Stem Cell Research (2014) 12, 610–621

Available online at www.sciencedirect.com

ScienceDirect www.elsevier.com/locate/scr

Developmental insights from early mammalian embryos and core signaling pathways that influence human pluripotent cell growth and differentiation Kevin G. Chen a,⁎, Barbara S. Mallon a , Kory R. Johnson b , Rebecca S. Hamilton a , Ronald D.G. McKay c , Pamela G. Robey d a

NIH Stem Cell Unit, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892, USA b Information Technology and Bioinformatics Program, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892, USA c The Lieber Institute for Brain Development, 855 North Wolfe Street, Baltimore, MD 21205, USA d Craniofacial and Skeletal Diseases Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD 20892, USA

Received 26 September 2013; received in revised form 10 January 2014; accepted 4 February 2014

Abstract Human pluripotent stem cells (hPSCs) have two potentially attractive applications: cell replacement-based therapies and drug discovery. Both require the efficient generation of large quantities of clinical-grade stem cells that are free from harmful genomic alterations. The currently employed colony-type culture methods often result in low cell yields, unavoidably heterogeneous cell populations, and substantial chromosomal abnormalities. Here, we shed light on the structural relationship between hPSC colonies/embryoid bodies and early-stage embryos in order to optimize current culture methods based on the insights from developmental biology. We further highlight core signaling pathways that underlie multiple epithelial-to-mesenchymal transitions (EMTs), cellular heterogeneity, and chromosomal instability in hPSCs. We also analyze emerging methods such as non-colony type monolayer (NCM) and suspension culture, which provide alternative growth models for hPSC expansion and differentiation. Furthermore, based on the influence of cell–cell interactions and signaling pathways, we propose concepts, strategies, and solutions for production of clinical-grade hPSCs, stem cell precursors, and miniorganoids, which are pivotal steps needed for future clinical applications. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Abbreviations: FGF, fibroblast growth factor; hPSCs, human pluripotent stem cells; hESCs, human embryonic stem cells; hiPSCs, human induced pluripotent stem cells; NCM, non-colony type monolayer; ROCK, Rho-associated kinase; TGF-β, transforming growth factor β ⁎ Corresponding author at: NIH Stem Cell Unit, National Institute of Neurological Disorders and Stroke, National Institutes of Health, 37 Convent Drive, Room 1000, Bethesda, MD 20892, USA. E-mail address: [email protected] (K.G. Chen). http://dx.doi.org/10.1016/j.scr.2014.02.002 1873-5061/Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 3.0/).

Pluripotent stem cell growth models

Introduction Human pluripotent stem cells (hPSCs), including human embryonic stem cells (hESCs) and induced pluripotent stem cells (iPSCs), are potentially important resources for regenerative medicine-based cell replacement. However, this approach has not been efficiently implemented, due in part to currently used inefficient cell culture and differentiation strategies, which are heavily based on cell culture as colonies and aggregated embryoid bodies (EBs) (Chen et al., 2014; Mallon et al., 2006; Thomson et al., 1998; Xu et al., 2001). These methods have led to time-consuming production of relatively low numbers of cells (Hartung et al., 2010; Kehoe et al., 2010; Serra et al., 2012), heterogeneous cellular and molecular states (Bendall et al., 2007; K.G. Chen et al., 2012; Moogk et al., 2010; Stewart et al., 2006), and frequent reports of chromosomal abnormalities (reviewed in Baker et al., 2007; Lee et al., 2013). Certain chromosomal alterations such as trisomies 12, 17, and 20 have the capacity to confer growth advantages as indicated by comparative genetic analyses (Amps et al., 2011; Catalina et al., 2008; Draper et al., 2004; Lefort et al., 2008). The relationship between heterogeneous cellular states and the incidence of chromosomal abnormalities remains unclear. Thus far, there is no direct proof that the colonytype culture method is responsible for the reported chromosomal abnormalities. However, it is conceivable that the heterogeneity of hPSC colonies (induced by cellular stress) might result in the selection of cells with growth advantages, which are usually bestowed by mutations or chromosomal alterations (Baker et al., 2007; Lee et al., 2013). This cellular stress may be derived from unwanted growth factors or conditions, which could be categorized as abnormal metabolic pathways, deficiency in nutrient or growth factors, and excessive apoptotic or differentiation signals. These suboptimal conditions may directly or indirectly contribute to insufficient cell production, cellular heterogeneity, and chromosomal instability of hPSCs. It is imperative that we develop efficient culture methods that can overcome the obstacles found in the current culture systems. Moreover, new culture methods should be optimized for production of clinically relevant cell numbers, tissue morphogenesis, high-throughput drug discovery, and efficient assays of cellular stress. A comprehensive analysis of different cell culture protocols for hPSC is available in a recent review (Chen et al., 2014). Here, we will focus on the fundamental principles of developmental biology and core signaling pathways that underlie different growth methods. We will further discuss the relationship between the development of early stage embryos and the mechanisms underlying hESC heterogeneity and chromosomal instability. Finally, we will provide new insights into the development of optimal growth models for hPSCs.

Human pluripotent stem cell colonies and aggregates recapitulate some properties of early-stage embryo development Unlike mouse embryonic stem cells (mESCs), which possess high single-cell plating efficiency under dissociated conditions,

611 both hESCs and iPSCs cannot grow as single cells under commonly used culture conditions. Human pluripotent stem cells are conventionally cultured as condensed colonies and must be passaged as cell clumps to avoid cumulative apoptotic and differentiation stress. These cells are routinely differentiated toward derivatives of the three germ layers through three-dimensional (3D) EBs. In general, colony-type and aggregated cultures often produce heterogeneous colonies and EBs. Hence, we frequently observe more apoptotic activity in the center of a colony (with more spontaneously differentiated cells at the periphery) (Figs. 1 and 2) and complicated cystic structures after EB agglomeration (Fig. 3). The unique structures of hESC colonies and EBs are reminiscent of the inner cell mass (ICM) and post-implantation blastocysts observed during early embryo development. In the following sections, we will discuss the links between the structures of hPSC colonies and EBs and their developmental counterparts.

Human ES colony structure inherits epithelial features of the ICM The ICM is comprised of a group of epithelial-like cells, the epithelial features of which are reflected at the intercellular junctions and in their basement membranes (Fig. 1A). Typical epithelial cells are tightly polarized monolayers with a distinct apical surface, basal membrane, and lateral intercellular conjunctions (Figs. 2A, B). At day 4 of compacting human embryos, E-cadherin (E-cad), an important component of epithelial cells, is located on the membranes where cell– cell contacts occur. However, in the ICM, E-cad is diffusely expressed in the cytoplasm, but not on the plasma membrane (Alikani, 2005). Thus, the distribution of E-cad in human preimplantation embryos appears to be an embryonic stage-dependent process. Electron microscopic study has also revealed similar epithelial features in hESCs (Ullmann et al., 2007). Intercellular interactions in both trophectoderm and ICM cells are implemented through gap junctions, tight junctions, and desmosome-like structures (Dale et al., 1991; Gualtieri et al., 1992). Laminin is another epithelial component of the basement membrane that is expressed in the different stages of early embryos. Matrigel, a basement membrane preparation extracted from a murine Englebreth–Holm–Swarm sarcoma (Kleinman et al., 1982), containing laminins, as well as collagen IV, fibronectin, and vitronectin, was found to support robust hESC growth in defined medium (Ludwig et al., 2006) and is used routinely for the feeder-free culture of hPSCs. Laminin-111, the most abundant form in the Reichert's membrane, was found under the mouse trophoblasts, which was proposed to interact with ICM cells and induce differentiation (Domogatskaya et al., 2012; Miner et al., 2004). At least two laminin isoforms (i.e., laminin-511 and -521) have been detected in the basement membrane of the ICM of mammalian pre-implantation embryos (Domogatskaya et al., 2012; Miner et al., 2004). It has been shown that laminin-511, but not -111, supports mESC self-renewal in vitro without leukemia inhibitory factor (LIF) or mouse embryonic fibroblasts (MEFs) (Domogatskaya et al., 2008). Recently, the use of human recombinant laminin-511 has enabled long-term self-renewal of hPSCs (Rodin et al., 2010). Thus, both E-cad

612

K.G. Chen et al.

Figure 1 Early human embryo development, hESC culture, and acquisition of cellular heterogeneity. (A) Anatomical structure of the inner cell mass (ICM) at day 8 with major cell types indicated. (B) Derivation and growth of hESCs from the ICM: A typical hESC colony (e.g., hESCs grown on feeders) has three distinct zones in terms of its growth patterns, cell morphology, densities, differential marker expression, and differences in response to small molecules and signaling factors. The center of a colony represents highly compacted epiblast-like cells (EpiBL), which frequently experience either apoptotic or differential stress. The periphery is the region that undergoes epithelial-to-mesenchymal transitions (EMTs), whereas some differentiated fibroblast-like cells (FBL) express differential markers SSEA-1 and ABCG2 and are sensitive to BMP-4 signaling (K.G. Chen et al., 2012; Padmanabhan et al., 2012). The region between the center and periphery of the colony represents an intermediate state of cells that undergo reversible EMTs. Undifferentiated hESCs could be selected or adapted after serial passaging through growth-differentiation selection cycles in in vitro growth conditions. (C) Lineage commitment from blastocyst (BC) to gastrulation, trophectoderm (TE), and extraembryonic endoderm (EEE). The flow chart depicts the relationship among the cell types presented in panel A. EMT occurs at the primitive streak. (D) Anatomical structure and various cell types in hESC colonies under colony-type culture conditions: hESCs undertake either reversible EMT (that maintain the undifferentiated states) or irreversible EMT (that induce terminal differentiation). Abbreviations: BC, blastocyst; BM, basement membrane; CTB, cytotrophoblast; EEE, extraembryonic endoderm; EEM, extraembryonic mesoderm; EM, endometrium; EpiB, epiblast; EpiBL, epiblast-like cells; FBL, fibroblast-like cells; HB, hypoblast; HBL, hypoblast-like cells; ICM, inner cell mass; TE, trophectoderm; STB, syncytiotrophoblast; TB, trophoblast; and TBL, trophoblast-like cells.

and some specific laminins are indispensable in maintaining the epithelial features of undifferentiated hESCs. Alterations of these epithelial components in vitro during embryonic development may induce differentiation and cellular heterogeneity.

Within the ICM, the embryoblast has the ability to form the epiblast, the outer layer of a blastula that gives rise to the ectoderm, and the hypoblast, which in turn generates extraembryonic endoderm (Fig. 1C) (Carlson, 2009). Developmentally,

Figure 2 Growth models of hPSCs under various growth conditions. (A) Growth models of hPSC colonies in hESC medium that contains growth factors FGF2 and TGFβ on Matrigel-based extracellular matrices with deduced anatomical structures and molecular complexes: Laminin (LM), fibronectin (FN), and vitronectin (VNT) mediate their effect via the α6β1, α5β1, and αVβ5 integrins respectively (Braam et al., 2008; Brafman et al., 2012). Different zones of out-growing hPSC colonies have differential epithelial-to-mesenchymal transition (EMT) states, in which the cells in the center of the colonies possess more epithelial characteristics and cells at the periphery develop mesenchymal-like phenotypes. The cells between the center and the periphery of the colonies define an intermediate state that undergoes transient EMT and sustains pluripotency. Loss of the epithelial markers (e.g., E-cad) with degradation of ECM confers cells with an irreversible EMT state with fibroblast- or trophoblast-like phenotypes. (B) Growth models of hPSCs grown as non-colony type monolayer (NCM) have uniform intercellular interactions under the same growth conditions as indicated in panel A. (C) Core pathways that support hPSC growth: the model illustrates the dynamics of cellular states among epithelial hPSCs, cells with an intermediate ES state, and differentiated cells, in which two phases of EMTs (i.e., EMT1 and EMT2) or METs (MET1 and MET2) are likely to mediate these conversions. The potential regulators and pathways that control both EMTs and METs are proposed and indicated with an emphasis on the differential roles of TGFβ at different stages. Abbreviations: AKT, v-akt murine thymoma viral oncogene homolog; ALK5, transforming growth factor, beta receptor 1, known as activin receptor-like kinase 5; BM, basement membrane; CDH1, the gene encoding E-cadherin (epithelial); ColV, collagenase IV; E-cad, E-cadherin; ERK, extracellular signal-regulated kinase; FGF2, fibroblast growth factor 2 (basic); FN, fibronectin; GSK3, glycogen synthase kinase 3; INT, intergrins; LM, laminin; MAPK, mitogen-activated protein kinase; MET, mesenchymal-to-epithelial transition; MMPs, matrix metallopeptidases; Nu, nucleus of the cell; p53, tumor protein p53; PPJ, plastic-protein junction; PS, polystyrene surface of the cell culture dishes for hPSCs; RTK, receptor tyrosine kinases; Smad, SMAD family member; Snail, snail homolog 1 (Drosophila); Slug, snail homolog 2 (Drosophila); TGFβ, transforming growth factor beta; VNT, vitronectin; β-cat, β-catenin.

Pluripotent stem cell growth models embryonic epithelial cells can change their morphology and motility to resemble fibroblast-like mesenchymal cells, a process known as epithelial-to-mesenchymal transition (EMT) (Carlson, 2009). The peripheral cells of the embryoblast differentiate into

613 the cytotrophoblast and syncytiotrophoblast, resembling the EMTs required for implantation (Fig. 2A). The reverse process of EMT, termed mesenchymal-to-epithelial transition (MET), can be repeatedly observed during normal embryonic development

614

K.G. Chen et al.

Figure 3 Early mammalian embryo development and embryoid body (EB) differentiation. (A) Structural presentation of mammalian embryonic differentiation after the implantation of a blastocyte. Syncytiotrophoblasts are not shown in this figure. The figure is adapted from the book Human Embryology and Developmental Biology (Carlson, 2009). (B) Delineation of early- and late-stage EB formation with differentiation medium in a non-adherent culture dish. Abbreviations: CTB, cytotrophoblast; EpiB, epiblast; EpiBL, epiblast-like cells; HB, hypoblast; HBL, hypoblast-like cells; ICM, inner cell mass; and TB, trophoblast.

(Carlson, 2009). Thus, hESCs derived from the ICM may be considered to mimic epiblast-stage embryos before the occurrence of the primitive streak, where the first EMT event takes place (Figs. 1C, 2C). It is clear now that multiple phases of EMT and MET play pivotal roles in early embryo development and in the generation of various types of tissues (Carlson, 2009; Thiery et al., 2009). It is feasible that hESC lines, derived from the ICM, have predominantly epithelial characteristics and, to a lesser extent, resemble cell types with mesenchyme-like propensity. Consistently, expression of the epithelial marker genes CDH1, DSP and KRT19 is significantly increased in hESCs, as opposed to that observed in fibroblasts and partially reprogrammed iPSCs (Ullmann et al., 2007). Moreover, EMT has also been found in colonies of monkey ESCs and in feeder-free cultures that frequently induce spontaneous differentiation (Behr et al., 2005; Pratama et al., 2011). Thus, the heterogeneity of a hESC colony might be an inherent property of the ICM, which at least is partially attributed to irreversible EMTs (Figs. 1 and 2). It is conceivable that the heterogeneity of the ICM is required for consecutive three-germ layer differentiation after embryo implantation. In vitro, the counterpart of the three germ-layer lineage commitment is EB-mediated differentiation.

EB differentiation shares some fates of implanted blastocysts Without growth factors that support the pluripotent state, hESC clumps form 3D multilayer aggregates under suspension culture conditions, usually in a non-adherent culture dish. EB differentiation has been widely used as an intermediate step in multilineage differentiation. As the name suggests, these aggregates mimic many aspects of cell differentiation during the development of early embryos (Fig. 3). Based on

the stage and structural properties, EBs can be classified into simple and cystic EBs (Fig. 3B) (Magyar et al., 2001; Ungrin et al., 2008). Simple EBs are 2- to 4-day spherical aggregates (usually 100–400 μm), resembling morula-like and, to some extent, the ICM-stage embryos. With respect to cystic EBs, there are one or more central cavities after day 4 in suspension culture, which are similar to an implanted mammalian embryo from day 7 to 12. At this stage, the embryos have established an inner parietal endoderm enclosed cavity (i.e., primary yolk sac) and an outer layer termed extraembryonic mesoderm (Fig. 3A) (Carlson, 2009). EBs are developed through intercellular adhesion, which may involve the cadherin–β-catenin–Wnt pathway, a potent regulatory mechanism for tissue morphogenesis (Gumbiner, 1996). Indeed, E-cad, expressed in undifferentiated hESCs, is the initiator of early-stage EB formation and also plays a role in later-stage EB differentiation (Dang et al., 2004). Modulation of a specific cadherin isoform is associated with a definite cell lineage commitment, in which expression of E-cad leads to epithelial differentiation whereas up-regulation of N-cad prefers a neuroepithelial fate (Twiss and de Rooij, 2013). Moreover, spontaneous formation of primitive endoderm was observed at the exterior surface of EBs after cell aggregation (Maurer et al., 2008). The primitive endoderm cells were subsequently differentiated into visceral and parietal endoderm, and also generated a laminin–collagen-IV-enriched basement membrane (Fig. 3B) (Li et al., 2001). Similarly, the exterior domain of hESC EBs was stained positively for the endoderm marker FOXA2 (Ungrin et al., 2008). Unlike their embryonic counterpart, regular EBs do not have a thin layer of trophoblasts (Fig. 3). Thus, the exterior surface is directly exposed to soluble factors and nutrients in culture medium. The permissive diffusion of soluble factors (b 1000 Da) in culture medium from the exterior to the interior of EBs enables the formation of concentration gradients that

Pluripotent stem cell growth models influence the differentiation outcomes (Bratt-Leal et al., 2009). As EB development proceeds, asynchronous differentiation of EBs toward the three germ-layer lineages may be achieved. However, this heterogeneous differentiation presents a big challenge for large-scale production of homogeneously differentiated cells. Numerous efforts were made to control the size of both hESC colonies and aggregates in order to direct differentiation trajectories (Bauwens et al., 2008). To ensure the homogeneity of EBs, Zandstra and colleagues also utilized a new technique to generate uniform EBs from dissociated single-cells. The size-controlled EB formation with the ROCK inhibitor Y-27632 ensured its homogeneity and highly ordered structures (Ungrin et al., 2008). In addition, strategies that influence differentiation from both the exterior and interior of EBs can be used to efficiently direct differentiation by controlling the extracellular microenvironment and chemical signals (Bratt-Leal et al., 2009). Collectively, EB-based methods may offer several benefits for stem cell research, which include (i) use of EBs as a routine differentiation protocol, (ii) application of EBs to study early differentiation events of human embryos, (iii) use of genetically engineered EBs to determine the functionality of gene mutations that are lethal to early embryos, (iv) coupling of EB methods with new growth modules such as bioreactors to produce large amounts of differentiated cells for therapeutic use, and (v) combination of EB methods with microfluidic bioreactors to study morphogenesis in 4D culture systems (Chen et al., 2014). Nevertheless, we also need to consider the disadvantages of EB methods (i.e., time-consuming nature, cell loss, lower differentiation efficiency, and heterogeneous differentiation) in future pharmaceutical and clinical applications. Thus, both colony-type hPSC growth and EB aggregated differentiation share many structural features of early embryonic development, including the heterogeneous nature of cell growth and differentiation. It has been appreciated that generation of heterogeneity is actually required for lineage commitment and tissue maturation (Carlson, 2009). However, such a heterogeneous cellular state is an unwanted outcome for hPSC-based therapies, which usually need transplantation of a pure cell population. Understanding the precise signaling pathways that drive early embryonic development and hESC growth would provide useful tools to control these heterogeneous states during cell expansion and differentiation in vitro.

Core signaling pathways depict multiple EMT phases and cellular heterogeneity in human pluripotent stem cells The molecular mechanisms that maintain the epithelial identity of hPSCs depend on apical ligand-receptor signal transductions, intercellular adhesion (mediated by molecules such as E-cad-conjugated β-catenin), and basement membrane and ECM interactions (Figs. 2A, B). With regard to intercellular connections, glycogen synthase kinase-3 (GSK-3) is a potent regulator of E-cad and plays a pivotal role in EMT and in cell fate commitment. In general, GSK-3 phosphorylates Snail and initiates its proteasomal degradation, hence upregulating E-cad and concurrently inhibiting EMT (Doble and Woodgett, 2007). However, in the case of

615 hESCs, when GSK-3 inhibitor (6-bromoindirubin-3′-oxime) is added as a supplement to hESC medium to support hESC self-renewal, EMT is inhibited, likely through activation of the Wnt pathway (Sato et al., 2004; Ullmann et al., 2008). Nevertheless, the role of Wnt/β-catenin in the maintenance of undifferentiated hESCs has been questioned by a recent study in which the authors revealed that Wnt/β-catenin signaling actually promotes hESC differentiation, not self-renewal (Davidson et al., 2012). In general, concurrent upregulation of E-cad and down-regulation of metallopeptidases (such as MMP2 and MMP9), N-cadherin, vimentin, Snail, and Slug may be essential to inhibit EMT and differentiation (Fig. 2C) (Ullmann et al., 2008). Interestingly, E-cad may coordinate with many molecular events to reduce the motility and the occurrence of the fibroblast-like cells at the periphery of hESC colonies in various growth conditions (Ullmann et al., 2008; Ullmann et al., 2007). E-cad may be also associated with the arrangements of the core actin cytoskeleton and the EMT phenotype. This E-cad-EMT reverse correlation may be assayed by the presence of the membrane oncofetal protein 5T4, which is triggered by the addition of anti-adhesive antibody SHE78.7 that disrupts E-cad (Eastham et al., 2007). These data suggest that loss of E-cad and its regulators is a critical step that initiates EMT in hESCs. Several other signaling pathways that include the TGFβ and phosphoinositide 3-kinase (PI3K) pathways have also been implicated in the regulation of EMT. However, the role of TGFβ in EMT (the differentiation vector) and maintenance of undifferentiated hESCs is somewhat complicated. It is believed that FGF2, activin-A, TGFβ, and their associated receptors represent the well-characterized signaling pathways that maintain the undifferentiated states of epithelial hESCs (Fig. 2) (James et al., 2005; Ludwig et al., 2006). TGFβ maintains the pluripotent state of hESCs via receptor binding to ALK5, which phosphorylates Smad2/3 and mediates the nuclear localization of the Smad2/3/4 complex on the target genes (James et al., 2005). Indeed, TGFβ alone was able to sustain the expression of hESC markers in TeSR1 medium (Ludwig et al., 2006). Nonetheless, TGFβ is also involved in the regulation of multiphase differentiation such as mesodermal induction, myoblast proliferation, and invasion of cardiac jelly (Carlson, 2009). It controls the EMT that drives gastrulation through the activation of Snail (Thiery et al., 2009), which then increases vimentin and concurrently decreases the expression of epithelial genes such as DSP, KRT19, and the E-cad encoding gene CDH1. In addition, withdrawal of TGFβ from bovine serum and feeders enhances reprogramming to pluripotency in mouse cells (Li et al., 2010), likely through the MET (Fig. 2C). TGFβ most likely modulates transient EMT events that do not inhibit the pluripotent state of hESCs, whereas irreversible EMTs may be the major factor that initiates differentiation both in vitro and in vivo (Fig. 2). Taken together, these studies highlight the complexity of TGFβ in stem cell biology, suggesting differential roles of TGFβ in reprogramming somatic cells, maintaining pluripotency, and mediating lineage differentiation (Fig. 2C). A thorough understanding of the mechanisms that control intercellular interactions would enable us to optimize growth conditions, thereby minimizing cellular heterogeneity and abolishing potential chromosomal abnormalities in hPSC culture.

616

Chromosomal instability is still a problem in human pluripotent stem cell culture During the development of early human embryos, chromosome instability (e.g., aneuploidies, uniparental disomies, segmental deletions, and amplifications) is a common feature in human cleavage-stage embryonic cells (Vanneste et al., 2009). Thus, the chromosomes of hPSCs and their derivatives might be similarly altered at any stage during in vitro cell processing. Although most lines were reported to be cytogenetically normal, there was a substantial number of cell lines that acquire chromosomal changes after long-term culture, particularly on chromosomes 1, 12, 17 and 20 (reviewed in Baker et al., 2007; Lee et al., 2013). Recently, Benvenisty and colleagues analyzed chromosomal changes in 104 hPSC lines, including 66 hiPSC and 38 hESC samples from 18 different studies. They identified ~20% abnormal lines, of which ~9% of the lines had at least one trisomy (Mayshar et al., 2010). In a second cohort study, Andrews and colleagues reported 34% abnormal karyotypes in 129 hPSC lines, in which 20% of cytogenetically normal lines carried a conserved amplicon at 20q11.12 (Amps et al., 2011). These studies define a putative rate of chromosomal abnormalities, ranging from 20% to 54%. The estimated rates for genomic changes could be even higher if we could include undefined small genomic changes such as point mutations. It is believed that prolonged culture in vitro triggers selection of variant cells that have a higher tendency for survival and self-renewal (Baker et al., 2007; Lee et al., 2013). Notably, the methods employed for hPSC passaging may have a significant impact on chromosomal changes. It has been shown that hPSCs with enzymatic mass-passaging have a significant increase (i.e., 2-fold) in chromosomal abnormalities when compared with those cells using a manual cut-and-paste passaging technique (P = 0.017) (Amps et al., 2011). Nevertheless, the molecular basis that leads to chromosomal instability in hPSCs remains unclear. Recent studies have linked chromosomal abnormalities to impaired DNA damage response and cell cycle checkpoint control (e.g., failure to activate CHK1 in hPSCs) (Desmarais et al., 2012; Hyka-Nouspikel et al., 2012). These studies highlight a potential important protective mechanism that hPSCs maintain genomic stability by eradicating damaged cells through apoptosis, rather than DNA repair or DNA homologous recombination. Not surprisingly, Andrews and colleagues were able to map the anti-apoptotic gene BCL2L1 on the conserved amplicon at 20q11.12, which may confer enhanced antiapoptotic activity in hPSCs with a normal karyotype (Amps et al., 2011). Genomic changes at the chromosomal or single nucleotide level likely contribute to the variability of the pluripotent states and differentiation potential. These changes can be acquired and accumulated during cell culture because of suboptimal growth media, inappropriate extracellular matrices, and dysregulated cell signals (Baker et al., 2007). With regard to hPSC culture, our goal is to avoid or minimize the heterogeneity and the occurrence of these potentially harmful alterations that may possess tumorigenic effects. Notably, the chromosomal instability rates observed in hPSC are predominantly derived from prolonged hPSC culture as colonies on MEFs and Matrigel. To improve upon colonytype-based culture methods, considerable efforts have been

K.G. Chen et al. made to develop alternative culture platforms for efficient hPSC expansion, differentiation, and high-throughput assays.

Robust growth models for pluripotent stem cell growth and assay The need for clinically relevant quantities of hPSCs used for stem cell therapies and pharmaceutical applications promotes the development of compatible hPSC growth platforms. Stem-cell based therapies usually demand desired cell numbers ranging from 107 to 1010 or beyond. For example, repair of a 20-mm bone defect requires about 2 × 107 osteoblasts. Approximately 2 × 109 differentiated cardiomyocytes or insulin-producing beta cells are needed to replace the damaged tissues after myocardial infarction or for islet transplantation (Kehoe et al., 2010). The development of suspension culture and single-cell based growth protocols for manufacturing hPSCs is an active area of research. We have recently presented a comprehensive analysis of different culture protocols (Chen et al., 2014). In the following sections, we will focus on several new methods based on the variability, heterogeneity, chromosomal instability, and potential applications. These methods include suspension culture of undifferentiated cells, single-cell passaging, and non-colony type monolayer (NCM).

Suspension culture in stirred-suspension bioreactors Current suspension culture methods require that hESC colonies are dissociated into small clumps to maintain the viability and pluripotency of the cells. To overcome the size differences and variability of cell seeding, several groups have adapted their methods from seeding small aggregates to single-cells in stirred-suspension vessels (V.C. Chen et al., 2012; Kehoe et al., 2010; Krawetz et al., 2010; O'Brien and Laslett, 2012; Steiner et al., 2010). Typically, single cells dissociated from hPSC colonies are incubated with MEF-conditioned medium containing 10% Matrigel and 10 μM Y-27632 for 30 min to facilitate initial intercellular aggregation and introduced into spinner flasks with an agitation rate at 60 rpm for approximately 6 days followed by passaging (Kehoe et al., 2010). The suspension cell clusters were passaged weekly by mechanically triturating the cells into smaller aggregates in the presence of Y-27632. These investigators believed that the suspended “EB-like” clusters of hESCs maintained their pluripotency and did not turn into differentiated EBs (Steiner et al., 2010). Benefits, pitfalls, and challenges of suspension culture The benefits of culturing hPSCs by stirred-suspension bioreactors could be substantial: (i) rational scalability, (ii) reasonable hPSC expansion rates, (ii) elimination of feeder and matrices, (iv) long-term maintenance of pluripotency (over 20 passages), and (v) compatibility with several formats such as microcarriers and microencapsulation (Chen et al., 2014). Not surprisingly, Chen and colleagues demonstrated that they could maintain the pluripotency of hESCs and expanded H9 (WA09) cells over 20 passages in suspension culture, with a total fold expansion over 1 × 1013 (V.C. Chen et al., 2012). This represents approximately 4.3-fold

Pluripotent stem cell growth models expansion per passage in 3–4 days, which is much better than adherent colonies (K.G. Chen et al., 2012). Encouragingly, the karyotype appears to be normal after long-term suspension culture in at least 10 different laboratories. Large cohort analyses of chromosomal stability at a higher resolution are necessary to ensure its superiority to other culture methods. Nonetheless, the current methods still have several pitfalls that need to be overcome prior to clinical applications, which include: (i) reduced cellular viability in bioreactors due to agitation-induced shear, (ii) dependence on the constant use of Y-27632, which induces aggregation, to improve cell survival rates, and (iii) increased differentiation and cellular heterogeneity.

Single-cell passaging and short-term assays Single-cell passaging using ROCK inhibitors is now considered a routine technique to promote single-cell survival (Watanabe et al., 2007). Ding and coworkers demonstrated that several novel synthetic molecules could dramatically promote cloning efficiency of dissociated hESCs by enhancing ECM-mediated integrin activity (Xu et al., 2010). Wu and colleagues assessed the effects of small molecule combinations on colonies derived from dissociated single cells and found several small molecule cocktails support long-term maintenance of hESCs (Tsutsui et al., 2011). This method can be used to dissociate hPSC colonies to perform single-cells assays, cryopreservation, and genetic manipulation. However, the majority of studies using single-cell passaging techniques employed low-density cell seeding and hence produced colonies. In addition, long-term single cell culture may induce chromosomal abnormalities.

Non-colony type monolayer (NCM) culture We and others have recently developed a culture method based on single-cell passaging and NCM growth (K.G. Chen et al., 2012; Kunova et al., 2013). The key element in this technology is high-density single-cell seeding, thereby permitting uniform multicellular association and avoiding the formation of colonies (Fig. 2B). There are many potential advantages to applying NCM for hPSC biology. An immediate benefit of the NCM method is to accurately control growth rates, which yield approximately 3–4 fold average cell expansion per passage in 3–4 days, comparable to the outcomes from some suspension cultures (Chen et al., 2014). The demonstrated scalability of NCM culture to different sizes of culture dishes may ensure hPSC expansion using currently compatible cell culture systems to produce clinically relevant cell quantities. Importantly, the NCM method aims to homogenize heterogeneous cellular states in hPSC colonies. Indeed, cells under NCM conditions showed a more homogeneous response to signal molecules such as BMP-4 compared to the same cells in colony-type culture (K.G. Chen et al., 2012). In general, global gene expression patterns, determined by cDNA microarray assays, were found to be similar to each appropriate colony culture in three hPSC lines, with correlation ranges from 0.91 to 0.94. No EMT-related gene clusters were significantly increased in hPSCs under NCM conditions (K.G. Chen et al., 2012). Notably, the cell yields under NCM conditions are better than colony-type culture, but less than suspension culture methods (Chen et al., 2014).

617 In our previous study, there were approximately 13% (2/16) lines showing chromosomal abnormalities (K.G. Chen et al., 2012). Since hPSCs under NCM conditions were adapted from previous colony-type culture, it is unclear whether the chromosomal abnormalities observed in NCM culture were due to the selection of preexisting mutant cells of the colony-type culture or generated de novo under NCM. Further comparative studies are needed to clarify this issue.

Current challenges and future directions Obviously, colony type, non-colony type, and aggregated cultures show many advantages in hPSC research. However, each method has its own limitations, which might undermine its capacity to produce clinical-grade hPSCs. In the near future, we need to completely control the variability, heterogeneity, and chromosomal instability to ensure faithful hPSC expansion and differentiation. A systematic understanding of the mechanisms that contribute to the refractory properties of hPSC culture would facilitate the achievement of these goals. Moreover, development of hierarchical growth models for organogenesis would also have great clinical potential.

Control hPSC survival and fates with small molecules The use of small molecules that inhibit or activate core signaling pathways provides a powerful approach to support hPSC survival and direct lineage commitment. However, the long-term effects of small molecules on hPSC genomic integrity and differentiation capacities have not been carefully evaluated. For example, several culture protocols (as discussed above) require the use of the ROCK inhibitor Y-27632 to enhance single-cell plating and stabilize intercellular aggregations at every cell passage. Thus, the effects of Y-27632 on global gene expression and genomic stability should be carefully evaluated. Under enzymatically dissociated conditions, damaged single cells are supposed to be eliminated by programmed cell death or apoptosis. It is unclear if ROCK inhibition-mediated single-cell survival also preserves apoptotic cells that might harbor potential genomic mutations. Verification of these mechanistic details would provide new strategies to maintain genomic integrity of hPSCs. Nevertheless, further modifications and characterization of single-cell-based methods (such as suspension culture and NCM) are necessary to ensure its wide application for directed differentiation. First, to induce the cells to a specific precursor stage would facilitate directed differentiation of hPSCs under various culture conditions. It is likely that Y-27632-based NCM culture would enhance the differentiation efficiency of hPSCs toward neuroectodermal cells (K.G. Chen et al., 2012). We have also tested the effects of various small molecules that interact with ROCK signaling pathways on both single-cell plating and NCM growth. We found that Y-39983 (ROCK I inhibitor), phenylbenzodioxane (ROCK II inhibitor), and thiazovivin (a novel ROCK inhibitor) can also modulate single-cell plating and promote NCM growth. Theoretically, these small molecules may be also used for suspension culture and single-cell based assays. Furthermore, high-throughput screening would enable us to identify the next-generation of small molecules that drive directed differentiation of hPSCs. Second, the development of single-cell based methods without the use of small molecules,

618 such as Y-27632, would be potentially important for cell expansion and drug screening. The use of human recombinant laminin-521 enables single-cell plating, thus permitting us to establish NCM and suspension culture without the aid of Y-27632 and to produce homogeneous hPSCs in a chemically defined environment (Chen et al., unpublished data). Finally, it would be also possible to integrate optimized NCM and suspension cultures into a well-controlled multidimensional environment for tissue morphogenesis or organogenesis.

K.G. Chen et al.

Modulation of intercellular interactions for organogenesis Stem-cell niches, a hub for stem cell renewal and differentiation, are spatially oriented, temporally controlled, and usually heterogeneous micro-environments in mammalian tissues. ES cell fate determination can be traced in a 3D stem cell niche (e.g., in hair-follicles) in animal models (Rompolas et al., 2013). It is unclear how hPSC niches are

Figure 4 Intercellular interactions mediated by autocrine and paracrine signalings maintain stem cell niches and induce morphogenesis. (A) Stem cell niche maintenance exemplified by the interaction between IGF-II (from hdFs) and IGF1R of hESCs. Various intercellular interactions are also depicted. (B) Modes of intercellular interactions induce hPSC differentiation, maturation, and self-organization (morphogenesis). Abbreviations: hdFs, hESC-derived fibroblast-like cells; IGF-II, insulin-like growth factor 2; IGF1R, insulin-like growth factor 1 receptor.

Pluripotent stem cell growth models created in vitro microenvironment. Knowledge from both developmental biology and various hPSC culture models (Figs. 1–4) revealed that both autocrine and paracrine signaling may play an important role in hPSC self-renewal, lineage differentiation, and tissue specification in an artificial stem cell niche. Direct cell–cell interactions mediated by autocrine signaling This type of interaction defines self-secretory ligands binding at corresponding receptors of the same cells or the same types of cells (Fig. 4). These direct interactions may be exemplified by Notch signaling which promotes neural differentiation of hESCs. Activation of a highly conserved Notch pathway endorses neural lineage commitment in both mESCs and hESCs (Lowell et al., 2006) and improves motor skills after ischemic injury in adult rat brain (AndroutsellisTheotokis et al., 2006). Stromal cells that express Notch ligands also encourage neural specification of hESCs. Hence, incomplete activation of endogenous Notch results in heterogeneous lineage commitment. This neural commitment also necessitates a parallel signaling through the FGFR, highlighting the significance of cross-talk between core signaling pathways. Indirect cell–cell interactions via paracrine signaling Pluripotent stem cell fates are certainly influenced by extrinsic signals produced from heterogeneous types of cells. Paracrine signaling from different types of cells may have differential effects on hPSC fates. It has been shown that interplay between IGF-II (secreted from hESC-derived fibroblast-like cells, i.e., hdFs) and IGF1R (constitutively expressed in hESCs) promotes hESC self-renewal and the pluripotent state (Fig. 4A) (Bendall et al., 2007). Moreover, when hESCs were cocultured with human mesenchymal and umbilical vein endothelial cells, they formed self-organized miniorganoids (i.e., liver buds) (Takebe et al., 2013). Thus, heterogeneous signaling appears to be required to induce lineage differentiation and subsequent tissue morphogenesis (Fig. 4B). The generation of miniorganoids provides a proof-of-concept experiment that organogenesis can proceed in a well-defined 3D environment. In the coming years, we anticipate using completely defined growth factors and small molecules to replace heterogeneous types of cells and to generate pure populations of desired cells for transplantation therapy in patients.

Conclusions Human pluripotent stem cell growth models based on colonytype culture are widely used in stem cell research. However, these low efficient methods induce EMT, multicellular heterogeneity, and substantial genomic instability. Emerging hPSC growth models (such as NCM and suspension culture of undifferentiated hPSCs) provide robust and scalable expansion of hPSCs as well as many other utilities, making them ideal methods to complement colony-type culture. With further understanding of core signaling pathways that control intercellular interactions, efficient production of clinicalgrade hPSCs and 3D-tissue morphogenesis in vitro could become available for regenerative medicine and pharmaceutical development.

619

Disclosure of potential conflicts of interest The authors indicate no potential conflict of interest.

Acknowledgments This work was supported by the Intramural Research Program of the National Institutes of Health (NIH) at the National Institute of Neurological Disorders and Stroke. We would like to thank Dr. Peter Zandstra for his comments and suggestions.

References Alikani, M., 2005. Epithelial cadherin distribution in abnormal human pre-implantation embryos. Hum. Reprod. 20, 3369–3375. Amps, K., Andrews, P.W., Anyfantis, G., Armstrong, L., Avery, S., Baharvand, H., Baker, J., Baker, D., Munoz, M.B., Beil, S., et al., 2011. Screening ethnically diverse human embryonic stem cells identifies a chromosome 20 minimal amplicon conferring growth advantage. Nat. Biotechnol. 29, 1132–1144. Androutsellis-Theotokis, A., Leker, R.R., Soldner, F., Hoeppner, D.J., Ravin, R., Poser, S.W., Rueger, M.A., Bae, S.K., Kittappa, R., McKay, R.D., 2006. Notch signalling regulates stem cell numbers in vitro and in vivo. Nature 442, 823–826. Baker, D.E., Harrison, N.J., Maltby, E., Smith, K., Moore, H.D., Shaw, P.J., Heath, P.R., Holden, H., Andrews, P.W., 2007. Adaptation to culture of human embryonic stem cells and oncogenesis in vivo. Nat. Biotechnol. 25, 207–215. Bauwens, C.L., Peerani, R., Niebruegge, S., Woodhouse, K.A., Kumacheva, E., Husain, M., Zandstra, P.W., 2008. Control of human embryonic stem cell colony and aggregate size heterogeneity influences differentiation trajectories. Stem Cells 26, 2300–2310. Behr, R., Heneweer, C., Viebahn, C., Denker, H.W., Thie, M., 2005. Epithelial–mesenchymal transition in colonies of rhesus monkey embryonic stem cells: a model for processes involved in gastrulation. Stem Cells 23, 805–816. Bendall, S.C., Stewart, M.H., Menendez, P., George, D., Vijayaragavan, K., Werbowetski-Ogilvie, T., Ramos-Mejia, V., Rouleau, A., Yang, J., Bosse, M., et al., 2007. IGF and FGF cooperatively establish the regulatory stem cell niche of pluripotent human cells in vitro. Nature 448, 1015–1021. Braam, S.R., Zeinstra, L., Litjens, S., Ward-van Oostwaard, D., van den Brink, S., van Laake, L., Lebrin, F., Kats, P., Hochstenbach, R., Passier, R., et al., 2008. Recombinant vitronectin is a functionally defined substrate that supports human embryonic stem cell selfrenewal via alphavbeta5 integrin. Stem Cells 26, 2257–2265. Brafman, D.A., Phung, C., Kumar, N., Willert, K., 2012. Regulation of endodermal differentiation of human embryonic stem cells through integrin–ECM interactions. Cell Death Differ. 20, 369–381. Bratt-Leal, A.M., Carpenedo, R.L., McDevitt, T.C., 2009. Engineering the embryoid body microenvironment to direct embryonic stem cell differentiation. Biotechnol. Prog. 25, 43–51. Carlson, B.M. (Ed.), 2009. Human Embryology and Developmental Biology, 4th edn. Mosby, Inc., Philadelphia. Catalina, P., Montes, R., Ligero, G., Sanchez, L., de la Cueva, T., Bueno, C., Leone, P.E., Menendez, P., 2008. Human ESCs predisposition to karyotypic instability: is a matter of culture adaptation or differential vulnerability among hESC lines due to inherent properties? Mol. Cancer 7, 76. Chen, K.G., Mallon, B.S., Hamilton, R.S., Kozhich, O.A., Park, K., Hoeppner, D.J., Robey, P.G., McKay, R.D., 2012a. Non-colony type monolayer culture of human embryonic stem cells. Stem Cell Res. 9, 237–248.

620 Chen, V.C., Couture, S.M., Ye, J., Lin, Z., Hua, G., Huang, H.I., Wu, J., Hsu, D., Carpenter, M.K., Couture, L.A., 2012b. Scalable GMP compliant suspension culture system for human ES cells. Stem Cell Res. 8, 388–402. Chen, K.G., Mallon, B.S., McKay, R.D., Robey, P.G., 2014. Human pluripotent stem cell culture: considerations for maintenance, expansion, and therapeutics. Cell Stem Cell 14, 13–26. Dale, B., Gualtieri, R., Talevi, R., Tosti, E., Santella, L., Elder, K., 1991. Intercellular communication in the early human embryo. Mol. Reprod. Dev. 29, 22–28. Dang, S.M., Gerecht-Nir, S., Chen, J., Itskovitz-Eldor, J., Zandstra, P.W., 2004. Controlled, scalable embryonic stem cell differentiation culture. Stem Cells 22, 275–282. Davidson, K.C., Adams, A.M., Goodson, J.M., McDonald, C.E., Potter, J.C., Berndt, J.D., Biechele, T.L., Taylor, R.J., Moon, R.T., 2012. Wnt/beta-catenin signaling promotes differentiation, not selfrenewal, of human embryonic stem cells and is repressed by Oct4. Proc. Natl. Acad. Sci. U. S. A. 109, 4485–4490. Desmarais, J.A., Hoffmann, M.J., Bingham, G., Gagou, M.E., Meuth, M., Andrews, P.W., 2012. Human embryonic stem cells fail to activate CHK1 and commit to apoptosis in response to DNA replication stress. Stem Cells 30, 1385–1393. Doble, B.W., Woodgett, J.R., 2007. Role of glycogen synthase kinase-3 in cell fate and epithelial–mesenchymal transitions. Cells Tissues Organs 185, 73–84. Domogatskaya, A., Rodin, S., Boutaud, A., Tryggvason, K., 2008. Laminin-511 but not -332, -111, or -411 enables mouse embryonic stem cell self-renewal in vitro. Stem Cells 26, 2800–2809. Domogatskaya, A., Rodin, S., Tryggvason, K., 2012. Functional diversity of laminins. Annu. Rev. Cell Dev. Biol. 28, 523–553. Draper, J.S., Smith, K., Gokhale, P., Moore, H.D., Maltby, E., Johnson, J., Meisner, L., Zwaka, T.P., Thomson, J.A., Andrews, P.W., 2004. Recurrent gain of chromosomes 17q and 12 in cultured human embryonic stem cells. Nat. Biotechnol. 22, 53–54. Eastham, A.M., Spencer, H., Soncin, F., Ritson, S., Merry, C.L., Stern, P.L., Ward, C.M., 2007. Epithelial–mesenchymal transition events during human embryonic stem cell differentiation. Cancer Res. 67, 11254–11262. Gualtieri, R., Santella, L., Dale, B., 1992. Tight junctions and cavitation in the human pre-embryo. Mol. Reprod. Dev. 32, 81–87. Gumbiner, B.M., 1996. Cell adhesion: the molecular basis of tissue architecture and morphogenesis. Cell 84, 345–357. Hartung, O., Huo, H., Daley, G.Q., Schlaeger, T.M., 2010. Clump passaging and expansion of human embryonic and induced pluripotent stem cells on mouse embryonic fibroblast feeder cells. Curr. Protoc. Stem Cell Biol. (Chapter 1, Unit 1C 10). Hyka-Nouspikel, N., Desmarais, J., Gokhale, P.J., Jones, M., Meuth, M., Andrews, P.W., Nouspikel, T., 2012. Deficient DNA damage response and cell cycle checkpoints lead to accumulation of point mutations in human embryonic stem cells. Stem Cells 30, 1901–1910. James, D., Levine, A.J., Besser, D., Hemmati-Brivanlou, A., 2005. TGFbeta/activin/nodal signaling is necessary for the maintenance of pluripotency in human embryonic stem cells. Development 132, 1273–1282. Kehoe, D.E., Jing, D., Lock, L.T., Tzanakakis, E.S., 2010. Scalable stirred-suspension bioreactor culture of human pluripotent stem cells. Tissue Eng. Part A 16, 405–421. Kleinman, H.K., McGarvey, M.L., Liotta, L.A., Robey, P.G., Tryggvason, K., Martin, G.R., 1982. Isolation and characterization of type IV procollagen, laminin, and heparan sulfate proteoglycan from the EHS sarcoma. Biochemistry 21, 6188–6193. Krawetz, R., Taiani, J.T., Liu, S., Meng, G., Li, X., Kallos, M.S., Rancourt, D.E., 2010. Large-scale expansion of pluripotent human embryonic stem cells in stirred-suspension bioreactors. Tissue Eng. Part C Methods 16, 573–582.

K.G. Chen et al. Kunova, M., Matulka, K., Eiselleova, L., Salykin, A., Kubikova, I., Kyrylenko, S., Hampl, A., Dvorak, P., 2013. Adaptation to robust monolayer expansion produces human pluripotent stem cells with improved viability. Stem Cells Transl. Med. 2, 246–254. Lee, A.S., Tang, C., Rao, M.S., Weissman, I.L., Wu, J.C., 2013. Tumorigenicity as a clinical hurdle for pluripotent stem cell therapies. Nat. Med. 19, 998–1004. Lefort, N., Feyeux, M., Bas, C., Feraud, O., Bennaceur-Griscelli, A., Tachdjian, G., Peschanski, M., Perrier, A.L., 2008. Human embryonic stem cells reveal recurrent genomic instability at 20q11.21. Nat. Biotechnol. 26, 1364–1366. Li, X., Chen, Y., Scheele, S., Arman, E., Haffner-Krausz, R., Ekblom, P., Lonai, P., 2001. Fibroblast growth factor signaling and basement membrane assembly are connected during epithelial morphogenesis of the embryoid body. J. Cell Biol. 153, 811–822. Li, R., Liang, J., Ni, S., Zhou, T., Qing, X., Li, H., He, W., Chen, J., Li, F., Zhuang, Q., et al., 2010. A mesenchymal-to-epithelial transition initiates and is required for the nuclear reprogramming of mouse fibroblasts. Cell Stem Cell 7, 51–63. Lowell, S., Benchoua, A., Heavey, B., Smith, A.G., 2006. Notch promotes neural lineage entry by pluripotent embryonic stem cells. PLoS Biol. 4, e121. Ludwig, T.E., Levenstein, M.E., Jones, J.M., Berggren, W.T., Mitchen, E.R., Frane, J.L., Crandall, L.J., Daigh, C.A., Conard, K.R., Piekarczyk, M.S., et al., 2006. Derivation of human embryonic stem cells in defined conditions. Nat. Biotechnol. 24, 185–187. Magyar, J.P., Nemir, M., Ehler, E., Suter, N., Perriard, J.C., Eppenberger, H.M., 2001. Mass production of embryoid bodies in microbeads. Ann. N. Y. Acad. Sci. 944, 135–143. Mallon, B.S., Park, K.Y., Chen, K.G., Hamilton, R.S., McKay, R.D., 2006. Toward xeno-free culture of human embryonic stem cells. Int. J. Biochem. Cell Biol. 38, 1063–1075. Maurer, J., Nelson, B., Cecena, G., Bajpai, R., Mercola, M., Terskikh, A., Oshima, R.G., 2008. Contrasting expression of keratins in mouse and human embryonic stem cells. PLoS One 3, e3451. Mayshar, Y., Ben-David, U., Lavon, N., Biancotti, J.C., Yakir, B., Clark, A.T., Plath, K., Lowry, W.E., Benvenisty, N., 2010. Identification and classification of chromosomal aberrations in human induced pluripotent stem cells. Cell Stem Cell 7, 521–531. Miner, J.H., Li, C., Mudd, J.L., Go, G., Sutherland, A.E., 2004. Compositional and structural requirements for laminin and basement membranes during mouse embryo implantation and gastrulation. Development 131, 2247–2256. Moogk, D., Stewart, M., Gamble, D., Bhatia, M., Jervis, E., 2010. Human ESC colony formation is dependent on interplay between self-renewing hESCs and unique precursors responsible for niche generation. Cytometry A 77, 321–327. O'Brien, C., Laslett, A.L., 2012. Suspended in culture—human pluripotent cells for scalable technologies. Stem Cell Res. 9, 167–170. Padmanabhan, R., Chen, K.G., Gillet, J.P., Handley, M., Mallon, B.S., Hamilton, R.S., Park, K., Varma, S., Mehaffey, M.G., Robey, P.G., et al., 2012. Regulation and expression of the ATP-binding cassette transporter ABCG2 in human embryonic stem cells. Stem Cells 30, 2175–2187. Pratama, G., Vaghjiani, V., Tee, J.Y., Liu, Y.H., Chan, J., Tan, C., Murthi, P., Gargett, C., Manuelpillai, U., 2011. Changes in culture expanded human amniotic epithelial cells: implications for potential therapeutic applications. PLoS One 6, e26136. Rodin, S., Domogatskaya, A., Strom, S., Hansson, E.M., Chien, K.R., Inzunza, J., Hovatta, O., Tryggvason, K., 2010. Long-term self-renewal of human pluripotent stem cells on human recombinant laminin-511. Nat. Biotechnol. 28, 611–615. Rompolas, P., Mesa, K.R., Greco, V., 2013. Spatial organization within a niche as a determinant of stem-cell fate. Nature 502, 513–518.

Pluripotent stem cell growth models Sato, N., Meijer, L., Skaltsounis, L., Greengard, P., Brivanlou, A.H., 2004. Maintenance of pluripotency in human and mouse embryonic stem cells through activation of Wnt signaling by a pharmacological GSK-3-specific inhibitor. Nat. Med. 10, 55–63. Serra, M., Brito, C., Correia, C., Alves, P.M., 2012. Process engineering of human pluripotent stem cells for clinical application. Trends Biotechnol. 30, 350–359. Steiner, D., Khaner, H., Cohen, M., Even-Ram, S., Gil, Y., Itsykson, P., Turetsky, T., Idelson, M., Aizenman, E., Ram, R., et al., 2010. Derivation, propagation and controlled differentiation of human embryonic stem cells in suspension. Nat. Biotechnol. 28, 361–364. Stewart, M.H., Bosse, M., Chadwick, K., Menendez, P., Bendall, S.C., Bhatia, M., 2006. Clonal isolation of hESCs reveals heterogeneity within the pluripotent stem cell compartment. Nat. Methods 3, 807–815. Takebe, T., Sekine, K., Enomura, M., Koike, H., Kimura, M., Ogaeri, T., Zhang, R.R., Ueno, Y., Zheng, Y.W., Koike, N., et al., 2013. Vascularized and functional human liver from an iPSC-derived organ bud transplant. Nature 499, 481–484. Thiery, J.P., Acloque, H., Huang, R.Y., Nieto, M.A., 2009. Epithelial– mesenchymal transitions in development and disease. Cell 139, 871–890. Thomson, J.A., Itskovitz-Eldor, J., Shapiro, S.S., Waknitz, M.A., Swiergiel, J.J., Marshall, V.S., Jones, J.M., 1998. Embryonic stem cell lines derived from human blastocysts. Science 282, 1145–1147. Tsutsui, H., Valamehr, B., Hindoyan, A., Qiao, R., Ding, X., Guo, S., Witte, O.N., Liu, X., Ho, C.M., Wu, H., 2011. An optimized small molecule inhibitor cocktail supports long-term maintenance of human embryonic stem cells. Nat. Commun. 2, 167.

621 Twiss, F., de Rooij, J., 2013. Cadherin mechanotransduction in tissue remodeling. Cell. Mol. Life Sci. 70, 4101–4116. Ullmann, U., In't Veld, P., Gilles, C., Sermon, K., De Rycke, M., Van de Velde, H., Van Steirteghem, A., Liebaers, I., 2007. Epithelial– mesenchymal transition process in human embryonic stem cells cultured in feeder-free conditions. Mol. Hum. Reprod. 13, 21–32. Ullmann, U., Gilles, C., De Rycke, M., Van de Velde, H., Sermon, K., Liebaers, I., 2008. GSK-3-specific inhibitor-supplemented hESC medium prevents the epithelial–mesenchymal transition process and the up-regulation of matrix metalloproteinases in hESCs cultured in feeder-free conditions. Mol. Hum. Reprod. 14, 169–179. Ungrin, M.D., Joshi, C., Nica, A., Bauwens, C., Zandstra, P.W., 2008. Reproducible, ultra high-throughput formation of multicellular organization from single cell suspension-derived human embryonic stem cell aggregates. PLoS One 3, e1565. Vanneste, E., Voet, T., Le Caignec, C., Ampe, M., Konings, P., Melotte, C., Debrock, S., Amyere, M., Vikkula, M., Schuit, F., et al., 2009. Chromosome instability is common in human cleavage-stage embryos. Nat. Med. 15, 577–583. Watanabe, K., Ueno, M., Kamiya, D., Nishiyama, A., Matsumura, M., Wataya, T., Takahashi, J.B., Nishikawa, S., Nishikawa, S., Muguruma, K., et al., 2007. A ROCK inhibitor permits survival of dissociated human embryonic stem cells. Nat. Biotechnol. 25, 681–686. Xu, C., Inokuma, M.S., Denham, J., Golds, K., Kundu, P., Gold, J.D., Carpenter, M.K., 2001. Feeder-free growth of undifferentiated human embryonic stem cells. Nat. Biotechnol. 19, 971–974. Xu, Y., Zhu, X., Hahm, H.S., Wei, W., Hao, E., Hayek, A., Ding, S., 2010. Revealing a core signaling regulatory mechanism for pluripotent stem cell survival and self-renewal by small molecules. Proc. Natl. Acad. Sci. U.S.A. 107, 8129–8134.