Regenerative Medicine, Invitrogen Corp., Carlsbad, California, USA
Neurosciences, Johns Hopkins School of Medicine, Baltimore, Maryland, USA
Maureen L. Condic, Ph.D., University of Utah School of Medicine, Department of Neurobiology and Anatomy, 401 MREB, 20 North 1900 East, Salt Lake City, Utah 84132-3401, USA. Telephone: 801-585-3482; Fax: 801-581-4233
The development of personalized pluripotent stem cells for research and for possible therapies holds out great hope for patients. However, such cells will face significant technical and regulatory challenges before they can be used as therapeutic reagents. Here we consider two possible sources of personalized pluripotent stem cells: embryonic stem cells derived from nuclear transfer (NT-ESCs) and induced pluripotent stem cells (iPSCs) derived from direct reprogramming of adult somatic cells. Both sources of personalized pluripotent stem cells face unique regulatory hurdles that are in some ways significantly higher than those facing stem cells derived from embryos produced by fertilization (ESCs). However, the outstanding long-term potential of iPSCs and their relative freedom from the ethical concerns raised by both ESCs and NT-ESCs makes direct reprogramming an exceptionally promising approach to advancing research and providing therapies in the field of regenerative medicine.
Disclosure of potential conflicts of interest is found at the end of this article.
Author contributions: M.L.C.: conception and design, financial support, administrative support, manuscript writing, final approval of manuscript, production of artwork; M.R.: conception and design, financial support, manuscript writing, final approval of manuscript
Immune rejection of transplanted tissues presents a serious challenge for regenerative medicine. Just as for organ transplant, if stem cell-derived tissues are not a close major histocompatibility complex match to the patient, lifelong immune suppression is required (reviewed in ). Immune rejection is an even more serious concern for potential stem cell-based therapies than for conventional tissue transplant, because the dispersed integration of transplanted stem cell derivatives precludes their surgical removal should immune issues arise.
Estimates of the number of embryonic stem cell lines (Fig. 1A; hereafter, embryonic stem cell lines derived from embryos produced by fertilization are referred to as ESCs) required to provide a good match for specific patient populations vary enormously, from several hundred [2, 3] to several million [1, 4] lines, depending on the genetic diversity of the population under consideration and on the definition of what constitutes an “acceptable” degree of immune mismatch. In all likelihood, the number of ESC lines required to treat the genetically diverse U.S. patient population would be large  and would far exceed the number of lines that could be generated from excess embryos currently available from assisted reproductive clinics (in vitro fertilization [IVF] “spares”; http://www.rand.org/news/press.03/05.08.html). Producing ESCs from donated oocytes and sperm could expand the genetic diversity of available stem cell lines but would not substantially reduce the number of lines required for therapeutic purposes.
In contrast to ESCs derived from embryos produced by fertilization, patient-specific stem cell lines would offer compatibility at both major and minor antigenic sites. Thus, developing an efficient method of generating patient-matched pluripotent stem cells for possible therapeutic applications is a key component of developing effective stem-cell therapies. Here we compare the two major methods of generating of patient-matched pluripotent stem cells: somatic cell nuclear transfer (SCNT) to produce stem cells from cloned embryos (NT-ESCs) and direct reprogramming to yield induced pluripotent stem cells (iPSCs).
There have been various attempts over the last 10 years to obtain patient-matched pluripotent stem cells by reprogramming adult cells to an embryonic state (reviewed in ). The feasibility of reprogramming was first (and most dramatically) demonstrated by SCNT or cloning (Fig. 1B; ). Although SCNT has proven to be more challenging in primates than in other mammalian species, two embryonic stem cell lines have been isolated from cloned macaque monkey blastocysts , and SCNT-derived human blastocysts have also recently been reported , suggesting that it will soon be possible to derive patient-specific lines from nuclear-transfer human embryos.
As an alternative to reprogramming by nuclear transfer, Takahashi and Yamanaka  recently demonstrated that adult mouse skin cells can be reprogrammed into a state that is similar to an embryonic stem cell by expression of only four factors (Oct4, Klf4, Sox2, and c-Myc). Such reprogrammed cells (Fig. 1C) have been termed iPSCs. Rapid advances in the reprogramming technique have generated pluripotent cells that are not only similar to mouse ESCs, but functionally identical: capable of self renewal and germ-line transmission [11, –13]. Recent work has demonstrated that similar to ESCs, mouse iPSCs are able to differentiate into multiple cell types in vitro [14, 15]. Four independent groups have recently produced human iPSCs[16, , –19], making patient-specific pluripotent stem cell lines a possibility for the first time.
Producing patient-specific stem cell lines by direct reprogramming is relatively straightforward, albeit inefficient; initial studies reported that less than 1% of adult cells are transformed, although recent work reports increased efficiency when additional factors are used [20, 21]. A strong advantage of this approach compared with SCNT is that direct reprogramming circumvents not only the ethical concerns over the use of human embryos for research but also the ethical and practical concerns raised by human egg donation [6, 22]. However, iPSCs present significant technical challenges and regulatory issues that must be addressed before they can be used for treatment of human patients, and in several respects, these challenges are distinct from those presented by SCNT.
The scientific hurdles that must be overcome in developing human ESC-based therapies have been extensively reviewed . Recently, Halme and Kessler  (also described in ) have considered the regulatory requirements likely to be applied to ESC-based therapies in light of four important questions: Does the donor pose a risk of transmitting infectious or genetic diseases? Does cell-processing pose a risk of contamination or damage? What is the purity of the cells in the final product? Will the product be safe and effective in vivo? These same questions must be asked of iPSCs and NT-ESCs if they are to be developed as therapeutic agents. Here we consider the hurdles confronting the development of patient-specific stem cells for human therapies and contrast the challenges facing iPSCs to those facing NT-ESCs (Table 1).
Table Table 1.. Relative advantages and disadvantages of current stem cell approaches
ESCs can be obtained from either excess IVF embryos (IVF spares) or from embryos produced by fertilization of donated oocytes.
aFor all three methods, expansion, subcloning, and characterization of the cells to establish a stem cell line takes at least an additional 3 weeks.
Because SCNT requires the use of donated oocytes, NT-ESCs would be subject to regulatory requirements for assessment of donor-associated risk, and oocyte donors would need to be screened for both genetic and transmissible disease. Moreover, although NT-ESCs would be patient-specific, in most cases they would not be autologous (i.e., they would have nuclear DNA identical to that of the patient, but the mitochondrial genes would be from the oocyte donor). Because mitochondrial gene defects cause a large number of diseases [26, –28], egg donors would need to be screened to ensure the safety of donated oocyte mitochondria. It is also likely that mitochondria from the patient's somatic cell would be retained following nuclear transfer (NT), resulting in mitochondrial heteroplasmy. Such multiparental mitochondria have already been demonstrated in ooplasm transfer cases [29, –31] and in sheep derived from NT , making the health risks associated with heteroplasmy (mitochondrial dysfunction due to aberrant interactions between mismatched nuclear and mitochondrial genes)  a reasonably legitimate concern.
One of the strong advantages of direct reprogramming is the ability to generate patient-specific cell lines with relative ease [16, 17]. Because such patient-matched lines are autologous, they would be subject to significantly simpler donor screening requirements; donors would not have to be rigorously screened for genetic or infectious disease if the reprogrammed cells were intended only for transplant back to the patient from whom they originated, although the quality of the cells will still need to be evaluated, as discussed below.
Cell Processing-Associated Risks
Generating both iPSCs and NT-ESCs requires significant manipulation in the laboratory. Thus, both types of pluripotent cells would be subject to the provisions of the Code of Federal Regulations (CFR) governing transplant of laboratory-processed human cells and tissues (21 CFR 1271) and would have to be screened for contamination that may have been introduced during all stages of preparation and processing.
An added concern regarding iPSCs it that they are currently generated by retroviral insertion of at least three gene products, with an estimated 20 integration events for each cell-line founder , although recent results have suggested that this number can be lowered considerably (T.M. Townes, unpublished data). The use of retrovirally transformed cells is permitted in human clinical trials under established NIH guidelines (http://www4.od.nih.gov/oba/rac/guidelines/guidelines.html), yet such genetically modified cells carry an increased risk for patients. As transformed cells, iPSCs would be regulated by both the U.S. Food and Drug Administration and also by the NIH Office of Biotechnology Activities Recombinant DNA Advisory Committee. Consequently, developing potential therapies using iPSCs would carry increased technical and regulatory challenges associated with establishing the safety of the methods used for virus production and determining the safety of the site of viral insertion and the safety of the inserted genes themselves.
The additional regulatory concerns facing iPSCs may ultimately be obviated by modifications of the reprogramming technique that avoid the use of retroviral vectors (i.e., nonintegrating vectors or non-DNA-based reprogramming). Indeed, recent work has shown that transgene expression is required only for the first 10–12 days of the reprogramming process [34, 35], suggesting that expression of reprogramming factors from episomal viruses may be sufficient to transform cells. Moreover, treating cells with molecular inhibitors of either the G9a histone methyltransferase, specific protein kinases , or histone deacetylase  can both increase reprogramming efficiency and reduce the number of transforming factors required. However, at the current time, the use of retroviral vectors presents additional technical and regulatory challenges for iPSCs, compared with NT-ESCs.
A possible application of both of iPSCs and NT-ESCs would be to derive patient-specific cells and subsequently use genetic engineering methods to correct an inherited genetic defect. Genetically corrected cells could then be differentiated into the appropriate cell type and transplanted back to the patient to treat (or correct) the underlying genetic disease. The feasibility of this approach has been tested in animal models for both NT-ESCs  and iPSCs , with various degrees of success. Such a gene-correction approach would involve a second stage of genetic manipulation and therefore raise additional safety concerns, requiring additional regulation and testing.
Cell Type and Cell Purity Risks
Both SCNT and direct reprogramming use nuclei that have aged and therefore have undergone significant developmental changes. These changes include alteration in telomerase activity, shortening of telomeres, tissue- and lineage-specific methylation, imprinting, and other epigenetic modulation. DNA damage is also likely to have occurred in aged nuclei, and it has been shown that such damage varies with the exposure to environmental agents and with tissue type . Given that each of these changes can have potential deleterious effects on the behavior of cells generated from such nuclei, it is likely that both NT-ESCs and iPSCs will require much more stringent scrutiny than ESCs derived from embryos produced by fertilization. Moreover, it is well established that ESCs accumulate genetic and epigenetic abnormalities over time in culture [41, –43], and this problem is likely to face iPSCs and NT-ESCs as well. Therefore, pluripotent lines derived from SCNT or reprogramming would need to be tested for possible effects both of aging and of instability in the laboratory (i.e., DNA damage, changes in imprinting, and epigenetic changes).
Patient-specific lines from either reprogramming or SCNT would not be universal therapeutic agents. Consequently, regulatory authorities are likely to treat each line as a separate product, and testing will have to be rigorous for each line generated. Specific concerns about the safety of NT-ESCs and iPSCs that stem from the methods used to produce these cells are discussed below.
The efficiency of generating NT-ESCs is typically low (between 1% and 10%), with the live birth rate of cloned animals being even lower [44, 45]. Recent work (S. Wakayama, personal communication) suggests that the live birth rate can be increased to nearly 40% by serial transfer of nuclei (first producing a NT-ESC line and then using the NT-ESC nuclei for a second round of SCNT), although other studies have not seen an improvement in developmental capability using this approach [46, 47]. Although human NT-ESC lines have not yet been derived, the efficiency of producing genetically normal lines is likely to be relatively poor compared with the efficiency of generating lines from embryos produced by fertilization. In the single report of successfully deriving NT-ESCs from nonhuman primates, two lines were derived from 304 monkey oocytes (0.7% overall derivation efficiency), with one line being karyotypically abnormal . Work in other species also raises concern regarding the normalcy of NT-ESC lines. Although some studies have demonstrated that NT-ESC lines are indistinguishable from lines derived from fertilized oocytes [48, 49], others have documented significant abnormalities in cloned animals and in NT-ESC lines [42, 50, , –53]. For patient-matched NT-ESCs, it would therefore be necessary to establish a comprehensive screening procedure to select lines with normal karyotype and normal epigenetic characteristics.
An added concern for both NT-ESCs and iPSCs is large offspring syndrome (LOF). LOF is one of several abnormalities that lead to a reduction in the number of viable births following successful implantation. LOF is associated with a range of embryo manipulations in culture [54, –56], including SCNT [57, –59]. The basis of LOF appears to be culture-induced epigenetic changes that are not readily detected at early stages in vitro . It is therefore a reasonable concern that such epigenetic abnormalities will affect NT-ESC-derived cells. Although detailed testing of epigenetic state has not been done for multiple iPSC lines, the preliminary results are encouraging [11, 61, 62]. Nonetheless, epigenetic changes may also affect iPSCs. Importantly, assessing which iPSC or NT-ESC line is epigenetically normal may prove difficult but would be a necessary regulatory requirement.
The nature of somatic cell reprogramming raises unique concerns regarding iPSC type and cell purity. Recent work has shown that integration of retrovirus into specific sites in the genome is not required for reprogramming . Therefore, as long as retroviral vectors are used for reprogramming, each iPSC line will have a unique genetic constitution, and each iPSC line will need to be characterized independently to determine its precise properties. Although NT-ESCs from mouse have been extensively compared with conventional mouse ESC lines (as noted above), these data have not yet been obtained for mouse (or human) iPSCs. It will be important to determine whether every specific iPSC line is indeed ESC-like, to make use of the large number of data obtained from blastocyst-derived ESCs. Alternatively, safety and efficacy data, as well as stability and heritability of the induced pluripotent state and absence of bias in lineage differentiation, will have to be generated independently for every iPSC line.
In light of the low efficiency of reprogramming, an additional concern regarding iPSCs is the method used for selecting reprogrammed cells. Currently, mouse iPSCs have been selected on the basis of activation of a neomycin resistance gene inserted into the Nanog or Oct4 locus [11, –13] or by morphological criteria that can be applied to cells without introducing selectable markers that allow for identification of the altered cells [64, 65]. All four laboratories isolating human iPSCs based their selection on morphology [16, , –19], but this approach will need to be tested and validated prior to approval in any good manufacturing practice protocol.
Safety and Efficacy In Vivo
A challenge common to all pluripotent cells is the difficulty of differentiating them into stable, mature cell types. ESCs have been studied since the early 1980s, and yet it has proven challenging to induce them to form functional replacement tissues that are therapeutically useful for the treatment of disease and injury. Although ESCs are clearly capable of generating multiple tissue types in chimeric embryos and in tumors, producing clinically useful cells in the laboratory has been not been routinely achieved. iPSCs and NT-ESCs are likely to face similar challenges.
In most cases, survival of tissue derived from pluripotent stem cells is poor following transplantation. Although there has been some degree of success for long-term survival of hematopoietic progenitors generated either from NT-ESCs [38, 66] or from iPSCs , success with blood appears to be the exception rather than the rule. For example, a recent study using monkey ESCs in a Parkinson's disease model showed only 1% survival of transplanted neuronal cells at 14 weeks . Similar results have been seen for multiple tissues differentiated from ESCs, including (among others) hepatocytes , cardiomyocytes , and pancreatic islet cells . Although data from NT-ESCs and iPSCs are limited at this time, it is likely that patient-specific stem cells will face similar challenges in terms of producing tissues that survive long-term following transplantation. The efficacy in vivo of tissue derived from pluripotent stem cells therefore remains questionable for most cell types other than blood.
Similarly, safety in vivo presents a significant challenge for development of pluripotent-stem cell therapies. It is well established that undifferentiated pluripotent stem cells form tumors when transplanted to mature tissues, whereas stably differentiated cells do not. However, numerous studies have shown tumor formation even in cases where ESCs have been subjected to rigorous differentiation protocols, suggesting that the purity and stability of differentiated cell populations can be difficult to predict or to control . For example, human ESCs differentiated into dopaminergic neurons and transplanted into a rat model of Parkinson's disease showed substantial phenotypic instability and reverted to undifferentiated, tumorigenic neuroepithelial cells . Similarly, animals transplanted with mature dopaminergic neurons generated from ESCs showed frequent tumor-related deaths unless the transplanted population was first presorted to remove sox1-positive neuronal progenitors that persisted despite an extended differentiation protocol . Indeed, predifferentiation of human ESCs for up to 23 days in vitro reduced but did not entirely eliminate tumor formation . Teratoma formation has been observed for ESCs differentiated along multiple lineages, including (among others) hematopoietic , pancreatic [70, 76], hepatic [68, 77], and retinal  lineages.
Together, these results suggest that efficacy and safety in vivo have not yet be rigorously established for most tissues derived from ESCs. Although similar data are not currently available for human iPSC or NT-ESC lines, it is fair to assume that until contrary data have been accumulated, the transplant issues identified with ESCs will apply to NT-ESCs and iPSCs as well. Indeed it is likely that the frequency of tumor formation will be higher with patient-specific lines, as there is no immune mismatch that would co-opt the immune system in rejecting cells with immature or inappropriate phenotypes. Increased numbers of ESC-derived tumors were noted in SCID mice when human fetal tissue containing human embryonic stem cells (hESCs) was transplanted, compared with mice receiving hESCs alone, suggesting that the lack of immune rejection can enhance tumor formation [79, 80]. In addition it has been shown for mouse iPSCs that some of the genes used to induce transformation are oncogenic; chimeric mice containing iPSC-derived tissues formed head and neck tumors that were attributed to the myc gene used to reprogram fibroblasts into iPSCs. Subsequent experiments have suggested that this may not be true when other cell types are used for reprogramming  and that reprogramming can be accomplished without the use of the myc gene [18, 81], yet increased propensity toward tumor formation remains a concern for iPSCs, as they are currently generated.
Although patient-specific stem cells (derived both from SCNT and from direct reprogramming) have clear advantages over ESCs derived from embryos produced by fertilization, the very nature of the processes used to create such lines will create additional regulatory hurdles. Embryonic stem cells produced by SCNT would be patient-specific and thereby avoid the serious complications of immune rejection. However, because NT-ESCs require the use of human eggs, they would be subject to regulatory requirements governing egg-donor selection and screening. In addition, NT-ESCs raise serious ethical and practical issues that are not easily resolved. Although many Americans accept the generation of stem cell lines from excess IVF embryos (embryos that will “die anyway”), the production of embryos intended solely for research (through either fertilization or NT) has far less public support (for example, a poll conducted in early 2008 showed that 47% are in favor of using excess IVF embryos for research, whereas only 18%–30% support creation of embryos solely for research purposes; see questions 9 and 13–15 in ). The medical risks to women associated with egg donation also raise significant ethical concerns for NT-ESCs (as well as for ESCs produced by fertilization of donated oocytes). Human cloning itself presents ethical issues not raised by other forms of stem cell research. Finally, because generation of human ESCs and NT-ESCs requires destruction of human embryos, many oppose this research on either ethical or religious grounds . On a purely practical level, the Human Embryo Research Ban (the Dickey Amendment), which has been continuously authorized by Congress since 1995, prevents the use of federal funds for generation of human NT-ESC lines (text for the current fiscal year is available under section 509 of H.R.2764; [Section 509 of Division G of the Consolidated Appropriations Act of 2008 (Public Law 110-161), 121 Stat. 2209]), and this situation seems unlikely to change in the foreseeable future.
In contrast, iPSCs offer several advantages over both conventional ESCs and NT-ESCs, including the relative ease with which multiple cell lines can be obtained from a single biopsy, the eligibility of iPSC research for federal funding, and the freedom from the ethical and practical burdens associated with either embryo destruction or egg donation [6, 22]. Since iPSCs are entirely patient-derived, the burdens of demonstrating immune compatibility and donor safety are much lower than for ESCs or NT-ESCs. However, it is unlikely that other regulatory requirements can be lowered. Because the use of iPSCs is a new and still unproven technology that incorporates both genetic manipulation and pluripotent cell propagation, the regulatory hurdles for iPSCs will be high, particularly in the light of previous experience with gene therapy.
Despite the remaining technical and regulatory challenges, none of the hurdles facing iPSCs are insurmountable. Direct reprogramming technology represents a fundamental breakthrough that provides a novel path for generating multiple differentiated cell types with an isogenic gene profile that would be virtually impossible to obtain in therapeutically relevant numbers otherwise. Direct reprogramming thus offers an exceptionally promising method for advancing research and providing therapies in the field of regenerative medicine.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
M.R. gratefully acknowledges the contributions of the Packard Foundation, California Institute of Regenerative Medicine, and Amyotrophic Lateral Sclerosis Association. M.R. is an employee of Invitrogen Corp., which funds research of the stem cell group. This work was supported by grants from the Craig H. Neilsen Foundation and NIH Grant R01-NS048382 (to M.L.C.).