Induced Pluripotent Stem Cells
Published Online: 30 OCT 2012
Copyright © 2006 Wiley-VCH Verlag GmbH & Co. KGaA
Encyclopedia of Molecular Cell Biology and Molecular Medicine
How to Cite
Takahashi, K. and Yamanaka, S. 2012. Induced Pluripotent Stem Cells. Encyclopedia of Molecular Cell Biology and Molecular Medicine.
- Published Online: 30 OCT 2012
1.1 What Is Nuclear Reprogramming?
It was during the mid-twentieth century when Dr John Gurdon first demonstrated that the memory of somatic cells in tadpoles could be initialized in the eggs . The fact that cloned frogs hatched successfully suggested that undefined factors present in the unfertilized egg could convert the nuclei of the somatic cells to an embryonic state. In 1997, the birth of the cloned sheep, Dolly, some 50 years after this first reprogramming, confirmed that not only flexible animals such as amphibians but also more rigid animals, such as mammals, possessed reprogramming factors within their oocytes . Such findings demonstrated very clearly that unfertilized eggs do indeed incorporate reprogramming factors.
Fresh from the successful nuclear transfer in mammals, the establishment of human embryonic stem (ES) cells was reported shortly afterwards . Following the first description of their existence in mice in 1981, the ES cells were shown to be transformed from the inner cell masses of blastocyst-stage embryos, and to continue expanding without losing their capacity for differentiation, a property termed pluripotency [4, 5]. In structure, the human ES cells more closely resembled the cynomolgus ES cells that had been established several years earlier, rather than mouse ES cells [6, 7]. In addition, the optimum conditions for the maintenance of both human and primate ES cells (e.g., cytokine requirements) were shown to be totally different from those required for rodent ES cells [7, 8].
In some countries and regions the application of human ES cells is prohibited, despite the fact that these cells possess enormous therapeutic potential . For example, mouse ES cells are frequently used in laboratories worldwide on the basis of their legality, their availability and ease of handling, and their benefits as a tool for the generation of transgenic and gene-targeted animals. Today, based on knowledge acquired using mouse ES cells, the potential of pluripotent stem cells (PSCs) is widely recognized not only by basic scientists, but also by physicians and, increasingly, by patients. Consequently, there is today an extreme interest in and need for pluripotent cells that can be used for experimental purposes, without involving the ethical concerns that have become associated with ES cells. Thus, a major goal of ongoing nuclear reprogramming research is to generate ES cell-like cells, via the conversion of somatic cells.
The process of in vitro nuclear reprogramming was first demonstrated in 1983, following the observation of X-chromosome reactivation derived from female somatic nuclei after cell fusion with a mouse teratocarcinoma stem cell . Subsequently, reprogrammed T lymphocytes and fibroblasts that had been generated by cell fusion with ES cells showed similar characteristics as ES cells, such as gene expression and a high differentiation potential, not only in mouse cells but also when using human cells (although the latter cells were tetraploid) [11, 12]. Taken together, the results of these important studies have clearly suggested that such reprogramming factors were present not only in oocytes but also in the ES cells, and this has provided the basis for the generation of embryo-free PS cells. Moreover, even if the many ethical issues that have been encountered in the past were to be removed from the equation, this process would still be ideal for use by research groups, mainly because the treatment of embryos is generally unsuited to biochemical analyses. Clearly, the “road” to the discovery of reprogramming factors has been greatly broadened, despite a paucity of relevant information.
1.2 Knowledge Obtained Using ES Cells
Although the actual existence of reprogramming factors was clearly evident, the exact nature of these factors and how they mediated their effects remained unknown. Possible identities for the reprogramming factors included transcription factors, growth hormones, or epigenetic elements. It was also clear that, in many cases, proteins did not play important roles in dynamic changes of cell fate; consequently, despite the likelihood that hundreds of factors would cooperate to enforce reprogramming, attempts to understand the molecular mechanisms of pluripotency seemed to provide a means of unmasking the nature and properties of the reprogramming factors.
By the end of the twentieth century, the transcription factor network for the self-renewal of ES cells was slowly becoming clearer. For example, Oct3/4 – one of key players for pluripotency – is expressed predominantly in undifferentiated cells such as ES cells, in embryonic carcinoma (EC) cells, and in germ cells, rather than in differentiated cells. Subsequently, mice lacking octamer-binding transcription factor 3/4 (Oct3/4) clearly showed that this gene was absolutely imperative for pluripotency in blastocyst-stage embryos, primordial germ cells, and ES cells [13-15]. Another important transcription factor, Sox2, is known to bind directly to Oct3/4 and to cooperate in the regulation of the expression of target genes such as Fbx15, Utf1, and Fgf4. The expression of Sox2 can be detected in the undifferentiated cells of early embryos, germ cells, and neural lineages. Conventional Sox2 knockout mice die due to growth retardation of the epiblasts at the immediate post-implantation period . However, Sox2 is also required for neural development in the brain, and for the maturation of retinal cells, rather than simply for pluripotent cells [17, 18]. In mouse ES cells, Stat3 acts as a downstream molecule of leukemia inhibitory factor (LIF) signaling and, indeed, LIF was identified as being the first differentiation inhibitor for mouse ES cells . A simple combination of LIF and fetal bovine serum is sufficient to maintain the undifferentiated state of mouse ES cells. Stat3 can be activated via the stimulation of LIF, and subsequent phosphorylation by Jak kinase. The phosphorylated Stat3 then forms homodimers and is translocated into the nucleus, where it plays important roles in the self-renewal of mouse ES cells as a transcription factor, although its downstream target genes remain unclear. The indispensability of Stat3 was demonstrated by the forced expression of dominant active or negative mutants of Stat3 in mouse ES cells [20, 21].
Unfortunately, each of the above-mentioned essential transcription factors is insufficient to maintain pluripotency alone, or to reprogram the somatic cells. This, in turn, indicated that another as-yet undiscovered element(s) was required for pluripotency and, on that basis, Nanog was distinguished from the other known key molecules by its characteristics [22, 23]. Nanog expression can be observed in early embryos between the morula and early epiblast stages, and in primordial germ cells rather than in somatic tissues. Typically, Nanog-deficient mice die by embryonic day 6.5 at the latest, such characteristics being not too dissimilar to those of Oct3/4 and Sox2. It is particularly noteworthy that mouse ES cells expressing Nanog transgenes can self-renew, even without LIF, although when Nanog-transgenic ES cells are starved of LIF signals their expression of endogenous Nanog remains static [22, 23]. These findings suggest that Nanog is critical for the pluripotency of mouse ES cells, and that many other molecules – including LIF/Stat3 – are also related to the regulation of Nanog, though the mechanisms involved remain unclear.
Clearly, the outstanding abilities of ES cells go far beyond their pluripotency; moreover, their value is also boosted by their immortality . Under optimum culture conditions, ES cells are able to grow infinitely, and without abundant chromosomal abnormalities such as the deletion of tumor suppressors. Although, unlike most tissue stem cells, such growth superiority would lead to ES cells becoming a promising source for therapeutic use, their immortality is inextricably associated with the risks of tumor formation. In fact, after transplantation, ES cells readily produce benign tumors (teratomas) that incorporate various mixtures of cell types. Consequently, since any residual undifferentiated cells that remained following in vitro differentiation would present a risk for teratoma formation, effective methods were required both for differentiation and for the removal of any undifferentiated cells. Thus, the issue of safety proved to be a major obstacle that prevented the start of clinical trials of cell therapy using ES cells. Nevertheless, it was the presence of such problems – and the need for them to be overcome in order to provide cell therapy – that offered valuable insights into conducting investigations of the phenomenon of reprogramming. Notably, not only key molecules for pluripotency (e.g., Oct3/4, Sox2, and Nanog) but also tumor-related factors were shown to play parts in the maintenance and rebuilding of pluripotency.
1.3 The Blind Side of ES Cells
It soon became clear that the reprogramming of a patient's somatic cells could provide a means to overcome the problems related to ES cells. In particular, two major issues must be addressed before ES cells can be used for cell therapies:
The first issue relates to the ethical concern that an embryo must be broken up to establish ES cells. Although, in the past, most human ES cells have been generated from surplus embryos tendered by volunteer couples at fertility clinics, concerns have been expressed that embryos with the potential to be born are being used to generate ES cells. Thus, in an attempt to overcome this problem, a technology was developed by which pluripotent ES cell-like cells could be generated from a single-cell biopsy of blastomeres [25-27]. This methodology did not require embryo destruction because, even after the removal of a blastomere from the eight-cell stage, the remaining seven-cell embryo could still give rise to a normal infant. In fact, such biopsies have been conducted when performing preimplantation genetic diagnoses. Nonetheless, this technology is still considered unfavorable by many opponents, and its future approval seems unlikely.
A second issue associated with the use of ES cells is the risk of immunological rejection after transplantation, since it is almost impossible to match genetic types between patients and embryos as sources of ES cells. The development of an ES cell bank could perhaps resolve this issue. In fact, the Nakatsuji group have claimed an ability to calculate the proportion of patients that carry, at minimum, one human leukocyte antigen (HLA)-matched donor at the HLA-A, -B, and -DR loci . The group calculated that 170 randomly chosen embryos would be able to fulfill the demands of 80% of all patients. Furthermore, 80% of the patients would be expected to have an available ES cell line with complete confirmation at the three HLA loci only when 55 independent ES cell lines were available carrying parthenogenetic homozygous of these loci. Yet, even if this 80% scenario could be reached, an incredibly large number of embryos would be required because HLA typing before the establishment of ES cells is next to impossible.
One way in which the above-mentioned issues might be circumvented would be to use ES cells that had been generated using somatic cell nuclear transfer (SCNT). This is a technology whereby the nucleus of a somatic cell, such as a fibroblast or lymphocyte, is injected into an enucleated oocyte. As a consequence, SCNT-derived ES cells (now termed ntES cells) should inherit exactly the same DNA sequence from the donor, other than their mitochondrial DNA. Consequently – at least in theory – a donor could receive cells that had been differentiated from ntES cells, without any risk of immunological rejection following transplantation. In fact, by using a combination of this breakthrough technology and gene targeting, a therapeutic model with a correction of genetic defects was demonstrated in mice . Subsequently, ntES cells have been established not only in mice but also in cynomolgus monkeys [30, 31], although as yet there have been no reports of successful SCNT to establish human ntES cell lines. Nonetheless, the establishment of a cell bank of HLA homozygous ES cell lines and the generation of made-to-order ntES cells represent (again, in theory) ideal approaches to resolve those problems related to the clinical application of ES cells. Unfortunately, the major drawback here is that both methods would require a large number of human oocytes or fertilized eggs. Hence, although SCNT techniques have contributed greatly to the generation of new methods and techniques, and have also provided concepts for the epigenetic dedifferentiation of somatic nuclei, their successful clinical application remains difficult. Although the induction of pluripotency by the direct reprogramming of a patient's somatic cells had been expected to create a fundamental solution to the problems posed by ES cells, this proved not to be the case.
2.1 Discovery of the Reprogramming Factors
In 2000, the public database of gene expression – including expressed sequencing tags and unidentified transcripts – seemed to make quantum leaps each day. Notably, in silico quests made by using such databases allowed a much easier discovery of the ES cell-specific genes than did laboratory bench-top experiments (e.g., the subtraction method), and this resulted in 24 genes being selected as candidates for reprogramming factors. Subsequently, by narrowing down the number of candidates that included ES cell-associated transcripts, and some oncogenes by functional screening with an indicator mouse system, only four factors were ultimately identified as reprogramming factors that were capable of converting mouse embryonic fibroblasts to ES cell-like cells. The quartet consisted of Oct3/4, Sox2, Klf4, and c-Myc, and the resultant reprogrammed cells were termed induced pluripotent stem cells (iPSCs) . Although the remaining 20 of the 24 factors, such as Nanog and Sall4, were also recognized as playing important roles in pluripotency, they did not serve as triggers of reprogramming, at least in mice. In addition, not only was the germline transmission of mouse iPSCs reported, but also the creation of generations of cloned live pups by tetraploid complementation [33-38]. Generally, PSCs are injected into normal blastocysts in order to generate chimeric mice. However, as the original inner cell mass of the blastocysts also contributes to mouse development, they may also have non-cell autonomous effects on the injected cells. In contrast, the cells contained in tetraploid embryos do not contribute to the body formation of mice, although they can be transformed into extraembryonic tissues, such as the placenta.
Consequently, in order for tetraploid complementation to become useful as a highly accurate test of pluripotency, it must be capable of identifying the extremely high-grade differentiation potentials of PSCs. The results of these studies, when conducted in mice, confirmed strongly that, in terms of their differentiation potential, the iPSCs were essentially comparable to ES cells.
In 2007, based on the success of reprogramming in mice and an accumulated knowledge of human ES cells, human iPSCs were generated by using different two sets of reprogramming factors ([39, 40] (Fig. 1). One group demonstrated that Klf4 and c-Myc could be replaced with Nanog and Lin28 for the conversion of human fibroblasts into a pluripotent state, although Oct3/4 and Sox2 were commonly used by both groups. In this combination, Lin28 was shown to be effective – but not essential – for iPSC generation. Subsequently, most human iPSCs have been established by using either of these methods.
Oct3/4 is a central player in direct reprogramming. However, by using neural stem cells or neural progenitor cells which express predominantly endogenous Sox2, the forced-expression of only Oct3/4 allows the conversion of these cells to iPSCs from both mice and humans, although the efficiencies of reprogramming were quite low [41, 42]. Recently, the reprogramming of neonatal human epithelial keratinocytes by using only an Oct3/4 transgene along with bioactive compound mixtures including sodium butyrate (an inhibitor of histone deacetylase), PS48 (an activator of 3′-phosphoinositide-dependent kinase-1), A-83-01 (an inhibitor of transforming growth factor β; TGFβ), and PD0325901 (an inhibitor of mitogen-activated protein kinase; MAPK), was reported .
In contrast, three reports have also been to date made regarding the generation of iPSCs without exogenous Oct3/4:
The first report involved the reprogramming of mouse epiblast stem cells, which express Oct3/4 and Sox2, endogenously, into iPSCs , with only exogenous Klf4 being required for such conversion. In this situation, endogenous Oct3/4 and Sox2 are used instead of transgenes.
The second report involved a chemical compound that could be substituted for the Oct3/4 transgenes. Hence, supplementation with BIX-01294 (a specific inhibitor of histone methyltransferase G9a) led to the conversion of mouse neural progenitor cells into iPSCs when used in combination with Sox2, Klf4, and c-Myc, even in the absence of Oct3/4, although the frequency of reprogramming was <0.001% [45, 46].
In the third report, Oct3/4 was substituted with another gene, after which 19 nuclear receptors were screened for their activities in improving reprogramming frequency. As a result, two of the receptors – Nr1i2 and Nr5a2 – were identified as reprogramming enhancers that functioned with Oct3/4, Sox2, Klf4, and/or c-Myc. In addition, it was found that Nr5a2 could replace only Oct3/4, but not Sox2 and Klf4, in the reprogramming of mouse fibroblasts .
In mouse ES cells, the functional redundancy of Sox2 by other Sox family genes such as Sox4, Sox11, and Sox15, was demonstrated [48, 49]. Thus, the dispensability of Sox family genes in the generation of iPSCs could not be demonstrated, since most cell types express at least one of the Sox genes. However, inhibition of the TGFβ pathway by chemical compounds can replace Sox2 transgenes for the generation of iPSCs from mouse embryonic fibroblasts . Moreover, exogenous Sox2 is not required for the conversion of neural stem cells, melanocytes, and melanoma cells to iPSCs [51-54].
It is well appreciated that it is not only chemicals and cell types, but also the environment of the cell culture, that can affect the need for Sox2. Cultivation under hypoxic conditions can improve the reprogramming efficiency , and in such a situation the transduction of Oct3/4 and Klf4 reprogrammed mouse embryonic fibroblasts into iPSCs, although the efficiency was lower than in the presence of Sox2 transgenes. When these results are taken together, and despite case-to-case variation, it appears that exogenous Sox2 is most likely dispensable for the generation of iPSCs. However, the demonstration of Sox dispensability with full evidence may well be difficult, since at least one of the Sox family genes is likely to be expressed in all cell types. In contrast, it is clear that after becoming iPSCs, these cells require Sox family genes for self-renewal in the pluripotent state.
Klf4 is expressed not only in ES cells but also in various tissues, such as the gut and skin; consequently, it may appear contrary that Klf4 was included as one of the candidate reprogramming factors. In fact, Klf4 had been expected to be an important player for pluripotency since it first attracted attention in the PSC field as one of several genes that, in mouse ES cells, could undergo many changes depending on the LIF/Stat3 signaling activity. As the withdrawal of LIF from the culture medium is known to induce the differentiation of mouse ES cells, the discovery of the downstream molecules of the LIF/Stat3 pathway in the pluripotent state proved to be quite difficult. Nonetheless, Yoshimi Tokuzawa performed a series of critical microarray analyses, whereby ES cells with a forced-expression of Nanog were used that could maintain pluripotency, even in the absence of LIF. A comparison of the gene expression of these Nanog-expressing ES cells, with or without LIF, showed that the expression of Klf4 was drastically altered by LIF. Tokuzawa also found that a forced expression of Klf4 could mimic the effects of LIF, and consequently attention became focused on the role of Klf4 in pluripotency and its nomination as a candidate reprogramming factor. Typically, mouse ES cells express not only Klf4 but also other Klf family genes, such as Klf2 and Klf5, all of which share a functional redundancy in their roles for pluripotency . On a more relevant note, the role of Klf4 as a reprogramming factor can also be fulfilled by Klf2 and Klf5 . In addition, the cooperation of Klf4 with Oct3/4 and Sox2 was also demonstrated . For example, Klf4 is important for the expression of a well-known gene regulated by Oct3/4, referred to as Lefty 1.
A direct interaction of Klf4 with Oct3/4 and Sox2 during the generation of iPSCs was also observed through immunoprecipitation assays , such that not only reprogramming only with Oct3/4 (as noted above) but also the complete dispensability of Klf4 was suggested in some cases. In human cells, Nanog can substitute for Klf4 ; such substitution can be explained by the finding that Klf4 binds directly to the promoter region of the Nanog gene, thus regulating its expression in human ES cells, in cooperation with the homeobox protein, PBX1 . Subsequent treatment with valproic acid (an inhibitor of histone deacetylase) not only improved the efficiency of reprogramming, but also allowed the removal of Klf4 from the reprogramming factors in both mouse and human cell cultures [61, 62]. Furthermore, in mice with p53 null backgrounds, exogenous Klf4 could be eliminated from the set of reprogramming factors . The direct suppression of p53 expression (one of the most famous tumor suppressor genes) by Klf4 was reported previously . Taken together, these data suggest that the role of Klf4 (at least in part) involved suppression of the p53 gene, but in fact an inactivation of the p53 gene may lead to a dramatic increase in the efficiency of iPSC generation, from both mouse and human cells [63, 65-69]. The suppression of p53 may also provide differentiation-resistant characteristics to ES cells from both mice and humans [70, 71]. Similarly, at least in mouse ES cells, p53 directly binds to the promoter region of the Nanog gene and suppresses its expression . In addition, these data suggest that a blockade of p53 occurs not only in established pluripotent stem cells, but also during the reprogramming process.
It has been shown that Klf4 can be replaced with estrogen-related receptor-beta or -gamma for the conversion of mouse embryonic fibroblasts into iPSCs, although there is at present no critical evidence available to support a direct relationship . On the other hand, kenpaullone, a well-known bioactive compound that inhibits cyclin-dependent kinases, can remove the need for Klf4 – but not for other factors – from the requisite set of reprogramming factors.
Whilst Myc transgenes are not essential for the generation of iPSCs, the frequency of reprogramming has been shown to be dramatically reduced when Myc is removed [57, 74]. The effects of c-Myc can be substituted with N-Myc or L-Myc; notably, in human cells L-Myc is more effective for reprogramming compared to c-Myc, although the mechanisms underlying its role in reprogramming remain unclear . The essentiality of Myc in the generation of iPSCs also remains unknown, because most cell types express at least one of the Myc family genes.
The reason why c-Myc is a candidate reprogramming factor is based on the importance of c-Myc in PSCs, which was demonstrated even before iPSCs were first reported. The forced expression of c-Myc allows mouse ES cells to maintain their undifferentiated state even after the withdrawal of LIF, and c-Myc is regulated by Stat3 . Because Myc is not essential for reprogramming, but is important for pluripotency, the demonstration of its relationship to the other factors is difficult. Meanwhile, Wnt can make up for the absence of Myc transgenes . c-Myc is a substrate of phosphorylation by glycogen synthase kinase 3 (GSK3), which is negatively regulated by Wnt activity; the phosphorylated c-Myc is then rapidly degraded via the ubiquitin–proteosome pathway. The treatment of GSK3 with chemical inhibitors, such as CHIR99021 and 6-bromoindirubin-3′-oxime, can enhance the self-renewal of mouse ES cells and the generation efficiency of iPSCs [78-80]. Thus, the inhibition of GSK3, and the subsequent stabilization of c-Myc, most likely serves as an enhancer for both the self-renewal of mouse PSCs and the reprogramming of somatic cells. Whilst the true roles of c-Myc during reprogramming remain unclear , one of its more well-known roles is an enhancement of cell proliferation. Interestingly, p53 improves the efficiency of iPSC generation via an acceleration of somatic cell proliferation during reprogramming . Moreover, by mapping the binding sites of c-Myc on the genome of mouse ES cells, c-Myc was found to be associated with changing and maintaining the chromatin status in mouse ES cells . Because the effects of Myc and p53 are greater than those of other genes and chemicals, the ability to obtain a better understanding of their roles represents a reliable route to revealing details of the reprogramming process.
Nanog is critical for pluripotency not only in mouse, but also in human ES cells . The combination of chromatin immunoprecipitation and a microarray analysis or deep sequencing has revealed that the co-occupation of target sites on the genome by Oct3/4, Sox2, and Nanog is important not only for the activation of ES cell-specific genes, but that this also suppresses the genes associated with development [85-87]. Therefore, Nanog cooperates with Oct3/4 and Sox2 as a key player in the transcriptional network in PSCs, although the downstream targets which are associated with the self-renewal of ES cells are largely unknown. On the other hand, Nanog can also sustain the reprogramming efficiency of somatic cells by fusion with ES cells . Together, these data indicate that Nanog plays a positive role in the recapture of pluripotency.
Lin28 is known as an RNA-interacting zinc finger protein, and is associated with the stability of microRNA (miRNA) . Lin28 can interfere with the maturation of miRNAs and promote their degradation by the uridylation of pre-miRNA [90, 91]. One such Lin28-associated miRNA, let-7, can regulate the translation of several genes including c-Myc, K-Ras, Cyclin D1, and Hmga2 [92, 93]. However, the relationship between the effects of Lin28 in reprogramming and downstream molecules including miRNA and their targets remain unclear.
2.2 The Sources of iPS Cells
At least in mice, iPSCs can be generated from various cell types such as fibroblasts, hepatocytes, pancreatic β-cells, and muscles [94-96]. In the case of humans, fibroblasts, keratinocytes, neural stem cells, peripheral blood cells, and cord blood cells have all been reported as being sources of iPSCs [42, 97-100].
For both therapeutic use and the generation of disease models, the sources of human iPSCs should be collected using minimally invasive procedures to the greatest extent possible. For a patient, a skin biopsy to obtain fibroblasts and keratinocytes can be acquired during surgery, but for healthy subjects great care must be exercised when identifying possible harvest sites, as a biopsy has an associated risk of infection and will leave a visible scar.
An alternative, and very promising, source of iPSCs is that of cord blood, notably because its collection is patient-friendly for the donor, and there are no associated health or cosmetic risks for either the mother or infant. Also, because the donors are neonates, the cells from cord blood will include fewer acquired genetic mutations than would adult somatic cells. Currently, banks of cord blood have been established worldwide, and samples are freely available. One disadvantage with cord blood is that is virtually impossible to determine whether the original donors have continued their healthy status when older. In addition, cord blood may not be suitable for the generation of disease-specific iPSCs, as the blood banks will stock only those cells that have been derived from seemingly healthy donors.
If peripheral blood samples were to be collected from healthy people as a source of iPSCs then, in contrast to cord blood-derived cells and provided that adequate inspections were performed, congenital disorders would pose very little risk. An additional advantage is that peripheral blood can be easily collected from many patients with a minimal degree of invasion, thus allowing the generation of large quantities of cells required for research investigations. Recently, it has been reported that iPSCs can be established from T lymphocytes taken from peripheral blood samples [101-104]. Consequently, peripheral blood represents one of most promising sources of iPSCs, both for cell therapy and for the generation of disease-specific analyses.
Cells derived from dental pulp are also available for the generation of iPSCs [105-107]. Although, typically, extracted wisdom teeth are discarded at the dental clinic, many mesenchymal immature cells remain located within the pulp of the tooth roots. The determination of the HLA-types of 107 dental pulp cell lines led to two independent lines being identified that carried parthenogenetic homozygosity at these loci . Consequently, dental pulp cells were proposed as an excellent source for iPSC banking.
Whilst each of these sources has both drawbacks and advantages, the use of iPSCs is not limited. Paradoxically, the current status of these sources is expected to be resolved through further analyses of iPSCs derived from a variety of sources and the standardization of human PSCs.
2.3 The Generation of iPSCs
Both, mouse and human iPSCs were first generated by using retroviral vectors to introduce reprogramming factors into somatic cells [32, 39]. Retroviruses have at least two advantages for reprogramming:
high efficiency of transduction that allows not only a strong and stable expression of transgenes, but also multiple gene transductions into each cell. These points contribute to easy and reproducible methods for iPSC generation.
The silencing of transgene expression from retroviruses. When the cells have been reprogrammed the retroviral transgenes can be quickly and effectively silenced . This phenomenon is beneficial for iPSCs because the duration of transgene expression not only limits their differentiation potential, but also induces tumorigenicity.
Lentiviral vectors are also available for generation of iPSCs but, unlike retroviruses, the machinery of lentiviral infection is independent of cell division. Lentiviruses are relatively insusceptible to silencing compared to retroviruses; thus, drug-inducible transgene expression systems based on lentiviral vectors have been developed for basic studies of reprogramming [108, 109].
Both, retroviral and lentiviral vectors integrate randomly into the genome of host cells, with the insertion possibly stochastically disrupting tumor suppressor genes or hyperactivating proto-oncogenes. Either of the latter processes can increase the risk of tumorigenicity . Such risks should be precluded with regards to clinical usage, but they are less important in basic research. In order to overcome these issues, other approaches have been developed that can be used to generate integration-free iPSCs. The first two such reports, which involved integration-free mouse iPSCs, employed adenoviral vectors and conventional expression vectors, respectively [110, 111], and both methods included multiple transduction of vectors during iPSC generation. Although the efficiency of iPSC generation was lower than had been achieved using retroviral methods, mouse iPSCs were successfully established, without employing transgenes.
Another possibility would be to remove the transgenes after having established the iPSCs, and a Cre-mediated recombination represents a promising tool for removing transgenes from the iPSC genome. Lentiviral constructs that include loxP sequences in their long terminal regions (LTRs) have been used by one group to generate human iPSCs ; in this case, the established iPSCs were treated with Cre recombinase in order to excise the lentivirus cassettes. However, although this technology resolved the issue of transgene reactivation, one LTR sequence remained. Other groups have utilized a transposon vector, piggyBac, to deliver the reprogramming factors into somatic cells [113-115]. Transposase has activities for both the insertion and excision of transposon vectors by recognition of the TTAA tetra-nucleotide sequence in the host genome. Unlike the Cre-loxP system, however, transposon-mediated excision is seamless.
Several transient expression methods have been developed for the generation of human iPSCs. In a first report, episomal vectors were used in order to transduce reprogramming factors ; here, the vector contained both the replication origin and the nuclear antigen of the Epstein–Barr virus. Because the Epstein–Barr nuclear antigen (EBNA) can recognize the origin sequence and initiate replication, the EBNA vector can self-replicate, so as to allow long-term transgene expression without chromosomal integration. However, the efficiency of reprogramming was more than 10-fold lower than that of integrated vectors, even if additional factors were used to enhance the frequency of iPSC generation. In addition, there is a possibility for chromosomal insertion by these episomal vectors although, unlike viral vectors, they are unable to integrate into the genome. Recently, it was shown that episomal vectors encoding Oct3/4, Sox2, Klf4, L-Myc, Lin28, and short-hairpin RNA against p53 could generate iPSCs with high efficiencies . Consequently, this technology is expected to become a standard for the generation of integration-free iPSCs for clinical applications.
The Sendai virus has also been used for the establishment of human iPSCs [103, 118, 119], there being no risks of the virus becoming integrated into the genome because of its RNA-based structure. The most noteworthy characteristic of the Sendai virus is its high transduction efficiency into T lymphocytes ; hence, when the virus is used to deliver reprogramming factors, only a very small peripheral blood sample (<1 ml) is required for the generation of human iPSCs. Notably, as the frequencies of iPSC establishment with Sendai viruses are comparable to those using retroviruses, this technology may offer promising prospects for the generation of clinical-grade iPSCs.
The latest technology in the field of reprogramming involves the transduction of reprogramming factors via the administration of synthetic messenger RNA . In this case, the RNA employed is modified to overcome its instability, and to avoid inducing an interferon response in the transduced cells. It has been claimed that the efficiency of such iPSC generation was 10-fold higher than with viral methods. Although the risk of insertion into genomic DNA should be very low (because the strategy employs RNA), the method requires multiple transfections to be conducted daily over a period of two weeks. Therefore, further modifications of the procedure are required not only to improve the quality of the iPSCs generated, but also to make the process less stressful to the cells by reducing the number of transfections.
At present, it is difficult to determine which of the above-described procedures is the best, because the most important characteristics of the iPSCs are heavily dependent on their intended purposes. In addition, direct comparisons between iPSC clones established by various methods should be carried out to determine whether the different methods provide equivalent cell populations.
3 Application of iPS Cells
3.1 iPS Cells as Disease Models
Human ES cells with genetic disorders such as cystic fibrosis and Huntington's disease have been established [121-123]. However, although these represent powerful tools to aid in the understanding of the pathogenesis of these diseases, their derivation and use is not always permitted. If somatic cells derived from patients carrying genetic mutations are used, then the resulting iPSCs will inherit the same mutations. Subsequently, iPSCs carrying such mutations would be expected to reproduce the disease-specific phenotypes, either before or after differentiation. This would lead to patient-derived iPSCs, such as ES cells, becoming powerful tools for providing an understanding of disease pathogenesis. In fact, models of many diseases have already been generated, based on the premise that the quality and/or safety of iPSCs is less important for research investigations than for clinical applications.
Disease-specific iPSCs were first derived from two patients with familial amyotrophic lateral sclerosis (ALS) . In this case, in order to assess the degree of neural degeneration in ALS patients, attempts were made to differentiate patient-derived iPSCs into spinal motor neurons and glial cells. Although neural degeneration could not be reproduced in vitro by using ALS-iPSCs, large quantities of differentiated cells carrying mutations derived from the donor were also obtained from iPSCs, including cells derived from older patients. A second report described the generation of iPSCs derived from patients with diseases such as adenosine deaminase deficiency-related severe combined immunodeficiency, Shwachman–Bodian–Diamond syndrome, Gaucher disease type III, Duchenne and Becker muscular dystrophy, Parkinson disease, Huntington disease, juvenile-onset type 1 diabetes mellitus, Down syndrome, and Lesch–Nyhan syndrome . Subsequently, many reports have been made detailing the generation of disease-specific iPSCs .
The in vitro reproduction of disease phenotypes was first accomplished using iPSCs derived from a patient with spinal muscular atrophy (SMA) . When using motor neurons differentiated from patient-derived iPSCs, not only was a low expression of the SMN protein (the cause of SMA) observed, but also specific deficits such as a reduction in the number and size of motor neurons. Consequently, the effects of known drugs (e.g., valproic acid and tobramycin) on the accumulation of SMN protein were evaluated. Taken together, these data confirmed that the reproduction of pathological conditions by using patient-derived iPSCs was indeed possible, but dependent on the disease involved. Nonetheless, the potential application of disease-specific iPSCs for drug screening was confirmed by the results of these studies.
Similar outcomes have been demonstrated for other diseases , an example being familial dysautonomia (FD), which is caused by splicing abnormalities of the IKBKAP gene, resulting from a point mutation. With the main characteristic of FD being a loss of autonomic and sensory neurons, defects were observed in neurogenic differentiation and migration of neural crest precursor cells that had been differentiated from patient-derived iPSCs. Moreover, treatment with a candidate molecule corrected the aberrant splicing of the gene. A pathological reproduction with patient-specific iPSCs has also been reported for myeloproliferative disorders, as well as for LEOPARD syndrome, Rett syndrome, long QT syndrome, and other disorders [129-133]. Although, in the past, the collection of sufficient amounts of target cells from the patients in order to conduct analyses was near-impossible, today plentiful quantities of cells are available by using patient-derived iPSCs – at least as far as finances permit.
3.2 iPSCs for Regenerative Medicine
Because PSCs can, in theory, differentiate into all cell types of the body, future applications for cell therapy are to be expected, although it remains unclear as to when ES cells and/or iPSCs might become effective for cell therapy. The most common issue preventing the clinical use of ES cells and iPSCs is the risk of teratoma formation following transplantation. It is recognized that any residual undifferentiated cells present in differentiated cell cultures used for transplantation can cause a teratoma, and should be removed before use. Hence, effective methods for the removal of undifferentiated cell contamination (e.g., flow cytometry), as well as more efficient procedures for differentiation, are currently being developed. During late 2010, two U.S. companies – Geron and Advanced Cell Technology (both of which have invested a great deal of time and labor to confirm the safety of ES cell-derived differentiated cells) – announced the initiation of clinical trials that involved the transplantation of cells derived from ES cells for spinal cord injury and muscular degeneration, respectively. This development represented a major breakthrough, as well as having important implications for the availability and acceptance of PSCs for therapeutic use.
Compliance with Good Manufacturing Practices (GMPs) is essential for the clinical use of PSCs. One goal to fulfill these criteria is to achieve a xeno-free culture, but this is no longer possible for the already-produced ES cell lines. The removal of any animal-derived components from cell culture is far from straightforward, there being many important materials (e.g., albumin, insulin, and trypsin) present in the cultures. Nevertheless, the establishment of iPSCs under xeno-free conditions has been reported, based on the recent development of animal-free media for PSC culture [134, 135].
Typically, human ES cells and iPSCs are maintained on feeder cell layers, with mitomycin C-treated or gamma-irradiated primary mouse embryonic fibroblasts having been widely used for such purpose. In an attempt to eliminate any animal components, the use of human fibroblasts as feeder cells has been examined on a speculative basis [136-139]. Isogenic iPSCs can be supported by using autologous fibroblasts as the feeder cells, since treatment with mitomycin C and transduction with reprogramming factors can cause fibroblasts to be changed into feeder cells and iPSCs, respectively. Clearly, before putting a xeno-free system to practical use for the generation of clinical-grade iPSCs, the stabilities of such cells should be strictly observed on a long-term basis.
Among other hurdles recently encountered regarding clinical applications, the variability of differentiation potentials remains a common issue for both ES cells and iPSCs . The original somatic cells may have epigenetic memories that occur as an iPSC-specific phenomenon, at least in the case of the mouse [141, 142]; moreover, these memories – which might include DNA methylation and gene expression – can affect the differentiation potentials of iPSCs . In addition, certain reports have identified various iPSC-specific genetic abnormalities [144-146] such that, because the sources of iPSCs are somatic cells, the risks of aberrant genetic errors may be higher than are observed in embryo-derived PSC lines. Clearly, whilst the question of whether (or not) these iPSC-specific elements may affect the safety of cell therapy with iPSCs remains to be elucidated, any such problems must be fully resolved before clinical trials can be undertaken.
The phenomenon of reprogramming, which can be triggered by just four transcription factors, envelops a mysterious aura as to whether it is special, or not. As the efficiency of reprogramming somatic cells into iPSCs is typically <1% , the questions to be raised then is, “What are the other 99% of cells doing?” Whilst some cells die, and others probably senesce, even so the frequencies of reprogramming are too low. But then again, 1% is better than the rate of differentiation of PSCs into pancreatic cells and kidney cells, which is currently more difficult to achieve than is the generation of iPSCs.
Some twenty years before the “birth” of iPSCs, it was demonstrated in a groundbreaking study that MyoD expression alone could convert fibroblasts into myoblasts , while the direct reprogramming of somatic cells to specific lineages, such as pancreatic β-cells, neural cells, and cardiomyocytes, was reported during the next few years [149-151]. Together, these advanced technologies have provided an ability to avoid the risks of teratoma, as the cells are produced without PSCs along the path. Moreover, the accumulation of such knowledge suggests that, even after differentiation, these cells are still astonishingly flexible. Indeed, it appears that, in time, it will become possible to translate reprogramming into direct differentiation in pluripotent cells. Such natural flexibility may underlie the low frequencies of cells that become iPSCs, as these are able to adapt naturally to different conditions, without undergoing any major change in their phenotype.
Since the “dawn of reprogramming” more than 50 years ago, the major progress and rapid advances encountered in this field have been tightly associated with investigations into ES cells. It is likely that direct reprogramming with defined factors will soon be achieved, as the means to characterize the machinery that underlies pluripotency are inevitably discovered. Hopefully, within the next decade the use of PSCs should transition from theoretical knowledge to actual utilization. Clearly, iPSC techniques will undoubtedly increase the overall knowledge in this field, and thereby contribute positively to a new view of stem cell technologies as a whole.
The authors thank the members of the Yamanaka laboratory and CiRA. They are especially grateful to Dr Yoshimi Tokuzawa for the sharing of data, to Haruka Hasaba, Rie Kato, Eri Nishikawa, and Yuko Ohtsu for administrative support, and to Nanako Takizawa for providing the image of the iPSCs.
In the stem cell field, this previously indicated the conversion of differentiated cells into undifferentiated cells. More recently, however, it is difficult to distinguish between terms such as de-differentiation and trans-differentiation.
The ability of germ cells to differentiate into all three germ layers such as endoderm, mesoderm and ectoderm.
Pluripotent stem cells derived from somatic differentiated cells by reprogramming factors and/or small molecules.
- 1312010) A model for neural development and treatment of Rett syndrome using human induced pluripotent stem cells. Cell, 147, 523–539., , , et al. (
- 1342009) Reprogramming of human fibroblasts to induced pluripotent stem cells under xeno-free conditions. Stem Cells, 28, 36–44., , , et al. (