 Chang, T.C. et al. (2008) Widespread microRNA repression by Myc contributes to tumorigenesis. Nat. Genet. 40, 43–50 58 Viswanathan, S.R. et al. (2008) Selective blockade of microRNA processing by Lin28. Science 320, 97–100 59 Xie, H. et al. (2004) Stepwise reprogramming of B cells into macrophages. Cell 117, 663–676 60 Thattai, M. and van Oudenaarden, A. (2004) Stochastic gene expression in fluctuating environments. Genetics 167, 523–530 61 Macarthur, B.D. et al. (2009) Systems biology of stem cell fate and cellular reprogramming. Nat. Rev. Mol. Cell Biol. 10, 672–681 62 Dimos, J.T. et al. (2008) Induced pluripotent stem cells generated from patients with ALS can be differentiated into motor neurons. Science 321, 1218–1221 63 Ebert, A.D. et al. (2009) Induced pluripotent stem cells from a spinal muscular atrophy patient. Nature 457, 277–280 64 Soldner, F. et al. (2009) Parkinson’s disease patient-derived induced pluripotent stem cells free of viral reprogramming factors. Cell 136, 964–977

Trends in Biotechnology

Vol.28 No.4

65 Wang, Y. et al. (2009) Generation of induced pluripotent stem cells from human beta-thalassemia fibroblast cells. Cell Res. 19, 1120– 1123 66 Lee, G. et al. (2009) Modelling pathogenesis and treatment of familial dysautonomia using patient-specific iPSCs. Nature 461, 402–406 67 Park, I.H. et al. (2008) Disease-specific induced pluripotent stem cells. Cell 134, 877–886 68 Hanna, J. et al. (2007) Treatment of sickle cell anemia mouse model with iPS cells generated from autologous skin. Science 318, 1920– 1923 69 Raya, A. et al. (2009) Disease-corrected haematopoietic progenitors from Fanconi anaemia induced pluripotent stem cells. Nature 460, 53– 59 70 Buffo, A. et al. (2008) Origin and progeny of reactive gliosis: A source of multipotent cells in the injured brain. Proc. Natl. Acad. Sci. U. S. A. 105, 3581–3586 71 Foust, K.D. et al. (2009) Intravascular AAV9 preferentially targets neonatal neurons and adult astrocytes. Nat. Biotechnol. 27, 59–65 72 Rossi, J.J. (2007) Transcriptional activation by small RNA duplexes. Nat. Chem. Biol. 3, 136–137 73 Sul, J.Y. et al. (2009) Transcriptome transfer produces a predictable cellular phenotype. Proc. Natl. Acad. Sci. U. S. A. 106, 7624– 7629 74 Bae, K.H. et al. (2003) Human zinc fingers as building blocks in the construction of artificial transcription factors. Nat. Biotechnol. 21, 275– 280 75 Aoi, T. et al. (2008) Generation of pluripotent stem cells from adult mouse liver and stomach cells. Science 321, 699–702 76 Miura, K. et al. (2009) Variation in the safety of induced pluripotent stem cell lines. Nat. Biotechnol. 27, 743–745

223