Extraordinary advances in pluripotent stem cell research have initiated an era of hope for regenerative strategies to treat human disease. Besides embryonic stem cells, the discovery of induced pluripotent stem cells widened the possibility of patient-specific cell therapy, drug discovery, and disease modeling. Although similar, it has become clear that these two pluripotent cell types display significant differences. In this review, we explore current knowledge of the molecular and functional similarities and differences between these two cell types to emphasize the necessity for thorough characterization of their properties as well as their differentiation capabilities in the pluripotent state. Such comparative studies will be crucial for determining the more suitable cell type for future stem cell-based therapies for human degenerative diseases. STEM CELLS 2012;30:10–14
Since the advent of human embryonic stem cells (hESCs) in 1998 , stem cell research has been developing at a breathtaking pace. The pluripotent nature of these cells renders them the ability to differentiate into any cell type—including into those with therapeutic potential—after practically unlimited self-renewal in the stem cell state. ESCs hold enormous promise as tools for understanding normal development and disease, and just as importantly, for cell therapy applications to treat devastating and currently incurable disorders, such as spinal cord injury, neurological disease, blindness, and type 1 diabetes. On the other hand, the use of human embryos to derive these cells has ignited a diverse ethical debate rooted in the complex background of human historical, cultural, and religious differences. In this review, we are not going to pursue a discussion of ethical issues but rather focus on the potential of pluripotent cells in general to cure disease and eliminate human suffering.
Following the characterization of the first hESC lines in the late 1990s, standard protocols have steadily been developed to accommodate future clinical applications, including maintenance of these cells in the absence of animal-derived culture components. Furthermore, guided by insights gained from decades of research on the molecular genetic basis of mammalian development, detailed protocols have emerged for the reproducible generation of enriched populations of various differentiated cell lineages in mouse and human, including neurons, cardiomyocytes, and hematopoietic cells . Numerous preclinical animal studies have demonstrated that the differentiated derivatives of ESCs can provide functional replacements for diseased tissues, such as for Parkinson's disease , and clinical trials are currently underway for hESC-based cellular therapy for spinal cord injury and macular degeneration in the U.S. and U.K. .
Six years ago, Takahashi and Yamanaka astonished the world by showing that enforced expression of four key transcription factors, Oct4, Sox2, Klf4, and c-Myc, can reprogram mouse somatic cells such as fibroblasts to pluripotency, and achieve similar developmental potential as ESCs, without the requirement for an embryo . They named these new cells “induced pluripotent stem cells” or iPSCs. A year later, several groups, including Yamanaka's, reported the successful generation of iPSCs from human somatic cells [6, 7]. With this step forward, a race was initiated. The expectation that iPSCs will offer the same therapeutic potential as hESCs and the robust, reproducible method of deriving iPSCs have spawned hundreds of studies addressing in vitro disease modeling and cell therapy strategies in preclinical animal models. Indeed, iPS cell lines have now been generated from patients of several monoallelic and complex genetic disorders (reviewed in ). These developments have brought the field a hopeful step closer to the promises of in vitro disease modeling, disease-specific pharmacological treatment testing, and in some cases individualized cell replacement therapy. Several examples of the differentiation of disease-specific iPSCs into the cell types that are implicated in the disorder's pathogenesis have been reported, and therefore this technology is particularly attractive for the diseases for which animal models are either not available or do not accurately represent the human disease etiology.
The following question thus arises can iPSCs replace ESCs in clinical application and disease modeling? Although all of us would likely welcome a “yes,” it is clear that we are not yet in the position to answer this question, despite the unprecedented speed of development in the iPSC research area. Our understanding of the full characteristics of iPSCs and the mechanistic details of reprogramming to pluripotency is far from complete . Although several analyses indicate that iPSCs share many key properties with ESCs including morphology, pluripotency, self-renewal, and similar gene expression profiles, there are just as many published examples that point out their differences. There is an urgent need to shed more light on the complex landscapes of safety, efficacy, economy, and disease coverage associated with clinical use of these new and exciting pluripotent cell types. At the current state of knowledge, we promote the view that parallel, direct studies on iPSCs and ESCs, including detailed characterization of mechanisms of pluripotency and differentiation are required to make the promise of stem cell-based therapeutics for human disease a reality.
During expansion and prolonged passage, hESC lines frequently acquire abnormal karyotypes such as trisomy 12 and 17 [10, 11] as well as genetic amplification at 20q11.21, which has been associated with oncogenic transformation [12, 13]. iPSC lines are also subjected to similar selective forces, resulting in cell culture adaptation frequently manifested in karyotypic abnormalities . So in this respect, ESCs and iPSCs are likely equivalent.
The distinct cellular origin, however, could lead to significant differences between these two pluripotent stem cell types. ESCs are derived from the inner cell mass of a blastocyst-stage embryo before the soma and the germ cell lineages separate. iPSCs, however, are derived from somatic cells.
It was August Weismann who in 1889 first recognized that in most organisms, the somatic and germ cell lineages separate very early in development and pointed out the evolutionary consequences of this separation . Weismann postulated that hereditary information moves only from germ cells to somatic cells. The reverse direction, soma to germ line, would be impossible. Moreover, the genome of the germ cell lineage is passed to the new generation and in this respect it is immortal. On the other hand, the genome of the somatic cells is mortal, as it is discontinued with the death of the organism. Therefore, mutations generated in the soma are not subjected to evolutionary selective forces, such as natural selection or genetic drift. Instead, the immortal germ cell genome remains mostly locked into the germ cell lineage with a brief passing through the very early developmental stages (preimplantation and early postimplantation in mouse and human), prior to the germ-soma separation (Fig. 1).
As evolutionary selective forces act only on mutations in the germ line genome, the expectation is that the strength of genome integrity protection might therefore be different between germ line and soma. A putative differential genome protection could have significant consequence regarding the genome integrity of ESCs versus iPSCs. ESCs derived from the inner cell mass of the blastocyst have never had a journey through a stage in the soma.
John Gurdon's somatic cell nuclear transfer (SCNT) in frogs showed that with experimental manipulation it is possible to return the genome of a somatic cell to the germ line . This discovery was later followed by success in sheep , mice , and numerous agricultural species . SCNT reprograms the somatic cell genome into a totipotent cell state (Fig. 1). As SCNT has become a routine procedure in many mammalian species, it has become evident that cloned animals suffer increased risk of abnormalities ranging from prenatal death to altered development . It is still not completely clear what proportion of these abnormalities is due to incomplete epigenetic reprogramming or due to permanent genetic changes occurring during somatic cell development or during the reprogramming process (see below).
The generation of iPSCs by reprogramming using enforced expression of a finite number of transcription factors is similar in this respect. The genome of a fully differentiated somatic cell is returned to pluripotency, which theoretically includes germ line competence (Fig. 1). Therefore, iPSCs can acquire genetic alterations at two additional phases: during somatic differentiation and during reprogramming. It is likely that none of these phases have developed genome-protecting mechanisms responding to evolutionary pressure.
Several recent studies have demonstrated that the reprogramming process leads to genomic instability and genomic abnormalities, with a notable proportion of lesions mapping to known cancer causative loci [14, 21, 22]. Reprogramming causes genomic copy number variations (CNVs) to occur early in iPSC passage leading to mutations and a mosaic iPSC population [21, 22]. During passage, iPSCs undergo strong selection pressure against most of the mutations and reach a CNV load similar to that of ESCs. Nevertheless, hiPSCs contain de novo mutations that are not detected in hESCs, suggesting that certain mutations are selected for and are advantageous to reprogramming [14, 21, 22]. Taken together, the available data suggest that reprogrammed cells indeed likely pose a greater risk for accumulation of deleterious genomic mutations. Furthermore, when the reprogramming factors are not silenced, iPSCs are predisposed to additional genomic instability . These findings underscore the critical requirement for detailed characterization of the genome integrity of iPSCs in comparison to that of ESCs and the human genome for correct interpretation of experimental results using these cell lines, and also for safe future therapeutic applications.
GENETIC AND EPIGENETIC REGULATION OF THE PLURIPOTENT STATE
Comparisons of iPSCs and ESCs have indicated that major features of the ESC epigenome are reproduced in iPSCs, including genome-wide methylation patterns and the establishment of bivalent histone marks at specific loci [24–26]. However, some analyses of reprogramming in mouse cells have shown that differences in gene expression and differentiation potential are observed specifically in early passage iPSCs and have led to the concept that an “epigenetic memory” of previous fate persists in these cells [27–31]. Epigenetic memory has been attributed to the incomplete removal of somatic cell-specific DNA methylation at regions in proximity to CpG islands known as “shores” [28, 32]. The residual DNA methylation pattern and resulting gene expression of the somatic cell of origin are lost upon continued serial passage of derived iPSCs and after treatment with molecular inhibitors of DNA methyltransferase activity [28, 29] suggesting that epigenetic memory also identifies cells that are incompletely reprogrammed. On the other hand, these findings suggest that cell type of origin could affect results in disease modeling as iPSCs show distinct cellular and molecular characteristics based on the cell type of origin. However, it has been noted that this property may improve the prospects of generating some cell types for cell replacement therapy, in particular for those that are difficult to generate by differentiation from ESCs, including insulin producing pancreatic beta cells .
In agreement with the epigenetic similarity of the two pluripotent cell types, comparative transcriptome analyses using microarray also indicate that hESCs and hiPSCs are highly alike on a global scale, with gene expression patterns clustering together, and separate from the originating somatic cells . iPSCs may retain, however, a unique gene expression signature, including that of microRNAs and long noncoding RNAs [33–37]. In addition, a few studies have noted that some transcriptional differences can also be attributed to latent expression of the four reprogramming factors, to genetic background, and to differences in in vitro microenvironment and handling conditions in different laboratories [24, 38]. These findings collectively suggest that detailed analyses and standardization of reprogramming and cell culture protocols will be required to validate whether small variations in gene expression seen between iPSCs and ESCs have biological significance.
DEVELOPMENTAL POTENTIAL VERSUS DISEASE RISK
As mouse ESCs have the capacity to generate an entire normal adult mouse, they are considered as the gold standard against which all other cell types are compared with respect to pluripotency. The ability to significantly contribute to chimeras is considered the most stringent test of pluripotency for mouse iPSCs. Interestingly, available data suggest that compared with ESCs, only a small percentage of mouse iPS cell lines can contribute to strong chimeras or quite infrequently form completely iPSC-derived animals in tetraploid embryo complementation . Furthermore, the earliest studies on iPSC-derived chimeric mice demonstrated that they were prone to cancer and attributed this property to the re-expression of the c-myc reprogramming factor . C-myc is a well-studied oncogene, and the expression of the other three reprogramming factors has been associated with several forms of human cancer . For this reason, substantial efforts have been made to find reprogramming methods that do not require permanent transgene integrations. During the last 3 years, several such factor delivery methods have been developed using adenovirus, the piggyBac transposon, as well as direct protein transduction among others .
Pluripotency of ESCs and iPSCs, as defined by the ability to differentiate into tissues of all three germ layers, is also assessed using the in vivo teratoma assay, the only pluripotency test available for the study of human pluripotent cells. Detailed pathological characterization of teratomas in immunocompromised mice has recently revealed surprising differences between hESCs and iPSCs. iPSC-induced teratomas were more aggressive, with a shorter latency than ESCs and frequently contained areas with more aggressive teratocarcinoma characteristics . It remains to be determined whether such pathological features can be directly attributed to alterations at the genome level during reprogramming and prolonged passage in vitro. Recent analyses suggest that the pluripotent and tumorigenic capacity of ESCs may be governed by different cell signaling pathways , a property that most likely also applies to iPSCs. This necessitates a thorough molecular understanding of the differences between ESCs and iPSCs with respect to their developmental potential and risk of ill behaving if their derivatives were grafted into an individual.
Ironically, the vast proliferation and tissue differentiation potential of iPSCs and ESCs in vivo is considered to be one of their main obstacles for clinical use. For example, formation of teratoma-like tumors was observed in one of the tests for the efficacy of hESCs in a mouse model of Parkinson's disease and interfered with the ability of grafted cells to restore dopaminergic neural function . Furthermore, a survey of teratoma formation by grafted neural tissue obtained from iPSCs that were derived from different cellular sources and with different methods has identified another important aspect of the safety of cellular therapy. Tumor formation was positively correlated only with the residual presence of undifferentiated cells but, interestingly, not with the presence of c-myc or with other variables in the iPSC derivation process . These reports demonstrate that the elimination of residual pluripotent cells is a major challenge and an issue that is equally potent for ESCs as is for iPSCs. With current protocols, it is very difficult to produce completely pure populations of differentiated derivatives from ESC or iPSC cultures for transplantation. In the future, stringent cell surface marker-based cell separations, or depletion of undifferentiated cells, or modifications of the starting iPSC or ESC populations that permit deletion of undifferentiated cells in vivo will have to be considered.
DIFFERENTIATION AND DISEASE MODELING
For clinical applications, reprogramming is the first step with the ultimate goal being reproducible differentiation and maximum enrichment to specific cell lineages. While this property is established for ESCs, albeit still with technical barriers, very recent studies have begun to address the differentiation capacity of human iPSCs and the functionality of their differentiated derivatives. Although multiple protocols have been developed to derive specific cell types in vitro, there is considerable variability in the efficiency of generating differentiated lineages among independent hESC and iPSC lines . The production of hemangioblast cells and other derivatives occurred at a much lower efficiency from hiPSCs than from hESCs . Similarly, hiPSCs differentiate to neural lineages at a much lower frequency than ESCs regardless of the means of derivation . The molecular signature of iPSCs can be influenced by the cell type of origin, and in one case, can explain this biased differentiation potential . Premature senescence of differentiated endothelial cells and retinal pigment epithelium from iPSCs have also been observed [49, 50] suggesting that the differentiated progeny of iPSCs may also display significant functional differences that could undermine their therapeutic utility. Thus, it is important to consider that genetic or epigenetic features that affect iPSCs during differentiation could also do so after transplantation, generating cells with gene expression patterns or phenotypic characteristics that are different from ESC-derived transplants.
Permanent cell lines of pluripotent ESCs and iPSCs and our increasing ability to direct them into any cell type for therapeutic potential holds enormous promise for future regenerative medicine. ESCs are considered to be the gold standard of pluripotency, while iPSCs offer the development of cells from any adult individual, which advances the possibility of curing devastating degenerative diseases using cell or tissue grafts with perfect histocompatibility match. This potential calls for efforts to characterize and compare the nature of these pluripotent cell types in great detail. Only such deep studies can give us sufficient insight into the potential, efficacy, and safety to reach a decision; which one will be more favorable for future clinical applications. At the current state of knowledge, we are not in a position to make such a decision. The game between ESCs and iPSCs is still on with no obvious indication of the winner.
We thank Kristina Nagy, Peter Tonge, and Samer Hussein for valuable input on the manuscript. The authors acknowledge the support of the Stem Cell Network (Canada) and the Ontario Ministry of Research and Innovation, Genome and Life Sciences (GL2) Program.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.