Amnion-derived stem cells: in quest of clinical applications

In the promising field of regenerative medicine, human perinatal stem cells are of great interest as potential stem cells with clinical applications. Perinatal stem cells could be isolated from normally discarded human placentae, which are an ideal cell source in terms of availability, the fewer number of ethical concerns, less DNA damage, and so on. Numerous studies have demonstrated that some of the placenta-derived cells possess stem cell characteristics like pluripotent differentiation ability, particularly in amniotic epithelial (AE) cells. Term human amniotic epithelium contains a relatively large number of stem cell marker-positive cells as an adult stem cell source. In this review, we introduce a model theory of why so many AE cells possess stem cell characteristics. We also describe previous work concerning the therapeutic applications and discuss the pluripotency of the AE cells and potential pitfalls for amnion-derived stem cell research.


Introduction
Th e emerging fi eld of regenerative medicine requires a reliable cell source in addition to biomaterial scaff olds and cytokine/growth factors. Th e 'cell' is a particularly critical element for cell replacement therapies in order to provide a safe and suffi cient cell supply for clinical applications. Eff orts to search for an adequate cell type and cell source have been conducted and have continued along with the discussions for their use in clinical application.
Th ere are many potential cell sources for regenerative medicine, including bone marrow-derived mesenchymal stem cells, tissue-specifi c progenitor cells, embryonic stem (ES) cells, and induced pluripotent stem (iPS) cells. Although their biological potentials have been demonstrated, none of these cells is widely accepted as a defi nitive cell source for clinical applications. Each cell type possesses diff erent advantages as well as limitations for their use, such as safety or availability. It will be helpful to search for a potential stem cell source from the perspective of its potential for clinical application. What is the sine qua non of the cells for clinically applicable regenerative medicine? At the end of this review, this question will be discussed further.
Th ere is increasing evidence that the human placenta contains pluripotent or multipotent stem cells or both. Various multipotent stem cells have been isolated from diff erent parts of the human placenta, such as the amnion, chorion, umbilical cord, and fetal blood. As placenta-derived cells, these stem cells have common advantages (Figure 1). Specifi c types of placenta-derived stem cells, such as trophoblastic, hematopoietic, and mesen chymal stroma cells, have been discussed elsewhere [1-3]. Here, we will review stem cells derived from the amnion of human placentae, specifi cally amniotic epithelial (AE) cells. First, we will summarize previous studies that have demonstrated the unique stem cell characteristics of AE cells. On the basis of these fi ndings, we introduce a model theory that explains why some AE cells, unlike other adult somatic stem cells, may possess pluripotent features. Second, we will discuss topics and pitfalls that are currently under discussion. Th ird, previous works that are leading the therapeutic application of AE cells will be summarized. Last, the potential of the clinical application of AE-derived stem cells and the future direction of the research are discussed.
AE cells express stem cell surface markers, such as stagespecifi c embryonic antigen-3 (SSEA-3) and SSEA-4 and tumor rejection antigen 1-60 (TRA1-60) and TRA1-81, which are known to be expressed on human ES cells [6]. About 15%, 50%, and 5% to 10% of naïve human AE (hAE) cells are positive for SSEA-3, SSEA-4, and TRA stem cell markers, respectively [7]. Normally, undiff er entiated stem cells homogeneously express these stem cell markers [6]. Th e variance of the ratio of stem cell markerpositive cells indicates that naïve AE cell populations contain cells in various stage of 'stemness' . Interestingly, the ratios of stem cell marker-positive AE cells (5% to 50%) are considerably higher than for other somatic/ tissue stem cells. Most of the somatic/tissue stem cells are 0.1% to 0.01% of the residing tissue. For instance, the hematopoietic stem cell population is only 0.01% to 0.05% of all bone marrow cells [8]. Th e relatively high ratio of stem cell marker-positive cells in AE cell populations as somatic stem cells could be explained by the model theory. Th e cell surface markers that are expressed by hAE cells are summarized and compared with the expression of other types of stem cells in Table 1 [2, 7,[9][10][11][12][13][14][15].

Stem cells 'left behind': developmental uniqueness of the amniotic epithelial cell
Unlike other parts of the placenta, the amniotic epithelium is a tissue of epiblastic origin. Human amnioblast is derived from the pluripotent epiblast around the eighth day following fertilization, whereas other parts of the placenta are derived from the trophectoderm. When the blastocyst is partially embedded in the endometrial stroma, the inner cell mass (or embryoblast) diff erentiates into two layers: the hypoblast and the epiblast. Th e epiblast is the source of all three germ layers and eventually forms the developing embryo. At the same time, a small cavity (amniotic cavity) appears within the epiblast. Epiblast cells adjacent to the amniotic cavity ( Figure 2) are called amnioblasts, which eventually form the amniotic epithelium. Concomitantly, some of the migrating hypoblasts transdiff erentiate into mesenchymal cells (extraembryonic mesoderm) and develop into the amniotic connective tissue. Th e epiblast-amnioblast segregation occurs before gastrulation, which is considered the fi rst dynamic event of organogenesis. All short-range organogenetic signals may not reach the segregated stem cells throughout gestation. For instance, cardiogenesis is a complex event that is orchestrated by short-range fi broblast growth factors (FGFs) and Hedgehog signals [16]. Th erefore, some epiblasts/amnioblasts that are spatially segregated by the amniotic cavity from the epicenter of organogenesis may escape from these diff erentiation cues. After 10 months, although most of the cells have diff erentiated by following the epithelial cell fate and have lost their stem cell characteristics, about 5% to 10% of the AE cells may retain the epiblastlike stem cell characteristics at term [7]. If this model theory is correct, fetal amniotic epithelium should contain more stem cell marker-positive cells than term amniotic epithelium. Izumi and colleagues [17] demonstrated that about 40% and 30% of fetal (early second trimester) AE cells are positive for stem cell markers TRA1-60 and TRA 1-81, respectively, whereas 5% of term AE cells are positive for these markers. Th e amnion is a fairly large tissue that may not be very uniform but is rather regionalized [18]. To exclude variances due to the regionalized stem cell localization, amnion samples were harvested from three diff erent parts of the amnion: the center of the disc, the edge of the disc, and the membrane part. Th ere was no signifi cant diff erence between samples by region, at least in these three parts [19]. On the other hand, the mechanism and signals that induce diff erentiation on 90% of amnioblasts of epiblast origin are unclear. It has been shown that cultured AE cells secrete various morphogens and growth factors such as epidermal growth factor, Noggin, Activin [20], platelet-derived growth factor, vascular endothelial growth factor, angiogenin, transforming growth factor-beta-2 (TGF-β2), and tissue inhibitors of metalloproteinases (TIMP-1 and TIMP-2) [21]. In addition to playing an important role in maintaining pregnancy, these factors may induce AE cell maturation or apoptosis to the epiblast-like immature AE cells. It must be noted that hAE cells are able to support the pluripotency of primate and mouse ES cells when primary hAE cells were used as feeder layer cells [22,23]. Th ese data indicate that some of the secreted factors or cell-to-cell signaling (or both) might play a role in maintaining epiblast-like stemness of some AE cells. Th ere are, however, no clear experimental data that indicate why the stem cell marker-positive AE cells unevenly diff erentiate even though all AE cells are exposed to the same environmental signals from the amniotic fl uid. One of the possible mechanisms is 'lateral inhibition' , which is a type of cell-to-cell inter action to regulate cell fate in the development of various cell types. Th is could be an interesting question for further investigation.

Amniotic epithelial cells possess pluripotency?
In addition to expressing stem cell-specifi c surface markers, AE cells express molecular markers of pluri potent stem cells: octamer-4 (OCT-4), NANOG, sex deter mining region Y-box 2 (SOX-2), Lefty-A, FGF-4, REX-1, and terato carcinoma-derived growth factor 1 (TDGF-1) (cripto-1). Among those molecular stem cell markers, OCT-4 is known as one of the transcription factors that play a critical role in maintaining pluri potency and selfrenewal. OCT-4 belongs to the POU family of trans criptional regulators [24][25][26] and regulates the pluripotency of human and mouse ES cells [27]. Expression of OCT-4 is decreased along with the stem cell diff erentiation and the loss of expression leading to diff erentiation [28]. At the epiblast stage, OCT-4 con tinues to be expressed as long as cells remain undiff er entiated [26]. Th e expression of OCT-4 is controlled epigenetically by hypermethylation of the enhancer/promoter region [29]. OCT-4 protein expression is observed in most AE cells. Some display nuclear-localized OCT-4, but for the majority of AE cells, the expression is cytoplasmic. Th ere is concern over the OCT-4 expression in somatic cells [30]. OCT-4 exists as two splice variants: OCT-4A and An amniotic cavity appears in the middle of the epiblast. As the cavity grows, the spatial segregation allows some amnioblasts to retain epiblast-like stem cells. Red stars indicate an amniotic cavity, and pink arrows indicate short-range organogenic signals that could not reach the amnioblasts.
OCT-4B [31]. Recent studies have suggested that it is the OCT-4A isoform that has the ability to confer and sustain pluripotency but that the OCT-4B may not be functional [32,33]. Lengner and colleagues [34] pointed out that published data describing positive results of OCT-4 expression in somatic stem cells might be erroneous because of investigator ignorance of the two isoforms. Primers or antibodies that recognize both isoforms might be misused to claim functional OCT-4 expression in some somatic stem cells. We have confi rmed OCT-4A expression in naïve hAE cells by using a commercially available primer and probe set (Hs0300511_g1; Applied Biosystems, Foster City, CA, USA) that matches OCT-4A-specifi c exons [17].
Although a number of investigations have provided evidence that suggests multipotency of AE cells, the pluripotency has not yet been proven. One of the critical issues is the diffi culty to establish clonal expansion from a single AE cell, a step that is essential to demonstrate pluri potency in vitro. Unlike mouse ES cells, human ES cells and mouse epiblast-derived stem (EpiSC) cells are intolerant to passaging as single cells. Like EpiSC cells, AE cells do not maintain their stem cell characteristics well or survive as a single cell in culture. AE cells easily fall into the senescence state or diff er entiate into palmshaped epithelial cells when cultured at low density. Th e teratoma formation assay has been used as a gold standard assay to prove pluripotency of ES or iPS cells. However, this assay cannot be applied to evaluate AE cells. Because of the genetically stable charac ter istics, the AE cell does not form a teratoma when injected into immunodeffi cient mice [7,14]. Th e ultimate approach to determine pluripotency of AE-derived stem cells is generating chimeric animals. If an AE cell that is injected into the blastocyst will contribute to all germ layer cells in the resulting chimeric embryo, the pluri potency will be confi rmed. In 2004, Tamagawa and colleagues [35] derived cell lines from human amnion and mixed them with mouse early embryonic stem cells to form an aggregation chimera. Th e authors succeeded in demonstrating that the human cells contributed to all three primordial germ layer formations in the xenogeneic chimera embryo [35]. Although the cell line cells are established from a mixed amniotic cell population that contains both AE cells and amniotic mesenchymal fi broblasts, this investigation suggested the pluripotency of the human amniotic cells. Further investigation will be required to clarify which cell population is responsible for the pluripotency.

Multipotency of amniotic epithelial cells and the therapeutic potential
Although the pluripotency of a single AE cell is not clarifi ed yet, AE cells diff erentiate into cells of all three germ layers under appropriate culture conditions [7,14]. Th e changes of gene expression and cell morphology of AE cells in these experiments demonstrated the AE cell plasticity that is induced by exposure to exogenous growth factors or chemicals. At present, it has not been confi rmed whether a single pluripotent amniotic stem cell diff erentiates into all three germ layers or whether there are various lineage-committed multipotent cells in the AE cell population. In spite of this critical question from a basic science point of view, it is a secondary concern from the perspective of clinical application. Since it is impossible to simultaneously induce the desired diff erentiation in 100% of the starting material of stem cells, some form of purifi cation process is essential prior to using stem cell-derived therapeutic cells in clinical application. Th erefore, the most impor tant question from a clinical point of view is whether therapeutically useful cells can be produced from the hAE cell population. Here, we summarize previous works that suggest the diff erentiation potential of AE cells and the therapeutic potential tested in animal models.
It must be noted that there are developmental and anatomical diff erences between rodent and human amnion. Rodent amniotic epithelium is clearly derived from epiblasts; therefore, the usage of rodent AE cells as a model could be appropriate. However, owing to the size and anatomical uniqueness, the isolation of AE cells must be done very carefully. Recently, Dobreva and colleagues [36] focused on the species diff erences of the amnion and comprehensively reviewed this topic. Th is review is strongly recommended to researchers who plan to conduct research with rodent amnion stem cells, including amniotic fl uid-derived stem cells.

Ectoderm lineage
Neurodegenerative diseases are among the most suitable target diseases for stem cell-based therapies. Since neuro degenerative diseases have many pathological processes in common, the cell transplantation approach could potentially ameliorate the symptoms of several distinct neurodegenerative diseases. Th ere are two expected mechanisms of cell transplantation. One is the diff erentiation potential of the transplanted stem/ progenitor cells to neural cells. Sakuragawa's group [37], pioneers in AE research, demonstrated that cultured AE cells express markers of glial and neuronal progenitor cells. Our group confi rmed that naïve hAE cells express various neural marker genes, including neuron-specifi c enolase, neuro fi lament-M, myelin basic protein, microtubule-associated protein 2, and glial fi brillary acid protein [7]. Under appropriate culture conditions, AE cells express or up regu late neuron-specifi c gene expressions such as nestin and glutamic acid descarboxylase. Using the adenoviral labeling system, Ishii and colleagues [38] demonstrated that about 20% of hAE cells express oligodendrocyte marker genes, myelin basic protein, proteolipid protein, and 2' ,3'-cyclic nucleotide 3'-phospho diesterase (CNPase). Th e neural diff erentiation capabilities of AE cells were confi rmed by various researchers [39,40]. Amazingly, the capabilities were preserved even after long-term cryopreservation. Human amniocytes that were originally isolated in 1974 adopted neuronal morphology and expression of neural genes, including β-III-tublin, Gap-43, NF-M, TAU, and synaptophysin, after more than 30 years [41].
Th e other expectation of cell transplantation for neurodegenerative diseases is the ability to secrete functional or protective factors from the transplanted cells such as dopamine or some other factors, which result in protective/trophic eff ects or immunomodulatory eff ects [42,43]. For instance, in Parkinson's disease, there is a loss of the dopaminergic neural population in the substantia nigra [44]. In clinical settings, it has been shown that dopamine-producing tissue (fetal mesen cephalic grafts) transplantation could ameliorate the symptoms [45]. hAE cells synthesize and release dopamine [46,47]. Th e dopamine synthesis responds to supplemented L-DOPA (L-3,4-dihydroxy phenyl alanine) concentration in a dosedependent manner. Furthermore, transplanted AE cells might release neuroprotective factors or induce neurogenesis to improve diseased or damaged environments or both. It has been shown that hAE cells produce and secrete various types of trophic factors such as nerve growth factor, neurotrophin-3, and brain-derived neurotrophic factor [48][49][50][51].
Th is neural diff erentiation and neurotrophic potential of hAE cells has been tested in animal models. Transplanted hAE cells alleviated Parkinson-like symptoms in a dopamine-denervated rat model [52]. In these experiments, the engrafted hAE cells showed paracrine or neurotrophic eff ects rather than a contribution via neural diff erentiation. However, when rat AE cells were transplanted into ischemic hippocampus of adult gerbils, the rat AE-derived neuron-like cells were observed after 5 weeks of the transplantation [53]. Th e neural diff erentiation and the therapeutic eff ect of AE cells were also tested in a rat stroke model. Th e transplanted hAE cells migrated to the ischemic area and reduced infarct volume and improved behavioral function [54]. Recently, Suh [55] reported that hAE cell transplantation restored memory function in a transgenic mouse model of Alzheimer's disease. Although the mechanism is under investigation, these data encourage the clinical applications of the hAE cells for neurodegenerative diseases.

Endoderm lineage
Two cell types, hepatocytes and insulin-producing pancreatic cells, are most desired among the endoderm lineage cells. Both cells have been used for cell replacement therapies and their therapeutic concept and effi ciency have been shown [56][57][58][59]. Th e insuffi cient supply of human hepatocytes or beta cells, however, is one of the reasons that prevent these promising therapies from becoming standard clinical applications. Th us, safe and constant supplies of these functional cells are urgently required. In addition, stem cell-derived hepatocytes will be useful not only for cell replacement therapy (hepatocyte transplantation) [57] but also for toxicology and drug development [60].
Sakuragawa and colleagues [61] reported that cultured hAE cells expressed and produced albumin and α-fetoprotein in vitro and in vivo. Th e hepatic character istics of hAE cells were extensively investigated by Takashima and colleagues [62]. Our group applied a step-wise exogenous growth factor stimulation protocol to induce further hepatic maturation in hAE cells [63]. Th e AE-derived hepatocyte-like cells expressed late-phase hepatic diff eren tiation markers, including various inducible cytohrome P450 genes, which are essential for drug metabolism as functional hepatocytes. Th ese cells were also transplanted into immunodefi cient mice, and human α-1 antitrypsin was detected circulating in the serum of recipient mice, and this confi rmed that the engrafted hAE cells function as hepatocytes in mouse liver. Recently, Manuelpillai and colleagues [64] transplanted hAE cells into drug-induced cirrhosis model animals and demonstrated the anti-fi brosis eff ect of hAE cells. Th e data indicate that the therapeutic eff ect of transplanted hAE cells is more likely the immunomodulatory eff ect by suppressing infl ammatory activation of hepatic stellate cells. On the other hand, the authors demonstrated human albumin in mouse sera that might be secreted from diff erentiated hAE-derived hepatic cells. Rat amniotic cells have been isolated and used to simulate allogeneic cell transplantation [65][66][67]. Th e transplanted rat AE cells survived in the liver following allogeneic transplantation for at least 30 days [65]. Although the rodent amniotic cell property might be diff erent from that of humans, the therapeutic effi ciency of AE cells, together with the basal advantages of the placenta-derived stem cells, suggests a treatment option for liver diseases.
Several groups demonstrated the pancreatic diff erentia tion potential of hAE cells [7,68,69]. Under appropriate culture conditions, pancreatic cell-related genes such as PDX-1, PAX-6, NKX2.2, insulin, and glucagon were upregulated in vitro [7]. Th e therapeutic potential was also demonstrated by the transplantation of cultured hAE cells in the spleen of diabetic mice. Th e serum glucose levels were normalized for several months after the transplant, suggesting that the transplanted AE cells diff erentiated into insulin-producing beta cells [68]. Th is fi nding was later confi rmed with comprehensive analyses that demon strated glucose-responsive c-peptide production [69]. In vitro diff erentiation and involvement of histamine nicotinamide-induced pancreatic diff erentiation were further investigated [70].
In addition to the hepatic and pancreatic diff erentiation, the capability of AE cells diff erentiating into other types of endoderm lineage cells has been reported. Moodley and colleagues [71] demonstrated that naïve human amnion epithelial cells diff erentiate into lung epithelium (type II pneumocyte) 2 weeks after parenteral injection into a bleomycin-induced lung injury SCID (severe combined immunodefi ciency) mice model. Th e transplanted hAE cells reduced infl ammation and abrogated fi brosis post-lung injury. Moritoki and colleagues [72] systemically transplanted EGFP (enhanced green fl uorescent protein)-transgenic mice AE cells into chemically induced cholestasis mouse model animals. Th e EGFP and cholangiocyte marker CK7 doublepositive cells formed a bile duct-like tubular structure in the chronic cholestatic mouse liver.

Mesoderm lineage
Because adult cardiomyocytes do not regenerate sufficiently, there is great interest in fi nding suitable cell sources for cellular cardiomyoplasty. hAE cells also possess the potential to diff erentiate into cardiac cells [7]. Although AE-derived cardiomyocyte-like cells expressed cardiac diff erentiation marker genes, immunocytochemistry analysis showed that the expression pattern of α-actinin was similar to that of immature cardiomyocytes. Th e therapeutic potential was demonstrated by using rat amniotic cells and a rat acute infarction model [73]. Although transplanted rat amniotic cells dramatically improved the cardiac function, only a few transplanted cells were diff erentiated into cardiomyocytes (α-actininpositive cells). Th e therapeutic eff ect was speculated to be due to paracrine or immuno modu latory eff ects of the rat amniotic cells. An interesting application of amniotic membrane was tested, and the therapeutic effi ciency was demonstrated. Cargnoni and colleagues [74] applied a fragment of human amniotic membrane as a cardiac patch on an infarction area of a rat heart. Th e postischemic cardiac function was signi fi cantly improved with the amnion patch. Th is investi gation importantly demonstrated that secondary cardiac ischemic injury could be prevented by humoral factors that are released from the amnion. Recently, functional cardiac diff erentiation of human amniotic cells was demonstrated [75]. Th e cardiomyogenic diff erentiation was induced by a co-culture system with murine fetal cardiomyocytes. Th e structure of sarcomeric α-actinin and the spontaneous beating and in vivo contribution of human amnion-derived cardiomyocytes were demon strated. Stem cell-derived cardiomyocytes are also expected to be an important new tool for drug development [60]. Th e in vitro functional hAE-derived cardiomyocytes could be a cell source for these assays. Further investigation for culture condition optimization or direct reprogramming will be required along with a defi nition of selection markers of functional mature cardiomyocytes. Th e studies that demonstrate the diff erentiation capability of AE cells into all three germ layer lineages are summarized in Table 2.

Advantages of human amniotic epithelial cells for clinical applications
From the view of clinicians and patients, the sine qua non of clinically applicable stem cells is fi rst, 'safety'; second, 'therapeutic effi ciency'; and, last, 'availability/suffi cient quantity' . Several types of stem cells could serve as cell sources for cellular therapy. Generally, stem cells are classifi ed according to their diff erentiation ability and origin. Pluripotent stem cells, such as ES cells and iPS cells, are considered to be the most promising stem cells because of their tremendous diff erentiation ability. Th e 'safety' , however, is always a concern. Th e pluripotency comes with genetical instability, which leads to concerns for tumorigenicity. Although the 'therapeutic effi ciency' is promised, the long-term effi ciency has not yet been proven. Furthermore, the expansion and maintenance to obtain a therapeutically suffi cient number of cells require time, eff ort, and cost.
Adult stem cells can be derived from virtually any tissue or organ. Most adult stem cells are tissue-specifi c lineage-committed multipotent cells. Some adult stem cells such as mesenchymal stem cells and hematopoietic stem cells are already applied in clinics and showed therapeutic effi ciency, mainly with their immuno modulatory property. Th erefore, the clinical applications are considerably safe, particularly in the case of autologous transplantation.
A similar immunomodulatory property has been demonstrated with hAE cells [76][77][78][79][80]. hAE cells inhibited allogeneic mixed lymphocyte reactions in a dosedependent manner with 66% to 93% inhibition [81]. Most of the report demonstrated the immunomodulatory eff ect by secretion of suppressive mediators such as TNFα, FasL, TRAIL, TGF-β, and MIF. On the other hand, Banas and colleagues [12] demon strated that the immunomodulatory eff ect of AE cells is dependent on cell-to-cell contact with responding T cells. By using non-serum culture conditions, the authors demonstrated that hAE cells inhibit peripheral blood mononuclear cell proliferative responses to mito gen, alloantigen, and recall antigen but preactivated T-cell blast response. Th e results suggested that the presence of HLA-G immunological cell surface molecules is responsible for the cell-to-cell immunosuppressive properties of AE cells. In addition to the HLA-G expres sion [82][83][84], the expressions of comple ment inhibitory proteins, CD59 antigen, decayaccelerating factor, mem brane attack complex, and Fas antigen/CD95/APO1 have been reported as potential immunoregulatory factors from hAE cells [85][86][87][88]. Never theless, further investigation is required to fully elucidate the underlying mechanisms of the immunomodu latory eff ect of hAE cells.
Importantly, the safety of AE upon transplantation has been shown in a clinical setting. hAE cells have been used in clinics to correct lysosomal storage disease [89][90][91]. Although the applications were not conducted as a stem cell therapy, more than 50 cases of AE cell/tissue transplantations have been performed in various institutes [90][91][92]. No tumor formation has been reported from these clinical trials. As it has been described, AE cells are clearly non-tumorigenic when transplanted into immunodefi cient animals [7,14]. A total of one to two million hAE cells was injected into more than 50 individual mice, which were observed for a maximum of 516 days. None of the AE cell transplants has led to the development of tumors by any route of administration in SCID-beige mice or Rag-2 knockout mice. In parallel, cytogenetic analysis confi rmed genetical stability of cultured AE cells [7]. AE cells do not express telomerase reverse trans criptase (TERT) mRNA [7]. A study demonstrated that immortalized cells by expression of TERT could exhibit some neoplastic transformation toward what seem to be cancer stem cells [93]. Missing TERT expression may be a safety advantage. Since more than 100 million cells can be isolated from one placenta, long-term culture and massive replication are not required to use AE cells as a cell source. For instance, only half a million cells will be suffi cient to improve the devastating symptoms of Parkinson's disease and Huntington's disease [94,95]. Furthermore, human placenta is a neonatal tissue that has less age-acquired and environmental DNA damage. Naturally, the neonatal cells should possess a life-long viability. Given these facts, the AE cell is clearly a safe cell for clinical application. A placenta is discarded after every live birth; objections to derive stem cells from placentae are not expected. Isolating and using stem cells from discarded human tissue are ultimate recycling biotechnologies, which would be acceptable in today's society. Human placenta is readily available tissue wherever human society exists. Th erefore, there will be no regional disparities for placenta-derived stem cell therapies. Current statistics indicate that there are over 4 million births and over 1 million cesarean sections performed in the US every year. Th e sheer volume of available placenta tissue leads us to the idea of establishing a biobank system for placentaderived stem cells. Nakatsuji and colleagues [96] estimated that a cell bank with only 30 stem cell lines could match the HLA-A, HLA-B, and HLA-DR haplotypes in 82.2% of the Japanese population. On the basis of these estimations and the theoretical number of available placentae, it is clearly feasible to establish a biobank that stores human placenta-derived stem cells with all HLA haplotypes. Clinically relevant stem cells should be easily and reproducibly cultured and manipulated. Th e AE cell isolation procedure is relatively easy and does not require a special laboratory set-up [4]. Th e biobanking system therefore could be established in any country and connected as a network to provide all HLA types, including race dominant types.
Ethical issues surrounding both embryonic and fetal stem cells do not apply to the use of discarded human placentae. However, once the therapeutic effi ciency of placenta-derived stem cells is demonstrated, this normally discarded medical waste may turn into valuable property. Th e cell-acquiring process, cost, and proprietary rights will be new ethical issues. A regulatorycompliant system will be required for cell-acquiring and -providing pro cesses. Th e precedent and current regulations for the umbilical cord blood cell usage could be useful to prepare a guideline for procure ments of placenta-derived stem cells [97].
Th e use of this ideal stem cell could take one of two directions in future clinical regenerative medicine. One direction will be in developing technology to derive pluripotent stem cell lines from AE cells, which possess a biological potential equivalent to that of ES cells and iPS cells. Th is direction, however, may eliminate the AE cell advantage that is discussed here. Th e other direction is to aim directly at diff erentiation to obtain functional target cells. Th e stem cell-derived cell therapy requires a selection step prior to the cell application to patients to ensure safety and effi ciency. Th e genetic stability and non-tumorigenicity of AE cells will be the advantage for this approach. Unlike in ES cells or iPS cells, leakage concerns at the selection step will not be a critical issue.

Conclusions
Herein, we reviewed the stem cell characteristics of amnion cells, especially AE cells. We introduced a model theory that may explain why so many cells with stem cell features are present in the amnion. Th e model theory has been proposed by several research teams, including ours [5,7,14,34]. Previous studies that demonstrate the diff eren tiation and therapeutic potential of AE cells were summarized. We described four major reasons why placentaderived cells are a signifi cant cell source for clinical applications. Th e AE cell meets two important conditions that are required for clinically relevant stem cells: safety and availability. So far, no stem cells are able to diff erentiate into therapeutically useful cell types in vitro, or their diff erentiation is not well controlled. As with other types of stem cells, further investigations will be required to induce AE cells to diff erentiate into therapeutically useful cells. Since AE cells are extremely safe and show thera peutic effi ciency in animal models, clinical applica tion should be considered in the near future.

Competing interests
The author owns stock in Stemnion, Inc. (Pittsburgh, PA, USA). He has received no payment for the preparation of this manuscript and declares that he has no other competing interests.