Cell-to-Cell Variability and Culture Conditions during Self-Renewal Reversibly Affect Subsequent Differentiation of Mouse Embryonic Stem Cells by Jit Hin Tan Bachelor of Science in Chemical Engineering Cornell University, Ithaca, NY, 2003 Master of Science in Chemical Engineering Practice Massachusetts Institute of Technology, Cambridge, MA, 2004 Submitted to the Department of Chemical Engineering in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Chemical Engineering Practice at the Massachusetts Institute of Technology June 2013 © 2013 Massachusetts Institute of Technology All Rights Reserved Signature of Author……………………………………………………………………………… Department of Chemical Engineering September 6, 2011 Certified by………………………………………………………………………………………... Clark K. Colton Professor of Chemical Engineering Thesis Supervisor Accepted by…………………………………………………………………………………………………... William M. Deen Professor of Chemical Engineering Chairman, Committee for Graduate Students 1 -Page Intentionally Left Blank- 2 Cell-to-Cell Variability and Culture Conditions during Self-Renewal Reversibly Affect Subsequent Differentiation of Mouse Embryonic Stem Cells by Jit Hin Tan Submitted to the Department of Chemical Engineering on September 6, 2011 in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Chemical Engineering Practice Cell-to-cell variability in clonal populations is reflected in a distribution of mRNA and protein levels among individual cells, including those of key transcription factors governing embryonic stem cell (ESC) pluripotency and differentiation. This may be a source of heterogeneity resulting in mixtures of cell types in differentiated populations despite efforts to control the differentiation conditions and the use of a clonal starting population. In addition, this distribution may be affected by the cell microenvironment during self-renewal. Prior studies on self-renewal culture of ESC, however, focused on long term proliferation and pluripotency. The effects of culture conditions during selfrenewal on the effectiveness of subsequent differentiation protocols remains understudied. Using a mouse ESC line harboring a GFP reporter, we examined cell-to-cell variability in clonal undifferentiated populations and how such variability affects subsequent differentiation. Subpopulations sorted according to their levels of Oct4-GFP expression displayed distinctly different expression levels of pluripotency and early differentiation markers and differentiated into cardiomyocytes at different efficiencies. However, when allowed to self-renew after sorting, the subpopulations regenerated the parental distributions of Oct4-GFP and subsequent differentiation after regeneration did not show differences. In addition to differences between cells in a clonal population, self-renewal conditions affecting Oct4 expression on the population-level were examined. Changes in culture conditions during self-renewal by low oxygen culture or small molecule dual inhibition (2i) of mitogen-activated protein kinase and glycogen synthase kinase reversibly affected levels of Oct4 expression in cells that were otherwise pluripotent. Effects of different self-renewal conditions immediately preceding differentiation are manifested by changes in subsequent differentiation to cardiomyocytes. This study demonstrates that manipulation of self-renewal culture conditions can lead to changes in the outcomes of defined differentiation protocols, a novel dimension to explore for directed differentiation of pluripotent stem cells. Thesis Supervisor: Clark K. Colton Title: Professor of Chemical Engineering 3 -Page Intentionally Left Blank- 4 Acknowledgements The single name on the front page belies the generous contributions of many, without whom this thesis would not have been possible. When I returned to MIT for the PhD program, a research area heavily underpinned by biology was not on my radar, due to a lack of prior training. In spite of this, I was welcomed with open arms into the research group by my advisor, Professor Clark Colton. As Clark mentioned during a tribute to his graduate advisor Ed Merrill, he “did not know that a graduate student was supposed to do the research that his advisor chooses for him”, and I am very grateful that Clark has allowed me the same ignorance during my thesis work to explore questions of my own choice. I must also thank the members of my thesis committee: Professor William Deen, for his invaluable advice in navigating the ins and outs of the doctoral process, and for reminding me of the value of my training in Chemical Engineering; Professor Rudolf Jaenisch, whom, despite his busy schedule, never refused a meeting, and is able, in a short discussion, to open up new horizons to explore; and Professor Narendra Maheshri, who has been generous in his suggestions on both experiments worth doing, and perhaps more importantly, not worth doing. To the members of the Colton Lab – Kevin Brower, Anna Coclite, Amanda DiIenno, Amy Johnson, Clarke Low, Michelle Miller, Jeff Millman: No one else understands the journey we all undertake as well as all of you do, and I am extremely grateful for your fellowship and for allowing me to commiserate during my time here. Also, I offer sincere thanks to Kellie Young, my UROP, who remained enthusiastic despite some of the mundane tasks I had assigned during her time with the lab. In particular, I am deeply indebted to Jeff, who had been instrumental in showing me the ropes and also holding long discussions on practically every aspect of my work throughout the entire time. Jeff inspires me with his insight and enthusiasm for research, and has taught by example hard work and perseverance in the face of fickle cell cultures and impenetrable protocols. I leave the lab in Amanda’s very capable hands, and am confident that she will take the research ever further. Many thanks must also go to the underappreciated Val Grimm, our administrative assistant, who has skillfully helped me to wrangle schedules and untangle paperwork. I would like to thank the many facilities and laboratories who have offered assistance, expertise and equipment that we were lacking and were essential to the completion of this work: Dr Malgorzata Borowiak (Melton Lab, Harvard University), for providing the Oct4-GFP mESC reporter line that made this work possible; Glenn Paradis, Michael Jennings, and Michele Griffin of the Koch Institute Flow Cytometry Core Facility, whose expertise and patience I have taken advantage of repeatedly; Han-Hwa Hung and Linda Bragman of the Centre of Biomedical Engineering’s qPCR Facility for allowing me to hog the equipment on many occasions, and T.L To, formerly of the Maheshri lab, for generously sharing equipment and expertise in the many common biological assays and protocols that I was regretfully short of. As Michael Mosley opined, “Science is a human endeavor. It’s been shaped as much by what’s outside the laboratory as inside.” To ignore the people who have been a part of 5 my life outside of research would be to ignore those who have provided an equal, though indirect, contribution to this thesis. I have been extremely blessed to be part of two warm and welcoming groups outside of the laboratory, who have provided safe haven and sanity when I have needed it most, shared in the highs and provided unquestioning support during the lows – my fellow Singaporeans, who I have subjected to many culinary experiments and hijacked conversations steered towards research; and the graduate community of Sidney-Pacific, who have given me free reign over the kitchen without the obligation of responsibility. When you are away from home, your friends become your family – as you all have. I will always be impressed and humbled by your great achievements and diverse abilities, honest humility and generous spirits. Sometimes things do not go according to plan, and the journey takes an unexpected detour. Looking back, I see a path that I may not have chosen at the start, but I am happy that I have taken, because of the people I would not have met, lessons I would not have learnt, and experiences I would have not have gained had I not done so. Half a world away, there are four people who continue to shower me with their unconditional love and support - my parents, Michael and Doreen, and my siblings Jit Quan and Michelle, who, despite not fully understanding why this process has taken me away from home for so long, remain patient and understanding. Sometimes the most important things are the hardest to say and I find myself at a loss for words, except Thank you. 6 Dedicated to my parents, whose unconditional love made this possible 7 -Page Intentionally Left Blank- 8 Table of Contents Chapter 1 Introduction ............................................................................................................................. 15 1.1 Pluripotent stem cells............................................................................................................... 15 1.2 Self-renewal and propagation in the pluripotent state ........................................................ 16 1.3 Directed differentiation............................................................................................................. 18 1.4 Cell-to-cell variability in clonal undifferentiated PSC populations ..................................... 20 1.5 Effects of oxygen on the undifferentiated phenotype ......................................................... 21 1.5.1 Derivation of new embryonic stem cell lines ...................................................................... 22 1.5.2 Pluripotency ............................................................................................................................ 23 1.5.3 Proliferation ............................................................................................................................. 26 1.5.4 Clonal recovery ....................................................................................................................... 26 1.5.5 Reprogramming ...................................................................................................................... 27 1.5.6 Control of pO2cell ..................................................................................................................... 27 1.5.7 Effects on subsequent differentiation .................................................................................. 29 1.6 Use of 2i in PSC derivation and propagation ....................................................................... 30 1.7 Objectives .................................................................................................................................. 31 1.8 Overview .................................................................................................................................... 32 Chapter 2 Cell-to-cell variability within an undifferentiated clonal population affects subsequent differentiation of mouse embryonic stem cells ..................................................................................... 35 2.1 Abstract ........................................................................................................................................... 35 2.2 Introduction ..................................................................................................................................... 37 2.3 Methods .......................................................................................................................................... 40 2.3.1 Self-renewing culture of mESC ............................................................................................ 40 2.3.2 Cell cycle analysis .................................................................................................................. 40 2.3.3 Fluorescence-Activated Cell Sorting (FACS) of undifferentiated O4G cells ................. 41 2.3.4 Flow cytometric analysis of Oct4-GFP expression ........................................................... 41 2.3.5 Differentiation of mESC to cardiomyocytes ....................................................................... 42 2.3.6 Real-time polymerase chain reaction (PCR) ..................................................................... 43 2.3.7 Cell enumeration .................................................................................................................... 44 2.3.8 Flow cytometric analysis of MF20+ cells ............................................................................ 44 2.3.9 Statistics .................................................................................................................................. 45 2.4 Results ............................................................................................................................................ 46 2.4.1 Large cell-to-cell variability in Oct4-GFP expression in undifferentiated mESC which corresponds to Oct4 mRNA expression ....................................................................................... 46 2.4.2 Distribution of Oct4-GFP is not entirely cell cycle dependent ......................................... 46 9 2.4.3 Levels of Oct4-GFP mark undifferentiated subpopulations with different levels of pluripotency and early differentiation markers ............................................................................. 47 2.4.4 Populations differentiated from cells with higher initial Oct4-GFP levels had a greater fraction and number of cardiomyocytes ........................................................................................ 47 2.4.5 Sorted subpopulations recapitulate the parental Oct4-GFP distribution when allowed to propagate in self-renewing culture ............................................................................................ 49 2.4.6 Recapitulation of the parental Oct4-GFP distribution erases subpopulation differences in pluripotency and early differentiation markers......................................................................... 50 2.4.7 Recapitulation of the parental Oct4-GFP distribution removes the differences in subsequent differentiation to cardiomyocytes ............................................................................. 50 2.5 Discussion....................................................................................................................................... 52 2.6 Future Work .................................................................................................................................... 55 2.7 Tables and Figures........................................................................................................................ 56 Chapter 3 Culture oxygen levels and use of small molecule inhibitors during self-renewal reversibly affect subsequent differentiation of mouse embryonic stem cells .................................. 68 3.1 Abstract ........................................................................................................................................... 69 3.2 Introduction ..................................................................................................................................... 71 3.3 Methods .......................................................................................................................................... 73 3.3.1 Self-renewing culture of mESC ............................................................................................ 73 3.3.2 Flow cytometric analysis of Oct4-GFP expression ........................................................... 74 3.3.3 Differentiation of mESC to cardiomyocytes ....................................................................... 74 3.3.4 Control of culture oxygen levels (pO2cell) ............................................................................ 75 3.3.5 Real-time polymerase chain reaction (PCR) ..................................................................... 75 3.3.6 Cell enumeration .................................................................................................................... 76 3.3.7 Flow cytometric analysis of MF20+ cells ............................................................................ 76 3.3.8 Estimation of spontaneously contracting areas ................................................................. 77 3.3.9 Statistics .................................................................................................................................. 77 3.4 Results ............................................................................................................................................ 78 3.4.1 Self-renewal of mESC at low oxygen reduces expression of Oct4-GFP but does not increase rates of spontaneous differentiation .............................................................................. 78 3.4.2 Reduction of Oct4-GFP expression by low oxygen culture is reversed by continued self-renewal in 20% oxygen............................................................................................................ 78 3.4.3 Low oxygen preconditioning decreases subsequent differentiation of mESC to cardiomyocytes ................................................................................................................................. 79 3.4.4 Total cell number and fraction of residual undifferentiated Oct4-GFP+ cells are not affected by preconditioning, but by oxygen conditions during differentiation ......................... 80 3.4.5 Reduction in efficiency of cardiomyogenesis due to low oxygen preconditioning is reversible by self-renewal in 20% oxygen before differentiation .............................................. 81 10 3.4.6 2i use during self-renewal increases expression of naïve pluripotency markers, decreases early differentiation markers synergistically with high oxygen, and is reversible by subsequent self-renewing culture without 2i........................................................................... 82 3.4.7 High oxygen and 2i use during preconditioning separately, synergistically, and reversibly increases subsequent differentiation to cardiomyocytes ......................................... 84 3.4.8 Fraction of MF20+ cells in differentiated populations correlates positively with mean Oct-GFP levels of the starting populations................................................................................... 86 3.5 Discussion and future work .......................................................................................................... 88 3.6 Figures ............................................................................................................................................ 91 Chapter 4 Integrative Perspective Paper: Opportunities and Obstacles in the Commercialization of Pluripotent Stem Cell-Based Products and Services ................................................................... 114 4.1 Executive Summary .................................................................................................................... 115 4.2 Introduction ................................................................................................................................... 116 4.3 Pluripotent Stem Cell Research and Diagnostic Consumables and Equipment ............... 117 4.4 Pluripotent Stem Cell-Derived Tools for Disease Modelling, Drug Discovery, and Toxicology ........................................................................................................................................... 120 4.5 Replacement Cell Therapies Derived from Pluripotent Stem Cells ..................................... 124 4.6 Discussion..................................................................................................................................... 130 References .............................................................................................................................................. 132 11 -Page Intentionally Left Blank- 12 List of Figures Figure 2-1 Characterization of Oct4-GFP (O4G) mESC reporter line demonstrates large cell-tocell variability in Oct4-GFP expression which corresponds to Oct4 mRNA levels ......................... 56 Figure 2-2 Oct4-GFP variability cannot be explained by differences in cell cycle stage .............. 57 Figure 2-3 Real-time PCR analysis of sorted O4G mESC subpopulations reveal that levels of Oct4-GFP mark undifferentiated subpopulations with different levels of pluripotency and early differentiation markers ............................................................................................................................. 58 Figure 2-4 Populations differentiated from cells with higher initial Oct4-GFP levels had a greater fraction and number of cardimyocytes .................................................................................................. 60 Figure 2-5 Oct4-GFP fluorescence distributions of sorted subpopulations cultured in selfrenewing conditions for up to 30 days after initial sorting demonstrate recapitulation of the parental Oct4-GFP distribution over time. ............................................................................................ 62 Figure 2-6 Time course of mean Oct4-GFP fluorescence of the subpopulations relative to the All subpopulation............................................................................................................................................ 63 Figure 2-7 Effects of regeneration of parental Oct4-GFP expression on gene expression of subpopulations.......................................................................................................................................... 64 Figure 2-8 Recapitulation of the parental Oct4-GFP distribution removes the differences in subsequent differentiation to cardiomyocytes...................................................................................... 66 Figure 3-1 Effects of culture O2 in self-renewing culture on distributions of Oct4-GFP expression in O4G mESC. .......................................................................................................................................... 91 Figure 3-2 Mean Oct4-GFP fluorescence and fraction of Oct4-GFP+ cells from self-renewal at different O2. ............................................................................................................................................... 92 Figure 3-3 Effects of O2 preconditioning on subsequent differentiation. ......................................... 94 Figure 3-4 Effects of O2 preconditioning on total cell number and fraction of residual Oct4-GFP+ cells after differentiation. ......................................................................................................................... 96 Figure 3-5 Reversibility of O2 preconditioning. .................................................................................... 97 Figure 3-6 Correlation between mean Oct4-GFP fluorescence of the starting undifferentiated populations and the fraction of MF20+ cells in the resulting differentiated populations................ 99 Figure 3-7 Experimental design and nomenclature for preconditioning with O2 modulation and 2i supplementation. .................................................................................................................................... 100 Figure 3-8 Effects of 2i supplementation and O2 during self-renewal on Oct4-GFP fluorescence distributions of undifferentiated mESC. .............................................................................................. 101 Figure 3-9 Effects of 2i supplementation and culture O2 on mean Oct4-GFP fluorescence and fraction of Oct4-GFP+ cells during self-renewal................................................................................ 102 Figure 3-10 Effects of 2i supplementation and culture O2 on gene expression of mESC after 18 days in self-renewing culture. ............................................................................................................... 104 Figure 3-11 Effects of 2i supplementation and culture O2 on gene expression of mESC after 24 days in self-renewing culture. ............................................................................................................... 106 Figure 3-12 Effects of 2i supplementation and culture O2 on subsequent differentiation after 18 days of preconditioning. ........................................................................................................................ 108 Figure 3-13 Effects of 2i supplementation and culture O2 on subsequent differentiation after 24 days of preconditioning. ........................................................................................................................ 109 13 Figure 3-14 Correlation between the fraction of MF20+ cells in the resulting differentiated populations for cells preconditioned for 18 (red triangles) and 24 days (blue circles) before differentiation and the mean Oct4-GFP fluorescence of the starting undifferentiated populations. Cells were differentiated at 20% O2. For cells preconditioned for 18d, <R2>= 0.63; for cells preconditioned for 24d, <R2>= 0.93, both from a linear least squares regression....................... 111 Figure 3-15 Linear and semi-log correlations between mean Oct4-GFP fluorescence of starting undifferentiated populations and the fraction of MF20+ cells in the resulting differentiated populations for all cells differentiated at 20% O2. .............................................................................. 112 Figure 3-16 Model fit for correlation between mean Oct4-GFP fluorescence of starting undifferentiated populations and the fraction of MF20+ cells in the resulting differentiated populations for all cells differentiated at 20% O2. .............................................................................. 113 14 Chapter 1 Introduction 1.1 Pluripotent stem cells First isolated from mouse [1], then more recently human blastocysts [2], pluripotent embryonic stem cells (ESC) have the ability to self-renew in the undifferentiated state indefinitely and differentiate into all cell types of the adult organism. They hold immense clinical and research potential for drug screening, disease modeling, as a model to study mammalian development, and to generate replacement cells for application in regenerative medicine [3–9]. Cardiomyocytes [10], bone marrow [11], dopamine neurons [12], and β-cells [13] have been differentiated from ESC as potential therapies for cardiac infarctions, leukemia, Parkinson’s disease, and diabetes respectively. Other studies have also demonstrated the ability to generate self-organized tissues [14] and to perform gene and cell therapy using ESC [15]. The discovery of another source of pluripotent cells, induced pluripotent stem cells (iPSC) [16–18] gave further hope for personalized therapies by autologous transplantation of cells and tissues differentiated from iPSC derived from the patients themselves [19], and for recapitulation of disease development by generating disease specific iPSC [20], in addition to all of the potential applications of ESC [21–25]. iPSC and ESC have shown similar potential to differentiate into cell types of interest and fulfill other criteria for pluripotency [26–28], however, some differences remain [29]. For pluripotent stem cells (PSC) of both types, many obstacles remain before their immense potential can be fulfilled. This includes removing the danger of tumorigenicity 15 of residual undifferentiated cells [30], immunogenicity of the transplanted cells [31], the development of methods for propagation in their undifferentiated state, and efficient protocols for directed differentiation in order to generate sufficient numbers and purities of the desired cell types [32]. 1.2 Self-renewal and propagation in the pluripotent state In order to generate sufficient numbers of cells, PSC can be propagated in their undifferentiated state until differentiation is initiated. For mouse ESC (mESC), this is usually accomplished by either co-culture on mouse embryonic fibroblasts (MEFs) or the addition of leukemia inhibitory factor (LIF) [33], a cytokine which acts by activating the JAK-STAT3 pathway [34]. Human ESC (hESC) can be kept in the pluripotent state by culture medium supplemented with basic Fibroblast Growth Factor (bFGF) [35] on a layer of various feeder cells [36] or an animal-derived protein matrix [37], although recently a fully defined self-renewal culture system has been discovered [38]. The main concern of undifferentiated PSC culture is to ensure the maintenance of the pluripotent state and prevent spontaneous differentiation while sustaining normal karyotype and cell viability. To do so, it must support the highly complex cellular machinery for maintaining pluripotency that is underpinned at its core by three pluripotency genes – Oct4, Sox2, and Nanog [39–43]. Disruption of these core transcription factors caused spontaneous differentiation of PSC and loss of the pluripotent state. 16 An ectopic increase in Oct4 expression in mESC directed spontaneous differentiation down the mesodermal lineage [44, 45] while a decrease in Oct4, or increase in Cdx2 expression relative to Oct4, directed differentiation to trophectoderm [45, 46]. Similar behavior occurred in hESC with an ectopic decrease in Oct4 expression resulting in spontaneous differentiation to ectoderm and mesoderm, while an ectopic increase caused differentiation to endoderm [47]. Both Oct4 and Sox2 determine germ layer fate decisions during early differentiation in opposite directions, with Oct4 promoting differentiation down the mesodermal and endodermal lineages, while Sox2 promoted differentiation down the ectodermal lineage, and each suppressed the promoted lineage of the other [48]. Expression of Nanog prevented differentiation, and mESC can exhibit transient down-regulation of Nanog expression. This down-regulation has been hypothesized to generate a necessary intermediate state between maintaining pluripotency and differentiation [49, 50]. In hESC, Nanog affected cell fate choice, with an extended expression of Nanog during differentiation leading to increased mesendodermal specification [51]. Recent studies demonstrated that multiple metastable pluripotent states for PSC exist as distinct subpopulations within a clonal population. In addition to Nanog, Rex1 [52] and Stella [53] are also heterogeneously expressed in mESC. These subpopulations were able to reversibly interconvert when the population was kept in conditions supporting self-renewal of the pluripotent state. Given that the culture conditions that maintain pluripotency can also affect the transition of mESC between the different pluripotent states, and that pluripotency factors can specify differentiation to different 17 cell fates [48, 54], it is thus important to assess the effects of the self-renewal conditions on PSC and on the efficiency of subsequent differentiation to different lineages. Two aspects in particular – culture oxygen levels, and the use of small molecule inhibitors supporting undifferentiated culture – are discussed in greater detail in sections 1.5 and 1.6. 1.3 Directed differentiation The design of directed differentiation protocols for pluripotent stem cells has employed many different strategies to enrich the resulting differentiated population with the desired cell type [55, 56]. To direct differentiation to cardiomyocytes from both mouse and human PSC [57], for example, protocols have included the use of growth factors such as Activin or Nodal and BMP4 [58], the use of small molecules 5-aza-2’deoxycytidine [59] or ascorbic acid [60], genetic manipulation to ectopically express a cytokine, TNF-α [61], co-culture with a visceral endoderm-like cell line (END-2) [62], or, most recently, synthetic mRNA transfection [63]. Combinations of techniques, such as the addition of stauprimide with Activin A for directed differentiation to definitive endoderm [64], or the step-wise protocol of sorting Flk1+ cells differentiated from miPSC, then co-culturing the purified intermediate population on OP9 stroma cells to obtain cardiomyocytes [65], have also been documented. The designs of directed differentiation protocols have grown in sophistication, inspired in part by insights from developmental biology. Strict temporal control of the cellular microenvironment in order to mimic developmental cues, optimizing concentrations of 18 growth factors (Activin A for example, can induce both endodermal and mesodermal specification depending on concentration), and optimizing levels of growth factors depending on the stage of differentiation have yielded success [13, 58, 66]. The focus, however, has been predominantly on the time period and process during which the cells are already differentiating, and much remains unknown about how the culture conditions before the initiation of differentiation affects the cells subsequently. This is in spite of the many studies (summarized above in section 1.2) showing that pluripotency factors, which are highly active when PSCs are in their undifferentiated state, can play roles in specifying cell fate during differentiation. The focus of this thesis is on cell-to-cell variability (population heterogeneity), culture oxygen levels, and the use of small molecules during self-renewing culture on subsequent differentiation. It examines the relationship between starting undifferentiated state of the ESC and the resulting differentiated population, an aspect of self-renewal that has not received much attention. Experiments on cell-to-cell variability examine differences between clonal undifferentiated cells within a population, while studies combining culture oxygen levels and the use of small molecules in self-renewal culture examine their effects on self-renewal and the efficiency of subsequent directed differentiation of undifferentiated populations as a whole. More detailed treatments of these topics follow. 19 1.4 Cell-to-cell variability in clonal undifferentiated PSC populations Cell-to-cell variability or cellular heterogeneity in clonal populations has received increasing attention with the development of methods to measure characteristics of single cells, rather than traditional methods yielding population-averaged results. Studies showing heterogeneity in single mammalian cells of a clonal population in mRNA [67, 68] and protein levels [69] that can have important biological functions suggest that traditional ensemble measurements insufficiently capture the characteristics of single cells [70, 71]. Numerous studies have reported heterogeneity in clonal stem cell populations that correlate with differences in subsequent cell fate choice. Hematopoietic stem cells have transcriptome-wide differences in subpopulations sorted by expression of a surface antigen, Sca-1, despite being clones generated from a single cell [72], and levels of Sca-1 correlated to differentiation efficiencies to different cell lineages. Within clonal populations of mESC, there were heterogeneities in expression of Nanog [50], Rex1 [52], and Stella [53], all normally associated with pluripotency, and in an early endodermal marker Hex [73]. Subpopulations exhibiting high or low levels of expression of these genes exhibited different differentiation proclivities when allowed to differentiate. Levels of these genes in mESC allowed to self-renew in the undifferentiated state, however, are dynamic, and single cells can switch between high and low levels of expression. In both human ESC and iPSC, heterogeneities occurred in mRNA levels of multiple genes, although single hiPSC exhibited higher levels of heterogeneity than hESC [74]. The multitude of metastable pluripotent states has led to a proposal that 20 pluripotency is not a set of distinct states, but instead a continuum of states [75]. This heterogeneity, however, is viewed as undesirable in directed differentiation protocols because it can lead to heterogeneity in the differentiated population, which can be detrimental to the production of clinically useful cells and tissues [76]. Cell-to-cell variability in mRNA expression levels can occur due to gene expression noise, arising from transcription occurring in bursts, rather than at a continuous steady rate [68, 77, 78]. Together with a positive feedback loop, for example, a gene that activates its own transcription, such stochastic gene expression is able to generate stable distinct phenotypes [79]. Many theories have been put forward to describe the mechanisms behind such cell-to-cell variability, and it has been proposed that the large interconnected network of genes and epigenetic states in mammalian systems, together with transcriptional noise, create multiple stable gene expression profiles that clonal single cells may inhabit [80]. More importantly, as evidenced by effects on subsequent differentiation of cells in different “states”, such cell-to-cell variability can have important developmental functions. 1.5 Effects of oxygen on the undifferentiated phenotype Adapted and updated from Millman, J.R., J.H. Tan, and C.K. Colton, The effects of low oxygen on self-renewal and differentiation of embryonic stem cells. Current Opinions in Organ Transplantation, 2009. 14(6): p. 694-700 with permission from Lippincott Williams & Wilkins 21 One aspect of the culture microenvironment that has mostly been ignored in conventional cell culture is the oxygen partial pressure experienced by the cells. All cells of the developing embryo are exposed to oxygen levels in vivo far below that of atmospheric levels, and they differentiate and undergo organogenesis in a low oxygen environment [81–87]. The local oxygen environment is potentially an important variable for manipulation by researchers hoping to recapitulate specific differentiation pathways in vitro. Mammalian cellular responses to oxygen levels are predominantly mediated by Hypoxia Inducible Factors (HIFs), a family of heteromeric transcription factors. HIF-α subunits are constitutively expressed but are post-translationally regulated through rapid degradation in the presence of oxygen levels higher than 5% [88–90]. The HIF-1β subunit (also known as the aryl hydrocarbon receptor nuclear translocator or ARNT) is oxygen insensitive, but is only active as a transcription factor when dimerized with the appropriate stabilized HIF-α subunit under sufficient cellular hypoxia. The mammalian hypoxic response includes up-regulation of a large number of genes, including the vascular endothelial growth factor (VEGF), the adipose differentiation related protein (ADRP), and RTP801, which is known to be regulated in development, amongst others [91]. 1.5.1 Derivation of new embryonic stem cell lines Derivation of new ES cell lines, both mouse [92, 93] and human [38, 94], benefit from being performed in low oxygen levels. Studies examining the establishment of new ES 22 cell lines from blastocysts under 5% oxygen have had greater success than under the 20% more commonly used. The benefits of low oxygen conditions included an increase in the number of individual colonies exhibiting alkaline phosphatase activity [93], an increase in the number of proliferating cells [92], and removal of the need for undefined animal-derived products such as serum or matrices in the process [38], the last of which can be crucial in the applicability of these derived cells to clinical applications. Most recently, it has been found that maintenance of constant low oxygen during derivation of hESC enables the establishment of pre-X inactivated cells, which are similar to mESC, and exposure to 20% oxygen or oxidative stress irreversibly causes X inactivation [94]. 1.5.2 Pluripotency Low oxygen affected the pluripotent phenotype of mESCs when they were cultured with LIF [33]. Normally, mESC colonies have smooth edges, but this morphological phenotype was lost with low oxygen exposure, instead causing the colonies to have fibroblast-like morphologies [95] or to form jagged edges [96], both signs of early differentiation despite the presence of LIF. These morphological changes were also associated with reduced levels of alkaline phosphatase activity and pluripotency gene (Oct4, Nanog, and Sox2) expression [95, 96]. One study [95] proposed that HIF-1α protein, the concentration of which increases at low oxygen, mediated suppression of the LIF-STAT3 pathway as evidenced by a reduction in the expression of the LIF receptor under hypoxia. Interestingly despite these observations, the cells maintained some potential to differentiate into cells of all three germ layers [96]. 23 In contrast to mESCs, expression of Oct4, Nanog, and Sox2 in hESCs were apparently not affected by hypoxic culture (2% to 5% oxygen) [94, 97, 98]. In one study, however, Nanog and Notch1 mRNA expression increased in hESCs cultured over a four-week period at 5% oxygen compared to those cultured under 20% oxygen. This difference was not maintained after long term culture under the hypoxic condition, since there was no significant difference in Nanog and Notch1 mRNA levels between cells cultured for 18 months under either 5% or 20% oxygen. Consistent with earlier results, no differences in Oct4 expression were observed under both oxygen conditions with either short or long term culture [99]. However, a recent report showed down-regulation of Oct4, Nanog and Sox2 in HIF-2α silenced hESC, together with reduced proliferation [100]. Hypoxic culture resulted in wide-spread transcriptome differences in hESCs when compared to culture under ambient oxygen conditions, including downstream targets of the core pluripotency network Oct4, Nanog, Sox2. Genes marking early differentiation to all cell lineages were upregulated at 20% oxygen [94, 98]. In addition, Lefty2, a member of the TGFβ family of proteins, which is hypothesized to play a role in preventing differentiation, was down-regulated at 20% oxygen compared to culture at 4% [98]. This may lead to a greater proclivity for the hESC to differentiate. In studies on the transcriptome-wide effects of hypoxic culture on hESCs kept under self-renewing conditions, culture at 20% O2 resulted in greater heterogeneity amongst different hESC lines than when cultured under hypoxic (2-4% oxygen) conditions [101]. This 24 heterogeneity may arise from subpopulations of cells already undergoing early differentiation when cultured at 20% oxygen. While low oxygen appears not to affect pluripotency gene expression directly, differentiated areas of each colony expressing SSEA-1, a marker of early differentiation for hESCs, were reduced when cultured under 3% and 1% oxygen compared to culture under 20% oxygen [102]. In a different study, however, culture at 2% oxygen did not affect expression of either SSEA-4 or SSEA-1 in hESCs when compared to ambient oxygen conditions [97]. HIF-1α has been investigated as an important gene mediating ESC hypoxic response. One study [103] suggests a possible link connecting HIF-1α action with Notch signaling. In mouse embryonic teratocarcinoma cells and neural stem cells, hypoxia allowed for the accumulation of HIF-1α, which stabilized the Notch intracellular domain, in turn promoting the undifferentiated phenotype. The action of Notch in promoting self-renewal appears to be conserved in hESCs in view of a recent demonstration that self-renewal is mediated by activation of the Notch pathway and that benefits of hypoxic culture were eliminated with inhibition of Notch signaling [99]. Interestingly, although another hypoxia-inducible factor in the same family as HIF1α, HIF-2α has not been shown to be active in mESC [91], other studies have shown that it is the HIF-2α heterodimer and not HIF-1α that can induce transcription of Oct4 [104], providing a potential link between culture oxygen levels and pluripotency gene expression in ESC. Several studies suggested that the hypoxic response mediated by HIF-1α is transient and not sustained 25 with chronic hypoxia or normoxia [94, 99]. However, it is the HIF-2α response that may influence pluripotency gene expression after long-term continuous culture under hypoxic conditions [100]. 1.5.3 Proliferation Many studies have found that low oxygen affects proliferation, but this effect varied between studies and cell lines. In one study [93], proliferation of mESCs increased under 5% compared to 20% oxygen, while in another [96], no difference was observed in one cell line and a small increase in another. Culture at 1% oxygen or lower, however, reduced the rate of proliferation compared to 20% [95, 96]. mESCs were able to adapt and continue to proliferate under anoxia, albeit with a longer doubling time compared to 20% oxygen, and had the ability to utilize anaerobic respiration exclusively for their energy requirements [96]. Similarly disparate results have been reported for hESCs: In one study, proliferation rates at 5% oxygen were reduced compared to 20% [99], in another, no differences in the rates were observed at 5% and 3% and only a slight reduction at 1% [102]. A conflicting recent report showed that low oxygen culture increased proliferation of hESC [100]. 1.5.4 Clonal recovery Low oxygen culture improved clonal recovery after passaging. mESCs passaged at 5% oxygen [34] and hESCs at 2% [101, 105] had improved clonal survival. Cytogenetic instability at high passages for mESCs was not reduced in culture at 5% oxygen [92], and further reductions in oxygen to even 0% did not affect the amount of oxidative DNA 26 damage incurred by the cells [96]. In contrast for hESCs, 2% oxygen reduced chromosomal abnormalities after multiple passages [101], suggesting a further benefit of hypoxic culture. 1.5.5 Reprogramming Moderate hypoxic conditions improved reprogramming of somatic cells to iPSC for both mouse and human lines. Performing reprogramming at 5% oxygen increased the number of iPSC colonies generated over both 20% and 1% oxygen conditions [106]. It is possible that this increase in reprogramming efficiency may be due to increased proliferation under 5% oxygen, as reprogramming is a stochastic process that can be accelerated by increasing rates of cell division [107]. 1.5.6 Control of pO2cell Most of the studies reviewed in this section implicitly make the assumption that the oxygen levels in the bulk gas phase (pO2gas) was equal to the actual levels of oxygen experienced by the cell (pO2cell). However, pO2cell can differ drastically from pO2gas,depending on the cell density, medium height, cellular oxygen consumption rate, and oxygen permeability of the culture substrate [108]. Furthermore, step changes in pO2gas due to removal of culture dishes from the incubators may result in transient changes in pO2cell with long time scales for equilibration. These problems are ameliorated by culturing the cells on the surface of a membrane with very high oxygen permeability, such as silicone rubber. 27 Differences in pO2cell between different experiments carried out at the same pO2gas provide a possible explanation for the inconsistencies in data reported. For instance, one study [99] reported reduced proliferation of hESCs cultured at 5% oxygen, and another [102] reported no reduction in proliferation at the same oxygen condition. In the first study, there were four weeks between passages (initial and final cell numbers not reported). In the second study, 6x105 cells were plated per well of standard six-well tissue culture plates, and there were only two weeks between passages, with final cell numbers of 1.5-2.5x106 cells per well. It is possible that the pO2cell of the first study was significantly lower than the pO2gas because of higher cell densities, in particular after extended culture without passage, while pO2cell in the latter study was closer to pO2gas. This would help explain the differences in results because the authors of the second study report a reduced proliferation at 1% gas phase oxygen level, which may be closer in cellular oxygen level to the 5% gas phase oxygen level of the first study that also demonstrated a reduced proliferation rate. Moreover, the first group used the Xvivo system (BioSpherix, Lacona, NY), which allowed for continuous culture low pO2gas without interruption for media changes and passaging. The second group used a conventional sealed chamber from which cells were periodically removed for media changes and passaging, allowing for periods where pO2gas is at ambient levels, disrupting the low pO2cell conditions. Without information regarding the experimental conditions, especially the media heights and cell densities, it is impossible to draw any conclusions. 28 The lack of recognition that pO2cell is often different from pO2gas, sometimes by a substantial amount, is exacerbated because values of cell density and media height necessary for retrospective estimation of pO2cell are often not reported. Knowledge of pO2cell is necessary for interpreting the results of studies and comparing them to data from other studies. Efforts to estimate or control pO2cell are essential for rational advancement of this field. 1.5.7 Effects on subsequent differentiation Given that low oxygen partial pressure (pO2) is an important physiological cue during development [109], it is not surprising that low pO2 during differentiation markedly influences resulting cell fates [110–116]. One example is the increase in fraction and number of cardiomyocytes by low oxygen culture during the first 6 days of differentiation [117]. However, despite the many studies on the effects of oxygen levels employed during self-renewal culture, its effect on subsequent differentiation of PSC is largely unknown. A naïve hypothesis based on the previous two observations may be that since low oxygen is beneficial to cardiomyogenesis in early differentiation, low oxygen culture during self-renewal immediately before differentiation would have a synergistic positive effect as well. Given the effects of oxygen levels on self-renewal culture, it is important to study the effects of modulating oxygen on subsequent differentiation, since it is the differentiated cells that are of therapeutic importance. 29 1.6 Use of 2i in PSC derivation and propagation Small molecules that are able to target specific signaling pathways or proteins have been invaluable in advancing stem cell culture. They allowed for defined conditions and media to be used in the generation, propagation, and differentiation of the desired cell types [118]. In particular, this thesis focuses on the use of a pair of small molecule inhibitors – PD0325901, a Mitogen-Activated Protein Kinase (MEK) inhibitor [119], and CHIR99021, a Glycogen Synthase Kinase 3β (GSK3β) inhibitor [120], together known as ‘2i’ [121]. PD0325901 acts as a MEK1/2 inhibitor that prevents the phosphorylation of ERK1/2, a kinase that can activate transcriptional activity associated with activation of the RTK (Receptor Tyrosine Kinase) signal transduction pathway associated with the Fibroblast Growth Factor (FGF). CHIR99021 acts as a GSK3β inhibitor that mimics the activation of the canonical Wnt pathway. Activation of the Wnt pathway or application of CHIR99021 prevents the degradation of β-catenin and allows for its translocation to the nucleus, dimerization with LEF/TCF, and subsequent transcriptional activity [122]. 2i was first identified as a replacement for LIF to allow for undifferentiated propagation of mESC [34]. It improves the efficiency of reprogramming of mouse iPSC to the naïve state when used together with LIF [121]. Along the same vein, 2i use in culture of mouse epiblast stem cells (mEpiSC), considered a “primed” state of pluripotency, can convert the cells to a “mESC-like” naïve state [123]. More recently, 2i, when used with ectopic gene expression, LIF, and Forskolin, is able to convert hESC to become a naïve 30 LIF-responsive state [124]. Furthermore, a cocktail of 2i with PD173074, a Fibroblast Growth Factor Receptor (FGFR) inhibitor, and Y27632, a Rho Kinase (ROCK) inhibitor, was able to maintain the undifferentiated state of hESC without feeder cells with singlecell passaging [125]. The increasing use of highly specific small molecules in defined stem cell culture, in particular, in propagating undifferentiated cells, raises the question about their downstream effects on subsequent differentiation of the cells. Studies have shown effects on expression of core pluripotency genes including Oct4, which affect cell fate during differentiation. It is thus not inconceivable that small molecule use during selfrenewal would have effects on cell fate after subsequent differentiation. 1.7 Objectives In order for PSC to fulfill their immense clinical and research potential, there must exist efficient methods for directing differentiation to the desired cell types. Two aspects that have been largely ignored have been differences within a clonal population that may exist due to cell-to-cell variation, and differences between clonal populations that have been subjected to different culture conditions that support undifferentiated PSC selfrenewal but may manifest differences in subsequent differentiation. The central hypothesis behind this thesis is that the starting “state” of a mESC affects the resulting cell fate after subsequent differentiation, despite using undifferentiated mESC from the same cell line that were also subjected to the same differentiation protocol. The objectives of this research were to identify effects of cell-to-cell variability and self- 31 renewing culture conditions (including culture oxygen levels and the use of small molecule inhibitors) immediately before the initiation of differentiation on the efficiency of established directed differentiation protocols in order to further improve them. This research would establish a novel dimension to explore in the design of future directed differentiation protocols. 1.8 Overview Chapter 2: Cell-to-cell variability within an undifferentiated clonal population affects subsequent differentiation of mouse embryonic stem cells This chapter describes the effects of differences within cells of an undifferentiated population of mESC, specifically Oct4 expression in single cells, on their efficiency of subsequent differentiation to cardiomyocytes. Using a mESC line harboring a GFP reporter driven by the Oct4 distal enhancer, subpopulations of mESC with high, medium, or low Oct4-GFP levels were sorted by FACS and assessed for pluripotency and early differentiation markers. Sorted cells were differentiated immediately or allowed to selfrenew before subsequent differentiation. (1) Differentiation immediately after sorting examined the effects of the initial levels of Oct4 expression on the differentiated cell fate; (2) self-renewal after sorting examined if the subpopulations are stably distinct or if the differences are transient; and (3) differentiation after the subpopulations have been allowed to self-renew examines if subpopulations with different differentiation efficiencies as found in (1) could be stably isolated and propagated. Differentiated cells were assessed for fractions of residual undifferentiated Oct4-GFP+ cells, 32 immunostained with MF20 antibodies for presence of sacromeric myosin heavy chain, and assessed for cardiac gene expression by qRT-PCR. Chapter 3: Culture oxygen levels and use of small molecule inhibitors during selfrenewal reversibly affect subsequent differentiation of mouse embryonic stem cells This chapter describes the effects of culture oxygen levels and use of a small molecule inhibitor cocktail (‘2i’) during self-renewal (termed ‘preconditioning’) on subsequent differentiation of mESC. mESC were kept in undifferentiated culture with strict control of cellular oxygen levels using silicone rubber membrane-based culture plates in conditions of 20%, 5%, or 1% oxygen, with or without addition of 2i in LIF supplemented medium. Cells self-renewed under different conditions were assessed for levels of Oct4GFP expression and gene expression of pluripotency and early lineage markers. Cells were subsequently differentiated and assessed for efficiency of differentiation to cardiomyocytes by immunostaining for sacromeric myosin heavy chain protein, qRTPCR assessment of cardiac gene expression, and semi-quantitative assessments of spontaneously contracting cells. In addition, the reversibility of different preconditioning effects was examined – for cells that received low oxygen preconditioning, further selfrenewal at 20% oxygen before differentiation; and for cells that received 2i supplementation during the initial 9 day preconditioning, further self-renewal without 2i supplementation before differentiation. In addition, the synergistic effects of combining preconditioning that produces the most efficient differentiation to cardiomyocytes with low oxygen during differentiation that has previously been shown to improve 33 differentiation were examined. Finally, a mathematical model describing the relationship between the fraction of cardiomyocytes in the differentiated populations with the initial mean Oct4-GFP levels was developed. 34 Chapter 2 Cell-to-cell variability within an undifferentiated clonal population affects subsequent differentiation of mouse embryonic stem cells 2.1 Abstract Differentiation of embryonic stem cells (ESC) typically results in a mixture of cell types despite use of a clonal starting population of genetically identical undifferentiated cells and of efforts to control the culture microenvironment. One cause may be cell-to-cell variability in gene expression, which results in a distribution of mRNA and protein levels, including those of key transcription factors governing ESC pluripotency and differentiation. Here we investigate how different levels of Oct4 expression resulting from cell-to-cell variation in a clonal undifferentiated population of mouse ESC (mESC) affect their proclivities to differentiate into cardiomyocytes, chosen here as a model system. Using a mESC line harboring a GFP reporter driven by the distal Oct4 enhancer, we found large cell-to-cell variability in Oct4-GFP expression in a clonal population of undifferentiated cells. These differences could not be attributed to cell cycle variation alone. Subpopulations with different levels of Oct4-GFP exhibited large differences in expression of pluripotency and early differentiation markers. Subpopulations sorted through FACS with higher starting Oct4-GFP levels differentiated into populations with greater fractions and numbers of cardiomyocytes. However, if the sorted populations were allowed to self-renew, they recapitulated the parental distribution of Oct4-GFP expression over a time period measured in days, and this erased the differences in expression of pluripotency and early differentiation markers 35 found immediately after sorting. Sorted populations that were allowed to recapitulate the parental Oct4-GFP distribution before subsequent differentiation did not exhibit differences in the resulting differentiated populations. These results indicate that cell-tocell variation within a clonal population during self-renewal prior to differentiation of mESC can affect the outcome of differentiation. 36 2.2 Introduction Pluripotent stem cells (PSC), including embryonic stem cells (ESC) and induced pluripotent stem cells (iPSC), with their ability to differentiate into all cell types of the adult organism, have enormous potential as a source of replacement tissue for regenerative medicine [4, 24], for studying disease development [22], and for drug screening [126]. One of the main obstacles to their successful application, however, is the need to produce sufficient quantities and purities of viable terminally differentiated or progenitor cells of the desired type [5]. Differentiation of PSC in vitro typically results in a heterogeneous mixture of different cell types [49], despite a starting population of genetically-identical cells and efforts to control the culture microenvironment. Prior studies on directing differentiation of PSC have focused on the process of differentiation, with much effort going to optimizing cellular microenvironments during differentiation [127]. However, heterogeneity in the resulting differentiated cells may have been caused by cell-to-cell variability in the undifferentiated population, even before differentiation had been initiated. This cell-tocell variability manifests as a distribution of mRNA and protein levels, including those of key transcription factors governing ESC pluripotency and differentiation. Indeed, this has shown to be true in clonal hematopoietic stem cells which exhibit cellto-cell transcriptome-wide differences and also correspondingly different proclivities to differentiate into different cell lineages [72]. Studies with ESC have shown that there is heterogeneous expression of Nanog within a clonal undifferentiated population, and its 37 transient down-regulation allows for cells to differentiate [49], [50]. More recently, heterogeneous expression of Rex1 and Stella have also been shown in mESC, and mESC when kept in the undifferentiated state, can interconvert between Rex1+ and Rex1-, and Stella+ and Stella- states. More significantly, when these subpopulations are differentiated, Oct4+/Rex1- mESC showed increased efficiency in differentiation to somatic cells of mesodermal and ectodermal lineages, while Oct4+/Rex1+ mESC differentiated more efficiently to primitive endoderm [52]. Similarly, Oct4+/Stella/Pecam- mESC showed increased efficiency in neuronal and trophectodermal differentiation compared to Oct4+/Stella+/Pecam+ mESC [53]. Prior studies have noted the role of Oct4 in fate choice in differentiating ESC. Ectopic, quantitative increases or decreases in expression of Oct4 control lineage commitment [45, 46, 128]. In addition, the interplay of Oct4, Sox2 and Sox17 during early differentiation affects hESC differentiation to cardiomyocytes [129]. Finally, upon initiation of differentiation, Oct4 and Sox2 play significant roles in germ layer determination [48]. Here, we investigate how different levels of Oct4 expression resulting from cell-to-cell variation at the initiation of differentiation of a clonal mouse ESC (mESC) population affect their proclivities to differentiate into cardiomyocytes, chosen here as a model system. We use a transgenic mESC reporter line harboring a GFP sequence driven by the Oct4 distal enhancer, active in the naïve pluripotent state [130, 131], to allow for non-invasive measurements of Oct4 expression levels. We also study the behavior of 38 subpopulations within a clonal undifferentiated population with different measured levels of Oct4-GFP expression when these subpopulations were allowed to self-renew in order to determine if cell-to-cell variations in Oct4-GFP are fixed or transient and if they retain their differences in affecting subsequent differentiation. 39 2.3 Methods 2.3.1 Self-renewing culture of mESC A transgenic R1 mESC reporter line harboring a GFP sequence driven by the Oct4 distal promoter (O4G) [132], generously donated by Professor Douglas Melton (Harvard University, Cambridge, MA, USA), was used for the experiments. Undifferentiated O4G mESC were propagated as previously described [96] in Dulbecco’s modified Eagles medium optimized for ES cells (ES-DMEM; SCRR-2010; American Type Culture Collection (ATCC), Manassas, VA, USA) supplemented with 10% (v/v) ES cell qualified fetal bovine serum (FBS; SCRR 30-2020; ATCC), 2 mM L-alanyl-L-glutamine (SCRR 30-2115; ATCC), 100 units/mL penicillin and 100 µg/mL streptomycin (30-002-CI; Mediatech, Manassas, VA, USA), 1x MEM non-essential amino acids (30-2116; ATCC), 100 µM 2-mercaptoethanol (M7522; Sigma-Aldrich, St. Louis, MO, USA), and 103 units/mL Leukemia Inhibitory Factor (ESGRO mLIF; ESG1106; Millipore, Billerica, MA, USA) on standard cell culture polystyrene plates (353043; BD Biosciences, Bedford, MA, USA) or flasks (353014; BD Biosciences) that have been previously treated for 30 min with a sterile 0.1% (w/v) solution of Type A gelatin (G-2500; Sigma-Aldrich) in tissue culture water (25-055-CM; Mediatech). Cells were plated at a density of 12,000 cells/cm2. Medium was replaced daily, and cells were detached with 0.25% trypsin (302101, ATCC) every 3 days and replated at 12,000 cells/cm2. 2.3.2 Cell cycle analysis Undifferentiated O4G mESC were detached and incubated as a single cell suspension at a concentration of 106 cells/mL in undifferentiated culture media supplemented with 40 fumitremorgin C to a final concentration of 10 µM (F9054; Sigma-Aldrich) for 30 min before live cell staining of cellular DNA content was performed using Vybrant DyeCycle Violet Stain (V35003; Molecular Probes, Eugene, OR, USA) according to the manufacturer’s instructions. Samples were analyzed using a FACS LSR HTS-2 (Beckton Dickinson) using 405 nm and 488 nm excitation sources (Koch Institute (KI), MIT, Cambridge, MA, USA). All flow cytometry data was analyzed using FlowJo (Tree Star, Ashland, OR, USA). Fluorescence was taken to be proportional to DNA content in individual cells, and the population distribution of DNA content was fitted to a Watson cell cycle model [133] in FlowJo, allowing for separation of the cell population into subpopulations of cells in G1, S, or G2 phases. 2.3.3 Fluorescence-Activated Cell Sorting (FACS) of undifferentiated O4G cells Undifferentiated O4G mESC were sorted according to Oct4-GFP fluorescence in individual cells using a FACS Aria (Beckton Dickinson) or MoFlo (Beckman Coulter) into the highest 15% (‘High’), middle 15% (‘Mid’), and lowest 15% (‘Low’). As a control, cells were also passed through the sorter and all Oct4-GFP-positive cells were collected (‘All’). R1 mESC without the reporter transgene were used as the negative gating control for Oct4-GFP expression. 2.3.4 Flow cytometric analysis of Oct4-GFP expression Oct4-GFP fluorescence in individual cells was measured using a FACScan (Beckton Dickson). The fraction of Oct4-GFP positive cells was calculated as the number of cells above the negative gating threshold as defined above, divided by the number of cells in 41 the entire population at each specific measurement time and condition. For undifferentiated cells, the mean Oct4-GFP fluorescence was determined from all Oct4GFP positive cells in the population. 2.3.5 Differentiation of mESC to cardiomyocytes Cells were differentiated as previously described [108]. Briefly, embryoid bodies (EBs) were formed with 500 mESC per drop in 20µL of cell culture medium as used in selfrenewing culture but without LIF and including 0.2 mM ascorbic acid (A4034; SigmaAldrich). After 2 days, 30 EBs were transferred per well into fibronectin-coated (F1141; Sigma-Aldrich) silicone rubber membrane-based plates [108] filled with 2mL of the EB medium. After an additional 3 days, medium was replaced with 2.5mL of serum-free medium, consisting of 50% (v/v) DMEM (90-133-PB; Mediatech) in tissue culture water and 50% (v/v) Ham’s F-12 (10-080-CV; Mediatech) supplemented with 0.75% (w/v) sodium bicarbonate (25-035-CI; Mediatech), 9mM glucose (G8769; Sigma-Aldrich), 5 µg/mL human insulin (I9278; Sigma-Aldrich), 100 units/mL penicillin, 100 µg/mL streptomycin, and 0.2 mM ascorbic acid. Cells were differentiated at either 20% or 1% O2 throughout the entire protocol for 10 days with oxygen conditions maintained by flowing premixed gases containing 5% O2, and 20% or 1% O2 with a balance of N2 (certified medical gas from Airgas, Hingham, MA, USA), which corresponded under humidified conditions to 142 and 7 mmHg pO2gas inside chambers maintained at 37oC [108]. 42 2.3.6 Real-time polymerase chain reaction (PCR) Samples were collected as single cell suspensions and stored at 4oC in RNAprotect Cell Reagent (76526; Qiagen, Valencia, CA, USA) according to manufacturer’s instructions for no longer than 7 days before RNA isolation. Total RNA was isolated using the RNeasy Plus Mini Kit (74134; Qiagen), cDNA synthesized using the High Capacity cDNA Reverse Transcription Kit (4368814, Applied Biosystems, Foster City, CA, USA). Real-time PCR was performed on a Fast Real-Time PCR System (7900HT, Applied Biosystems), at the Center for Biomedical Engineering (MIT) using Power SYBR Green PCR Master Mix (4367659, Applied Biosystems) or Taqman Gene Expression Master Mix (43669016, Applied Biosystems) as appropriate. 18S ribosomal RNA (rRNA) was used as an endogenous control. Primer sequences used for Nanog, Oct4, Sox2, and Sox17 have been previously reported [96]. Additional primer sequences used (in order of forward and reverse primer, and 5’ to 3’) were cardiac α-Actin GCTTCCGCTGTCCAGAGACT and TGCCAGCAGATTCCATACCA, cardiac Troponin T (cTnT) AGATGCTGAAGAAGGTCCAGTAGAG and CACCAAGTTGGGCATGAAGA, Cdx2 TCACCATCAGGAGGAAAAGTGA and GCTCTGCGGTTCTGAAACCA, and GFP CTGCTGCCCGACAACCA and TGTGATCGCGCTTCTCGTT. Commercially available primers for 18S rRNA (Hs99999901_s1, Applied Biosystems) and Nestin (Mm03053244_s1, Applied Biosystems) were also used. A standard curve was constructed with cDNA pooled from all time points and conditions then serially diluted. 43 2.3.7 Cell enumeration Enumeration of viable cells in undifferentiated culture was performed by staining with Guava Viacount (4000-0041; Millipore) and data acquired using a Guava PCA flow cytometer (Millipore) according to manufacturer’s instructions. For differentiated cells, nuclei enumeration was performed by exposing dispersed cells to a lysis solution of 1% (v/v) Triton X-100 and 0.1 M citric acid (0627; Mallinckrodt Specialty Chemicals, Paris, KY, USA), then vortexed to obtain nuclei samples. Samples were then stained with Guava Viacount and data acquired [96]. 2.3.8 Flow cytometric analysis of MF20+ cells The fraction of MF20+ cells in dispersed cells was acquired with flow cytometry as previously described [96]. Briefly, samples were fixed by a 20 min incubation in 1% (w/v) paraformaldehyde (36606; Alfa Aesar, Ward Hill, MA, USA) in PBS. Fixed cells were permeabilized in 0.5% saponin (S-4521; Sigma-Aldrich) for 10 min, non-specific binding blocked by a 30 min incubation in 2% (v/v) FBS in PBS, and then incubated at room temperature with primary antibodies against sarcomeric myosin heavy chain (MF20; MF20 supernatant; Developmental Studies Hybridoma Bank, Iowa City, IA, USA) diluted 1:10 in 2% FBS solution for 1 hr. Samples were then incubated with goat anti-mouse phycoerythin-conjugated secondary antibodies (115-116-146; Jackson ImmunoResearch; West Grove, PA, USA) diluted 1:250 in 2% FBS solution for 30 min at room temperature in the dark, washed 3x in 2% FBS solution, then data acquired using a FACScan. 44 2.3.9 Statistics Statistical analysis was performed with unpaired two-tailed t-tests, with p<0.05 considered statistically significant. Error bars are ± 1 standard deviation calculated from biological replicates. 45 2.4 Results 2.4.1 Large cell-to-cell variability in Oct4-GFP expression in undifferentiated mESC which corresponds to Oct4 mRNA expression Flow cytometric analysis of Oct4-GFP fluorescence in a clonal population of undifferentiated O4G mESC revealed a large range of Oct4-GFP fluorescence intensities in individual cells, much larger the broadening attributable to instrument errors or resolution limitations as shown by measurements with reference fluorescence beads (SPHERO Rainbow Calibration Particles; RCP-30-5; Spherotech, Lake Forest, IL) (Figure 2-1A). Using FACS, cells were sorted into high, med, and low subpopulations (Figure 2-1B). Oct4-GFP fluorescence intensity, and Oct4 and GFP mRNA levels of the subpopulations correlated positively (Figure 2-1C & D). A clonal population of O4G mESC expressed cell-to-cell variation in Oct4-GFP fluorescence intensities that spanned over two orders of magnitude, which was reflected by the 60x difference in Oct4 mRNA levels between the high and low subpopulations. 2.4.2 Distribution of Oct4-GFP is not entirely cell cycle dependent We next tested the hypothesis that this variation is cell cycle dependent by analysis of DNA content and corresponding Oct4-GFP fluorescence intensity in individual cells (Figure 2-2). Fractions of cells in G1, S, and G2/M phases were similar to other published data for mESC [134]. The Oct4-GFP fluorescence profiles for cells in each cell cycle phase exhibited the same large range of intensities as the whole population with large overlap amongst all three phases. Although peak intensity for cells in G2 was 1.2x higher than cells in S phase, and both the aforementioned were higher than cells in 46 G1 (G2 by 1.5x, S by 1.2x), the differences in peak intensities were much smaller than the range of fluorescence intensities for cells in all three phases, which spanned more than 2 orders of magnitude. These differences between the subpopulations in different phases of the cell cycle were not significant at the 95% confidence interval. 2.4.3 Levels of Oct4-GFP mark undifferentiated subpopulations with different levels of pluripotency and early differentiation markers mRNA expression analysis by real-time PCR revealed cell-to-cell variability in expression of pluripotency and early lineage markers within the clonal undifferentiated population. The sorted subpopulations showed that expression of pluripotency associated genes Oct4, Nanog, and Sox2 [39–41] correlated positively with each other, with each gene showing a greater than an order of magnitude difference in expression between the high and low subpopulations (Figure 2-3A-C). Trophectodermal marker Cdx2 [135], early endodermal marker Sox17 [136] and neural stem cell marker Nestin [137] showed the reverse trend, with higher expression in the Low subpopulation. Expression of Cdx2 was 200x higher, Sox17 20x higher, and Nestin 10x higher in the Low subpopulation than in the High subpopulation (Figure 2-3D-F), despite all cells being Oct4-GFP+. 2.4.4 Populations differentiated from cells with higher initial Oct4-GFP levels had a greater fraction and number of cardiomyocytes Prior studies have shown that ectopically increasing levels of Oct4 in ESC cause spontaneous differentiation into the mesodermal lineage [45] and that Oct4 plays a role 47 in fate specification during early differentiation [48], [129]. We hypothesized that natural cell-to-cell variations in clonal mESC can affect the proclivity of individual cells in the population to differentiate into cells of different lineages, specifically that cells with higher Oct4-GFP at the initiation of differentiation would differentiate into cardiomyocytes with greater efficiency. To test this, we sorted the cells as described above and immediately initiated differentiation by EB formation. In addition, as a control, a 4th population of cells that had simply been passed through the FACS instrument and all Oct4-GFP+ cells collected (‘All’) was differentiated in the same manner as well to control for the effects of FACS. Each subpopulation was differentiated separately at 20% and 1% O2 throughout the entire differentiation protocol (Figure 2-4). Despite the clonal nature of the starting cells, the subpopulations differentiated into cells positively stained for MF20 (anti-sacromeric myosin heavy chain, a cardiac specific antibody [112], [138], [139]) with different efficiencies. Cells differentiated at 20% O2 from the high subpopulation were 8.6% MF20+ compared to 4.2% from the low subpopulation, and when differentiated at 1% O2, the high subpopulation produced 21.7% MF20+ compared to 14.1% MF20+ from the low subpopulation (Figure 2-4A). Following the trend, the med and all subpopulations both produced fractions in between the high and low subpopulations in both cases. These results also replicated earlier studies that low oxygen during differentiation increased cardiomyocyte formation [117], [140]. The positive correlation between initial Oct4-GFP fluorescence and differentiated MF20+ fractions was retained within each oxygen level by the total number of MF20+ 48 cells produced (Figure 2-4B), which was calculated by multiplying the fraction of MF20+ cells by the total number of cells in each differentiated subpopulation (Figure 2-4C). There were no statistically significant differences in the total cell numbers in the populations that were differentiated at the same O2, although differentiation at 1% O2 produced less than 50% of the total cell number than differentiation at 20% O2. In addition, the fraction of residual Oct4-GFP+ cells (Figure 2-4D) and average fluorescence of these cells (data not shown) in each subpopulation were also only dependent on the differentiation O2 (with low O2 decreasing residual fraction of undifferentiated cells as previously shown [139, 140]). Finally, the trend of increasing MF20+ cells with increasing initial Oct4-GFP fluorescence were also corroborated by real-time PCR analysis (Figure 2-4E&F) of expression of cardiac genes Cardiac α-Actin [142] and cTnT [143]. 2.4.5 Sorted subpopulations recapitulate the parental Oct4-GFP distribution when allowed to propagate in self-renewing culture To investigate the stability of the cell-to-cell variation in Oct4-GFP expression intensities, we cultured the sorted subpopulations in self-renewing conditions and measured Oct4GFP fluorescence for up to 30 days (Figure 2-5 & 6). Within 12 hr of sorting, the cells repopulated the entire range of fluorescence intensities of the parental population, and the high and med subpopulations had similar peak intensities 3 days after the sorting. However, the low subpopulation took much longer to reconstitute the parental distributions, and still exhibited small differences (not statistically significant) up to 30 49 days after sorting with a broader distribution and proportionally more cells at lower intensities. 2.4.6 Recapitulation of the parental Oct4-GFP distribution erases subpopulation differences in pluripotency and early differentiation markers We then hypothesized that the regeneration of parental Oct4-GFP levels would also remove differences in pluripotency and early differentiation markers observed in the subpopulation immediately after sorting. RNA was isolated from cells allowed to selfrenew for 30 days after sorting and levels of pluripotency and early differentiation genes were analyzed with real time PCR (Figure 2-7). The large differences in expression of all genes observed immediately after sorting (Figure 2-3) no longer existed in the subpopulations. 2.4.7 Recapitulation of the parental Oct4-GFP distribution removes the differences in subsequent differentiation to cardiomyocytes Finally, we tested the efficiency of differentiation to cardiomyocytes of the regenerated subpopulations by differentiation at 1% O2 (Figure 2-8) There were no significant differences in the differentiated cells of the subpopulations in terms of fraction and number of MF20+ cells, and in the expression of Cardiac α-Actin and cTnT. In addition, the fraction of MF20+ cells (averaging 23.7%) differentiated from cells that had undergone the 30d regeneration after sorting (Figure 2-8A) was higher than those differentiated immediately after FACS (Figure 2-4A). The fraction of MF20+ cells was 50 higher than in the cells differentiated from cells simply passed through FACS without sorting (all population) (14.1%) and comparable to the high subpopulation (21.7%). We postulate that this difference could be due to the harsh treatment of cells during FACS and the extended period that the cells were kept outside the incubators at the initiation of differentiation. 51 2.5 Discussion Using an Oct4-GFP reporter driven by the distal enhancer active when mESC are in the naïve pluripotent state, we studied Oct4 expression in single cells of clonal populations of mESC, and observed large cell-to-cell variations in Oct4 within the undifferentiated populations. Levels of Oct4-GFP expression correlated positively with pluripotency gene expression and negatively with early differentiation markers. These differences translated into a greater fraction and number of cardiomyocytes after differentiation from cells with higher Oct4 expression at the initiation of differentiation. The cell-to-cell variation in Oct4 was transient, and sorted subpopulations with different levels of Oct4GFP expression reconstituted the parental distributions when allowed to self-renew over a period of 30 days. Allowing reconstitution of the parental distributions of Oct4-GFP also erased differences in pluripotency and early differentiation marker expression and in efficiency of subsequent cardiomyocyte differentiation. The differences in differentiation efficiencies of the sorted subpopulations were not due to greater cell survival or proliferation of select subpopulations because the total cell numbers after the differentiation were not significantly different. It also cannot be attributed to a resistance to or slower differentiation by the low subpopulation because the fraction of undifferentiated cells were the same in all populations differentiated at the same O2. Prior studies have shown that levels of Oct4 during early differentiation can influence the differentiation pathway of both human and mouse ESC, on some occasions by ectopically increasing or decreasing expression levels [45, 128, 129]. This study has shown that differences in Oct4 within undifferentiated clonal populations do 52 occur without intervention and, significantly, affect differentiation proclivities in the same manner. It may seem that the results of this study contradict those of an early work which showed the heavy dependence of the pluripotent state of mESC on strict control over Oct4 expression [45], whereas this study has shown large variation in Oct4 expression in mESC. An explanation could be the coordinated expression of Sox2 with Oct4 [144] in this study, rather than an ectopic change in Oct4 expression alone, that allows the mESC to retain the pluripotent state, thereby suggesting that future experiments of such a nature would benefit from studies of multiple genes simultaneously, as it may be their relative expressions that are significant. The observed cell-to-cell variation in Oct4 expression cannot be attributed entirely to the normal passage of cells through the cell cycle due to two observations – the small differences in Oct4-GFP of cells in different phases of the cell cycle, and the long time scale (up to 30 days or 60 cell divisions) required for the regeneration of the parental Oct4-GFP expression after FACS. We also note the significantly longer time required for the low subpopulation to recapitulate the parental distribution, and postulate that transitions to and from a low Oct4 state is slow. There have been many recent studies on sources of heterogeneities within clonal populations (reviewed in [80]). While studies have shown the mechanisms behind gene expression noise and their relationship to phenotypic differences [68, 78, 79, 144], the time scale for relaxation observed in this study suggests another source for the heterogeneity, in addition to noise. These slow fluctuations (such as the >18 day period 53 for regeneration of the sorted low population in this study) in mammalian cells remain poorly understood. It has been proposed that this heterogeneity in gene expression is based in epigenetics, influenced by (yet unidentified) non-stochastic factors; and most importantly, appears to have a role in development [80]. Prior studies of hematopoietic stem cells have shown transcriptome-wide differences in subpopulations sorted by expression of a surface antigen [72] which correlated to differentiation efficiencies to different cell lineages. Earlier studies have shown heterogeneous expression of Nanog [50], Rex1 [52], and Stella [53] in mESC which corresponded to effects on subsequent fate choice. This study has demonstrated a similar phenomenon based on levels of Oct4 gene expression in undifferentiated mESC, which may be correlated with the heterogeneous behaviors of the aforementioned genes. 54 2.6 Future Work Future studies could focus on further characterization of the sorted subpopulations of mESC, which could illuminate processes and mechanisms of early cell fate decisions during which different levels of pluripotency factors affect induction of different germ layers as demonstrated by a recent publication [48]. We hypothesize that the different starting levels of pluripotency factors at the initiation of differentiation ‘carry over’ into early differentiation, hence causing heterogeneity in germ layer induction. Other possibilities include the study of differentiation to other cell types of different lineages, such as neurons; improving the resolution of the sorted subpopulation by decreasing the range of Oct4-GFP intensities within each subpopulation, perhaps even down to the single cell level; simultaneous evaluation of multiple markers [146] to measure the ratios of two markers, such as Oct4 and Sox2 as suggested previously; elimination of cell cycle differences in the undifferentiated populations, perhaps by reversibly trapping the cells in one phase before sorting and differentiation [134]; and non-transgenic methods for non-destructive single cell measurements, perhaps by identifying the appropriate surface antigens suitable for translation to human pluripotent stem cell studies (e.g. SSEA4). Much about the dynamics of genome-wide interactions in such complex systems on the single cell level for both the maintenance of pluripotency and during differentiation remains to be studied. These studies could lead to a better understanding of the sources of heterogeneity within populations differentiated in vitro and improve yields of desired cell products from directed differentiation protocols. 55 2.7 Tables and Figures 250 O4G beads B 200 Med Cell Count 200 Count 250 A 150 100 150 100 High Low 50 50 0 102 103 104 GFP Fluorescence [AU] C High 1 Med Low 0.1 103 104 102 Oct4-GFP Fluorescence [AU] 105 Relative GFP mRNA Expression Relative Mean Oct4-GFP Fluorescence 0 101 D Med 0.1 Low 0.01 0.1 1 Relative Oct4 mRNA Expression High 1 0.1 1 Relative Oct4 mRNA Expression Figure 2-1 Characterization of Oct4-GFP (O4G) mESC reporter line demonstrates large cell-to-cell variability in Oct4-GFP expression which corresponds to Oct4 mRNA levels (A) Comparison of distribution of Oct4-GFP fluorescence and reference fluorescence beads. (B) Undifferentiated O4G mESC sorted by FACS into High, Med, and Low subpopulations. (C) Correlation between Oct4 mRNA expression (real-time PCR) and mean Oct4-GFP fluorescence (flow cytometry) [High = 1]. (D) Correlation between Oct4 and GFP mRNA expressions (real-time PCR) [High = 1]. 56 A Watson Model Cell Count 1200 G1 900 S 600 G2 300 0 100 0 50 DNA Content (DyeCycle Violet) [AU] 100 B G1 Cell Counts 80 S G2 60 40 20 0 104 102 103 Oct4-GFP Fluorescence [AU] Figure 2-2 Oct4-GFP variability cannot be explained by differences in cell cycle stage (A) Cell cycle analysis of undifferentiated and unsorted O4G mESC by fitting to a Watson cell cycle model [133] to identify cells in different cell cycle phases. (B) Oct4GFP distribution of undifferentiated O4G mESC at G1, S, and G2 phases of the cell cycle. 57 100 A 103 Oct4 D Cdx2 E Sox17 F Nestin 102 101 Expression Relative to High Subpopulation 10-1 100 10-2 100 10-1 102 Nanog B 101 10-1 100 10-1 10-2 100 Sox2 C 101 10-1 100 10-2 10-1 Low Med Low High Med High Sorted Subpopulation Figure 2-3 Real-time PCR analysis of sorted O4G mESC subpopulations reveal that levels of Oct4-GFP mark undifferentiated subpopulations with different levels of pluripotency and early differentiation markers 58 Cells were sorted as described in Figure 2-1B into High, Med, or Low subpopulations. Relative expression of pluripotency-associated genes (A) Oct4, (B) Nanog, and (C) Sox2, and trophectodermal marker (D) Cdx2, and early differentiation markers (E) Sox17 and (F) Nestin [High = 1] (N = 3). 59 * Number MF20+ Cells (x10-5) Fraction MF20+ Cells [%] A Starting Subpopulation All 20 Low Med 15 High 25 * * * 10 5 6 * * * 4 2 D 100 6 4 10-1 2 10-2 0 Expression Relative to Undifferentiated mESC (x10-4) * Fraction Residual Oct4-GFP+ Cells [%] Total Cell Number (x10-6) C B * 0 101 0 8 8 E 3 2 * Cardiac α-Actin F * 6 * 4 * 1 0 cTnT * 2 20 0 1 20 1 Differentiation O2 [%] Figure 2-4 Populations differentiated from cells with higher initial Oct4-GFP levels had a greater fraction and number of cardimyocytes Cells were sorted as described in Figure 2-1B into High, Med, or Low subpopulations, or simply passed through FACS without selection (All subpopulation) as a control, and 60 immediately differentiated using the cardiomyocyte protocol at 20% or 1% O2 for 10 days, then analyzed. (A) Fraction of MF20+ cells. (B) Final number of MF20+ cells. (C) Total cell count. (D) Fraction of residual Oct4-GFP+ cells. Real-time PCR analysis of cardiac markers (E) cardiac α-Actin, and (F) cTnT relative to undifferentiated mESC.(Error bars = standard deviation; N = 3; * indicates statistical differences at 95% confidence.) 61 150 100 4H 12H 1D 2D 3D 4D 6D 9D 12D 15D 18D 30D Low Med High All 50 0 150 100 Number of Cells 50 0 150 100 50 0 150 100 50 0 102 103 102 103 104 104 Oct4-GFP Fluorescence [AU] 102 103 104 Figure 2-5 Oct4-GFP fluorescence distributions of sorted subpopulations cultured in self-renewing conditions for up to 30 days after initial sorting demonstrate recapitulation of the parental Oct4-GFP distribution over time. 62 Mean Oct4-GFP Fluorescence relative to All Subpopulation 2 High Med Low All 1.5 1 0.5 0 5 10 15 20 Period of Regeneration [Days] 25 30 Figure 2-6 Time course of mean Oct4-GFP fluorescence of the subpopulations relative to the All subpopulation. 63 101 A Oct4 D Cdx2 B Nanog E Sox17 C Sox2 F Nestin Expression Relative to All Subpopulation 100 10-1 101 100 10-1 101 100 10-1 All All Low Med High Low Med High Sorted and Regenerated Subpopulation Figure 2-7 Effects of regeneration of parental Oct4-GFP expression on gene expression of subpopulations. 64 Cells sorted as described in Figure 2-1B were grown in self-renewing conditions for 30 days before real-time PCR analysis of gene expression. Relative expression of pluripotency-associated genes (A) Oct4, (B) Nanog, and (C) Sox2, and trophectodermal marker (D) Cdx2, and early differentiation markers (E) Sox17 and (F) Nestin [High = 1]. (Error bars = standard deviation; N = 3; No statistically significant differences between gene expression of the regenerated subpopulations were found). 65 8 30 Number MF20+ Cells (x10-5) Fraction MF20+ Cells [%] A 20 10 Expression Relative to Undifferentiated mESC (x10-4) 6 4 2 0 6 0 C B Cardiac α-Actin D cTnT 3 4 2 2 1 0 All Low 0 All Low Med Med High Sorted and Regenerated Subpopulation High Figure 2-8 Recapitulation of the parental Oct4-GFP distribution removes the differences in subsequent differentiation to cardiomyocytes Cells were sorted and allowed to grow in self-renewing conditions for 30 days as described in Figure 2-7 were then subjected to cardiomyocyte differentiation for 10 days at 1% O2, then analyzed. (A) Fraction and (B) Number of MF20+ cells in the final differentiated populations. Realtime PCR analysis of cardiac markers (C) cardiac α-Actin and (D) cTnT in the 66 differentiated populations relative to undifferentiated mESC. (Error bars = standard deviation; N = 3) 67 -Page Intentionally Left Blank- 68 Chapter 3 Culture oxygen levels and use of small molecule inhibitors during self-renewal reversibly affect subsequent differentiation of mouse embryonic stem cells 3.1 Abstract Prior studies on self-renewal culture of embryonic stem cells (ESC) which had focused on long term proliferation and pluripotency, with assessment by differentiation to derivatives of the three germ layers, include the discovery of different states of pluripotency, some achieved by small molecule inhibition of signaling pathways. The effects of culture conditions during self-renewal, (which we term preconditioning), on the effectiveness of subsequent differentiation has not been examined. Here we show that changes preconditioning by low oxygen culture or small molecule dual inhibition (2i) of mitogen-activated protein kinase (using MEK inhibitor PD0325901) and glycogen synthase kinase (with GSK inhibitor CHIR99021) (1) reversibly affect levels of Oct4 expression in cells that are otherwise pluripotent, and (2) alters the efficiency of subsequent differentiation to cardiomyocytes (CM). Using a mESC line harboring a GFP reporter driven by the distal Oct4 enhancer, we performed preconditioning at different oxygen levels, with or without 2i supplementation. Under all conditions, the fraction of Oct4-GFP+ cells remained relatively constant (>90%) but the mean Oct4-GFP expression increased with 2i supplementation and decreased with low oxygen culture. Changes in preconditioning cells produced large differences in the resulting number and fraction of CM, despite use of identical differentiation protocols. Preconditioned cells that expressed higher Oct4-GFP levels before differentiation produced more CM after 69 differentiation. Low oxygen culture reduced Oct4-GFP expression during self-renewal and subsequent efficiency of CM differentiation while 2i supplementation increased both Oct4-GFP expression and subsequent efficiency of CM differentiation. These changes were reversed by continued self-renewal at high oxygen or removal of 2i supplementation respectively. Finally, the enhancement in CM generation from high oxygen culture and 2i supplementation during preconditioning is synergistic with low oxygen culture during differentiation, which had earlier been shown to improve differentiation of mESC to cardiomyocytes, demonstrating the importance of strict oxygen control during all aspects of ESC culture and differentiation. These results demonstrate that the manipulation of self-renewing culture conditions can lead to changes in the outcomes of defined differentiation protocols, a novel dimension to explore for directed differentiation of pluripotent stem cells. 70 3.2 Introduction In order to realize the immense clinical potential of pluripotent stem cells (PSC), which include both embryonic stem cells and induced pluripotent stem cells (iPSC), efficient protocols to differentiate these PSC into sufficient numbers and purities of the desired terminally differentiated or progenitor cell types are required. Much work in this area of directed differentiation has focused on the period during which differentiation is already occurring, and this includes differentiating ESC by co-culture with stromal cell lines [62], differentiation with various small molecule growth factors [60], and genetic manipulations [147]. However, the effects of the self-renewal culture environment used to propagate the undifferentiated cells before the initiation of differentiation on the effectiveness of subsequent differentiation protocols have not been examined. Most prior studies on self-renewal culture of embryonic stem cells (ESC) focused on long term proliferation and pluripotency and were assessed by differentiation to derivatives of the three germ layers. One recent study highlighted the importance of pluripotency factors Oct4 and Sox2 in germ layer determination during early differentiation [48]. Another study revealed that low oxygen levels during self-renewing culture decreased population-averaged levels of Oct4 mRNA expression, although cells remained pluripotent [96]. Taken together, this would suggest that different culture conditions can affect key pluripotency marker expression while still maintaining pluripotency, and these differences may affect the subsequent differentiation of the cells. 71 We are thus interested in examining different pluripotency maintaining culture conditions that result in changes in pluripotency marker expression. Two variables of interest are (1) culture oxygen levels, and (2) the use of two small molecule inhibitors (known together as ‘2i’) which have been shown to able to replace leukemia inhibitory factor (LIF) in self-renewing culture of mESC [34]. Would changes in oxygen levels or 2i use during self-renewal affect subsequent differentiation of mESC? In addition, would such changes affect cells permanently or were they reversible? In this study, we examined the self-renewing culture conditions for mouse ESC that affect the expression of Oct4, a key pluripotency marker, using a transgenic mESC reporter cell line harboring a GFP sequence driven by the Oct4 distal enhancer. In addition, we studied how different culture conditions during self-renewal immediately before differentiation, which we termed ‘preconditioning’, affect the efficiency of differentiation to cardiomyocytes, chosen here as a model system. Finally, we examined if changes in Oct4 expression due to different self-renewing conditions continued to be plastic when self-renewal conditions were changed. This work explored a novel dimension that had been previously ignored in designing directed differentiation protocols, resulting in an enhanced differentiation protocol which could be used in conjunction with and enhance the efficiency of other existing differentiation protocols. 72 3.3 Methods 3.3.1 Self-renewing culture of mESC A transgenic R1 mESC reporter line harboring a GFP sequence driven by the Oct4 distal promoter (O4G) [132], generously donated by Professor Douglas Melton (Harvard University, Cambridge, MA, USA), was used for the experiments. Undifferentiated O4G mESC were propagated as previously described [96] in Dulbecco’s modified Eagles medium optimized for ES cells (ES-DMEM; SCRR-2010; American Type Culture Collection (ATCC), Manassas, VA, USA) supplemented with 10% (v/v) ES cell qualified fetal bovine serum (FBS; SCRR 30-2020; ATCC), 2 mM L-alanyl-L-glutamine (SCRR 30-2115; ATCC), 100 units/mL penicillin and 100 µg/mL streptomycin (30-002-CI; Mediatech, Manassas, VA, USA), 1x MEM non-essential amino acids (30-2116; ATCC), 100 µM 2-mercaptoethanol (M7522; Sigma-Aldrich, St. Louis, MO, USA), and 103 units/mL Leukemia Inhibitory Factor (ESGRO mLIF; ESG1106; Millipore, Billerica, MA, USA) on plates with wells whose bottoms have been replaced by silicone rubber membranes [108] and previously treated for 30 min with a sterile 0.1% (w/v) solution of Type A gelatin (G-2500; Sigma-Aldrich) in tissue culture water (25-055-CM; Mediatech). In some cases, the medium was further supplemented with 1µM PD0325901 and 3 µM CHIR99021 (‘+2i’). Cells were plated at a density of 12,000 cells/cm2. Medium was replaced daily, and cells were detached with 0.25% trypsin (30-2101, ATCC) every 3 days and replated at 12,000 cells/cm2. 73 3.3.2 Flow cytometric analysis of Oct4-GFP expression Oct4-GFP fluorescence in individual cells was measured using a FACScan (Beckton Dickson). The fraction of Oct4-GFP positive cells was calculated as the number of cells above the negative gating threshold as defined above, divided by the number of cells in the entire population at each specific measurement time and condition. For undifferentiated cells, the mean Oct4-GFP fluorescence was determined from all Oct4GFP positive cells in the population. 3.3.3 Differentiation of mESC to cardiomyocytes Cells were differentiated as previously described [108]. Briefly, embryoid bodies (EBs) were formed with 500 mESC per drop in 20µL of cell culture medium as used in selfrenewing culture but without LIF and including 0.2 mM ascorbic acid (A4034; SigmaAldrich). After 2 days, 30 EBs were transferred per well into fibronectin-coated (F1141; Sigma-Aldrich) silicone rubber membrane-based plates [108] filled with 2mL of the EB medium. After an additional 3 days, medium was replaced with 2.5mL of serum-free medium, consisting of 50% (v/v) DMEM (90-133-PB; Mediatech) in tissue culture water and 50% (v/v) Ham’s F-12 (10-080-CV; Mediatech) supplemented with 0.75% (w/v) sodium bicarbonate (25-035-CI; Mediatech), 9mM glucose (G8769; Sigma-Aldrich), 5 µg/mL human insulin (I9278; Sigma-Aldrich), 100 units/mL penicillin, 100 µg/mL streptomycin, and 0.2 mM ascorbic acid. 74 3.3.4 Control of culture oxygen levels (pO2cell) Cells were self-renewed and differentiated at combinations of 20%, 5%, or 1% O2 as specified, with oxygen conditions maintained by flowing premixed gases containing 5% O2, and 20%, 5%, or 1% O2 with a balance of N2 (certified medical gas from Airgas, Hingham, MA, USA), which corresponded under humidified conditions to 142, 36, and 7 mmHg pO2gas inside chambers maintained at 37oC. Control of pO2cell equal to pO2gas was accorded by culture on custom-made silicone rubber membrane-based wells as previously described [108]. 3.3.5 Real-time polymerase chain reaction (PCR) Samples were collected as single cell suspensions and stored at 4oC in RNAprotect Cell Reagent (76526; Qiagen, Valencia, CA, USA) according to manufacturer’s instructions for no longer than 7 days before RNA isolation. Total RNA was isolated using the RNeasy Plus Mini Kit (74134; Qiagen), cDNA synthesized using the High Capacity cDNA Reverse Transcription Kit (4368814, Applied Biosystems, Foster City, CA, USA). Real-time PCR was performed on a Fast Real-Time PCR System (7900HT, Applied Biosystems), at the Center for Biomedical Engineering (MIT) using Power SYBR Green PCR Master Mix (4367659, Applied Biosystems) or Taqman Gene Expression Master Mix (43669016, Applied Biosystems) as appropriate. 18S ribosomal RNA (rRNA) was used as an endogenous control. Primer sequences used for Nanog and Sox17 have been previously reported [96]. Additional primer sequences used (in order of forward and reverse primer, and 5’ to 3’) were cardiac α-Actin GCTTCCGCTGTCCAGAGACT and TGCCAGCAGATTCCATACCA and cardiac Troponin T (cTnT) 75 AGATGCTGAAGAAGGTCCAGTAGAG and CACCAAGTTGGGCATGAAGA, and Cdx2 TCACCATCAGGAGGAAAAGTGA and GCTCTGCGGTTCTGAAACCA. Commercially available primers for 18S rRNA (Hs03928990_g1, Applied Biosystems), Blimp1 (Mm00476128_m1, Applied Biosystems), Dusp6 (Mm00518185_m1, Applied Biosystems), Fgf5 (Mm00438918_m1, Applied Biosystems), Rex1 (Mm03053975_g1, Applied Biosystems), Stella (Mm01184198_g1, Applied Biosystems), Titin (Mm00621005_m1, Applied Biosystems), and Wnt3a (Mm00437337, Applied Biosystems) were also used. A standard curve was constructed with cDNA pooled from all time points and conditions then serially diluted. 3.3.6 Cell enumeration Enumeration of viable cells in undifferentiated culture was performed by staining with Guava Viacount (4000-0041; Millipore) and data acquired using a Guava PCA flow cytometer (Millipore) according to manufacturer’s instructions. For differentiated cells, nuclei enumeration was performed by exposing dispersed cells to a lysis solution of 1% (v/v) Triton X-100 and 0.1 M citric acid (0627; Mallinckrodt Specialty Chemicals, Paris, KY, USA), then vortexed to obtain nuclei samples. Samples were then stained with Guava Viacount and data acquired with a Guava PCA flow cytometer as previously described [96]. 3.3.7 Flow cytometric analysis of MF20+ cells The fraction of MF20+ cells in dispersed cells was acquired with flow cytometry as previously described [96]. Briefly, samples were fixed by a 20 min incubation in 1% (w/v) 76 paraformaldehyde (36606; Alfa Aesar, Ward Hill, MA, USA) in PBS. Fixed cells were permeabilized in 0.5% saponin (S-4521; Sigma-Aldrich) for 10 min, non-specific binding blocked by a 30 min incubation in 2% (v/v) FBS in PBS, and then incubated at room temperature with primary antibodies against sarcomeric myosin heavy chain (MF20; MF20 supernatant; Developmental Studies Hybridoma Bank, Iowa City, IA, USA) diluted 1:10 in 2% FBS solution for 1 hr. Samples were then incubated with goat anti-mouse phycoerythin-conjugated secondary antibodies (115-116-146; Jackson ImmunoResearch; West Grove, PA, USA) diluted 1:250 in 2% FBS solution for 30 min at room temperature in the dark, washed 3x in 2% FBS solution, then data acquired using a FACScan. 3.3.8 Estimation of spontaneously contracting areas The fraction of well coverage by spontaneously contracting cells was averaged from measurements in 15 random 2 mm2 spots within each well by visual estimation as previously described [108]. 3.3.9 Statistics Statistical analysis was performed with unpaired two-tailed t-tests, with p<0.05 considered statistically significant. All data presented as mean ± standard deviation of biological replicates. 77 3.4 Results 3.4.1 Self-renewal of mESC at low oxygen reduces expression of Oct4-GFP but does not increase rates of spontaneous differentiation To examine the effect of low oxygen on self-renewing culture of mESC, we cultured O4G mESC under 20%, 5%, and 1% O2 with self-renewal media on silicone rubberbased plates for 9 days and measured their distributions of Oct4-GFP fluorescence at each cell passage (Figure 3-1). Culture at low O2 shifted the distribution of Oct4-GFP fluorescence towards reduced (but still Oct4-GFP+) values (Figure 3-2A and B) and reduced the population averaged value (Figure 3-2A). However, under all conditions the fraction of Oct4-GFP+ cells remained relatively constant and >90% (Figure 3-2B). A prior study reported a reduction in Oct4 mRNA expression in mESC cultured under low oxygen [96] but did not identify if it resulted from a reduction across individual cells in the population, or an increase in fraction of Oct4-negative cells. Our results indicate that there have not been increased differentiation (which would have resulted in an increase in the fraction of Oct4-GFP-negative cells) due to reduced oxygen culture, but rather, a reduction in the levels of Oct4-GFP in individual cells. 3.4.2 Reduction of Oct4-GFP expression by low oxygen culture is reversed by continued self-renewal in 20% oxygen We next tested if the reduction in Oct4-GFP expression was reversible by subsequent self-renewal cultured at 20% O2 (Figure 3-1C and 3-2) Cells that were cultured at 5% or 1% O2 for 9 days were cultured under self-renewing conditions for an additional 9 days 78 at either 20% O2 (indicated as 5 20 or 1 20) or the same low oxygen condition as the first 9 days. The additional 9-day self-renewing culture at 20% O2 reversed the reduction in Oct4-GFP expression caused by low O2 culture during the first 9 days and restored the Oct4-GFP distribution to that of cells that had been continuously cultured at 20% O2 for all 18 days. Cells continuously cultured under low O2 conditions did not undergo reversal of the reduction. 3.4.3 Low oxygen preconditioning decreases subsequent differentiation of mESC to cardiomyocytes We hypothesized that conditions of self-renewing culture immediately before differentiation (termed preconditioning) at different O2 would change the efficiency of differentiation of clonal mESC to different cell types. To test this, we preconditioned clonal mESC at 20%, 5% and 1% O2 for 9 days before differentiating each population of preconditioned cells at 20%, 5%, and 1% O2 for 10 days, for a total of 9 different combinations of preconditioning and differentiation conditions (Figure 3-3). At each differentiation condition, low oxygen preconditioning reduced the fraction and number of cells positively immunostained for MF20 (anti-sarcomeric myosin heavy chain, a cardiac specific antibody [112, 138, 139]). For cells differentiated at 5% O2, for example, about 15, 25, and 29% of the population were MF20+ in differentiated cells preconditioned at 1, 5, and 20% O2 respectively (Figure 3-3A). 79 These results are further supported by similar trends in cardiac gene expression (cardiac α-Actin [142], cTnT [143] and Titin [148]) by the differentiated populations, as well as semi-quantitative estimations of the fraction of spontaneously contracting tissue (Figure 3-3C to E). 3.4.4 Total cell number and fraction of residual undifferentiated Oct4-GFP+ cells are not affected by preconditioning, but by oxygen conditions during differentiation To investigate the possibility that the reduction in cardiomyogenesis in low oxygen preconditioned cells was due to changes in cell proliferation or a ‘resistance’ to any differentiation, we examined the total cell number and the fraction of residual undifferentiated mESC (identified by Oct4-GFP fluorescence) in differentiated populations (Figure 3-4). Total cell number after differentiation did not exhibit significant differences among populations differentiated at the same O2, but showed differences between populations preconditioned at the same O2 and differentiated at different O2. For cells preconditioned at 20% O2, for example, (white bars in Figure 3-4A), cells differentiated at 20% produced 5.6x106, 4.1x106, and 2.2x106 total cells when differentiated at 20, 5, and 1% O2 respectively. While changes in preconditioning O2 produced differences in the fraction of residual Oct4-GFP+ cells for populations differentiated at 20% O2, such differences were not observed for those differentiated at 5% or 1% O2. Differences between populations preconditioned at the same O2 but differentiated at either 20% O2 or low (5% or 1%) O2 80 were significant in all cases. For example, cells preconditioned at 1% (black bars in Figure 3-4B) had about 7.5%, 0.8%, and 1.1% of the differentiated population remaining as Oct4-GFP+ when differentiated at 20, 5, and 1% O2 respectively. Similar behavior observed for cells preconditioned at 20% and 5% O2. 3.4.5 Reduction in efficiency of cardiomyogenesis due to low oxygen preconditioning is reversible by self-renewal in 20% oxygen before differentiation To investigate if the change in differentiation efficiency of clonal mESC to cardiomyocytes by low oxygen preconditioning was reversible, we cultured cells under self-renewing conditions at either 20%, 5%, or 1% O2 for 18 days, or 5% or 1% for the first 9 days and then 20% for the next 9 days, before subjecting all five populations to cardiomyocyte differentiation at 5% O2 (Figure 3-5). We found that the changes produced by low oxygen preconditioning could be largely reversed by subsequent self-renewing culture at 20% O2 before differentiation. The fraction and number of MF20+ cells in the populations differentiated from either 20%, 520%, or 120% preconditioned cells were not significantly different, and this trend is also seen in cardiac α-Actin gene expression analysis and estimations of spontaneously contracting areas. The cells that were preconditioned at 5% and 1% O2 for all 18 days, however, continued to exhibit the reduced differentiation efficiencies to cardiomyocytes. 81 3.4.6 2i use during self-renewal increases expression of naïve pluripotency markers, decreases early differentiation markers synergistically with high oxygen, and is reversible by subsequent self-renewing culture without 2i The fraction of MF20+ cells in the differentiated populations correlated positively with the mean Oct4-GFP fluorescence of the starting populations (Figure 3-6). This correlation suggested that preconditioning mESC to increase their Oct4-GFP levels before differentiation would increase the yield in subsequent differentiation to cardiomyocytes. Prior work [96] had shown that increasing O2 levels in self-renewing culture above 20% did not further increase pluripotency gene expression but instead decreased them, possibly due to increased cell death due to increased reactive oxygen species (ROS). We then explored an alternative strategy of using small molecule inhibitors known to support the undifferentiated state and identified the combination of two small molecules: PD0325901, an inhibitor of Mitogen-Activated Protein Kinase (MEK), and CHIR99021, an inhibitor of Glycogen Synthase Kinase (GSK), referred together as ‘2i’, as a possibility. 2i has been used previously to maintain the undifferentiated state of mESC without LIF, and to improve the establishment of ESC and iPSC lines. Its use is thought to promote entry into the naïve or ground state of pluripotency [34, 149]. However, their effects during self-renewing culture on ESC’s subsequent differentiation have not been studied. We hypothesize that 2i supplementation in self-renewal media would increase Oct4 expression in undifferentiated mESC and consequently increase the efficiency of subsequent differentiation to cardiomyocytes. 82 To study the effects of 2i in self-renewing culture of mESC, we cultured O4G mESC under four different self-renewing conditions – 20% O2, with (20/2iL) or without 2i (20/L), or 1% O2 with (1/2iL) or without 2i (1/L) for 9 days (See Figure 3-7 for nomenclature for the different conditions). 2i supplementation increased the Oct4-GFP expression without significant changes in the fraction of Oct4-GFP+ cells compared to self-renewal without 2i supplementation at either O2. 2i supplementation acted synergistically with high oxygen during self-renewal to increase Oct4-GFP expression beyond levels of cells cultured at 1% O2 with 2i. After 9 days of culture, 20% O2 with 2i supplementation produced cells with a population-averaged Oct4-GFP fluorescence of nearly 3x higher than cells self-renewed with LIF only at 20% O2, while 1% O2 with 2i supplementation achieved an increase of 2x over the 20% O2 condition with LIF only, and a 4x increase over cells self-renewed at 1% O2 with LIF alone (Figure 3-8A, 3-9A). To study the reversibility of the increases in Oct4-GFP expression due to 2i supplementation during self-renewal, cells cultured with 2i at 20% or 1% O2 were then subcultured with or without 2i supplementation for up to an additional 15 days at the same oxygen level for a total of six culture conditions. The increase in Oct4-GFP expression was reversed by subsequent self-renewal culture without 2i supplementation (Figure 3-8C,E & 9A). In all cases, the fraction of Oct4-GFP+ cells were relatively constant and >90% (Figure 3-9B). Therefore, changes in mean Oct4-GFP levels resulted from changes in the levels of Oct4-GFP expression in individual cells rather than in the fraction of Oct4-GFP+ cells in the populations (Figure 3-8). 83 We further characterized the cell populations that had been self-renewed under different conditions for 18 (Figure 3-10) and 24 days (Figure 3-11 Effects of 2i supplementation and culture O2 on gene expression of mESC after 24 days in self-renewing culture. by real-time PCR analysis. Expression of Nanog and Rex1, genes marking the naïve pluripotent state [52, 149], were increased in self-renewing cultures supplemented with 2i at either O2 level, compared to the culture with the same O2 but without 2i. Markers of primordial germ cells Stella [150] and Blimp1 [151] were downregulated in the presence of 2i in self-renewal culture. Similarly, early lineage markers Fgf5 [152] and Sox17 [136] exhibited reduced expression in the presence of 2i during self-renewal. As a check, we also analyzed behaviors of Dusp6, a marker of functional ERK activity, which would be downregulated by MEK (Mitogen-activated protein kinase kinase) inhibitor PD035901 [119, 153, 154], and Wnt3a, which would be upregulated by use of CHIR99021 [120,155]; and found the expression behaviors were consistent with 2i activity. 3.4.7 High oxygen and 2i use during preconditioning separately, synergistically, and reversibly increases subsequent differentiation to cardiomyocytes To study the effects on subsequent differentiation of oxygen and 2i supplementation during preconditioning, we self-renewed the cells as described above for 18 (Figure 3-12) and 24 days (Figure 3-13) and performed subsequent differentiation at 20% O2 for all preconditioning conditions (& 13). We measured the fraction and number of MF20+ cells, the relative expression of cardiac associated genes cardiac α-Actin, cTnT, and Titin, and estimated the area fraction of wells covered by spontaneously contracting cells in the differentiated populations. 84 Similar trends were observed for both rounds of differentiation at 20% initiated after 18 and 24 days of preconditioning. In cell populations differentiated after 18 days of preconditioning, the fraction of MF20+ cells in the population with 2i (20/2iL) was about 26%, 2.4x higher than the control population that was preconditioned at 20% O2 without 2i for the entire preconditioning period (20/L, 11%). The population that received 2i supplementation for the first 9 days, and then only LIF for the next 9 days (20/(2iL/L)) generated 12% MF20+ cells, which was not significantly different from the 20/L control population. Preconditioning at 1% O2 (1/L) reduced the fraction of MF20+ cells in the differentiated population compared to 20% O2 preconditioning (20/L). As in the case with 20% O2 preconditioning, low oxygen with 2i supplementation (1/2iL) increased the fraction of the MF20+ cells compared to both the 1/L and 20/L conditions, and cells differentiated from the 1/(2iL/L) preconditioned cells did not show a significant difference in the fraction of MF20+ cells from the 1/L preconditioned cells. These trends are corroborated by the number of MF20+ cells generated, the fraction of contracting areas, and gene expression analyses of cardiac genes, as well as again by the results from differentiation of cells preconditioned for 24 days before initiation of differentiation (Figure 3-13). Prior work [117, 140] has established that low oxygen during differentiation increases cardiomyogenesis. We hypothesize that low oxygen during differentiation acts synergistically with high oxygen and 2i supplementation during precondition to further increase cardiomyogenesis. We tested this by differentiating 20/2iL preconditioned cells 85 after 24 days of preconditioning at 5% O2 (blue bars in Figure 3-13). Indeed, low oxygen further increased all aspects assessed – fraction and number of MF20+, fraction of spontaneously contracting areas, and gene expression of cardiac-associated genes. 20/2iL preconditioned cells produced about 31% MF20+ cells when differentiated at 5% O2, which was higher than the same cells differentiated at 20% O2 by 30%, 20/L preconditioned cells differentiated at 20% O2 by 160%. 3.4.8 Fraction of MF20+ cells in differentiated populations correlates positively with mean Oct-GFP levels of the starting populations For cells differentiated at 20% O2, the fraction of MF20+ cells correlated linearly with the mean Oct4-GFP fluorescence of the starting (Figure 3-14) with correlation coefficients <R2> of 0.63 and 0.93 for cells differentiated after 18 and 24 days of preconditioning respectively. When data from all differentiation experiments at 20% O2 were pooled into a single analysis, the correlation remained strong (<R2> = 0.84) in linear form (Figure 3-15A), and with a logarithmic dependence on the initial mean Oct4-GFP fluorescence (<R2> = 0.88) (Figure 3-15B). Both regressions predict >100% MF20+ cells if the starting Oct4-GFP fluorescence is sufficiently high, which is physically impossible. We therefore investigated a saturable function of the form 𝑦 = 𝑦𝑚𝑎𝑥 (1 − exp�𝐴(𝑥 − 𝑥𝑚𝑖𝑛 )�) (1) where y is the final fraction of MF20+ cells in the differentiated population, x the starting average Oct4-GFP fluorescence, ymax, is the highest fraction that can be achieved with a given cell line and protocol, xmin is a minimum value for a given setting on the flow cytometer, cells would not be considered to be Oct4-GFP+ at the initiation of 86 differentiation, and A, a scaling factor to account for differences between flow cytometer settings. The three parameters were fitted [156] by nonlinear regression minimizing the sum of the residuals in y squared. The result plotted in Figure 3-16 has physically meaningful parameter values of ymax = 29.5%, xmin = 166.5 AU, A = -7.4x10-4, with <R2> = 0.84. In general, we would expect that ymax would depend on the differentiation protocol used (using low oxygen during differentiation for example, would increase ymax), and xmin and A would be dependent on the actual flow cytometer and the settings employed. 87 3.5 Discussion and future work High oxygen and 2i supplementation during self-renewal separately and synergistically increased Oct4 expression, and subsequently increased differentiation to cardiomyocytes. This effect was reversible with continued self-renewal by using appropriate oxygen modulation and elimination of 2i supplementation before differentiation. Differentiation to cardiomyocytes was enhanced synergistically by using low oxygen during differentiation. Earlier studies determined that low oxygen is best employed during the first 6 days of differentiation, then high oxygen subsequently [117, 140]. Thus, oxygen conditions for synergistic enhancement of cardiomyogenesis are different depending on the stage of differentiation, and realization of the most enhancement of directed differentiation requires good control of O2 conditions to which the cultured cells are exposed. Future work could translate these findings to human ESC, as well as to differentiation protocols for other clinically relevant cell types. Other published studies have shown that human ESC are more similar to mouse epiblast stem cells (mEpiSC) than mESC, and requires ERK1/2 signaling for renewal [157], which 2i supplementation blocks. With the isolation of hESC that are in a naïve state more similar to mESC [124], it would be interesting to determine if these results can be directly translated to the differentiation of such hESC to cardiomyocytes. This study demonstrated a positive correlation between Oct4-GFP expression immediately before differentiation and the fraction of MF20+ cells in the differentiated populations. Recent studies [48, 128, 129] have provided a possible explanation and mechanism for this correlation through the action of Oct4 enhancing fate choice down 88 the mesodermal lineage during early differentiation. Future studies could include additional characterization of the differently self-renewed population via microarrays or RNA-Seq to determine transcriptome-wide differences generated in clonal populations. Results from such studies could direct investigations into better predictive methods for differentiation into other therapeutically relevant cell types. In addition, this work also raises new questions on the nature of preconditioning on clonal ESC populations – is preconditioning employing selective pressures on the population, and certain subpopulations favored by some culture conditions, or are all cells being affected equally? One possibility would be to use brainbow technology [158] or other similar methods to differentially mark otherwise clonal cells and to track the composition of the marked cells under different self-renewal conditions. Finally, to understand the role of Oct4 in the naïve, primed, and early differentiating ESC states, researchers could generate a double reporter ESC line with the distal and proximal promoter regions of Oct4 driving the expression of two different fluorophores in order to determine their relative activities in individual cells during self-renewal and differentiation. It is interesting to note that both Stella and Blimp1 expressions were downregulated in the presence of 2i in self-renewing culture, although Stella expression is associated with the naïve pluripotent state [53], while Blimp1 is associated with the primed pluripotent state [150]. Both genes are markers for primordial germ cells, and this suggests that 2i use is suppressing differentiation of ESC to germ cells. Prior studies on self-renewal culture of PSC have focused on the maintenance and propagation of cells in the undifferentiated state, and pluripotency of the cultured cells 89 was assessed by testing for the ability to differentiate into cells of all three germ layers, the ability to contribute to chimeras, or form teratomas. However, there have not been studies on the effects of self-renewal culture on the efficiency of subsequent differentiation to specific cell types – a focus of directed differentiation studies. Similarly, efforts to direct differentiation of PSC have focused on the period of differentiation, and have not taken into consideration the effects self-renewal conditions that have occurred before differentiation. This study has taken the direction of studying the effects of selfrenewing culture conditions immediately before directed differentiation of PSC, and found that it indeed affects the efficiency of differentiation to cardiomyocytes. These results provide a novel direction for consideration in design of directed differentiation protocols as different self-renewing culture conditions employed immediately before differentiation can affect the efficiency of subsequent differentiation of clonal PSC. Studies on undifferentiated culture systems of PSC would benefit from assessments of the efficiency of subsequent differentiation to one or more cell types of interest, in addition to the standard pluripotency assays, to aid future translational work. 90 3.6 Figures 150 A Sd9 O2 [%] 100 20 5 1 Cell Counts 50 0 1 10 150 B 103 102 104 150 Sd18 C Sd18 O2 [%] 100 100 520 120 50 0 101 50 102 0 104 101 103 102 Oct4-GFP Fluorescence [AU] 103 104 Figure 3-1 Effects of culture O2 in self-renewing culture on distributions of Oct4-GFP expression in O4G mESC. Oct4-GFP distributions in populations self-renewed at 20%, 5%, or 1% O2 for (A) 9 days, (B) 18 days, and (C) 5% or 1% for the first 9 days, then 20% for the next 9 days. 91 Relative Mean Oct4-GFP Fluorescence 1.2 A 1.0 0.8 O2 [%] 20 5 1 520 120 0.6 0.4 0.2 0 3 6 9 12 15 18 6 9 12 15 Days of Self-Renewal 18 100 Fraction Oct4-GFP+ Cells [%] B 80 60 40 20 0 0 3 Figure 3-2 Mean Oct4-GFP fluorescence and fraction of Oct4-GFP+ cells from selfrenewal at different O2. 92 (A) Time course of mean Oct4-GFP fluorescence of mESC self-renewed under different O2 conditions. (B) Fraction of Oct4-GFP+ cells in the self-renewing populations. Self-renewal of mESC at low O2 reduced cell-level expression of Oct4-GFP, but does not increase rates of spontaneous differentiation based on the fraction of Oct4-GFP+ cells in the population. 93 Differentiation after 9 days of Preconditioning A * 30 * * Preconditioning O2 * 20% * 5% 20 * * B 1.0 * * 45 * Fraction of spontaneously contracting areas [%] 10 0 1.5 1% * * * * 0.5 0 1.6 D 1.2 0.8 0.4 0 Cardiac α-Actin * * * * * * Expression relative to undifferentiated cells (x10-4) Expression relative to undifferentiated cells (x10-3) Number of MF20+ cells (x10-6) Fraction of MF20+ cells [%] 40 * 30 * 15 * 0 1.6 E cTnT * * 0.8 * 0.4 * 20 5 1 Differentiation O2 [%] Figure 3-3 Effects of O2 preconditioning on subsequent differentiation. 94 * * 1.2 0 20 5 1 Differentiation O2 [%] * C Cells were self-renewed at 20%, 5%, or 1% O2 for 9 days (preconditioning) before subsequent differentiation at 20%, 5%, or 1% O2. Each group of bars denote cells that have been differentiated at the same O2, with white bars representing cells that have been preconditioned at 20% O2, grey bars 5%, and black bars 1% O2. (A) Fraction of MF20+ cells in the differentiated populations. (B) Number of MF20+ cells. (C) Fraction of well that is spontaneously contracting. Real-time PCR analysis of (D) cardiac α-Actin and (E) cTnT relative to undifferentiated cells. Low O2 preconditioning decreases subsequent differentiation of mESC to cardiomyocytes. 95 10 A Fraction of Residual Oct4-GFP+ cells [%] Total Cell Number (x10-6) 8 6 4 2 0 8 6 * * Preconditioning O2 20% 5% 1% 4 2 0 20 5 1 Differentiation O2 [%] B 20 5 1 Differentiation O2 [%] Figure 3-4 Effects of O2 preconditioning on total cell number and fraction of residual Oct4-GFP+ cells after differentiation. Cells were differentiated as described in Figure 3-3 and assessed for (A) total cell number and (B) fraction of residual undifferentiated Oct4-GFP+ cells in the differentiated populations. 96 20 * 15 * 10 C Cardiac α-Actin 0.8 0.4 0 * Fraction of spontaneously contracting areas [%] 5 0 1.2 Expression relative to undifferentiated cells (x10-3) A Number of MF20+ cells (x10-6) Fraction of MF20+ cells [%] 25 Differentiation after 18 days of Preconditioning (All differentiation performed at 5% O2) 1.5 B 1.0 0 25 D 20 15 10 * * 5 0 1 20 5 1 5 20 20 Preconditioning O2 [%] * 0.5 1 20 5 1 5 20 20 Preconditioning O2 [%] Figure 3-5 Reversibility of O2 preconditioning. Cells were preconditioned at 20%, 5%, or 1% O2 for 18 days, or 5% or 1% for the first 9 days, then at 20% O2 for the next 9 days (denoted as 520, 120) before differentiation at 5% O2 for all conditions. *p<0.05 compared to cells preconditioned at 20% O2. 97 (A) Fraction of MF20+ cells in the differentiated populations. (B) Number of MF20+ cells. (C) Real-time PCR analysis of cardiac α-Actin relative to undifferentiated cells. (E) Fraction of cells that are spontaneously contracting. The reduction of resulting cardiomyocytes differentiated from low O2 preconditioned mESC can be reversed by self-renewal at high O2 before differentiation, although there is a non-significant decrease remaining in fraction of spontaneously contracting cells for 120 preconditioning. 98 Fraction MF20+ cells in differentiated population [%] Differentiation O2 (%) 30 20 5 <R2> 0.97 0.74 0.99 1 20 10 0 0 250 500 750 Starting Mean Oct4-GFP Fluorescence [AU] Figure 3-6 Correlation between mean Oct4-GFP fluorescence of the starting undifferentiated populations and the fraction of MF20+ cells in the resulting differentiated populations. 99 1000 5 Figure 3-7 Experimental design and nomenclature for preconditioning with O2 modulation and 2i supplementation. 100 Cell Number 150 A 100 Sd9 B Sd18 Sd24 D 20/2iL 20/(2iL/L) 20/L 20/2iL 1/2iL 20/L 1/L 50 102 103 104 150 Cell Number 0 1 10 100 102 C 103 102 104 Sd18 E 103 104 103 104 Sd24 1/2iL 1/(2iL/L) 1/L 50 0 101 102 103 104 Oct4-GFP Fluorescence [AU] 102 Figure 3-8 Effects of 2i supplementation and O2 during self-renewal on Oct4-GFP fluorescence distributions of undifferentiated mESC. Oct4-GFP distributions in populations of mESC self-renewed (A) at 20% or 1% O2 for 9 days, at (B) 20% or (C) 1% O2 for 18 days, or at (D) 20% or (E) 1% O2 for 24 days. 101 A 3 2 1 0 0 100 Fraction Oct4-GFP+ [%] Relative Mean Oct4-GFP Expression 4 5 10 15 20 25 B 80 60 40 20 0 0 Self-Renewal Condition 20/L 1/L 20/2iL 1/2iL 20/(2iL/L) 1/(2iL/L) 5 20 25 10 15 Days in Self-Renewing Culture Figure 3-9 Effects of 2i supplementation and culture O2 on mean Oct4-GFP fluorescence and fraction of Oct4-GFP+ cells during self-renewal. (A) Time course of mean Oct4-GFP fluorescence of the populations relative to cells selfrenewed continuously at 20% O2 with LIF only (20/L). (B) Fraction of Oct4-GFP cells in the undifferentiated populations. 102 1.5 A 1.5 * Nanog B Rex1 * # # 1.0 1.0 * # * 0.5 0.5 * 0 3.6 0 Dusp6 C * * D * Wnt3a # * 2.4 # 1 1.2 * # * 0 2.4 0 2.4 Stella E * 1.6 * 0.8 0 18 1.6 # * * # * * 0 3 Sox17 * G Blimp1 * F 0.8 # * 12 * Fgf5 H 2 * 6 1 Self-Renewal Condition 103 1/(2iL/L) 1/2iL 1/L * 20/(2iL/L) 20/2iL 0 # * 20/L 1/(2iL/L) 1/2iL 1/L 20/(2iL/L) 0 20/2iL # 20/L Expression relative to 20/L at Sd18 2 Figure 3-10 Effects of 2i supplementation and culture O2 on gene expression of mESC after 18 days in self-renewing culture. Real-time PCR analysis of cells after 18 days of self renewal. Each group of bars denote cells self-renewed at the same O2. White bars represent cells in culture with LIF only, black with LIF and 2i, and grey with LIF and 2i for the first 9 days, and then with LIF only for the next 9 days with expressions relative to that of the 20/L condition. Genes assessed: Naïve pluripotency markers (A) Nanog and (B) Rex1; (C) marker of functional ERK activity (inhibited by PD0325901) Dusp6; (D) Wnt3a (stimulated by CHIR99021); primordial germ cell markers (E) Stella and (F) Blimp1; and early lineage markers (G) Sox17 (primitive endoderm), and (H) Fgf5 (primitive ectoderm). *p<0.05 compared to 20/L; #p<0.05 compared to 1/L. 104 1.8 1.5 # Nanog A B 1.2 1.0 0.6 0.5 Rex1 # * * * * * 0 1.8 C 0 4 Dusp6 * D * Wnt3a # 1.2 # 2 * 0.6 * # * 0 2 * * Stella* E 0 1.5 Blimp1 F 1.0 1 * * * 0 9 # * 0.5 # * 0 * Sox17 G #* Fgf5 H 4 6 * Self-Renewal Condition 105 1/(2iL/L) * 1/2iL 1/L 20/(2iL/L) 0 # * 20/L 1/(2iL/L) 1/L 20/(2iL/L) 20/2iL 1/2iL # * 20/2iL 3 0 * 2 20/L Expression relative to 20/L at Sd24 * Figure 3-11 Effects of 2i supplementation and culture O2 on gene expression of mESC after 24 days in self-renewing culture. Real-time PCR analysis of cells after 24 days of self renewal, as described in Figure 3-10. 2i use during self-renewal increased naïve pluripotency markers and decreased early differentiation markers synergistically with high O2 and was reversible by subsequent self-renewing culture without 2i. 106 0 1.0 * # 0.5 # 0 30 C * None Detected 107 * 20 4 0 Self-Renewal Condition cTnT * F 20 1/(2iL/L) * E 1/2iL * 1/L B * 20/(2iL/L) D 20/2iL # 20/L * Expression relative to undifferentiated mES (x10-3) * 1/(2iL/L) A 1/2iL 10 1/L 20/(2iL/L) Fraction of MF20+ cells [%] 30 20/2iL Number MF20+ cells (x10-6) 0 1.5 20/L Fraction of spontaneously contracting areas [%] 6 Cardiac α-Actin 4 20 # 2 * 0 * # 10 0 Titin * # 10 2 Figure 3-12 Effects of 2i supplementation and culture O2 on subsequent differentiation after 18 days of preconditioning. Cells preconditioned for 18 days as described in Figure 3-10 are subsequently differentiated at 20% O2. (A) Fraction of MF20+ cells in differentiated population. (B) Number of MF20+ cells. (C) Fraction of well that is spontaneously contracting. Real-time PCR analysis of cardiac markers (D) cardiac α-Actin, (E) cTnT, and (F) Titin, relative to undifferentiated mESC. *p<0.05 compared to population differentiated from 20/L cells, # p<0.05 compared to population differentiated from 1/L cells. 108 Cardiac α-Actin D 30 * 20 * 2 * B 1.0 Expression relative to undifferentiated mES (x10-3) 10 0 1.5 * * * # 0.5 * * 0 30 C * # * 0 40 * E * cTnT 30 20 * # 10 * * 0 F * 12 Titin 20 8 # Self-Renewal Condition 1/2iL 1/(2iL/L) * * 1/L 0 * 20/L 1/(2iL/L) None Detected 1/2iL 1/L 20/(2iL/L) None Detected 20/2iL 0 4 20/2iL 10 20/L Fraction of spontaneously contracting areas [%] * 4 # 20/(2iL/L) Fraction of MF20+ cells [%] Number MF20+ cells (x10-6) 6 * A Figure 3-13 Effects of 2i supplementation and culture O2 on subsequent differentiation after 24 days of preconditioning. 109 Cells preconditioned for 24 days as described in Figure 3-11 are subsequently differentiated at 20% (all populations) or 5% (blue bars - 20/2iL only) O2. (A) Fraction of MF20+ cells in differentiated population. (B) Number of MF20+ cells. (C) Fraction of well that is spontaneously contracting. Real-time PCR analysis of cardiac markers (D) cardiac α-Actin, (E) cTnT, and (F) Titin, relative to undifferentiated mESC. *p<0.05 compared to population differentiated from 20/L cells, #p<0.05 compared to population differentiated from 1/L cells. High O2 and 2i use during preconditioning separately, synergistically, and reversibly increased subsequent differentiation to cardiomyocytes. This effect can be further enhanced by low O2 during differentiation. 110 Final mean fraction MF20+ cells [%] 30 20 18d preconditioning 24d preconditioning 10 0 0 1000 2000 Starting Mean Oct4-GFP Fluorescence [AU] 3000 Figure 3-14 Correlation between the fraction of MF20+ cells in the resulting differentiated populations for cells preconditioned for 18 (red triangles) and 24 days (blue circles) before differentiation and the mean Oct4-GFP fluorescence of the starting undifferentiated populations. Cells were differentiated at 20% O2. For cells preconditioned for 18d, <R2>= 0.63; for cells preconditioned for 24d, <R2>= 0.93, both from a linear least squares regression. 111 All cells differentiated at 20% O2 30 A Final mean fraction MF20+ cells [%] 20 Length of Preconditioning 9d 18d 24d 10 Linear correlation: r2 = 0.84 0 0 30 1000 2000 3000 B 20 10 Semi-log fit: r2 = 0.88 0 1000 10000 100 Starting Mean Oct4-GFP Fluorescence [AU] Figure 3-15 Linear and semi-log correlations between mean Oct4-GFP fluorescence of starting undifferentiated populations and the fraction of MF20+ cells in the resulting differentiated populations for all cells differentiated at 20% O2. 112 Final mean fraction MF20+ cells [%] 30 20 10 r2 = 0.84 0 0 1000 2000 3000 Starting Mean Oct4-GFP Fluorescence [AU] Figure 3-16 Model fit for correlation between mean Oct4-GFP fluorescence of starting undifferentiated populations and the fraction of MF20+ cells in the resulting differentiated populations for all cells differentiated at 20% O2. Model used was y = ymax(1 – exp(A(x-xmin)), where y = final fraction of MF20+ cells in the differentiated population; ymax = hypothesized maximum MF20+ fraction possible; x = starting average Oct4-GFP fluorescence; xmin = minimum fluorescence for cell to be considered Oct-GFP+; A = scaling factor. Fitted values: ymax = 29.5%, xmin = 166.5; A = -7.4x10-4. <R2> = 0.84. 113 -Page Intentionally Left Blank- 114 Chapter 4 Integrative Perspective Paper: Opportunities and Obstacles in the Commercialization of Pluripotent Stem Cell-Based Products and Services 4.1 Executive Summary The major commercialization opportunity resulting from the discovery of human pluripotent stem cells is their role in the production of replacement cells, tissues, and organs for cell therapy, with a single product potentially reaching blockbuster US market sizes of US$1 billion or more. However, this tremendous opportunity is equally matched by significant challenges, both technological and commercial. Investors and entrepreneurs willing to enter the market now can look to related fields of providing research consumables and equipment for pluripotent stem cell research or using the cells themselves to create research and diagnostic tools. While these markets are significantly smaller than that of cell therapies, there is already demonstrated profitability and fewer uncertainties. This also allows players to gain technologies, market knowledge, and the financial ability to support their moves into the cell therapy space in the long run. As the cell therapy market evolves from its current development phase to having actual products on the shelf, initial strategies taken by companies are expected to focus on high impact therapies, in order to attract financing with promises of high returns and to establish market credibility. Players pursuing this long-term goal will need to navigate issues of intellectual property, and regulatory and ethical challenges, before the oft-repeated potential of pluripotent stem cells in bringing new therapies to the clinic can be realized. 115 4.2 Introduction The discovery of human embryonic stem cells (hESCs) and their pluripotent nature rapidly led to widespread expectations that their contribution to human therapies and greatest opportunities for commercialization would arise from their potential to generate replacement cells, tissues, and even organs for use in regenerative medicine. The subsequent discovery of human induced pluripotent stem cells (hiPSCs) heralded a second path to the same goal, with the added advantage of significantly reducing the risk of rejection of the transplant, as well as sidestepping the ethical challenges associated with use of hESCs. Realizing this goal still requires many challenges to be overcome, and at this point in time, there are no FDA-approved hESC or hiPSC derived cell-based therapeutics commercially available, with only two hESC-based clinical trials in progress in the United States. The chasm between current reality and the vision is undeniably large and fraught with obstacles, both scientific and commercial. However, the commercialization options for human pluripotent stem cells (hPSCs, which include both hESCs and hiPSCs) are not limited to producing replacement tissues, with companies already bringing other hPSC-based products and services, such as laboratory consumables for use in hPSC research and hPSC-based tools for development research, drug discovery, disease modeling, and toxicology studies to the market. The different commercialization opportunities, associated obstacles, and potential business models for hPSC-based products and services are examined here. 116 4.3 Pluripotent Stem Cell Research and Diagnostic Consumables and Equipment As hPSC research activity grows in academic, medical, and commercial settings, so does demand for laboratory consumables and equipment that support such endeavors. Of course, much of the consumables and scientific equipment used are not specific to hPSC research activity, but there is a category that caters exclusively or predominantly to hPSC work. Examples include reprogramming ‘kits’ based on different techniques or services to perform the actual work of reprogramming the client’s cells, hPSC specific growth factors, culture reagents, and antibodies for maintenance of pluripotency or directed differentiation, and testing services to evaluate the genetic stability of stem cell lines. The global market for such products and services was estimated to be US$883M in 2011, growing to US$1.35B in 2015, with a 11% annual growth rate [159]. While this is only a small fraction of the entire US$20.5B market for the entire laboratory supply market in the US alone, it is certainly growing much quicker than the 0.1% annual growth the larger category achieved in the 5 years from 2007 to 2012, and the projected 3.4% in the next 5 years [160]. This market is heavily dependent on research funding availability which feeds demand from scientists and laboratories, and the anemic growth in the US market over the past 5 years reflects the changes in funding availability, both from federal and other government sources, private foundations, and commercial companies. 117 Companies in this market segment include dedicated small start-ups such as Stemgent (Cambridge, MA), and ArunA Biomedical (Athens, GA) as well as divisions within wellestablished large companies such as Life Technologies (Carlsbad, CA, soon to be acquired by Thermo Fisher Scientific (Waltham, MA)) and Becton Dickinson (Franklin Lakes, NJ). This market is considered to be mature, with well-established business models and a number of large operating firms with stable growth. Scale and a deep portfolio of products are keys to success in this model, with substantial revenue possible in the aggregate of small revenue streams of individual products, hence favoring large companies with the necessary sales and supply chain bandwidth. While smaller companies and start-ups may be sustainable with one or two blockbuster products, a typical business model for these companies would be to license their products to the larger companies rather than going it alone, or to be acquired [161]. An example would be ArunA’s agreement with Thermo Fisher Scientific to ride on their sales and distribution channels. Interestingly enough, the reverse has happened with the agreement between Pfizer and Stemgent, which will see custom reagents developed by Pfizer sold through the latter [162]. With the research landscape rapidly evolving, new tools and reagents come on to the market very frequently. Recent new products include mRNA-based reprogramming kits introduced in 2011, based on a 2010 publication [63]. The rapid translation of scientific discoveries to commercially available products and services in this market makes it attractive to investors and entrepreneurs. Stemgent, for example, raised $14M by 2009 from Venture Capital (VC) firms HealthCare Ventures and Morgenthaler Ventures [163]. 118 However, the challenge is to find and occupy a niche in this crowded market with large established companies. Establishing a presence in this market, while not directly involved in cell therapy, can also be a strategy to enter the regenerative medicine market more directly. This could be achieved by entering collaborations with firms developing hPSC-derived cell therapy products in the hope that when a product does succeed, the firm would have locked in future demand for the consumables used in the manufacture of the cell therapy product. Companies looking to supply to this market will need to develop or have access to cGMP-certified facilities, and satisfy other FDA requirements such as those for record keeping, and purity and potency of product [164]. 119 4.4 Pluripotent Stem Cell-Derived Tools for Disease Modelling, Drug Discovery, and Toxicology While it is understandable that the general public’s interest in hPSC stem from their potential for use in replacement cell therapies, researchers and pharmaceutical companies are currently equally or even more interested in hPSCs as research tools for use in developmental studies, disease modeling, drug discovery, and toxicology [165– 167]. While this is a market with evolving business models, researchers [166, 167], investors [161], and industry [165, 168] are bullish on its short to medium-term profitability and scalability. Using hPSC’s ability to differentiate into specific adult cell phenotypes, researchers are now able to test drug candidates in vitro for toxicological effects, displacing current technologies of using transformed cell lines, primary cell lines, or animal models. The market for such an application alone has been estimated to be US$100M per product, for a total of US$500M or more [161]. However, it is hiPSC’s ability to generate adult cells with specific disease phenotypes, by creating hiPSC from patients with those diseases, and differentiating the cells into the desired cell types, that hPSCs come into their own as research tools. This ‘disease-in-a-dish’ concept [166] overcomes the lack of appropriate animal models for certain human diseases, and also enables highthroughput screening activities not possible with use of animal subjects. Multiple disease-specific hiPSC lines have already been developed, such as those derived from patients with Type 1 diabetes, Parkinson’s disease, and Down’s syndrome [20]. 120 Companies already active in this new market space also include both small and large firms such as iPierian (San Francisco, CA), Fate Therapeutics (San Diego, CA), GlaxoSmithKlein (London, UK), and GE Healthcare (Piscataway, NJ). For example, iPierian focusses on drug discovery platforms for neurodegenerative diseases such as Alzheimer’s, while GE Healthcare produces human cardiomyocytes derived from ESC for both toxicology studies and drug discovery. We are only beginning to see profitability and hence market share and leadership are in flux, with the first mover advantage for many diseases still not taken. That said, there are several technical challenges that need to be addressed: (1) lack of reliable and efficient differentiation protocols to produce certain desired differentiated cell types; (2) heterogeneity in differentiated cell phenotypes after differentiation; and (3) difficulty in recapitulating the disease in the in vitro model. For (1), both purity of the final cell product and amount of cells produced are still lacking. Methods to purify for the desired cell types continue to be topics of research [24]. As for (2), it has been reported that cells derived from in vitro differentiation exhibit heterogeneity in their level of maturity, and can thus have different phenotypes from cell to cell. Using these cells to assess the effect of a drug may well result in multiple false positives and negatives arising from a lack of consistency in the cells used [24]. Finally, as certain diseases may not be cell-type specific, but arise instead from interactions between different cell types (autoimmune diseases are a prime example), the ‘disease-in-a-dish’ may not be able to capture the phenotype to be studied. 121 In spite of the aforementioned technical challenges to be addressed, investors and entrepreneurs are positive about this market, and it has been described to have “the most attractive commercial opportunities for companies” in the immediate future [161]. Compared to providing tools for PSC-related research, developing the PSC themselves as research tools allows companies active in this market to move more directly into therapeutic applications of hPSC products. Building competencies necessary for success in this market, especially manufacturing-scale methods for differentiation, will be crucial in entering the cell therapy market. In addition, the possibility of short to medium term profitability in this market will allow companies to obtain favorable valuations from the capital markets, as well as to build a brand and financial resources for future expansion into cell therapies, creating opportunities for long term success. Another application of hPSCs in this arena that is further away from commercial reality is the concept of personalized drug discovery and testing. Instead of a ‘disease-in-adish’, the possibility is now of a ‘patient-in-a-dish’ where each testing platform is genetically identical to the specific patient from which it was derived using iPSC technologies. If successfully implemented, individuals can now test the safety and efficacy of drugs on their own cells. However, the profitability of such a business model will inevitably be affected by the higher costs of having personalized testing platforms rather than a common one for each disease. The viability of such a service will depend on degree of heterogeneity in responses to the drugs being tested, and the prevalence of the disease. While consumers of the ‘disease-in-a-dish’ product would be commercial laboratories or academic researchers, it is less clear who would be paying for the 122 ‘patient-in-a-dish’ service. It is likely that individuals will have to bear this cost, through health insurance providers, although one could imagine that the drug manufacturer may choose to defray part or all of the cost as a means to avail themselves of future recurring purchases of their therapy. In fact, some of the challenges associated with personalized cell therapy manifest themselves here, where there needs to be a severe unmet need (such as a high likelihood of a severe adverse reaction to a drug which, on the other hand, could be therapeutically very effective) for such a product to be commercially viable. 123 4.5 Replacement Cell Therapies Derived from Pluripotent Stem Cells We now turn to the major goal of effective, possibly personal, cell-based therapeutics derived from hPSCs. Despite the fact that such technologies are still in their nascent state, there are already those in the general public who are seeking such therapies, usually driven by personal circumstances, fuelled by inaccurate depictions of stem cell based medicine perpetuated by unscrupulous providers [169]. In reality, the FDA has clarified that there are currently no approved hPSC-derived cell therapies, although two are currently undergoing clinical trials, not including Geron’s high profile exit from the field after citing financial concerns [170]. To investors and entrepreneurs, the larger field of cell therapy, which is agnostic to the source of the cells used, is itself just a few years ahead in maturity. This rapidly evolving field has both ardent supporters and firm detractors [171–173]. Adding on the uncertainty of hPSC technology raises additional challenges, both technical and commercial, that need to be addressed for the successful commercialization of hPSCderived cell therapies. As with using hPSC-derived cells as drug discovery tools, a primary technological challenge is the lack of robust and scalable protocols capable of generating the desired cell types in the required numbers and purity. That they will be used for transplants raises additional constraints from the US Food & Drug Administration (FDA) such as additional investigations needed for those that have ex vivo contact with live non-human cells, on top of the already understandably strict guidelines for cell therapies [174–176]. 124 A second technological challenge is the difficulty of tracking cells after their introduction into the patient, and the lack of possible safeguards to reverse the procedure should anything go wrong [177]. The field requires major advances in biomonitoring and novel technologies capable of selectively removing or destroying rogue cells after transplantation in order to properly address safety concerns arising from the use of such therapeutics. Thirdly, the issue of tumorigenicity of residual, undifferentiated hPSC is a major cause for concern. Uncontrolled differentiation of pluripotent cells introduced into a patient, together with the hPSC’s ability to proliferate indefinitely, creates a safety issue unlike that of any other cellular therapeutic. In addition, genetic abnormalities in the hPSC could potentially cause the development of teratocarcinomas, which are highly malignant tumors with a persisting core of proliferating stem cells, raising the level of concern yet again [178]. A related problem arises from the use of hiPSC around the issue of safety of the method of reprogramming used to generate the pluripotent cells. While recent advances have moved away from genetic manipulations using viral vectors, favoring transient, non-integrative methodologies [179], difficulties in assessing the quality and efficiency of the reprogramming, heterogeneities in effectiveness of the techniques depending on the starting cell populations, and finally, concerns regarding the possible post-differentiation tumorigenicity of the cell products, make this a closely studied field. 125 Finally, while it is commonly recognized that allogeneic transplants using hPSC-derived products will necessarily have to address the immune response of the recipient, the assumption that autologous hPSC-derived cellular therapies will not face the same issues has been cast into doubt. Multiple recent publications reached conflicting conclusions [31, 180, 181], and this is not surprising, given that these cells have been exposed to factors absent in normal human physiology which may have altered their phenotype sufficiently to evoke an immune response. Even as researchers are tackling these technological challenges, investors, industry players, and entrepreneurs are finding that the rules of this new market are far from being written, and are contending with challenges of their own. Firstly, the patent landscape for intellectual property (IP) associated with PSCs has been described as a “minefield” [182], with companies interested in entering the field having to invest upwards of hundreds of millions of dollars on IP protection in order to legally operate. As a result, successful companies in this field will need to have comprehensive and robust IP strategies, as well as the ability to either license the necessary technologies from their holders, or operate around them. This issue is compounded by the fact that there is still no harmonization world-wide regarding the patentability of hPSC-associated technologies, with the example of Germany restricting stem cell patents [183]. The original patents issued to James Thomson and the Wisconsin Alumni Research Fund (WARF) covering hESC survived legal challenge in 126 the US, but with the rapidly evolving science and intense interest in the field, we should not expect that we will hear the definitive last word on this complex issue any time soon [184–186]. Looking beyond the issue of IP, business models and their associated sustainability have not been verified. Successful models will arise from the simultaneous evolution of technology and business strategy. A prior study recommended five possible strategies for value capture in the regenerative medicine market: (1) Orphan Designation, where the company focuses on a small market with unmet needs to take advantage of favored regulatory treatment resulting in prolonged market exclusivity, then using this opportunity to expand into other products with larger demands in the long run; (2) High Impact Therapies, where there are high unmet needs for treatments for chronic diseases in order to make oneself attractive for financing; (3) Platform Solution, where a single product serves multiple disease markets in order to build a diverse revenue stream; (4) Therapy Kit, for therapies with minimal cell manipulations but depend more on devices for administration, are for those firms with the goal of transforming into a device-based company; and (5) Incremental Solution, where the company introduces an incrementally better product than an existing offering, making use of regulators’ and physician’s familiarities with the product category [187]. For hPSC products, it is not possible for a company to adopt strategy (5), for obvious reasons, and there does not seem to be current endeavors to develop products suitable for strategies (3) and (4). Most companies entering the field are expected to adopt 127 strategy (2), given the need to attract long term and high levels of financing, and the obligation to generate high returns. As the market matures in the future, newer companies will attempt to find their own niche by adopting (1), or gain traction using (5). However, this is projecting far into the future, assuming continued successes for hPSCderived cell therapies for a wide variety of conditions. It is also worthwhile to consider the impact of pursuing allogeneic or autologous products. As previously mentioned, autologous cell therapies would naturally attract greater operating costs due to the need for individualized processing of cells from each patient, while autologous treatments can take advantage of the economies of scale of bulk processing. This advantage is reduced with lower numbers of units produced, and suggests that for companies pursuing strategy (1) of targeting orphan diseases, autologous treatments may succeed, especially if the increased cost is lower than the potential gain of the autologous product over the allogeneic equivalent. A proposed hybrid model, where the patient requires multiple doses of the same product, can take advantage of economies of scale at the individual level, where the multiple doses of the autologous product are manufactured simultaneously. This would reduce the average costs of goods compared to single dose autologous products, but still suffers from a cost disadvantage when compared to allogeneic equivalents [161]. It is well known that there are political uncertainties and ethical controversies associated with the generation and use of hESCs, resulting in regulatory restrictions on their derivation and use in many parts of the world, including the US [188]. From the Bush 128 era restrictions to Sherley v. Sebelius, the research community has faced uncertainties in funding, and this may well have pushed researchers away from the field [189, 190]. As hESC research is still far from commercialization, a large proportion of the advances made have come from academic research funded through federal support. With these restrictions, investors may view hESC endeavors as being more risky than other avenues, with reduced risk sharing by the government and academia. Even hiPSCs may not be free from controversy, as questions around ownership of derived cell lines created from donated tissue arise, and issues of personal privacy around information indirectly obtained as a result of undergoing treatment become known, as the field advances [191]. Companies will need to remain aware of the sentiment of policy makers and the general public with regards to such issues and anticipate that their business models and strategies will need to adapt to this audience. With a US market conservatively estimated at US$1B just for a hPSC-derived product for Type 1 diabetes alone [161], and similar blockbuster potential expected for other diseases with levels of morbidity and mortality that cannot be managed with current therapies, it is not surprising that, despite the nascent state of the technology and significant risks and uncertainties, there continues to be widespread efforts to translate hPSC-based research into clinical applications. 129 4.6 Discussion While the major commercialization potential of hPSCs is that of generating replacement cells, tissues, and possibly organs for regenerative medicine, we are still one to two decades away from a first commercial product, extrapolating from the historical development timelines of other biotechnologies such as protein-based pharmaceuticals and therapeutic antibodies [192]. In the meantime, it is important to temper the expectations of both supporters and detractors of the field, given the far-reaching claims of the potential of hPSC technology, in order for the field to continue to receive sustainable levels of financial, regulatory, and public support [193]. Entrepreneurs and investors eyeing the substantial prize on the long term horizon can and already are making moves in the supporting market for consumables and equipment, as well as developing hPSC-based diagnostic and research tools to position themselves for successful transitions into the business of hPSC-based therapeutics [163, 165, 168]. The road to successful commercialization will not be smooth. Geron’s high profile exit from the cell therapy market, after the fanfare of being the first hPSC-derived cell therapy to reach clinical trials (a hESC-derived product for spinal cord injuries), was seen by many as a setback. However, after the announcement, while Geron’s share price dropped by 20%, the two other players with significant investments in the field, Advanced Cell Technology, currently the only firm with on-going Phase I trials for hESCderived therapies for macular degeneration, and BioTime, saw modest gains instead [170]. Prices of companies in the wider biotechnology field, as a whole, did not react to the news. This signified that investors are confident in the future earnings potential of 130 companies pursuing hPSC-based therapies. BioTime, led by Geron’s former CEO Dr. Tom Okarma, subsequently purchased Geron’s hPSC assets, fulfilling the underlying expectation that Geron’s exit was an opportunity for other players to pick up technologies and IP that have already been significantly de-risked by Geron’s years of investments [170, 194]. Geron’s decision was reported to be financially driven, which underscores the large levels of investments and long incubation times before results can be reasonably expected. With the current VC exit timelines of between five and ten years, and the first commercial product expected to be at least twice that amount of time away, there is a financial gap that needs to be filled for the field to continue to move forward. Governments and non-profit foundations have typically filled the gap for research in this field, and more recently, the industry has also stepped in, many with realistic short to medium term strategies for sustainable operations (including those described above), with an eye on the long-term goal. Savvy investors and entrepreneurs who are willing to be patient, have the knowledge and ability to navigate the complexities of the hPSC markets, and the intellectual and technological capital for high impact therapies, will be the ones who deliver on the oft-repeated promises made at the discovery of hESCs. 131 -Page Intentionally Left Blank- 132 References [1] M. J. Evans and M. H. Kaufman, “Establishment in culture of pluripotent cells from mouse embryos,” Nature, vol. 292, pp. 154–156, 1981. [2] J. A. Thomson, J. Itskovitz-Eldor, S. S. Shapiro, M. A. Waknitz, J. J. Swiergiel, V. S. Marshall, and J. M. Jones, “Embryonic Stem Cell Lines Derived from Human Blastocysts,” Science, vol. 282, no. 5391, pp. 1145–1147, Nov. 1998. [3] K. Hochedlinger and R. Jaenisch, “Nuclear Transplantation, Embryonic Stem Cells and the Potential for Cell Therapy,” The New England Journal of Medicine, vol. 349, pp. 275–286, Jan. 2003. [4] G. Q. Daley, M. A. Goodell, and E. Y. Snyder, “Realistic Prospects for Stem Cell Therapeutics,” Hematology, pp. 398–418, Jan. 2003. [5] G. Keller, “Embryonic stem cell differentiation: emergence of a new era in biology and medicine,” Genes & Development, vol. 19, pp. 1129–1155, May 2005. [6] G. Q. Daley and D. T. Scadden, “Prospects for Stem Cell-Based Therapy,” Cell, vol. 132, no. 4, pp. 544–548, 2008. [7] S. Yamanaka, J. Li, G. Kania, S. Elliott, R. P. Wersto, J. Van Eyk, A. M. Wobus, and K. R. Boheler, “Pluripotency of embryonic stem cells,” Cell and Tissue Research, vol. 331, no. 1, pp. 5–22, Jan. 2008. [8] J. Rossant, “Stem Cells and Early Lineage Development,” Cell, vol. 132, no. 4, pp. 527–531, Feb. 2008. [9] C. J. Lengner, F. D. Camargo, K. Hochedlinger, G. G. Welstead, S. Zaidi, S. Gokhale, H. R. Scholer, A. Tomilin, and R. Jaenisch, “Oct4 Expression Is Not Required for Mouse Somatic Stem Cell Self-Renewal,” Cell Stem Cell, vol. 1, pp. 403–415, Oct. 2007. [10] M. A. Laflamme, K. Y. Chen, A. V Naumova, V. Muskheli, J. A. Fugate, S. K. Dupras, H. Reinecke, C. Xu, M. Hassanipour, S. Police, C. O’Sullivan, L. Collins, Y. Chen, E. Minami, E. A. Gill, S. Ueno, C. Yuan, J. Gold, and C. E. Murry, “Cardiomyocytes derived from human embryonic stem cells in pro-survival factors enhance function of infarcted rat hearts,” Nature Biotechnology, vol. 25, pp. 1015– 1024, Sep. 2007. [11] R. K. Burt, L. Verda, D.-A. Kim, Y. Oyama, K. Luo, and C. Link, “Embryonic Stem Cells As an Alternate Marrow Donor Source: Engraftment without Graft-VersusHost Disease,” The Journal of Experimental Medicine, vol. 199, no. 7, pp. 895– 904, Apr. 2004. 133 [12] J.-H. Kim, J. M. Auerbach, J. a Rodríguez-Gómez, I. Velasco, D. Gavin, N. Lumelsky, S.-H. Lee, J. Nguyen, R. Sánchez-Pernaute, K. Bankiewicz, and R. McKay, “Dopamine neurons derived from embryonic stem cells function in an animal model of Parkinson’s disease,” Nature, vol. 418, pp. 50–56, Jul. 2002. [13] K. A. D’Amour, A. G. Bang, S. Eliazer, O. G. Kelly, A. D. Agulnick, N. G. Smart, M. a Moorman, E. Kroon, M. K. Carpenter, and E. E. Baetge, “Production of pancreatic hormone-expressing endocrine cells from human embryonic stem cells,” Nature Biotechnology, vol. 24, no. 11, pp. 1392–1401, Nov. 2006. [14] M. Eiraku, K. Watanabe, M. Matsuo-Takasaki, M. Kawada, S. Yonemura, M. Matsumura, T. Wataya, A. Nishiyama, K. Muguruma, and Y. Sasai, “SelfOrganized Formation of Polarized Cortical Tissues from ESCs and Its Active Manipulation by Extrinsic Signals,” Cell Stem Cell, vol. 3, no. 5, pp. 519–532, Nov. 2008. [15] W. M. Rideout, K. Hochedlinger, M. Kyba, G. Q. Daley, and R. Jaenisch, “Correction of a Genetic Defect by Nuclear Transplantation and Combined Cell and Gene Therapy,” Cell, vol. 109, no. 1, pp. 17–27, Apr. 2002. [16] K. Takahashi, K. Tanabe, M. Ohnuki, M. Narita, T. Ichisaka, K. Tomoda, and S. Yamanaka, “Induction of Pluripotent Stem Cells from Adult Human Fibroblasts by Defined Factors,” Cell, vol. 131, no. 5, pp. 861–872, Nov. 2007. [17] J. Yu, M. A. Vodyanik, K. Smuga-Otto, J. Antosiewicz-Bourget, J. L. Frane, S. Tian, J. Nie, G. A. Jonsdottir, V. Ruotti, R. Stewart, I. I. Slukvin, and J. A. Thomson, “Induced Pluripotent Stem Cell Lines Derived from Human Somatic Cells,” Science, vol. 318, no. 5858, pp. 1917–1920, Dec. 2007. [18] M. Wernig, A. Meissner, R. Foreman, T. Brambrink, K. Hochedlinger, M. Ku, B. E. Bernstein, and R. Jaenisch, “In vitro reprogramming of fibroblasts into a pluripotent ES-cell-like state,” Nature, vol. 448, no. 7151, pp. 318–324, Jul. 2007. [19] J. Hanna, M. Wernig, S. Markoulaki, C.-W. Sun, A. Meissner, J. P. Cassady, C. Beard, T. Brambrink, L.-C. Wu, T. M. Townes, and R. Jaenisch, “Treatment of Sickle Cell Anemia Mouse Model with iPS Cells Generated from Autologous Skin,” Science, vol. 318, pp. 1920–1923, Dec. 2007. [20] I.-H. Park, N. Arora, H. Huo, N. Maherali, T. Ahfeldt, A. Shimamura, M. W. Lensch, C. Cowan, K. Hochedlinger, and G. Q. Daley, “Disease-Specific Induced Pluripotent Stem Cells,” Cell, vol. 134, no. 5, pp. 877–886, Sep. 2008. [21] N. Maherali, R. Sridharan, W. Xie, J. Utikal, S. Eminli, K. Arnold, M. Stadtfeld, R. Yachechko, J. Tchieu, R. Jaenisch, K. Plath, and K. Hochedlinger, “Directly Reprogrammed Fibroblasts Show Global Epigenetic Remodeling and Widespread Tissue Contribution,” Cell Stem Cell, vol. 1, no. 1, pp. 55–70, Jun. 2007. 134 [22] S. Nishikawa, R. A. Goldstein, and C. R. Nierras, “The promise of human induced pluripotent stem cells for research and therapy,” Nature Reviews. Molecular Cell Biology, vol. 9, pp. 725–729, 2008. [23] M. Stadtfeld and K. Hochedlinger, “Induced pluripotency: history, mechanisms, and applications,” Genes & Development, vol. 24, no. 20, pp. 2239–2263, Oct. 2010. [24] S. M. Wu and K. Hochedlinger, “Harnessing the potential of induced pluripotent stem cells for regenerative medicine,” Nature Cell Biology, vol. 13, no. 5, pp. 497– 505, May 2011. [25] F. Soldner, J. Laganière, A. W. Cheng, D. Hockemeyer, Q. Gao, R. Alagappan, V. Khurana, L. I. Golbe, R. H. Myers, S. Lindquist, L. Zhang, D. Guschin, L. K. Fong, B. J. Vu, X. Meng, F. D. Urnov, E. J. Rebar, P. D. Gregory, H. S. Zhang, and R. Jaenisch, “Generation of Isogenic Pluripotent Stem Cells Differing Exclusively at Two Early Onset Parkinson Point Mutations,” Cell, vol. 146, no. 1, pp. 318–331, Jul. 2011. [26] M. Wernig, J.-P. Zhao, J. Pruszak, E. Hedlund, D. Fu, F. Soldner, V. Broccoli, M. Constantine-Paton, O. Isacson, and R. Jaenisch, “Neurons derived from reprogrammed fibroblasts functionally integrate into the fetal brain and improve symptoms of rats with Parkinson’s disease,” Proceedings of the National Academy of Sciences of the United States of America, vol. 105, no. 15, pp. 5856– 5861, Apr. 2008. [27] C. N. Mayhew and J. M. Wells, “Converting human pluripotent stem cells into beta cells: recent advances and future challenges,” Current Opinion in OrganTransplantation, vol. 15, no. 1, pp. 54–60, 2010. [28] D. Taura, M. Sone, K. Homma, N. Oyamada, K. Takahashi, N. Tamura, S. Yamanaka, and K. Nakao, “Induction and Isolation of Vascular Cells From Human Induced Pluripotent Stem Cells - Brief Report,” Arteriosclerosis, Thrombosis, and Vascular Biology, vol. 29, no. 7, pp. 1100–1103, Jul. 2009. [29] M. Stadtfeld, E. Apostolou, H. Akutsu, A. Fukuda, P. Follett, S. Natesan, T. Kono, T. Shioda, and K. Hochedlinger, “Aberrant silencing of imprinted genes on chromosome 12qF1 in mouse induced pluripotent stem cells,” Nature, vol. 465, no. 7295, pp. 175–181, May 2010. [30] “CTGTAC Meeting # 45 Cellular Therapies Derived from Human Embryonic Stem Cells – Considerations for Pre-Clinical Safety Testing and Patient Monitoring,” 2008. [31] T. Zhao, Z.-N. Zhang, Z. Rong, and Y. Xu, “Immunogenicity of induced pluripotent stem cells,” Nature, vol. 474, no. 7350, pp. 212–215, Jan. 2011. 135 [32] K. Fukuda and S. Yuasa, “Stem Cells as a Source of Regenerative Cardiomyocytes.,” Circulation Research, vol. 98, pp. 1002–1013, Apr. 2006. [33] A. G. Smith, J. K. Heath, D. D. Donaldson, G. G. Wong, J. Moreau, M. Stahlt, and D. Rogers, “Inhibition of pluripotential embryonic stem cell differentiation by purified polypeptides,” Nature, vol. 336, pp. 688–690, 1988. [34] Q.-L. Ying, J. Wray, J. Nichols, L. Batlle-Morera, B. Doble, J. Woodgett, P. Cohen, and A. Smith, “The ground state of embryonic stem cell self-renewal,” Nature, vol. 453, no. 7194, pp. 519–523, 2008. [35] C. Xu, E. Rosler, J. Jiang, J. S. Lebkowski, J. D. Gold, C. O’Sullivan, K. DelavanBoorsma, M. Mok, A. Bronstein, and M. K. Carpenter, “Basic Fibroblast Growth Factor Supports Undifferentiated Human Embryonic Stem Cell Growth Without Conditioned Medium,” Stem Cells, vol. 23, no. 3, pp. 315–323, Mar. 2005. [36] J. B. Lee, J. M. Song, J. E. Lee, J. H. Park, S. J. Kim, S. M. Kang, J. N. Kwon, M. K. Kim, S. Il Roh, and H. S. Yoon, “Available human feeder cells for the maintenance of human embryonic stem cells,” Reproduction, vol. 128, pp. 727– 735, Dec. 2004. [37] C. Xu, M. S. Inokuma, J. Denham, K. Golds, P. Kundu, J. D. Gold, and M. K. Carpenter, “Feeder-free growth of undifferentiated human embryonic stem cells,” Nature Biotechnology, vol. 19, no. 10, pp. 971–974, Oct. 2001. [38] T. E. Ludwig, M. E. Levenstein, J. M. Jones, W. T. Berggren, E. R. Mitchen, J. L. Frane, L. J. Crandall, C. A. Daigh, K. R. Conard, M. S. Piekarczyk, R. A. Llanas, and J. A. Thomson, “Derivation of human embryonic stem cells in defined conditions,” Nature Biotechnology, vol. 24, no. 2, pp. 185–187, Feb. 2006. [39] L. A. Boyer, D. Mathur, and R. Jaenisch, “Molecular control of pluripotency,” Current Opinion in Genetics & Development, vol. 16, no. 5, pp. 455–462, 2006. [40] L. A. Boyer, T. I. Lee, M. F. Cole, S. E. Johnstone, S. S. Levine, J. P. Zucker, M. G. Guenther, R. M. Kumar, H. L. Murray, R. G. Jenner, D. K. Gifford, D. A. Melton, R. Jaenisch, and R. A. Young, “Core Transcriptional Regulatory Circuitry in Human Embryonic Stem Cells,” Cell, vol. 122, no. 6, pp. 947–956, Sep. 2005. [41] Y.-H. Loh, Q. Wu, J.-L. Chew, V. B. Vega, W. Zhang, X. Chen, G. Bourque, J. George, B. Leong, K.-Y. Wong, J. Liu, K. W. Sung, X.-D. Zhao, C. W. H. Lee, K.-P. Chiu, L. Lipovich, V. A. Kuznetsov, P. Robson, L. W. Stanton, C.-L. Wei, Y. Ruan, B. Lim, and H.-H. Ng, “The Oct4 and Nanog transcription network regulates pluripotency in mouse embryonic stem cells,” Nature Genetics, vol. 38, pp. 431– 440, Apr. 2006. 136 [42] H.-H. Ng and M. A. Surani, “The transcriptional and signalling networks of pluripotency,” Nature Cell Biology, vol. 13, no. 5, pp. 490–496, May 2011. [43] R. A. Young, “Control of the Embryonic Stem Cell State,” Cell, vol. 144, no. 6, pp. 940–954, Mar. 2011. [44] D. Zeineddine, E. Papadimou, K. Chebli, M. Gineste, J. Liu, C. Grey, S. Thurig, A. Behfar, V. A. Wallace, I. S. Skerjanc, and M. Pucéat, “Oct-3/4 dose dependently regulates specification of embryonic stem cells toward a cardiac lineage and early heart development.,” Developmental Cell, vol. 11, no. 4, pp. 535–546, Oct. 2006. [45] H. Niwa, J. Miyazaki, and A. G. Smith, “Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells,” Nature Genetics, vol. 24, no. 4, pp. 372–376, Apr. 2000. [46] H. Niwa, Y. Toyooka, D. Shimosato, D. Strumpf, K. Takahashi, R. Yagi, and J. Rossant, “Interaction between Oct3/4 and Cdx2 Determines Trophectoderm Differentiation,” Cell, vol. 123, no. 5, pp. 917–929, Dec. 2005. [47] R. T. Rodriguez, J. M. Velkey, C. Lutzko, R. Seerke, D. B. Kohn, K. S. O’Shea, and M. T. Firpo, “Manipulation of OCT4 Levels in Human Embryonic Stem Cells Results in Induction of Differential Cell Types,” Experimental Biology and Medicine, vol. 232, no. 10, pp. 1368–1380, Nov. 2007. [48] M. Thomson, S. J. Liu, L.-N. Zou, Z. Smith, A. Meissner, and S. Ramanathan, “Pluripotency Factors in Embryonic Stem Cells Regulate Differentiation into Germ Layers,” Cell, vol. 145, no. 6, pp. 875–889, Jun. 2011. [49] A. M. Singh, T. Hamazaki, K. E. Hankowski, and N. Terada, “A Heterogeneous Expression Pattern for Nanog in Embryonic Stem Cells,” Stem Cells, vol. 25, no. 10, pp. 2534–2542, Oct. 2007. [50] I. Chambers, J. Silva, D. Colby, J. Nichols, B. Nijmeijer, M. Robertson, J. Vrana, K. Jones, L. Grotewold, and A. Smith, “Nanog safeguards pluripotency and mediates germline development,” Nature, vol. 450, no. 7173, pp. 1230–1234, 2007. [51] P. Yu, G. Pan, J. Yu, and J. A. Thomson, “FGF2 Sustains NANOG and Switches the Outcome of BMP4-Induced Human Embryonic Stem Cell Differentiation,” Cell Stem Cell, vol. 8, no. 3, pp. 326–334, Mar. 2011. [52] Y. Toyooka, D. Shimosato, K. Murakami, K. Takahashi, and H. Niwa, “Identification and characterization of subpopulations in undifferentiated ES cell culture,” Development, vol. 135, no. 5, pp. 909–918, Mar. 2008. 137 [53] K. Hayashi, S. M. C. D. S. Lopes, F. Tang, and M. A. Surani, “Dynamic Equilibrium and Heterogeneity of Mouse Pluripotent Stem Cells with Distinct Functional and Epigenetic States,” Cell Stem Cell, vol. 3, pp. 391–401, Oct. 2008. [54] K. M. Loh and B. Lim, “A Precarious Balance: Pluripotency Factors as Lineage Specifiers.,” Cell Stem Cell, vol. 8, pp. 363–369, Apr. 2011. [55] D. A. F. Loebel, C. M. Watson, R. A. De Young, and P. P. L. Tam, “Lineage choice and differentiation in mouse embryos and embryonic stem cells,” Developmental Biology, vol. 264, pp. 1–14, Dec. 2003. [56] D. E. Cohen and D. Melton, “Turning straw into gold: directing cell fate for regenerative medicine,” Nature Reviews. Genetics, vol. 12, no. 4, pp. 243–252, Apr. 2011. [57] A. M. Becker, M. Rubart, and L. J. Field, “Inducing Embryonic Stem Cells to Become Cardiomyocytes,” in Regenerating the Heart, I. S. Cohen and G. R. Gaudette, Eds. Totowa, NJ: Humana Press, 2011, pp. 7–24. [58] S. J. Kattman, A. D. Witty, M. Gagliardi, N. C. Dubois, M. Niapour, A. Hotta, J. Ellis, and G. Keller, “Stage-Specific Optimization of Activin/Nodal and BMP Signaling Promotes Cardiac Differentiation of Mouse and Human Pluripotent Stem Cell Lines,” Cell Stem Cell, vol. 8, pp. 228–240, Feb. 2011. [59] C. Xu, S. Police, N. Rao, and M. K. Carpenter, “Characterization and Enrichment of Cardiomyocytes Derived From Human Embryonic Stem Cells,” Circulation Research, vol. 91, no. 6, pp. 501–508, Aug. 2002. [60] T. Takahashi, B. Lord, P. C. Schulze, R. M. Fryer, S. S. Sarang, S. R. Gullans, and R. T. Lee, “Ascorbic Acid Enhances Differentiation of Embryonic Stem Cells into Cardiac Myocytes,” Circulation, vol. 107, no. 14, pp. 1912–1916, Apr. 2003. [61] A. Behfar, C. Perez-Terzic, R. S. Faustino, D. K. Arrell, D. M. Hodgson, S. Yamada, M. Puceat, N. Niederländer, A. E. Alekseev, L. V Zingman, and A. Terzic, “Cardiopoietic programming of embryonic stem cells for tumor-free heart repair.,” The Journal of Experimental Medicine, vol. 204, no. 2, pp. 405–420, Feb. 2007. [62] R. Passier, D. Ward-van Oostwaard, J. Snapper, J. Kloots, R. J. Hassink, E. Kuijk, B. Roelen, A. B. de la Riviere, and C. Mummery, “Increased Cardiomyocyte Differentiation from Human Embryonic Stem Cells in Serum-Free Cultures,” Stem Cells, vol. 23, pp. 772–780, 2005. [63] L. Warren, P. D. Manos, T. Ahfeldt, Y.-H. Loh, H. Li, F. Lau, W. Ebina, P. K. Mandal, Z. D. Smith, A. Meissner, G. Q. Daley, A. S. Brack, J. J. Collins, C. Cowan, T. M. Schlaeger, and D. J. Rossi, “Highly Efficient Reprogramming to 138 Pluripotency and Directed Differentiation of Human Cells with Synthetic Modified mRNA,” Cell Stem Cell, vol. 7, no. 5, pp. 618–630, Nov. 2010. [64] S. Zhu, H. Wurdak, J. Wang, C. A. Lyssiotis, E. C. Peters, C. Y. Cho, X. Wu, and P. G. Schultz, “A Small Molecule Primes Embryonic Stem Cells for Differentiation,” Cell Stem Cell, vol. 4, no. 5, pp. 416–426, May 2009. [65] G. Narazaki, H. Uosaki, M. Teranishi, K. Okita, B. Kim, S. Matsuoka, S. Yamanaka, and J. K. Yamashita, “Directed and Systematic Differentiation of Cardiovascular Cells From Mouse Induced Pluripotent Stem Cells,” Circulation, vol. 118, no. 5, pp. 498–506, Jul. 2008. [66] E. Kroon, L. A. Martinson, K. Kadoya, A. G. Bang, O. G. Kelly, S. Eliazer, H. Young, M. Richardson, N. G. Smart, J. Cunningham, A. D. Agulnick, K. A. D’Amour, M. K. Carpenter, and E. E. Baetge, “Pancreatic endoderm derived from human embryonic stem cells generates glucose-responsive insulin-secreting cells in vivo,” Nature Biotechnology, vol. 26, no. 4, pp. 443–452, Apr. 2008. [67] M. Bengtsson, A. Ståhlberg, P. Rorsman, and M. Kubista, “Gene expression profiling in single cells from the pancreatic islets of Langerhans reveals lognormal distribution of mRNA levels,” Genome Research, vol. 15, no. 10, pp. 1388–1392, Oct. 2005. [68] A. Raj, C. S. Peskin, D. Tranchina, D. Y. Vargas, and S. Tyagi, “Stochastic mRNA Synthesis in Mammalian Cells,” PLoS biology, vol. 4, no. 10, pp. 1707–1719, Oct. 2006. [69] A. Sigal, R. Milo, A. Cohen, N. Geva-Zatorsky, Y. Klein, Y. Liron, N. Rosenfeld, T. Danon, N. Perzov, and U. Alon, “Variability and memory of protein levels in human cells,” Nature, vol. 444, no. 7119, pp. 643–646, 2006. [70] J. M. Levsky and R. H. Singer, “Gene expression and the myth of the average cell,” Trends in Cell Biology, vol. 13, no. 1, pp. 4–6, 2003. [71] S. J. Altschuler and L. F. Wu, “Cellular Heterogeneity: Do Differences Make a Difference?,” Cell, vol. 141, no. 4, pp. 559–563, May 2010. [72] H. H. Chang, M. Hemberg, M. Barahona, D. E. Ingber, and S. Huang, “Transcriptome-wide noise controls lineage choice in mammalian progenitor cells,” Nature, vol. 453, no. 7194, pp. 544–547, 2008. [73] M. A. Canham, A. A. Sharov, M. S. H. Ko, and J. M. Brickman, “Functional Heterogeneity of Embryonic Stem Cells Revealed through Translational Amplification of an Early Endodermal Transcript,” PLoS Biology, vol. 8, no. 5, p. e1000379, May 2010. 139 [74] K. H. Narsinh, N. Sun, V. Sanchez-Freire, A. S. Lee, P. Almeida, S. Hu, T. Jan, K. D. Wilson, D. Leong, J. Rosenberg, M. Yao, R. C. Robbins, and J. C. Wu, “Single cell transcriptional profiling reveals heterogeneity of human induced pluripotent stem cells,” The Journal of Clinical Investigation, vol. 121, no. 3, pp. 1217–1221, Mar. 2011. [75] M. F. Pera and P. P. L. Tam, “Extrinsic regulation of pluripotent stem cells,” Nature, vol. 465, no. 7299, pp. 713–720, Jun. 2010. [76] A. M. Singh and N. Terada, “Bypassing Heterogeneity: The Road to Embryonic Stem Cell-Derived Cardiomyocyte Specification,” Trends in Cardiovascular Medicine, vol. 17, no. 3, pp. 96–101, 2007. [77] M. Elowitz, A. Levine, and E. Siggia, “Stochastic gene expression in a single cell,” Science, vol. 297, pp. 1183–1186, 2002. [78] M. Kaern, T. C. Elston, W. J. Blake, and J. J. Collins, “Stochasticity in gene expression: from theories to phenotypes.,” Nature Reviews. Genetics, vol. 6, pp. 451–464, Jun. 2005. [79] T.-L. To and N. Maheshri, “Noise Can Induce Bimodality in Positive Transcriptional Feedback Loops Without Bistability,” Science, vol. 327, no. 5969, pp. 1142–1145, Feb. 2010. [80] S. Huang, “Non-genetic heterogeneity of cells in development: more than just noise,” Development, vol. 136, pp. 3853–3862, Dec. 2009. [81] M. Intaglietta, P. C. Johnson, and R. M. Winslow, “Microvascular and tissue oxygen distribution,” Cardiovascular Research, vol. 32, pp. 632–643, Oct. 1996. [82] P. W. Soothill, K. H. Nicolaides, C. H. Rodeck, and S. Campbell, “Effect of Gestational Age on Fetal and Intervillous Blood Gas and Acid-Base Values in Human Pregnancy,” Fetal Diagnosis and Therapy, vol. 1, no. 4, pp. 168–175, 1986. [83] O. Siggaard-Andersen and R. Huch, “The oxygen status of fetal blood,” Acta Anaesthesiologica Scandinavica, vol. 39, no. Supplementum 107, pp. 129–135, Jan. 1995. [84] E. Jauniaux, B. Gulbis, and G. J. Burton, “Physiological implications of the materno–fetal oxygen gradient in human early pregnancy,” Reproductive BioMedicine Online, vol. 7, no. 2, pp. 250–253, 2003. [85] Y. M. Lee, C.-H. Jeong, S.-Y. Koo, M. J. Son, H. S. Song, S.-K. Bae, J. A. Raleigh, H.-Y. Chung, M.-A. Yoo, and K.-W. Kim, “Determination of Hypoxic Region by Hypoxia Marker in Developing Mouse Embryos In Vivo: A Possible Signal for 140 Vessel Development,” Developmental Dynamics, vol. 220, pp. 175–186, Feb. 2001. [86] S. C. Land, “Hochachka’s ‘Hypoxia Defense Strategies’ and the development of the pathway for oxygen,” Comparative Biochemistry and Physiology. Part B, Biochemistry & Molecular Biology, vol. 139, no. 3, pp. 415–433, Nov. 2004. [87] E. Jauniaux, B. Gulbis, and G. J. Burton, “The Human First Trimester Gestational Sac Limits Rather than Facilitates Oxygen Transfer to the Foetus - A Review,” Placenta, vol. 24, no. Suppl A, pp. S86–S93, Apr. 2003. [88] R. H. Wenger, I. Kvietikova, A. Rolfs, M. Gassmann, and H. H. Marti, “Hypoxiainducible factor-1alpha is regulated at the post-mRNA level,” Kidney International, vol. 51, pp. 560–563, 1997. [89] G. L. Wang, B.-H. Jiang, E. A. Rue, and G. L. Semenza, “Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension,” Proceedings of the National Academy of Sciences of the United States of America, vol. 92, no. 12, pp. 5510–5514, Jun. 1995. [90] B.-H. Jiang, G. L. Semenza, C. Bauer, and H. H. Marti, “Hypoxia-inducible factor 1 levels vary exponentially over a physiologically relevant range of O2 tension,” The American Journal of Physiology, vol. 271, no. 4, pp. C1172–C1180, Oct. 1996. [91] C.-J. Hu, L.-Y. Wang, L. A. Chodosh, B. Keith, and M. C. Simon, “Differential Roles of Hypoxia-Inducible Factor 1alpha (HIF-1alpha) and HIF-2alpha in Hypoxic Gene Regulation,” Molecular and Cellular Biology, vol. 23, no. 24, pp. 9361–9374, 2003. [92] F. Wang, S. Thirumangalathu, and M. R. Loeken, “Establishment of New Mouse Embryonic Stem Cell Lines Is Improved by Physiological Glucose and Oxygen,” Cloning and Stem Cells, vol. 8, no. 2, pp. 108–116, Jan. 2006. [93] J. Gibbons, E. Hewitt, and D. K. Gardner, “Effects of Oxygen Tension on the Establishment and Lactate Dehydrogenase Activity of Murine Embryonic Stem Cells,” Cloning and Stem Cells, vol. 8, no. 2, pp. 117–122, 2006. [94] C. J. Lengner, A. a Gimelbrant, J. a Erwin, A. W. Cheng, M. G. Guenther, G. G. Welstead, R. Alagappan, G. M. Frampton, P. Xu, J. Muffat, S. Santagata, D. Powers, C. B. Barrett, R. a Young, J. T. Lee, R. Jaenisch, and M. Mitalipova, “Derivation of Pre-X Inactivation Human Embryonic Stem Cells under Physiological Oxygen Concentrations,” Cell, vol. 141, pp. 872–883, May 2010. [95] C.-H. Jeong, H.-J. Lee, J.-H. Cha, J. H. Kim, K. R. Kim, J.-H. Kim, D.-K. Yoon, and K.-W. Kim, “Hypoxia-inducible Factor-1alpha Inhibits Self-Renewal of Mouse 141 Embryonic Stem Cells in Vitro via Negative Regulation of the Leukemia Inhibitory Factor-STAT3 Pathway,” The Journal of Biological Chemistry, vol. 282, no. 18, pp. 13672–13679, May 2007. [96] D. E. Powers, J. R. Millman, R. B. Huang, and C. K. Colton, “Effects of Oxygen on Mouse Embryonic Stem Cell Growth, Phenotype Retention, and Cellular Energetics,” Biotechnology and Bioengineering, vol. 101, no. 2, pp. 241–254, Oct. 2008. [97] N. R. Forsyth, A. Musio, P. Vezzoni, A. H. R. W. Simpson, B. S. Noble, and J. McWhir, “Physiologic Oxygen Enhances Human Embryonic Stem Cell Clonal Recovery and Reduces Chromosomal Abnormalities,” Cloning and Stem Cells, vol. 8, no. 1, pp. 16–23, Jan. 2006. [98] S. D. Westfall, S. Sachdev, P. Das, L. B. Hearne, M. Hannink, R. M. Roberts, and T. Ezashi, “Identification of Oxygen-Sensitive Transcriptional Programs in Human Embryonic Stem Cells,” Stem Cells and Development, vol. 17, no. 5, pp. 869–881, Oct. 2008. [99] S. M. Prasad, M. Czepiel, C. Cetinkaya, K. Smigielska, S. C. Weli, H. Lysdahl, A. Gabrielsen, K. Petersen, N. Ehlers, T. Fink, S. L. Minger, and V. Zachar, “Continuous hypoxic culturing maintains activation of Notch and allows long-term propagation of human embryonic stem cells without spontaneous differentiation,” Cell Proliferation, vol. 42, no. 1, pp. 63–74, 2009. [100] C. E. Forristal, K. L. Wright, N. a Hanley, R. O. C. Oreffo, and F. D. Houghton, “Hypoxia inducible factors regulate pluripotency and proliferation in human embryonic stem cells cultured at reduced oxygen tensions,” Reproduction, vol. 139, pp. 85–97, Jan. 2010. [101] N. R. Forsyth, A. Kay, K. Hampson, A. Downing, R. Talbot, and J. McWhir, “Transcriptome alterations due to physiological normoxic (2% O2) culture of human embryonic stem cells,” Regenerative Medicine, vol. 3, no. 6, pp. 817–833, Nov. 2008. [102] T. Ezashi, P. Das, and R. M. Roberts, “Low O2 tensions and the prevention of differentiation of hES cells,” Proceedings of the National Academy of Sciences of the United States of America, vol. 102, no. 13, pp. 4783–4788, 2005. [103] M. V Gustafsson, X. Zheng, T. Pereira, K. Gradin, S. Jin, J. Lundkvist, J. L. Ruas, L. Poellinger, U. Lendahl, and M. Bondesson, “Hypoxia Requires Notch Signaling to Maintain the Undifferentiated Cell State,” Developmental Cell, vol. 9, pp. 617– 628, Nov. 2005. [104] K. L. Covello, J. Kehler, H. Yu, J. D. Gordan, A. M. Arsham, C.-J. Hu, P. A. Labosky, M. C. Simon, and B. Keith, “HIF-2alpha regulates Oct-4: effects of 142 hypoxia on stem cell function, embryonic development, and tumor growth,” Genes & Development, vol. 20, pp. 557–570, Mar. 2006. [105] Z. Hewitt, N. R. Forsyth, M. Waterfall, D. Wojtacha, A. J. Thomson, and J. McWhir, “Fluorescence-Activated Single Cell Sorting of Human Embryonic Stem Cells,” Cloning and Stem Cells, vol. 8, no. 3, pp. 225–234, Jan. 2006. [106] Y. Yoshida, K. Takahashi, K. Okita, T. Ichisaka, and S. Yamanaka, “Hypoxia Enhances the Generation of Induced Pluripotent Stem Cells,” Cell Stem Cell, vol. 5, no. 3, pp. 237–241, Sep. 2009. [107] J. Hanna, K. Saha, B. Pando, J. van Zon, C. J. Lengner, M. P. Creyghton, A. van Oudenaarden, and R. Jaenisch, “Direct cell reprogramming is a stochastic process amenable to acceleration,” Nature, vol. 462, pp. 595–601, Dec. 2009. [108] D. E. Powers, J. R. Millman, S. Bonner-Weir, M. J. Rappel, and C. K. Colton, “Accurate Control of Oxygen Level in Cells During Culture on Silicone Rubber Membranes with Application to Stem Cell Differentiation,” Biotechnology Progress, vol. 26, no. 3, pp. 805–818, 2009. [109] M. C. Simon and B. Keith, “The role of oxygen availability in embryonic development and stem cell function,” Nature Reviews. Molecular Cell Biology, vol. 9, no. 4, pp. 285–296, Apr. 2008. [110] T.-S. Kim, S. Misumi, C.-G. Jung, T. Masuda, Y. Isobe, F. Furuyama, H. Nishino, and H. Hida, “Increase in Dopaminergic Neurons From Mouse Embryonic Stem Cell-Derived Neural Progenitor/Stem Cells Is Mediated by Hypoxia Inducible Factor-1alpha,” Journal of Neuroscience Research, vol. 86, pp. 2353–2362, Aug. 2008. [111] P. Mondragon-Teran, G. J. Lye, and F. S. Veraitch, “Lowering Oxygen Tension Enhances the Differentiation of Mouse Embryonic Stem Cells into Neuronal Cells,” Biotechnology Progress, vol. 25, no. 5, pp. 1480–1488, 2009. [112] C. Bauwens, T. Yin, S. Dang, R. Peerani, and P. W. Zandstra, “Development of a Perfusion Fed Bioreactor for Embryonic Stem Cell-Derived Cardiomyocyte Generation: Oxygen-Mediated Enhancement of Cardiomyocyte Output.,” Biotechnology and Bioengineering, vol. 90, no. 4, pp. 452–461, May 2005. [113] K. A. Purpura, S. H. L. George, S. M. Dang, K. Choi, A. Nagy, and P. W. Zandstra, “Soluble Flt-1 Regulates Flk-1 Activation to Control Hematopoietic and Endothelial Development in an Oxygen-Responsive Manner,” Stem Cells, vol. 26, no. 11, pp. 2832–2842, Nov. 2008. [114] S. Niebruegge, C. L. Bauwens, R. Peerani, N. Thavandiran, S. Masse, E. Sevaptisidis, K. Nanthakumar, K. Woodhouse, M. Husain, E. Kumacheva, and P. 143 W. Zandstra, “Generation of Human Embryonic Stem Cell-Derived Mesoderm and Cardiac Cells Using Size-Specified Aggregates in an Oxygen-Controlled Bioreactor,” Biotechnology and Bioengineering, vol. 102, no. 2, pp. 493–507, Feb. 2009. [115] E. J. Koay and K. A. Athanasiou, “Hypoxic chondrogenic differentiation of human embryonic stem cells enhances cartilage protein synthesis and biomechanical functionality,” Osteoarthritis and Cartilage, vol. 16, pp. 1450–1456, Dec. 2008. [116] C. A. V Rodrigues, M. M. Diogo, C. L. da Silva, and J. M. S. Cabral, “Hypoxia Enhances Proliferation of Mouse Embryonic Stem Cell-Derived Neural Stem Cells,” Biotechnology and Bioengineering, vol. 106, no. 2, pp. 260–270, Jun. 2010. [117] J. R. Millman, D. E. Powers, R. E. Yu, J. Paul, G. M. Dula, B. Chevva, J. Paulo, M. M. Almeida, J. H. Tan, and C. K. Colton, “Differentiation of pluripotent stem cells under low oxygen increases generation of cardiomyocytes and influences differentiation to ectoderm, mesoderm, and endoderm,” (in preparation). [118] X. Yuan, W. Li, and S. Ding, “Small Molecules in Cellular Reprogramming and Differentiation,” in Epigenetics and Disease, S. M. Gasser and E. Li, Eds. Basel: Springer Basel, 2011, pp. 253–266. [119] J. Bain, L. Plater, M. Elliott, N. Shpiro, C. J. Hastie, H. McLauchlan, I. Klevernic, J. S. C. Arthur, D. R. Alessi, and P. Cohen, “The selectivity of protein kinase inhibitors: a further update,” The Biochemical Journal, vol. 408, no. 3, pp. 297– 315, Dec. 2007. [120] D. B. Ring, K. W. Johnson, E. J. Henriksen, J. M. Nuss, D. Goff, T. R. Kinnick, S. T. Ma, J. W. Reeder, I. Samuels, T. Slabiak, A. S. Wagman, M.-E. W. Hammond, and S. D. Harrison, “Selective Glycogen Synthase Kinase 3 Inhibitors Potentiate Insulin Activation of Glucose Transport and Utilization In Vitro and In Vivo,” Diabetes, vol. 52, pp. 588–595, 2003. [121] J. Silva, O. Barrandon, J. Nichols, J. Kawaguchi, T. W. Theunissen, and A. Smith, “Promotion of Reprogramming to Ground State Pluripotency by Signal Inhibition,” PLoS biology, vol. 6, no. 10, pp. 2237–2247, Oct. 2008. [122] S. F. Gilbert, Developmental Biology, Eighth Edi. Sinauer Associates Inc., 2006. [123] B. Greber, G. Wu, C. Bernemann, J. Y. Joo, D. W. Han, K. Ko, N. Tapia, D. Sabour, J. Sterneckert, P. Tesar, and H. R. Schöler, “Conserved and Divergent Roles of FGF Signaling in Mouse Epiblast Stem Cells and Human Embryonic Stem Cells,” Cell Stem Cell, vol. 6, pp. 215–226, Mar. 2010. [124] J. Hanna, A. W. Cheng, K. Saha, J. Kim, C. J. Lengner, F. Soldner, J. P. Cassady, J. Muffat, B. W. Carey, and R. Jaenisch, “Human embryonic stem cells with 144 biological and epigenetic characteristics similar to those of mouse ESCs,” Proceedings of the National Academy of Sciences of the United States of America, vol. 107, no. 20, pp. 9222–9227, 2010. [125] H. Tsutsui, B. Valamehr, A. Hindoyan, R. Qiao, X. Ding, S. Guo, O. N. Witte, X. Liu, C.-M. Ho, and H. Wu, “An optimized small molecule inhibitor cocktail supports long-term maintenance of human embryonic stem cells,” Nature communications, vol. 2, no. 167, pp. 1–8, Jan. 2011. [126] L. L. Rubin and K. M. Haston, “Stem cell biology and drug discovery,” BMC Biology, vol. 9, no. 42, Jan. 2011. [127] C. E. Murry and G. Keller, “Differentiation of Embryonic Stem Cells to Clinically Relevant Populations: Lessons from Embryonic Development,” Cell, vol. 132, pp. 661–680, 2008. [128] V. Camara-Clayette, F. Le Pesteur, W. Vainchenker, and F. Sainteny, “Quantitative Oct4 Overproduction in Mouse Embryonic Stem Cells Results in Prolonged Mesoderm Commitment During Hematopoietic Differentiation In Vitro,” Stem Cells, vol. 24, pp. 1937–1945, 2006. [129] S. Stefanovic, N. Abboud, S. Désilets, D. Nury, C. Cowan, and M. Pucéat, “Interplay of Oct4 with Sox2 and Sox17: a molecular switch from stem cell pluripotency to specifying a cardiac fate,” The Journal of Cell Biology, vol. 186, no. 5, pp. 665–673, Sep. 2009. [130] Y. Il Yeom, G. Fuhrmann, C. E. Ovitt, A. Brehm, K. Ohbo, M. Gross, K. Hubner, and H. R. Scholer, “Germline regulatory element of Oct-4 specific for the totipotent cycle of embryonal cells,” Development, vol. 122, no. 3, pp. 881–894, 1996. [131] J. H. Hanna, K. Saha, and R. Jaenisch, “Pluripotency and Cellular Reprogramming: Facts, Hypotheses, Unresolved Issues,” Cell, vol. 143, pp. 508– 525, Nov. 2010. [132] Q. Zhou, H. Chipperfield, D. A. Melton, and W. H. Wong, “A gene regulatory network in mouse embryonic stem cells,” Proceedings of the National Academy of Sciences of the United States of America, vol. 104, no. 42, pp. 16438–16443, Oct. 2007. [133] J. V Watson, S. H. Chambers, and P. J. Smith, “A Pragmatic Approach to the Analysis of DNA Histograms With a Definable G1 Peak,” Cytometry, vol. 8, no. 1, pp. 1–8, 1987. [134] E. Stead, J. White, R. Faast, S. Conn, S. Goldstone, J. Rathjen, U. Dhingra, P. Rathjen, D. Walker, and S. Dalton, “Pluripotent cell division cycles are driven by 145 ectopic Cdk2, cyclin A/E and E2F activities,” Oncogene, vol. 21, no. 54, pp. 8320– 8333, Nov. 2002. [135] D. Strumpf, C.-A. Mao, Y. Yamanaka, A. Ralston, K. Chawengsaksophak, F. Beck, and J. Rossant, “Cdx2 is required for correct cell fate specification and differentiation of trophectoderm in the mouse blastocyst,” Development, vol. 132, no. 9, pp. 2093–2102, May 2005. [136] M. Kanai-Azuma, Y. Kanai, J. M. Gad, Y. Tajima, C. Taya, M. Kurohmaru, Y. Sanai, H. Yonekawa, K. Yazaki, P. P. L. Tam, and Y. Hayashi, “Depletion of definitive gut endoderm in Sox17-null mutant mice,” Development, vol. 129, pp. 2367–2379, May 2002. [137] U. Lendahl, L. B. Zimmerman, and R. D. G. McKay, “CNS Stem Cells Express a New Class of Intermediate Filament Protein,” Cell, vol. 60, pp. 585–595, Mar. 1990. [138] D. Bader, T. Masaki, and D. A. Fischman, “Immunochemical Analysis of Myosin Heavy Chain during Avian Myogenesis In Vivo and In Vitro.,” The Journal of Cell Biology, vol. 95, no. 3, pp. 763–770, Dec. 1982. [139] J. Xavier-Neto, C. M. Neville, M. D. Shapiro, L. Houghton, G. F. Wang, W. Nikovits, F. E. Stockdale, and N. Rosenthal, “A retinoic acid-inducible transgenic marker of sino-atrial development in the mouse heart,” Development, vol. 126, no. 12, pp. 2677–2687, Jun. 1999. [140] J. R. Millman, “Effects of Low Oxygen Culture on Pluripotent Stem Cell Differentiation and Tumor Formation,” Massachusetts Institute of Technology, 2011. [141] J. R. Millman, R. E. Yu, G. M. Dula, J. Paul, B. Chevva, J. H. Tan, and C. K. Colton, “Extended low oxygen culture of pluripotent stem cells reduces the fraction of residual tumor-forming cells in differentiated populations,” (in preparation). [142] D. A. Sassoon, I. Garner, and M. Buckingham, “Transcripts of alpha-cardiac and alpha-skeletal actins are early markers for myogenesis in the mouse embryo,” Development, vol. 104, no. 1, pp. 155–164, Sep. 1988. [143] Q. Wang, R. S. Reiter, Q.-Q. Huang, J.-P. Jin, and J. J.-C. Lin, “Comparative Studies on the Expression Patterns of Three Troponin T Genes During Mouse Development,” The Anatomical Record, vol. 263, no. 1, pp. 72–84, May 2001. [144] S. Masui, Y. Nakatake, Y. Toyooka, D. Shimosato, R. Yagi, K. Takahashi, H. Okochi, A. Okuda, R. Matoba, A. A. Sharov, M. S. H. Ko, and H. Niwa, 146 “Pluripotency governed by Sox2 via regulation of Oct3/4 expression in mouse embryonic stem cells,” Nature Cell Biology, vol. 9, no. 6, pp. 625–635, Jun. 2007. [145] A. Raj and A. van Oudenaarden, “Nature, Nurture, or Chance: Stochastic Gene Expression and Its Consequences,” Cell, vol. 135, no. 2, pp. 216–226, 2008. [146] A. Peixoto, M. Monteiro, B. Rocha, and H. Veiga-Fernandes, “Quantification of Multiple Gene Expression in Individual Cells,” Genome Research, vol. 14, no. 10A, pp. 1938–1947, Oct. 2004. [147] V. Tropepe, S. Hitoshi, C. Sirard, T. W. Mak, J. Rossant, and D. van der Kooy, “Direct Neural Fate Specification from Embryonic Stem Cells: A Primitive Mammalian Neural Stem Cell Stage Acquired through a Default Mechanism,” Neuron, vol. 30, no. 1, pp. 65–78, Apr. 2001. [148] S. Labeit and B. Kolmerer, “Titins: Giant Proteins in Charge of Muscle Ultrastructure and Elasticity,” Science, vol. 270, no. 5234, pp. 293–296, Oct. 1995. [149] J. Silva, J. Nichols, T. W. Theunissen, G. Guo, A. L. van Oosten, O. Barrandon, J. Wray, S. Yamanaka, I. Chambers, and A. Smith, “Nanog is the Gateway to the Pluripotent Ground State,” Cell, vol. 138, no. 4, pp. 722–737, Aug. 2009. [150] M. Saitou, S. C. Barton, and M. A. Surani, “A molecular programme for the specification of germ cell fate in mice,” Nature, vol. 418, no. 6895, pp. 293–300, Jul. 2002. [151] Y. Ohinata, B. Payer, D. O’Carroll, K. Ancelin, Y. Ono, M. Sano, S. C. Barton, T. Obukhanych, M. Nussenzweig, A. Tarakhovsky, M. Saitou, and M. A. Surani, “Blimp1 is a critical determinant of the germ cell lineage in mice,” Nature, vol. 436, no. 7048, pp. 207–213, Jul. 2005. [152] O. Haub and M. Goldfarb, “Expression of the fibroblast growth factor-5 gene in the mouse embryo,” Development, vol. 112, pp. 397–406, Jun. 1991. [153] M. Ekerot, M. P. Stavridis, L. Delavaine, M. P. Mitchell, C. Staples, D. M. Owens, I. D. Keenan, R. J. Dickinson, K. G. Storey, and S. M. Keyse, “Negative-feedback regulation of FGF signalling by DUSP6/MKP-3 is driven by ERK1/2 and mediated by Ets factor binding to a conserved site within the DUSP6/MKP-3 gene promoter,” The Biochemical Journal, vol. 412, no. 2, pp. 287–298, Jun. 2008. [154] F. Lanner and J. Rossant, “The role of FGF/Erk signaling in pluripotent cells,” Development, vol. 137, pp. 3351–3360, Oct. 2010. [155] N. Sato, L. Meijer, L. Skaltsounis, P. Greengard, and A. H. Brivanlou, “Maintenance of pluripotency in human and mouse embryonic stem cells through 147 activation of Wnt signaling by a pharmacological GSK-3-specific inhibitor,” Nature Medicine, vol. 10, no. 1, pp. 55–63, Jan. 2004. [156] H. P. Garnir and F. S. Garnirmonjoie, “Fit of the function y(x) = A [1 − exp(−kx)] to data strongly embedded in noise,” Nuclear Instruments and Methods in Physics Research, vol. 190, pp. 333–336, Dec. 1981. [157] P. J. Tesar, J. G. Chenoweth, F. A. Brook, T. J. Davies, E. P. Evans, D. L. Mack, R. L. Gardner, and R. D. G. McKay, “New cell lines from mouse epiblast share defining features with human embryonic stem cells,” Nature, vol. 448, no. 7150, pp. 196–199, Jul. 2007. [158] J. Livet, T. A. Weissman, H. Kang, R. W. Draft, J. Lu, R. A. Bennis, J. R. Sanes, and J. W. Lichtman, “Transgenic strategies for combinatorial expression of fluorescent proteins in the nervous system,” Nature, vol. 450, pp. 56–62, Nov. 2007. [159] MarketResearch.com, “Market Research Projects Stem Cell Market at $1.3 Billion by 2015,” 2011. [Online]. Available: http://finance.yahoo.com/news/marketresearch-projects-stem-cell-120800551.html. [Accessed: 01-May-2013]. [160] A. Son, “IBISWorld Industry Report 42349: Laboratory Supply Wholesaling in the US,” 2012. [161] D. Smith, “Commercialization challenges associated with induced pluripotent stem cell-based products,” Regenerative Medicine, vol. 5, no. 4, pp. 593–603, Jul. 2010. [162] H. McCusker, “Stem Cell Research Tools, Services and Reagents,” Scimitar Equity Blog, 2010. [Online]. Available: http://www.scimitarequity.com/blog/2010/04/19/stem-cell-research-tools-servicesand-reagents/. [Accessed: 28-Apr-2013]. [163] S. Webb, “The gold rush for induced pluripotent stem cells,” Nature Biotechnology, vol. 27, no. 11, pp. 977–9, Nov. 2009. [164] “Guidance for Human Somatic Cell Therapy and Gene Therapy,” 1998. [165] R. McKernan, J. McNeish, and D. Smith, “Pharma’s Developing Interest in Stem Cells,” Cell Stem Cell, vol. 6, no. 6, pp. 517–20, Jun. 2010. [166] M. Csete, “Translational prospects for human induced pluripotent stem cells,” Regenerative Medicine, vol. 5, no. 4, pp. 509–19, Jul. 2010. [167] M. Grskovic, A. Javaherian, B. Strulovici, and G. Q. Daley, “Induced pluripotent stem cells--opportunities for disease modelling and drug discovery,” Nature Reviews. Drug Discovery, vol. 10, no. 12, pp. 915–29, Dec. 2011. 148 [168] S. Webb, “Burgeoning stem cell product market lures major suppliers,” Nature Biotechnology, vol. 28, no. 6, pp. 535–6, Jun. 2010. [169] D. Lau, U. Ogbogu, B. Taylor, T. Stafinski, D. Menon, and T. Caulfield, “Stem Cell Clinics Online: The Direct-to-Consumer Portrayal of Stem Cell Medicine,” Cell Stem Cell, vol. 3, no. 6, pp. 591–4, Dec. 2008. [170] D. Brindley and C. Mason, “Human embryonic stem cell therapy in the post-Geron era,” Regenerative Medicine, vol. 7, no. 1, pp. 17–8, Jan. 2012. [171] L. B. Giebel, “Stem cells - a hard sell to investors,” Nature Biotechnology, vol. 23, no. 7, pp. 798–800, Jul. 2005. [172] T. N. McAllister, N. Dusserre, M. Maruszewski, and N. L’heureux, “Cell-based therapeutics from an economic perspective: primed for a commercial success or a research sinkhole?,” Regenerative Medicine, vol. 3, no. 6, pp. 925–37, Nov. 2008. [173] C. Mason, D. a Brindley, E. J. Culme-Seymour, and N. L. Davie, “Cell therapy industry: billion dollar global business with unlimited potential,” Regenerative Medicine, vol. 6, no. 3, pp. 265–72, May 2011. [174] “Preclinical Assessment of Investigational Cellular and Gene Therapy Products,” 2012. [175] J. P. Siegel, “Human Cells or Tissues Intended for Transplant into a Human Recipient that Have Ex-vivo Contact with Live Non-human Animal Cells, Tissues or Organs Letter,” 2002. [Online]. Available: http://www.fda.gov/BiologicsBloodVaccines/SafetyAvailability/ucm136703.htm. [Accessed: 28-Apr-2013]. [176] D. G. Halme and D. A. Kessler, “FDA Regulation of Stem-Cell-Based Therapies,” The New England Journal of Medicine, vol. 255, no. 16, pp. 1730–1735, 2006. [177] C. E. P. Goldring, P. a Duffy, N. Benvenisty, P. W. Andrews, U. Ben-David, R. Eakins, N. French, N. a Hanley, L. Kelly, N. R. Kitteringham, J. Kurth, D. Ladenheim, H. Laverty, J. McBlane, G. Narayanan, S. Patel, J. Reinhardt, A. Rossi, M. Sharpe, and B. K. Park, “Assessing the Safety of Stem Cell Therapeutics,” Cell Stem Cell, vol. 8, no. 6, pp. 618–28, Jun. 2011. [178] U. Ben-David and N. Benvenisty, “The tumorigenicity of human embryonic and induced pluripotent stem cells,” Nature Reviews Cancer, vol. 11, no. 4, pp. 268– 77, Apr. 2011. [179] C. A. Sommer and G. Mostoslavsky, “The Evolving Field of Induced Pluripotency: Recent Progress and Future Challenges,” Journal of Cellular Physiology, vol. 228, no. 2, pp. 267–75, Feb. 2013. 149 [180] K. Okita, N. Nagata, and S. Yamanaka, “Immunogenicity of Induced Pluripotent Stem Cells,” Circulation Research, vol. 109, no. 7, pp. 720–1, Sep. 2011. [181] P. Guha, J. W. Morgan, G. Mostoslavsky, N. P. Rodrigues, and A. S. Boyd, “Lack of Immune Response to Differentiated Cells Derived from Syngeneic Induced Pluripotent Stem Cells,” Cell Stem Cell, vol. 12, no. 4, pp. 407–412, Jan. 2013. [182] K. Sheridan, “Rush for patents is choking US stem cell research,” Agence FrancePresse, Washington, 25-Jan-2011. [183] M. Lutticke, “Germany restricts stem cell patents,” 2012. [Online]. Available: http://dw.de/p/16quR. [Accessed: 23-Apr-2013]. [184] B. P. Georgieva and J. M. Love, “Human induced pluripotent stem cells: a review of the US patent landscape,” Regenerative Medicine, vol. 5, no. 4, pp. 581–91, Jul. 2010. [185] K. T. Vrtovec and C. T. Scott, “Patenting pluripotence: the next battle for stem cell intellectual property,” Nature Biotechnology, vol. 26, no. 4, pp. 393–5, Apr. 2008. [186] G. Bahadur and M. Morrison, “Patenting human pluripotent cells: balancing commercial, academic and ethical interests,” Human Reproduction, vol. 25, no. 1, pp. 14–21, Jan. 2010. [187] D. S. Couto, L. Perez-Breva, and C. L. Cooney, “Regenerative Medicine: Learning from Past Examples,” Tissue Engineering: Part A, vol. 18, no. 21–22, pp. 2386– 93, Nov. 2012. [188] STEMGEN, “Stem Cell World Map,” International Database on the Legal and Socio-Ethical Issues in Stem Cell Research. [Online]. Available: http://www.stemgen.org/stem-cell-world-map. [Accessed: 28-Apr-2013]. [189] A. D. Levine, “Policy Uncertainty and the Conduct of Stem Cell Research,” Cell Stem Cell, vol. 8, no. 2, pp. 132–5, Feb. 2011. [190] J. L. Furman, F. Murray, and S. Stern, “Growing Stem Cells: The Impact of Federal Funding Policy on the U.S. Scientific Frontier,” Journal of Policy Analysis and Management, vol. 31, no. 3, pp. 661–705, Jun. 2012. [191] T. Caulfield, C. Rachul, and A. Zarzeczny, “The Evolution of Policy Issues in Stem Cell Research: An International Survey,” Stem cell Reviews and Reports, vol. 8, no. 4, pp. 1037–42, Dec. 2012. [192] G. Q. Daley, “Stem cells : roadmap to the clinic,” The Journal of Clinical Investigation, vol. 120, no. 1, pp. 8–10, 2010. 150 [193] T. Bubela, M. D. Li, M. Hafez, M. Bieber, and H. Atkins, “Is belief larger than fact: expectations, optimism and reality for translational stem cell research,” BMC Medicine, vol. 10, no. 1, p. 133, Jan. 2012. [194] P. Krzyzanowski, “BioTime: A new Geron, without a decade of baggage,” Signals Blog, 2013. [Online]. Available: http://www.signalsblog.ca/biotime-a-new-geronwithout-a-decade-of-baggage/. [Accessed: 26-Apr-2013]. 151