Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

New approaches to the optimized maintenance of embryonic stem cells in culture Glover, Clive Hamilton 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2007-267226.pdf [ 24.04MB ]
Metadata
JSON: 831-1.0100378.json
JSON-LD: 831-1.0100378-ld.json
RDF/XML (Pretty): 831-1.0100378-rdf.xml
RDF/JSON: 831-1.0100378-rdf.json
Turtle: 831-1.0100378-turtle.txt
N-Triples: 831-1.0100378-rdf-ntriples.txt
Original Record: 831-1.0100378-source.json
Full Text
831-1.0100378-fulltext.txt
Citation
831-1.0100378.ris

Full Text

NEW APPROACHES TO THE OPTIMIZED MAINTENANCE OF EMBRYONIC STEM CELLS IN CULTURE by Clive Hamilton Glover B . S c , M c G i l l University 1999 M.Sc . , University of British Columbia 2001 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF D O C T O R OF P H I L O S O P H Y in The Faculty O f Graduate Studies (Genetics) T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A March 2007 © Clive Hamilton Glover, 2007 Abstract The realization of many stem cell-based therapies w i l l rely on the development of methods for the expansion and controlled differentiation of stem cells and the development of assays to rapidly detect the results of culture manipulation. Mouse embryonic stem cells (ESC) provide a relatively abundant and high purity model system to investigate environmental cues that influence stem cell fate. The dynamics of loss of pluripotency following the removal of leukemia inhibitory factor (LIF) were investigated using three functional assays (chimeric mouse formation, embryoid body generation and colony forming ability). A rapid loss (>70%) of pluripotent cells was detected within 24 hours, with very low residual activity by all criteria within 72 hours. Surprisingly, functional endpoints of pluripotency correlated poorly with two commonly used markers, expression of Oct4 and S S E A - 1 . The embryoid body assay was then used in factorial and central composite design experiments to define optimized levels of ascorbic acid ( A A ) , chondroitin sulphate and PD98059 (PD), three factors that enhance undifferentiated E S C maintenance. These experiments identified an unexpected negative interaction between A A and P D and allowed the development of a cocktail of these three factors that increased E S C yield 3-fold over untreated cultures. To identify more rapid endpoints for the detection of undifferentiated E S C , global gene expression profiling was performed on E S C induced to differentiate by removal of L I F , addition of retinoic acid and addition of D M S O . Gene expression profiles were determined using the Affymetrix platform and a new meta-analysis methodology was developed and applied to the resulting gene expression data sets. This analysis revealed that the expression of the stem cell factor receptor, c-kit, expressed on the cell surface, ii correlated with the frequency o f undifferentiated E S C making it an ideal marker for further studies. Furthermore, the expression changes of seven genes - 103728_at, 8430410A17Rik, Mf2, nrObl, sox2, tell and zfp42 - was able to predict undifferentiated E S C frequencies in both differentiation and maintenance cultures. These experiments should provide a useful framework for making further improvements in the culture of undifferentiated E S C . in Table of Contents Abstract i i Table of Contents iv List of Tables ix List of Figures x i List of Abbreviations x i i i Acknowledgements xiv Co-Authorship Statement xv 1. Introduction 1 1.1. Embryonic Stem Cells 2 1.1.1. Differences Between Mouse and Human Embryonic Stem Cells 5 1.1.2. Advantages of Mouse Embryonic Stem Cells ; 6 1.2. Culture and Characterization of Embryonic Stem Cells 7 1.2.1. Culture Conditions for the Propagation of Undifferentiated Embryonic Stem Cells 7 1.2.2. Molecular Control of Self-Renewal 9 1.2.3. Differentiation of Mouse Embryonic Stem Cells 11 1.2.4. Functional Assays to Detect Undifferentiated Embryonic Stem Cells 12 1.2.5. Phenotypic Markers of Undifferentiated Embryonic Stem Cells 14 1.2.6. Characterization of Mouse Embryonic Stem Cel l Responses 15 iv 1.3. Challenges in Embryonic Stem Cel l Culture '. 16 1.3.1. High Throughput Screening to Identify Factors Supporting the Maintenance of Embryonic Stem Cells 16 1.3.2. Statistical Design of Experiments to Understand Factor Interactions and Determine Non-Linear Responses 19 1.3.3. Culture Systems Suitable for High Throughput Screening 20 1.4. Thesis Objectives 21 1.5. References 25 2. Correlation of Mouse Embryonic Stem Cel l Gene Expression Profiles with Functional Measures of Pluripotency 36 2.1. Introduction 36 2.2. Materials and Methods 39 2.2.1. Culture of Mouse Embryonic Stem Cells and Embryonic Fibroblasts 39 2.2.2. Preparation of Embryonic Stem Cells for Differentiation 40 2.2.3. Blastocyst Injection 41 2.2.4. Embryoid Body Formation Assays 42 2.2.5. Colony Forming Cel l Assay 42 2.2.6. R N A Extraction and Array Hybridization 43 2.2.7. Data Analysis 43 2.2.8. Quantitative R T - P C R 44 2.2.9. F low Cytometry 45 2.2.10. Comparison with Previously Published Gene Expression Array Data 46 2.3. Results 47 2.3.1. Pluripotency During Early Differentiation 47 2.3.2. Array Analysis 50 2.3.3. Gene Expression Validation 52 2.3.4. Quantitative R T - P C R validation 55 2.3.5. Functional Classification of Differentially Expressed Genes 56 2.3.6. Molecular Markers of Pluripotency 59 2.3.7. Comparison with Previously Published Gene Expression Array Data 60 2.4. Discussion 64 2.5. References 83 3. Increased Capacity of Mouse Embryonic Stem Cells to Form Embryoid Bodies Following Optimization of Ascorbic A c i d , PD98059 A n d Chondroitin Sulphate Levels 91 3.1. Introduction 91 3.2. Materials and Methods 94 3.2.1. Embryonic Stem Cel l Maintenance Cultures ; 94 3.2.2. Embryonic Stem Cel l Experimental Cultures 94 3.2.3. Design of Experiments 95 3.2.4. F low Cytometry 96 3.2.5. R N A extraction and Quantitative R T - P C R 97 3.3. Results 97 3.3.1. Dose Response of Ascorbic A c i d , Chondroitin Sulphate or PD98059 97 3.3.2. Interactions between Ascorbic A c i d , Chondroitin Sulphate and PD98059 98 vi 3.3.3. Central Composite Design Experiments to Determine Non-Linear Responses....99 3.3.4. Verification of Cocktail to Maximize Expansion of Embryoid Body-Forming Cells 100 3.3.5. Influence of Ascorbic A c i d , Chondroitin Sulphate and PD98059 on Self-Renewal and Apoptosis 101 3.3.6. Differentiation Abi l i ty of Treated Cells 103 3.4. Discussion 104 3.5. References ...119 4. Meta-Analysis of Differentiating Mouse Embryonic Stem Cel l Gene Expression Kinetics Reveals Early Change of a Small Gene Set 121 4.1. Introduction 121 4.2. Materials and Methods 124 4.2.1. Embryonic Stem Cel l Maintenance Cultures 124 4.2.2. Embryonic Stem Cel l Experimental Cultures 124 4.2.3. Embryoid Body and Colony Forming Cel l assays 124 4.2.4. R N A Extraction and Array Hybridization 124 4.2.5. Gene Expression Data Sets 124 4.2.6. Gene Expression Analysis 126 4.2.7. Defining the Embryonic Stem Cel l Signature Change in Terms of Gene Expression 127 4.2.8. Confidence Values 129 4.2.9. Comparison of M O E 4 3 0 and M G _ U 7 4 v 2 Chips 130 vii 4.2.10. Gene Ontology 130 4.2.11. Quantitative R T - P C R 131 4.3. Results 131 4.3.1. Functional Assay Analysis 131 4.3.2. Gene Expression Analysis 132 4.3.3. Identification of a Robust Set of Early Gene Expression Changes that Indicate Decreased Frequency of Undifferentiated Embryonic Stem Cells 133 4.3.4. Confidence Values 134 4.3.5. Meta-Analysis via Pareto Optimization 135 4.3.6. Q - R T - P C R Verification of Array Results 136 4.4. Discussion 139 4.5. References 157 5. Conclusions and Future Directions 163 5.1. Conclusions 163 5.2. Future Directions 167 5.3. References 170 Appendix A 172 Appendix B 174 Appendix C 201 vi i i List of Tables Table 1.1 Similarities and differences between human and mouse embryonic stem cells 23 Table 2.1 Expression change of genes previously reported as enriched in undifferentiated embryonic stem cells or to be markers of embryonic stem cell differentiation 76 Table 2.2 Comparison of relative expression levels obtained by quantitative R T - P C R in the R l , J1, and E F C embryonic stem cell lines during differentiation 77 Table 2.3 Genes decreased during embryonic stem cell differentiation and most strongly correlated with the loss of pluripotency 78 Table 2.4 Differentially expressed genes with plasma membrane localization 80 Table 2.5 Summary of published gene expression studies in mouse and human embryonic stem cells used for comparison 82 Table 3.1 Design matrix for a two-level, three-factor factorial design experiment 116 Table 3.2 Design matrix for a two-factor central composite design experiment 117 Table 3.3 Factor concentrations used in the factorial design and central composite design experiments 118 Table 4.1 Summary of all microarray experiments used in this study 150 Table 4.2 Thresholds used in the definition of the embryonic stem cell signature change 151 Table 4.3 Genes identified on the first five Pareto fronts 152 Table A 1 Primers used for quantitative R T - P C R throughout this study 172 Table B l Genes decreased between 0 and 18 hours after leukemia inhibitory factor removal 174 ix Table B2 Genes increased between 0 and 18 hours after leukemia inhibitory factor removal 176 Table B3 Genes decreased between 18 and 72 hours after leukemia inhibitory factor removal 177 Table B4 Genes increased between 18 and 72 hours after leukemia inhibitory factor removal 181 Table B5 Genes decreased between 0 and 72 hours after leukemia inhibitory factor removal 183 Table B6 Genes increased between 0 and 72 hours after leukemia inhibitory factor removal 188 Table B7 Comparison of changed genes to previously published data sets 192 Table C I Function of the 88 genes on the first five Pareto fronts 201 Table C2 Comparison of genes on the first five Pareto Fronts with previously published data sets 202 x List of Figures Figure 1.1 Graphical representation of factorial and central composite design experiments for 3 factors 24 Figure 2.1 Embryonic stem cell morphology during leukemia inhibitory factor removal 68 Figure 2.2 Assays of embryonic stem cell pluripotency following removal of leukemia inhibitory factor 69 Figure 2.3 Hierarchical clustering of array samples 70 Figure 2.4 Correlation between array and quantitative R T - P C R results for multiple genes in multiple cells lines 71 Figure 2.5 Annotation of differentially expressed genes 72 Figure 2.6 Flow cytometry profile to test for cell surface expression of c-kit following removal of leukemia inhibitory factor 74 Figure 2.7 Correlation between c-kit expression and embryoid body formation ability 75 Figure 3.1 Dose response of embryonic stem cells to ascorbic acid 109 Figure 3.2 Dose response of embryonic stem cells to PD98059 110 Figure 3.3 Parameter estimates derived from factorial design experiments I l l Figure 3.4 Parameter estimates describing yield of embryoid body-forming cells in two central composite design experiments 112 Figure 3.5 Verification of factor combination that maximizes embryoid body-forming cell output compared to other factor combinations and untreated cells 113 xi Figure 3.6 Effect of ascorbic acid, chondroitin sulphate and PD98059 on self-renewal and apoptosis 114 Figure 3.7 Affect of differentiation by either leukemia inhibitory factor removal or retinoic acid addition on lineage marker genes of ascorbic acid, chondroitin sulphate and PD98059-treated cells 115 Figure 4.1 The effect of leukemia inhibitory factor removal with or without addition of D M S O or retinoic acid on the maintenance and differentiation of embryonic stem cells 145 Figure 4.2 Transcriptome plots of estimated expression changes, based on fitting models to each data set 146 Figure 4.3 Comparison of differently measured changes in gene expression within and between two embryonic stem cell lines grown for 0, 24,72 and 96 hours in +LIF±RA or - L I F ± D M S O 147 Figure 4.4 Quantitative R T - P C R profiles of transcript levels for seven genes that showed rapid decrease in embryonic stem cells subjected to several differentiating conditions 148 Figure 4.5 Comparison of biological and molecular changes in embryonic stem cells stimulated to differentiate by exposure to ascorbic acid 149 xi i List of Abbreviations A A ascorbic acid A P alkaline phosphatase C C D central composite design C F C colony forming cell CoVar, coefficient of variation CS chondroitin sulphate C V confidence value D M S O dimethylsulfoxide D M E M Dulbecco's Modified Eagles Medium E B embryoid body E S C embryonic stem cell F D factorial design F B S fetal bovine serum F S C forward scatter G F P green fluorescent protein H S C hematopoietic stem cell H F Hanks Buffered Saline Solution + 2% F B S I C M inner cell mass I M D M Iscove's Modified Dulbecco's Medium LIF leukemia inhibitory factor M A S MicroArray Suite M I A M E minimal information about a mieroarray experiment M - L R multiple cell line L IF removal M T G monothioglycerol PBS phosphate buffered saline P B S F B S phosphate buffered saline + 2% F B S P D PD98059 P E phycoerythrin P E F primary embryonic fibroblast PF Pareto front P F A Pareto front analysis Q - R T - P C R quantitative real time P C R R l - L R R l L IF removal R A retinoic acid SCF stem cell factor SSC side scatter T /E Tryps in -EDTA xii i Acknowledgements I would like to acknowledge my supervisors Dr. James Piret and Dr. Connie Eaves for their support and guidance through this thesis. Without their energy and enthusiasm this thesis would not have been completed. ' Thanks are also due to the many collaborators that I have worked with during the course of this thesis, in particular Dr. Jenny Bryan and Dr. Cheryl Helgason. M y education is the better for having the opportunity to work with both of them. I would also like to thank Dr. Arshad Chaudhry, Dr. Lars Palmqvist, Dr. M i n L u , Lien Hsu, Rewa Grewal, Bolette Bossen, Dr. Keith Humphries, Michael Marin , Dr. Nicolas Caron, Dr. Fabio Rossi and Dr Leah Keshet for advice and support at various stages during this thesis. Finally Dave Thomson has spent many evenings and weekends as a "science widower." Words cannot express how thankful I am for all the love and support that you have given me and the patience that you have shown over the course of this work. This work was funded by the Natural Sciences and Engineering Research Counci l , Canadian Institutes of Health Research, Michael Smith Foundation for Health Research, Genome Canada, the Stem Cel l Network and the Mathematics of Information Technology and Complex Systems Network. I personally have been supported by studentships from the Stem Cel l Network and the Canadian Institutes of Health Research. xiv Co-Authorship Statement Chapter 2 was co-authored with Lars Palmqvist, Bolette Bossen, L ien Hsu, M i n L u , James Piret, Keith Humphries and Cheryl Helgason. I conceived, designed and performed the experiments leading to Figures 2.1, 2.2, 2.6 and 2.7, performed the analysis to generate Table 2.5 and assisted in analysis of mieroarray data and preparation of the manuscript. Chapter 3 is co-authored with Connie Eaves and James Piret. I conceived, designed and performed all experiments and prepared the manuscript with assistance from both co-authors. Chapter 4 is co-authored with Michael Marin, Connie Eaves, Cheryl Helgason, James Piret and Jenny Bryan. I conceived, designed and performed all the experiments and developed and implemented the analysis techniques in collaboration with Michael Mar in and Jenny Bryan. I prepared the manuscript in collaboration with Jenny Bryan with input from the other authors. xv 1. Introduction Stem cells are defined by their ability to differentiate into multiple, specialized cell types, the function of which depends on the tissue of origin of the stem cell , and to self-renew this property. In vivo, stem cells are now recognized as being responsible for the generation of many tissues and organs during development and for their maintenance and repair into adulthood. To date, stem cells have been isolated from the early blastocyst (Evans and Kaufman 1981; Kunath et al. 2005; Mart in 1981; Tanaka et al. 1998; Thomson et al . 1998), several adult tissues including hematopoietic (T i l l and M c C u l l o c h 1961), central nervous system (Reynolds and Weiss 1992), the epidermis (Lavker and Sun 2000), the retina (Tropepe et al. 2000) and the breast (Shackleton et al. 2006; Stingl et al. 2006); as wel l as several cancers such as acute myeloid leukemia (Bonnet and Dick 1997), breast cancer (Al-Hajj et al. 2003) and brain tumours (Singh et al. 2003). Because of their regenerative potential, normal stem cells have been proposed as sources of cells useful for the replacement of defective tissue which otherwise does not have a renewal capacity (Zandstra and Nagy 2001). Application of clinical-scale stem cell based therapies requires a thorough understanding of the environmental clues that determine stem cell fate choices (i.e. proliferation or quiescence, self-renewal or differentiation and survival or apoptosis). Understanding the ways in which the cell integrates the multiple, and possibly conflicting, external signals that they receive w i l l help determine protocols for the expansion of undifferentiated cells or for the differentiation of cells towards specific clinically-relevant lineages. 1 1 . 1 . Embryonic Stem Cells In mice, the blastocyst develops at embryonic day 3.5 and contains cells from three different lineages. The trophoectoderm surrounds a cavity called the blastocoel, at one end of which lies the inner cell mass ( ICM). The I C M consists of cells from the epiblast and primitive endoderm lineages. Many studies have demonstrated that cells in the epiblast give rise to the embryo proper as wel l as several extraembryonic tissues, while the progeny of trophoectoderm and primitive endoderm are restricted to the trophoblast layers of the placenta and the visceral endoderm respectively (Lu et al. 2001). Mouse embryonic stem cells (ESC) can be derived from the epiblast of preimplantation embryos by culturing either intact blastocysts or immunosurgically-isolated epiblast (Solter and Knowles 1975) on a layer of mitotically inactivated primary embryonic fibroblasts (Abbondanzo et al. 1993). Fol lowing several days of culture, colonies of cells expand, some of which retain an undifferentiated morphology. These undifferentiated colonies can be further passaged until a continuous cell line is established that retains the ability to contribute to all tissues of chimeric mice including the germ cells, when injected back into recipient blastocysts. Subsequent application of similar methods to the human system resulted in the derivation of human E S C (Thomson et al. 1998). While their ability to contribute to all tissues in a developing embryo has not been assessed for ethical reasons, in vitro and in vivo testing indicates that they retain multilineage differentiation capacity (Reubinoff et al. 2000). 2 Techniques have been developed to differentiate E S C into a variety of adult cell types including hematopoietic (Palacios et al. 1995; Wang et al. 2006b), cardiac (Kattman et al. 2006; Moretti et al. 2006; W u et al. 2006), neural (Conti et al. 2005), osteogenic (zur Nieden et al. 2003) and germ cells (Geijsen et al. 2004; Hubner et al. 2003; Nayernia et al. 2006;.Toyooka et al. 2003). Because of this differentiation potential, it is hoped that these cells could form the basis of cellular therapies in many diseases where tissue damage or malfunction is severe and irreversible. While clinical-scale application of E S C based therapies is l ikely several years away, initial reports from animal studies suggest that ESC-derived cells have the ability to contribute to the repair of injury and disease. For example, transplant of ESC-der ived oligodendrocyte progenitor cells into a rat model of spinal chord injury was able to significantly improve remyelenation and locomotor recovery (Keirstead et al. 2005) and it is hoped that human clinical trials based on this result w i l l begin in 2007 (Geron website). Cardiovascular diseases, type-1 diabetes and Parkinson's disease are other examples of diseases where it is hoped that ESC-based therapies w i l l provide a cure. A number of obstacles remain before the clinical-scale application of ESC-based therapies can be realized. A t present, only limited testing of ESC-derived cells has been performed to ensure full functionality of differentiated cells. A l l of these tests have been performed in laboratory animals and have had mixed results with respect to efficacy of treatment (Menard et al. 2005). Furthermore, protocols for the differentiation of E S C to functionally relevant progeny are currently inefficient, resulting in low differentiated cell yields and contamination as a result of aberrant differentiation. O f greater concern is the possibi l i ty o f contamination of 3 undifferentiated E S C in transplanted populations which may form tumors known as teratomas (Bjorklund et al. 2002; Fujikawa et al. 2005). To counteract this, the molecular regulators of the tumorogenic potential of E S C are being explored (Takahashi et al. 2003a) with the hope that increased understanding of this may lead to methods of eliminating this possibility. Further concerns surround the possibility of immune rejection of transplanted cells either due to the expression o f different major histocompatability complex antigens on donor cells (Swijnenburg et al. 2005) or from the expression of foreign antigens as a result of culturing E S C in animal products (Martin et al. 2005). The possibility of the derivation of patient-specific E S C through somatic cell nuclear transfer would eliminate the possibility of immune rejection. However this process is currently very inefficient as well as difficult due to the requirement of a donor human egg (Gurdon et al. 2003). Although the possibility of rejection due to expression of foreign animal antigens is controversial (Cerdan et al. 2006; Mart in et al. 2006), scientists are devoting much effort to determining xeno-free culture conditions for the expansion of E S C (Ellerstrom et al. 2006). Because of their tissue of origin, it is also hypothesized that differentiating E S C recapitulate many aspects o f development and thus can be used as a model system to understand developmental biology. Traditionally, this has been difficult to study in vivo due to the small number of cells available early in development and the ethical questions surrounding scientific access to early embryos. Recently, three groups reported the discovery of a common cardiovascular progenitor which gives rise to cardiomyocytes and endothelial and smooth muscle 4 cells (Kattman et al. 2006; Moretti et al. 2006; W u et al. 2006). A l l three studies initially identified these progenitors in populations of differentiated E S C and used these findings to identify similar cell types in vivo. Conti et al. used mouse E S C to establish conditions for the expansion of a purified population o f neural stem cells derived from E S C (Conti et al. 2005). These conditions were readily applied to the derivation of neural stem cell lines from the fetal mouse and human brains demonstrating the usefulness of this in vitro system for understanding developmental cues present in vivo. 1.1.1. Differences Between Mouse and Human Embryonic Stem Cells Human and mouse E S C share many fundamental characteristics including the ability to proliferate as an undifferentiated, untransformed population for many passages in culture and to differentiate into all germ layers both in culture and upon formation of a teratoma when injected into recipient mice. In addition, they have high telomerase and alkaline phosphatase (AP) activity (Kirschstein and Skirboll 2001). Despite these similarities, several differences do exist. Reports have indicated that human E S C have the ability to differentiate into trophoblast (Odorico et al. 2001; Thomson et al. 1998), a lineage that mouse E S C have never been reported to generate (Niwa et al. 2000; Rossant 2001). Furthermore, while mouse E S C pluripotency is optimally retained by signaling through leukemia inhibitory factor (LIF) /gpl30/STAT3 pathway (Matsuda et al. 1999; N i w a et al. 1998; Smith et al. 1988), concurrent with BMP4/Smad signaling (Ying et al. 2003), S T A T 3 is dispensable for human E S C self-renewal (Sumi et al. 2004). Instead, optimal maintenance of human E S C occurs in the presence of b F G F (Levenstein et al. 2006) and TGFpVactivin/nodal (Beattie et al. 2005; 5 Ludwig et al. 2006; Val l ier et al. 2004) or noggin ( X u et al. 2005). Differences also exist in commonly used cell surface marker expression. Undifferentiated mouse E S C express high levels of the cell surface marker SSEA-1 and low levels of SSEA-3 (Kirschstein and Skirboll 2001). In contrast, human E S C express high levels of SSEA-3 and no SSEA-1 (Draper et al. 2002). Transcriptional analysis of mouse and human E S C has revealed further molecular differences. Ginis et al., studied 400 genes that were expressed by both human and mouse E S C and found that at least one quarter of them were differentially expressed between the two cell types (Ginis et al. 2004). A larger scale comparison revealed only a small set of commonly expressed genes and revealed large differences in expression of genes associated with the L I F , TGF(3, Wnt and F G F signaling pathways further confirming findings of differences in signaling between the two cell types. Interestingly the conserved genes were enriched with those previously identified as being involved in maintenance of E S C (Wei et al. 2005). 1.1.2. Advantages of Mouse Embryonic Stem Cells Mouse E S C provide a well-established system for investigation of stem cell properties due to their ease of culture and rapid growth rate. They can be readily differentiated into multiple cell types providing closely related progeny that have lost their stem cell potential. Furthermore, much is known about the specific factors that govern the self-renewal of these cells. They are also able to grow clonally allowing many biological questions to be asked at the single cell level. Experimental access to their tissue o f origin and the ability to test their pluripotency by blastocyst injection are also relatively easy, thus providing rigorous assays of their stem cell activity. 6 1.2. Culture and Characterization of Embryonic Stem Cells 1.2.1. Culture Conditions for the Propagation of Undifferentiated Embryonic Stem Cells Optimal culture conditions for mouse E S C requires growth on a feeder layer of mitogenically inactivated P E F and medium supplemented with fetal bovine serum (FBS) and L I F . Defined medium for the growth of E S C have been reported in which P E F are replaced with gelatin (Smith et al. 1988) and B M P addition replaces F B S (Ying et al. 2003). It has, however, been noted that growth of E S C on gelatin (Glover et al. 2006) or replacement of serum with a commercially available replacement (Chaudhry 2006) reduces the output of functionally defined E S C indicating that there are more factors that are responsible for their maintenance.' L IF is the primary mediator of self-renewal although it appears to have multiple effects on E S C . In addition to a role in inhibiting differentiation of undifferentiated cells, it also appears to confer a survival and growth advantage over differentiated cells present in the same culture (Viswanathan et al. 2003a). L IF binds to a heterodimeric receptor complex consisting of the L IF receptor and gpl30 (Davis et al. 1993). Binding of L I F causes a cascade of signaling events mediated by associated Janus-associated tyrosine kinases, culminating predominantly in phosphorylation of S T A T 3 (Niwa et a l .T998) and, to a lesser extent, activation of the E R K pathway (Burdon et al. 1999). Interestingly these two pathways have opposite effects on E S C self-renewal, S T A T 3 being necessary and sufficient for self-renewal (Matsuda et al. 1999) and E R K activation being inhibitory to self-renewal (Burdon et al. 1999). In addition to the 7 involvement of these pathways in E S C self-renewal, the P I 3 K pathway is involved in E S C propogation (Burdon et al. 2002). Evidence also suggests the involvement of the Wnt pathway in E S C self-renewal (Ding et al. 2003). In addition to soluble factors, there is also evidence that insoluble factors, such as those contained in the extracellular matrix secreted by P E F , play a role in modulating E S C fate. The extracellular matrix consists of a large variety of proteins, proteoglycans (such as chondroitin sulphate (CS) and heparin sulphate) and hyaluronic acid and can mediate cell fate both through direct interaction with cells as wel l as by sequestering soluble growth factors. Replacement of P E F with collagen in the form of gelatin is possible although, as noted above, self-renewal of E S C is compromised in this setting. Immunocytochemical characterization of mouse embryonic fibroblasts revealed a large amount of collagen IV and laminin as well as vimentin but very low levels of fibronectin. It has been reported that laminin, with addition of P E F conditioned media, can replace the requirement for P E F in human E S C culture for short periods, although P E F are more commonly replaced with matrigel, a complex of multiple extracellular matrix proteins (Xu et al. 2001). Inactivation of (31 integrin extracellular matrix receptor severely inhibited association with P E F (Fassler and Meyer 1995). (31 integrin expression is essential for the growth of teratomas (Bloch et al. 1997) although E S C lacking (31 integrin expression showed equal ability to integrate into the I C M (Fassler and Meyer 1995). Disruption of talin, an intracellular mediators of integrin signaling, leads to impaired interaction response to attachment in undifferentiated cells only (Priddle et al. 1998). Downstream signaling pathways activated by integrin receptors include P I3K and M A P K , the same pathways known to be activated by L I F . 8 Ultimately, cell fate decisions are mediated by the integration of multiple signals stimulated by both soluble and insoluble factors (Prudhomme et al. 2004). 1.2.2. Molecular Control of Self-Renewal Oct4 (Nichols et al. 1998), sox2 (Avi l ion et al. 2003) and nanog (Chambers et al. 2003; Mitsui et al. 2003) are thought to be central to the specification of stem cell identity due to their unique expression patterns and their essential roles in early development. Much effort has been dedicated to understanding the transcriptional state of undifferentiated stem cells using a variety of techniques (Rao and Stice 2004). Earlier studies focused on cataloguing the transcripts that were unique to undifferentiated cells by comparing the genes expressed to those of differentiated progeny (Ivanova et al. 2002; Ramalho-Santos et al. 2002). These initial studies were of limited value due to the large number of genes identified and small amount of agreement between the studies (Fortunel et al. 2003). Subsequent studies investigated more detailed time courses of differentiation (Sharov et al. 2003) or used multiple differentiation pathways to induce loss of pluripotency (Brandenberger et al. 2004; Rao et al. 2004). Overall there has been some consensus on a small set of genes that is expressed in undifferentiated cells (including oct4, nanog and sox2), but many differences exist between the studies indicating the sensitivity of the expression of genes to specific culture conditions used to treat the cells (Skottman et al. 2005). More recently, gene expression studies have focused on finding the genes downstream of oct4, sox2 and nanog (Boyer et al. 2005; L o h et al. 2006). These studies identified binding sites for 9 these transcription factors in many active and inactive genes. Since many of the inactive genes are developmental transcription factors, oct4, nanog and sox2 may be involved in blocking differentiation. Computational analysis o f the network formed by these three transcription factors also supports the hypothesis that they are centrally involved in maintaining pluripotency of E S C (Chickarmane et al. 2006). Further functional analysis of these downstream target genes has also started to identify a series of pathways which are involved in making specific developmental decisions (Boyer et al. 2006). Studies of nanog and its associated proteins have allowed determination of a putative protein interaction network which governs stem cell pluripotency (Wang et al, 2006a). This study identifies a specific interaction between nanog and oct4 as well as connections with several other proteins which may be anticipated from gene expression studies including zfp42 (Rogers et al. 1991), sall4 (Zhang et al. 2006), daxl (Swain et al. 1998) and rifl (Adams and McLaren 2004). In addition, several new.interactions are determined including the involvement of several proteins known to function as transcriptional repressors. Another intriguing study showed that overexpression of 24 carefully selected genes was able to recover a ESC- l ike state in fibroblasts. Upon subsequent investigation, the authors determined that only 4 of these factors - oct4, sox2, c-myc and klf4 - were absolutely required to produce this effect. Interestingly, nanog was not required indicating that it acts downstream of the required factors (Takahashi and Yamanaka 2006). 10 1.2.3. Differentiation of Mouse Embryonic Stem Cells Differentiation of mouse E S C can be easily induced by removal of L I F and P E F either when cultured in a monolayer or by aggregation in suspension culture where the cells form a heterogeneous mix of differentiated cells called an embryoid body (EB, Doetschman et al. 1985). In either of these formats, a variety of differentiated cell types develop that express specific lineage markers. It is possible to bias the type of differentiated cell formed by adding certain factors, for example retinoic acid (Rohwedel et al. 1999), however, inevitably a mixture of differentiated cell types w i l l be formed. Furthermore, although ESC-derived cell types express key lineage markers for a variety of cell types, true functional ability of these cells remain to be established for many lineages (Smith 2001). Ultimately, efficient generation of differentiated cell types requires knowledge of the molecular cues that drive formation of specific lineages. For example, constitutive activation of the notch pathway forces E S C towards a neural fate (Lowell et al. 2006). However, at present, little is known about these early cues and many protocols instead rely on the enrichment of desired cell types by selection. One selection strategy involves coupling lineage-specific markers to either a drug resistance or fluorescence gene and then consequent selection by the appropriate method. For example, cardiac cells can be selected from a mixed population of differentiated cells by incorporation of a drug resistance cassette coupled to a cardiac lineage-specific marker (Zandstra et al. 2003). Alternatively a survival advantage can be conferred upon desired cell types with appropriate culture conditions. For example, hematopoietic cells can readily be generated by culture in semi-solid with appropriate growth factors (Wiles and Keller 1991). 11 Several factors have been identified which promote differentiation towards specific cell types although at present it is not known whether these work by driving differentiation towards a specific lineage or by allowing the survival o f the differentiated cells! D M S O treatment promotes differentiation towards skeletal myoblasts (Dinsmore et al. 1996). Retinoic acid (RA) causes differentiation primarily towards neural cells (Dinsmore et al. 1996) although other lineages have also been reported (Dani et al. 1997). Ascorbic acid ( A A ) was identified in a screen of small molecules which caused differentiation towards cardiomyocytes (Takahashi et al. 2003b) although it has also been reported to enhance differentiation towards several lineages in conjunction with other compounds (Lee et al. 2000; Tsuneto et al. 2005; zur Nieden et al. 2003). Also , increasing intracellular reactive oxygen species concentration is hypothesized to increase differentiation of E S C towards cardiomyocytes (Sauer et al. 2000). 1.2.4. Functional Assays to Detect Undifferentiated Embryonic Stem Cells The most rigorous in vivo assay to establish functionality of cultured mouse E S C requires their contribution to all adult tissue, including germ cells, in chimeric mice established from the injection of E S C into pre-implantation mouse blastocysts. Here a defined number of test cells (~15) are injected into 3.5-day blastocysts before reintroduction into a pseudo-pregnant female. The blastocyst is allowed to come to term and the degree of chimerism is assessed in the newborn pup. Karyotypically normal, multipotent E S C are typically able to contribute to the germ line. In order to test this property, chimeric mice must be bred and a genetic contribution from the injected test cells must be observed in the resulting offspring. While this assay shows rigorous functional ability, it does so in a non-quantitative fashion and is time consuming and 12 difficult to perform. A second in vivo assay involves the transplantation of undifferentiated cells into recipient mice. Undifferentiated cells retain the ability to form teratocarcinomas containing cel l types from each o f the 3 germ layers. Whi l e this assay clearly determines multidifferentiation potential, it is again non quantitative and analysis of all differentiated tissues formed is time-consuming and difficult. Two in vitro assays, the E B formation and colony-forming cell (CFC) assays, lend themselves to more quantitative analysis. The E B formation assay (for review see (Keller 1995) relies on the assumption that only undifferentiated E S C are able to differentiate into multicellular, multilineage, three dimensional cellular clusters called embryoid bodies in optimal conditions. Defined numbers of test cells are plated in semi-solid methylcellulose in the absence of L IF and the resulting number of E B counted several days later. The C F C assay relies on the assumption that the ability to attach and generate colonies of undifferentiated E S C on gelatin-coated tissue culture dishes is restricted to undifferentiated stem cells. In this assay, a defined number of test culture output cells are plated on gelatin-coated dishes containing E S C maintenance media. Since it is known that undifferentiated E S C have high A P activity (Robertson et al. 1993), the number of colonies with high activity can be determined after several days. This assay specifically tests E S C self-renewal. Both of these assays rely on a specific definition of what constitutes positive colonies in each assay. This is done by defining a minimum colony size and optimal colony shape. 13 Little is known about the stringency of these in vitro assays although a significant correlation was found between cell-specific SSEA-1 and A P expression and the ability of cells to clonally form E B (Zandstra et al. 2000). 1.2.5. Phenotypic Markers of Undifferentiated Embryonic Stem Cells In addition to the functional assays mentioned above, E S C can be identified by a variety of phenotypic and gene expression markers. Oct4 has been shown to be an important player in the regulation of pluripotency in mouse E S C and it decreases in expression upon differentiation (Viswanathan et al. 2003b). A s a consequence, it is one of the most commonly used phenotypic markers of undifferentiated stem cells (see for example (Davey and Zandstra 2006; Viswanathan et al. 2003a)). However, unless it is specifically linked to a fluorescent marker such as G F P (Viswanathan et al. 2003a), its intracellular location reduces its value as a marker. Several other cell surface markers have been used to identify mouse E S C . The most prevalent of these is S S E A - 1 , a specific glycosylation modification of a surface marker of undifferentiated cells that is lost upon differentiation (Solter and Knowles 1978). Other cell surface markers whose expression on undifferentiated E S C is known to distinguish them from their differentiated progeny are 5T4 (Ward et al. 2003), P E C A M - 1 ( L i et al. 2005), C D 9 (Oka et al. 2002) and I C A M - 1 (Tian etal . 1997). Another potential assay for undifferentiated cells involves the assessment of changes in signal transduction pathways. Davey et al. have demonstrated concurrent radial organization of Oct4 and phosphorylated S T A T 3 (pSTAT3) indicating that the level of p S T A T 3 may be used to distinguish between differentiated and undifferentiated E S C (Davey and Zandstra 2006). 14 However, more work is required to determine i f this difference reflects a reversible or irreversible commitment of cells towards a differentiated fate. 1.2.6. Characterization of Mouse Embryonic Stem Cell Responses Verification of the performance of the different assays of undifferentiated cells described above requires conditions that are known to alter the developmental potential of E S C . B y comparing assay readouts in cells that have been cultured in these conditions, more can be understood about their relative performance. Several factors have been found to increase the maintenance and proliferation rates of undifferentiated E S C in the presence of L I F . Growth of E S C on P E F increases the frequency of EB-forming cells (Glover et al. 2006). Inhibition of the E R K pathway, which is activated by L I F signaling, increases E S C self-renewal (Burdon et al. 1999). Furthermore, several factors have been shown to either increase the proliferation or self-renewal rate of E S C in culture (Han et al. 2006; Heo and Han 2006; Heo et al. 2006; K i m et al. 2006b). E S C cultured in the absence of L IF rapidly lose their pluripotent state. When E S C are cultured in l iquid suspension or in semi-solid methylcelluose-based media in the absence of L I F , they form spheroid aggregates of proliferating cells known as E B . These colonies contain differentiated cells from all germ layers and some reports have advocated that these structures recapitulate in vivo development (Doetschman et al. 1985). 15 1.3. Challenges in Embryonic Stem Cell Culture A t present, optimal expansion of E S C requires the use of undefined factors such as F B S and PEF; This is of concern due to batch-to-batch variability as well as the increased potential for rejection of ESC-derived tissues following transplantatation due to the expression of foreign antigens (Martin et al. 2005). Furthermore, current methods of culturing human E S C result in excessive loss of desirable cells through spontaneous differentiation in culture and during standard passaging despite relatively rapid growth rates (Stewart et al. 2006). 1.3.1. High Throughput Screening to Identify Factors Supporting the Maintenance of Embryonic Stem Cells A variety of screening technologies are now being used to gain further insight into molecular mechanisms that govern E S C fate determination as well as to discover new factors that are able to direct stem cells to predetermined fates. Several gene screens of both overexpression and R N A i mediated knockdown have begun to determine the molecular machinery that maintains E S C in the undifferentiated state. Gene expression based studies (summarized above) have been performed that have identified candidate genes. Subsequent experiments have identified several of these that are specifically involved in E S C maintenance. Using an R N A i based approach, Ivanova identified 5 genes in addition to oct4, sox2 and nanog which were involved in self-renewal of E S C (Ivanova et al. 2006). B y performing further gene expression profiling of cells with these genes knocked down they were also able to propose a provisional control hierarchy map for these genes. In a related 16 study, the same group overexpressed a c D N A library in E S C and then by observing the effect on cell cycle identified clones that increased or decreased self-renewal. These identified 11 genes that were able to partially rescue E S C self-renewal in the absence of L IF (Pritsker et al. 2006). While the screens described above identify molecular pathways involved in E S C fate control, these have not yet helped to determine optimal culture conditions for the expansion of undifferentiated E S C . Other studies have focused on identifying small molecules and adherent factors that have an effect on E S C fate, from the perspective of both self-renewal and differentiation. Chen et al. used an Oct4-GFP reporter in E S C unable to sustain self-renewal in PEF-free conditions to screen a library of 50,000 small molecules (Chen et al. 2006). After 6 days of culture in the absence of feeders, they identified 28 molecules that supported E S C self-renewal. Further investigation of these factors revealed that 17 were able to maintain expression of S S E A - 1 , Oct4 and A P . . B y investigating the specific chemical structure o f these 17 compounds, they were able to determine a basic structure upon which chemical modifications had different levels of effect on E S C self-renewal. A similar screen determined that an inhibitor of GSK-3(3 increased self-renewal in E S C revealing a role for the Wnt pathway in E S C maintenance (Ding et al. 2003). Identification of non-soluble factors that support E S C self-renewal is also the subject of research. Anderson and colleagues designed a mieroarray which allows the screening of multiple factors for the attachment of E S C (Anderson et al . 2004). Similarly Flaim and colleagues present a similar device for investigating combinations of various extracellular matrix proteins on differentiation of human E S C (Flaim et al. 2005). While these 17 devices have only been applied to differentiation at present, this approach could be extended to self-renewal. Several limitations exist at present with these approaches. First a specific challenge remains the identification of a suitable assay for the evaluation of the effect of the screened compounds. The assays used need to give a rapid readout and to be amenable to scale up to the level where tens of thousands of variables are screened simultaneously. Existing functional assays are not suitable for this due to their long lead-time and labour-intensive nature. Fluorescent reporters are more amenable to high-throughput screening, but careful choices must be made to ensure that these give appropriate readouts. For example, it has been shown that Oct4 is a slow indicator of loss of pluripotency (Palmqvist et al. 2005) indicating that more rapid markers should be used. Another problem lies in determining effective concentration ranges at which to test chemical screens. In al l reported screening experiments, compounds are tested at one specific concentration. However, effective concentrations of these chemicals w i l l probably differ considerably and non-linear concentration responses are known to occur (Audet et al. 2002) Once compounds and substrates are identified, consideration must be given to interactions between multiple molecules when added simultaneously to culture. Complex interactions such as synergy and redundancy may take place between multiple factors when added simultaneously. Thus systematic methods are required to identify non-linear responses as wel l as factor interactions. 18 1.3.2. Statistical Design of Experiments to Understand Factor Interactions and Determine Non-Linear Responses Statistical design of experiments, specifically factorial design (FD) and central composite design ( C C D ) , provide rigorous statistical approaches to efficiently determine individual and multi-factor responses. They make use of geometric principles for statistical sampling and minimize the variance of estimated coefficients obtained by regression. In addition, the designs are able to minimize the impact of unimportant factors (e.g. reagent batch, errors in cell counts, volume addition etc.) by grouping experiments into homogeneous blocks and by inducing replication in treatment (Figure 1.1). First-order designs such as F D can thus be useful as screening tools to identify a few critical factors, including interactions, for subsequent study. This type of study can then be followed up by more in-depth investigation by CCD-type experiments to determine non-linear responses and to obtain a complete mathematical description of responses. This can then be used to determine factor levels that maximize the response of interest. Statistical design of experiments have previously been used to enhance stem cell expansion and differentiation for a variety of stem cell types. Prudhomme et al. used F D experiments to gain insight into the signaling responses of E S C to both soluble and insoluble factors (Prudhomme et al. 2004). Audet et al. used a F D experiment.to screen multiple factors for their influence on H S C expansion, and followed this with a C C D to optimize the level of 3 of these factors for maximal stem cell expansion (Audet et al. 2002). Statistical design of experiments has also been used in the context of differentiation. Chang et al., investigated the effect o f 5 factors on endoderm differentiation from E S C (Chang and Zandstra 2004). Experiments of this nature w i l l 19 become increasingly important as a clinical need for stem cell based therapies develops. Clinical application of these therapies w i l l require the development of xeno-free methods of culturing stem cells to avoid subsequent tissue rejection and immunological reaction. A s replacements for xeno-derived components of cell culture medium are identified, significant effort w i l l have to be made to incorporate these into cell culture medium and to ensure that there are no negative consequences of addition as a result of factor interactions. 1.3.3. Culture Systems Suitable for High Throughput Screening Specialized culture systems are required for efficient large-scale and multifactorial screening. Due to the large number of cultures that must be run in parallel, small volume, automatic control culture systems are ideal. Technology derived from the field of microfluidics provides the basis for culture systems of this type (Balagadde et al. 2005). Application of these technologies to E S C is just beginning. Recently K i m et al. demonstrated the growth and maintenance of mouse E S C in a microfluidic device with laminar flow ( K i m et al. 2006a). Further advantages of this technology are the other analytical tools that can be incorporated. For example, Warren and colleagues were able to monitor the expression of specific transcription factors in individual hematopoietic stem cells using microfluidic devices (Warren et al. 2006). Combinations of these technologies w i l l provide incredible power and high-throughput ability to allow the optimization of stem cell expansion as well as an increased understanding of the control of stem cell fate. 20 1.4. Thesis Objectives Clin ica l scale application of ESC-based therapies requires development of protocols for the expansion of undifferentiated E S C and their subsequent differentiation into relevant cell types in large numbers. Despite recent reports of expansion of undifferentiated cells in large-scale bioreactors (Cormier et al. 2006; Fong et al . 2005; Oh et al. 2005), more work is needed to understand the essential components of undifferentiated E S C as well as their tolerance of varying culture conditions. Understanding of these factors w i l l allow the development of protocols for the optimal expansion of undifferentiated E S C . The overall goal of this thesis was to develop tools to allow the optimal expansion of E S C . The expectation was that investigation and comparison of existing assays for undifferentiated E S C would allow the development of novel assays for E S C based of gene expression. Furthermore, the use of these methods would allow the optimization of E S C expansion. In Chapter 2, the goal was to compare existing assays and markers of E S C during differentiation to evaluate their relative ability to monitor loss of E S C as wel l as to select appropriate time points for gene expression analysis. It was hypothesized that analysis of gene expression changes during differentiation would a l low the identification o f novel markers o f undifferentiated E S C as well as provide candidate genes that are involved in the maintenance and differentiation of E S C . 21 In Chapter 3, the goal was to optimize the level of 3 different factors ascorbic acid ( A A ) , chondroitin sulphate (CS) and an inhibitor of E R K 1 / 2 phosphorylation, PD98059 (PD), with respect to the expansion of undifferentiated E S C . A A and P D have previously been shown to increase the self-renewal of E S C (Burdon et al. 1999; Glover et al. 2006). CS was hypothesized to have an influence on E S C fate due to its inclusion in the extracellular matrix. Expansion of E S C was measured using clonal E B formation as a measure of pluripotency. F D and C C D analysis was used to identify optimal culture conditions for the expansion of E S C as well as identify any factor interactions. The E B assay provides a reliable indicator of pluripotency within a population but is labour-intensive and does not lend itself easily to high-throughput type experiments that would facilitate a greater understanding of the influence of the cell culture environment on E S C fate. Thus a further series of experiments, described in Chapter 4, was undertaken to identify reliable gene expression indicators of E S C fate. Since the gene expression changes useful for monitoring E S C fate, should be independent of the type of lineage that E S C were differentiating towards, gene expression profiling was carried out on E S C induced to differentiate by multiple stimuli. Desired gene expression changes should be able to identify changes in the developmental potential of E S C regardless of the culture conditions. 22 Table 1.1 Similarities and differences between human and mouse embryonic stem cells Adapted from Kirschstein and Skirball, 2001. P r o p e r t y M o u s e ESC H u m a n ESC SSEA-1 rf -SSEA-3 -AP + + Oct4 + + Telomerase + + Feeder-cell dependent No Yes Self-renewal factors LIF, BMP4 FGF2, TGFp\ noggin Teratoma formation + + Chimera formation + + (with mouse) Tissues generated Mesoderm, endoderm, Mesoderm, endoderm, ectoderm ectoderm, trophectoderm Abbreviations: E S C - embryonic stem cel l , A P - alkaline phosphatase, L I F - leukemia inhibitory factor. 23 Figure 1.1 Graphical representation of factorial and central composite design experiments for 3 factors Factorial design experiments are represented by white points and central composite design experiments are represented by both white and yellow points. Figure provided by Julie Audet. (1.63, 0,0) 1.-1.-1) (-1,1,-1 (1,-1,1) (-1,-1,1) (-1.63, 0,0) 24 1.5. References Abbondanzo SJ, Gadi I, Stewart C L . 1993. Derivation of embryonic stem cell lines. Methods Enzymol 225:803-23. Adams IR, McLaren A . 2004. Identification and characterisation of m R i f l : a mouse telomere-associated protein highly expressed in germ cells and embryo-derived pluripotent stem cells. Dev D y n 229(4):733-44. Al-Haj j M , Wicha M S , Benito-Hernandez A , Morr ison SJ, Clarke M F . 2003. Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A 100(7):3983-8. Anderson D G , Levenberg S, Langer R. 2004. Nanoliter-scale synthesis of arrayed biomaterials and application to human embryonic stem cells. Nat Biotechnol 22(7):863-6. Audet J, Mi l le r C L , Eaves CJ , Piret J M . 2002. Common and distinct features of cytokine effects on hematopoietic stem and progenitor cells revealed by dose-response surface analysis. Biotechnol Bioeng 80(4):393-404. A v i l i o n A A , Nicol is S K , Pevny L H , Perez L , V i v i a n N , Lovell-Badge R. 2003. Multipotent cell lineages in early mouse development depend on S O X 2 function. Genes Dev 17(1): 126-40. Balagadde F K , Y o u L , Hansen C L , Arno ld F H , Quake SR. 2005. Long-term monitoring of bacteria undergoing programmed population control in a microchemostat. Science 309(5731):137-40. . Beattie G M , Lopez A D , Bucay N , Hinton A , Firpo M T , K i n g C C , Hayek A . 2005. Ac t iv in A maintains pluripotency of human embryonic stem cells in the absence of feeder layers. Stem Cells 23(4):489-95. Bjorklund L M , Sanchez-Pernaute R, Chung S, Andersson T, Chen I Y , McNaught K S , Brownell A L , Jenkins B G , Wahlestedt C, K i m K S , Isacson O. 2002. Embryonic stem cells develop into, functional dopaminergic neurons after transplantation in a Parkinson rat model. Proc Natl Acad Sci U S A 99(4):2344-9. Bloch W, Forsberg E , Lentini S, Brakebusch C, Mart in K , K r e l l H W , Weidle U H , Addicks K , Fassler R. 1997. Beta 1 integrin is essential for teratoma growth and angiogenesis. J Ce l l B i o l 139(l):265-78. Bonnet D , Dick JE . 1997. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat M e d 3(7):730-7. 25 Boyer L A , Lee TI, Cole M F , Johnstone SE, Levine SS, Zucker JP, Guenther M G , Kumar R M , Murray H L , Jenner R G , Gifford D K , Mel ton D A , Jaenisch R, Young R A . 2005. Core transcriptional regulatory circuitry in human embryonic stem cells. Ce l l 122(6):947-56. Boyer L A , Mathur D , Jaenisch R. 2006. Molecular control of pluripotency. Curr Opin Genet Dev 16(5):455-62. Brandenberger R, Wei H , Zhang S, L e i S, Murage J, Fisk GJ , L i Y , X u C, Fang R, Guegler K , Rao M S , Mandalam R, Lebkowski J, Stanton L W . 2004. Transcriptome characterization elucidates signaling networks that control human E S cell growth and differentiation. Nat Biotechnol 22(6):707-16. Burdon T, Smith A , Savatier P. 2002. Signalling, cell cycle and pluripotency in embryonic stem cells. Trends Cel l B i o l 12(9):432-8. Burdon T, Stracey C, Chambers I, Nichols J, Smith A . 1999. Suppression of SHP-2 and E R K signalling promotes self-renewal of mouse embryonic stem cells. Dev B i o l 210(l):30-43. Cerdan C, Bendall S C , Wang L , Stewart M , Werbowetski T, Bhatia M . 2006. Complement targeting of nonhuman sialic acid does not mediate cell death of human embryonic stem cells. Nat M e d 12(10): 1113-4; author reply 1115. Chambers I, Colby D , Robertson M , Nichols J, Lee S, Tweedie S, Smith A . 2003. Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. C e l l 113(5):643-55. Chang K H , Zandstra P W . 2004. Quantitative screening of embryonic stem cell differentiation: endoderm formation as a model. Biotechnol Bioeng 88(3):287-98. Chaudhry M A S . 2006. Analysis of Stem and Progenitor C e l l Responses to Variations in Bioprocess Culture Variables [Ph.D.]. Vancouver: University of British Columbia. 187 p. Chen S, Do JT, Zhang Q, Yao S, Y a n F, Peters E C , Scholer H R , Schultz P G , Ding S. 2006. Self-renewal of embryonic stem cells by a small molecule 10.1073/pnas.0608156103. P N A S 103(46): 17266-17271. Chickarmane V , Troein C, Nuber U A , Sauro H M , Peterson C. 2006. Transcriptional dynamics of the embryonic stem cell switch. PLoS Comput B i o l 2(9):e 123. Conti L , Pollard S M , Gorba T, Reitano E , Toselli M , Bie l l a G , Sun Y , Sanzone S, Y i n g Q L , Cattaneo E , Smith A . 2005. Niche-independent symmetrical self-renewal of a mammalian tissue stem cell. PLoS B i o l 3(9):e283. 26 Cormier JT, Nieden NI , Rancourt D E , Kallos M S . 2006. Expansion of Undifferentiated Murine Embryonic Stem Cells as Aggregates in Suspension Culture Bioreactors. Tissue Eng. Dani C, Smith A G , Dessolin S, Leroy P, Staccini L , Villageois P, Darimont C, Ailhaud G . 1997. Differentiation of embryonic stem cells into adipocytes in vitro. J Ce l l Sci 110 (P t 11): 1279-85. Davey R E , Zandstra P W . 2006. Spatial organization of embryonic stem cell responsiveness to autocrine gpl30 ligands reveals an autoregulatory stem cell niche. Stem Cells 24(11):2538-48. Davis S, Ald r i ch T H , Stahl N , Pan L , Taga T, Kishimoto T, Ip N Y , Yancopoulos G D . 1993. L I F R beta and gpl30 as heterodimerizing signal transducers of the tripartite C N T F receptor. Science 260(5115): 1805-8. Ding S, W u T Y H , Brinker A , Peters E C , Hur W , Gray N S , Schultz P G . 2003. Synthetic small molecules that control stem cell fate 10.1073/pnas.0732087100. P N A S 100(13):7632-7637. Dinsmore J, Ratliff J, Deacon T, Pakzaban P, Jacoby D , Galpern W , Isacson O. 1996. Embryonic stem cells differentiated in vitro as a novel source of cells for transplantation. Ce l l Transplant 5(2):131-43. Doetschman T C , Eistetter H , Katz M , Schmidt W , Kemler R. 1985. The in vitro development of blastocyst-derived embryonic stem cell lines: formation of visceral yolk sac, blood islands and myocardium. J Embryol Exp Morphol 87:27-45. Draper JS, Pigott C , Thomson J A , Andrews P W . 2002. Surface antigens of human embryonic stem cells: changes upon differentiation in culture. J Anat 200(Pt 3):249-58. Ellerstrom C, Strehl R, M o y a K , Andersson K , Bergh C, Lundin K , Ffyllner J, Semb H . 2006. Derivation of a xeno-free human embryonic stem cell line. Stem Cells 24(10):2170-6. Evans M J , Kaufman M H . 1981. Establishment in culture of pluripotential cells from mouse embryos. Nature 292(5819): 154-6. Fassler R, Meyer M . 1995. Consequences of lack of beta 1 integrin gene expression in mice. Genes Dev 9(15): 1896-908. Fla im C J , Chien S, Bhatia S N . 2005. A n extracellular matrix mieroarray for probing cellular differentiation. Nat Methods 2(2): 119-25. Fong W J , Tan H L , Choo A , Oh S K . 2005. Perfusion cultures of human embryonic stem cells. Bioprocess Biosyst Eng 27(6):381-7. 27 Fortunel N O , Otu H H , N g H H , Chen J, M u X , Chevassut T, L i X , Joseph M , Bailey C, Hatzfeld J A , Hatzfeld A , Usta F , Vega V B , Long P M , Libermann T A , L i m B.' 2003. Comment on " 'Sternness': transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature". Science 302(5644):393; author reply 393. Fujikawa T, Oh S H , P i L , Hatch H M , Shupe T, Petersen B E . 2005. Teratoma formation leads to failure of treatment for type I diabetes using embryonic stem cell-derived insulin-producing cells. A m J Pathol 166(6): 1781-91. Geijsen N , Horoschak M , K i m K , Gribnau J, Eggan K , Daley G Q . 2004. Derivation of embryonic germ cells and male gametes from embryonic stem cells. Nature 427(6970): 148-54. Ginis I, Luo Y , Miura T, Thies S, Brandenberger R, Gerecht-Nir S, Ami t M , Hoke A , Carpenter M K , Itskovitz-Eldor J, Rao M S . 2004. Differences between human and mouse embryonic stem cells. Dev B i o l 269(2):360-80. Glover C H , M a r i n M , Eaves C J , Helgason C D , Piret J M , Bryan J. 2006. Meta-Analysis of Differentiating Mouse Embryonic Stem Cel l Gene Expression Kinetics Reveals Early Change of a Small Gene Set. PLoS ComputBio l 2 ( l l ) : e l58 . Gurdon JB, Byrne J A , Simonsson S. 2003. Nuclear reprogramming and stem cell creation. Proc Natl Acad Sci U S A 100 Suppl 1:11819-22. Han H J , Heo JS, Lee Y J . 2006. Estradiol-17beta stimulates proliferation of mouse embryonic stem cells: involvement of M A P K s and C D K s as well as protooncogenes. A m J Physiol Ce l l Physiol 290(4):C1067-75. Heo JS, Han H J . 2006. A T P Stimulates Mouse Embryonic Stem Ce l l Proliferation via Protein Kinase C, Phosphatidylinositol 3-Kinase/Akt, and Mitogen-Activated Protein Kinase Signaling Pathways. Stem Cells 24(12):2637-48. Heo JS, Lee Y J , Han H J . 2006. E G F stimulates proliferation of mouse embryonic stem cells: involvement of Ca2+ influx and p44/42 M A P K s . A m J Physiol Cel l Physiol 290(l):C123-33. Hubner K , Fuhrmann G , Christenson L K , Kehler J, Reinbold R, De L a Fuente R, Wood J, Strauss JF, III, Boiani M , Scholer H R . 2003. Derivation of Oocytes from Mouse Embryonic Stem Cells 10.1126/science. 1083452. Science 300(5623):1251-1256. Ivanova N , Dobrin R, L u R, Kotenko I, Levorse J, DeCoste C, Schafer X , L u n Y , Lemischka IR. 2006. Dissecting self-renewal in stem cells with R N A interference. Nature 442(7102):533-8. Ivanova N B , Dimos JT, Schaniel C , Hackney J A , Moore K A , Lemischka IR. 2002. A stem cell molecular signature. Science 298(5593):601-4. 28 Kattman SJ, Huber T L , Keller G M . 2006. Multipotent flk-l+ cardiovascular progenitor cells give rise to the cardiomyocyte, endothelial, and vascular smooth muscle lineages. Dev C e l l l l(5):723-32. Keirstead H S , Nistor G , Bernal G , Totoiu M , Cloutier F , Sharp K , Steward O. 2005. Human embryonic stem cell-derived oligodendrocyte progenitor cell transplants remyelinate and restore locomotion after spinal cord injury. J Neurosci 25(19):4694-705. Keller G M . 1995. In vitro differentiation of embryonic stem cells. Curr Opin Cel l B i o l 7(6):862-9. K i m L , Vahey M D , Lee H Y , Voldman J. 2006a. Microfluidic arrays for logarithmically perfused embryonic stem cell culture. Lab Chip 6(3):394-406. K i m Y H , Heo JS, Han H J . 2006b. High glucose increase cell cycle regulatory proteins level of mouse embryonic stem cells v ia P I 3 - K / A k t and M A P K s signal pathways. J C e l l Physiol 209(1):94-102. Kirschstein R, Skirbol l L R . 2001. Stem Cells : Scientific Progress and Future Research Directions. Bethesda, M D : National Institutes of Health, Department o f Health and Human Services. Appendix E 1-11 p. Kunath T, Arnaud D , U y G D , Okamoto I, Chureau C , Yamanaka Y , Heard E , Gardner R L , .. Avner P, Rossant J. 2005. Imprinted X-inactivation in extra-embryonic endoderm cell lines from mouse blastocysts. Development 132(7): 1649-61. Lavker R M , Sun TT. 2000. Epidermal stem cells: properties, markers, and location. Proc Nat l Acad Sci U S A 97(25): 13473-5. Lee S H , Lumelsky N , Studer L , Auerbach J M , M c K a y R D . 2000. Efficient generation of midbrain and hindbrain neurons from mouse embryonic stem cells. Nat Biotechnol 18(6):675-9. Levenstein M E , Ludwig T E , X u R H , Lianas R A , VanDenHeuvel-Kramer K , Manning D , Thomson J A . 2006. Basic fibroblast growth factor support of human embryonic stem cell self-renewal. Stem Cells 24(3):568-74. L i ZJ , Wang Z Z , Zheng Y Z , X u B , Yang R C , Scadden D T , Han Z C . 2005. Kinetic expression of platelet endothelial cell adhesion molecule-1 ( P E C A M - 1 / C D 3 1 ) during embryonic stem cell differentiation. J Cel l Biochem 95(3):559-70. L o h Y H , W u Q, Chew J L , Vega V B , Zhang W , Chen X , Bourque G , George J, Leong B , L i u J, Wong K Y , Sung K W , Lee C W , Zhao X D , Chiu K P , L ipovich L , Kuznetsov V A , Robson P, 29 Stanton L W , Wei C L , Ruan Y , L i m B , N g H H . 2006. The Oct4 and Nanog transcription network regulates pluripotency in mouse embryonic stem cells. Nat Genet 38(4):431-40. Lowe l l S, Benchoua A , Heavey B , Smith A G . 2006. Notch promotes neural lineage entry by pluripotent embryonic stem cells. PLoS B i o l 4(5):el21. / L u C C , Brennan J, Robertson EJ . 2001. From fertilization to gastrulation: axis formation in the mouse embryo. Curr Opin Genet Dev 11(4):384-92. Ludwig T E , Levenstein M E , Jones J M , Berggren W T , Mitchen E R , Frane J L , Crandall L J , Daigh C A , Conard K R , Piekarczyk M S , Lianas R A , Thomson J A . 2006. Derivation of human embryonic stem cells in defined conditions. Nat Biotechnol 24(2): 185-7. Mart in G R . 1981. Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc Natl Acad Sci U S A 78(12):7634-8. Mart in M J , Muotr i A , Gage F , V a r k i A . 2005. Human embryonic stem cells express an immunogenic nonhuman sialic acid. Nat M e d 11(2):228-32. Martin M J , Muotr i A , Gage F H , Vark i A . 2006. Complement targeting of nonhuman sialic acid does not mediate cell death of human embryonic stem cells. Nat M e d 12(10): 1115. Matsuda T, Nakamura T, Nakao K , A r a i T, Katsuki M , Heike T, Yokota T. 1999. S T A T 3 activation is sufficient to maintain an undifferentiated state of mouse embryonic stem cells. E m b o J 18(15):4261-9. Menard C, Hagege A A , Agbulut O, Barro M , Morichetti M C , Brasselet C , Be l A , Messas E , Bissery A , Bruneval P, Desnos M , Puceat M , Menasche P. 2005. Transplantation of cardiac-committed mouse embryonic stem cells to infarcted sheep myocardium: a preclinical study. Lancet 366(9490).:1005-12. Mitsui K , Tokuzawa Y , Itoh H , Segawa K , Murakami M , Takahashi K , Maruyama M , Maeda M , Yamanaka S. 2003. The homeoprotein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells. Cel l 113(5):631-42. ' . Moretti A , Caron L , Nakano A , Lam JT, Bernshausen A , Chen Y , Qyang Y , B u L , Sasaki M , Martin-Puig S, Sun Y , Evans S M , Laugwitz K L , Chien K R . 2006. Multipotent Embryonic Isll(+) Progenitor Cells Lead to Cardiac, Smooth Muscle, and Endothelial Cel l Diversification. Cel l . Nayernia K , Nolte J, Michelmann H W , Lee J H , Rathsack K , Drusenheimer N , Dev A , W u l f G , Ehrmann IE, Elliott DJ . 2006. In Vitro-Differentiated Embryonic Stem Cells Give Rise to Male Gametes that Can Generate Offspring Mice . Developmental Ce l l 11(1):125-132. 30 Nichols J, Zevnik B , Anastassiadis K , N i w a H , Klewe-Nebenius D , Chambers I, Scholer H , Smith A . 1998. Formation of pluripotent stem cells in the mammalian embryo depends on the P O U transcription factor Oct4. Ce l l 95(3):379-91. N i w a H , Burdon T, Chambers I, Smith A . 1998. Self-renewal of pluripotent embryonic stem cells is mediated via activation of S T A T 3 . Genes Dev 12(13):2048-60. N i w a H , Miyazaki J, Smith A G . 2000. Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells. Nat Genet 24(4):372-6. Odorico JS, Kaufman D S , Thomson J A . 2001. Multi l ineage differentiation from human embryonic stem cell lines. Stem Cells 19(3): 193-204. Oh S K , Fong W J , Teo Y , Tan H L , Padmanabhan J, Chin A C , Choo A B . 2005. High density cultures of embryonic stem cells. Biotechnol Bioeng 91(5):523-33. Oka M , Tagoku K , Russell T L , Nakano Y , Hamazaki T, Meyer E M , Yokota T, Terada N . 2002. C D 9 is associated with leukemia inhibitory factor-mediated maintenance of embryonic stem cells. M o l B i o l Cel l 13(4): 1274-81. Palacios R, Golunski E , Samaridis J. 1995. In vitro generation of hematopoietic stem cells from an embryonic stem cell line. Proc Natl Acad Sci U S A 92(16):7530-4. Palmqvist L , Glover C H , Hsu L , L u M , Bossen B , Piret J M , Humphries R K , Helgason C D . 2005. Correlation of murine embryonic stem cell gene expression profiles with functional measures of pluripotency. Stem Cells 23(5):663-80. Priddle H , Hemmings L , Monkley S, Woods A , Patel B , Sutton D , Dunn G A , Zicha D , Critchley D R . 1998. Disruption of the talin gene compromises focal adhesion assembly in undifferentiated but not differentiated embryonic stem cells. J Cel l B i o l 142(4): 1121-33. Pritsker M , Ford N R , Jenq H T , Lemischka IR. 2006. Genomewide gain-of-function genetic screen identifies functionally active genes in mouse embryonic stem cells. Proc Natl Acad Sci U S A 103(18):6946-51. Prudhomme W , Daley G Q , Zandstra P, Lauffenburger D A . 2004. Multivariate proteomic analysis of murine embryonic stem cell self-renewal versus differentiation signaling. Proc Natl Acad Sci U S A 101(9):2900-5. Ramalho-Santos M , Y o o n S, Matsuzaki Y , Mul l igan R C , Mel ton D A . 2002. "Sternness": transcriptional profiling of embryonic and adult stem cells. Science 298(5593):597-600. Rao R R , Calhoun JD, Qin X , Rekaya R, Clark J K , Stice S L . 2004. Comparative transcriptional profiling of two human embryonic stem cell lines. Biotechnol Bioeng 88(3):273-86. 31 Rao R R , Stice S L . 2004. Gene expression profiling of embryonic stem cells leads to greater understanding of pluripotency and early developmental events. B i o l Reprod 71(6): 1772-8. Reubinoff B E , Pera M F , Fong C Y , Trounson A , Bongso A . 2000. Embryonic stem cell lines from human blastocysts: somatic differentiation in vitro. Nat Biotechnol 18(4):399-404. Reynolds B A , Weiss S. 1992. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255(5052):1707-10. Robertson M , Chambers I, Rathjen P, Nichols J, Smith A . 1993. Expression of alternative forms of differentiation inhibiting activity (DIA/LIF) during murine embryogenesis and in neonatal and adult tissues. Dev Genet 14(3): 165-73. Rogers M B , Hosier B A , Gudas L J . 1991. Specific expression of a retinoic acid-regulated, zinc-finger gene, Rex-1, in preimplantation embryos, trophoblast and spermatocytes. Development 113(3):815-24. Rohwedel J, Guan K , Wobus A M . 1999. Induction of cellular differentiation by retinoic acid in vitro. Cells Tissues Organs 165(3-4): 190-202. Rossant J. 2001. Stem cells from the Mammalian blastocyst. Stem Cells 19(6):477-82. Sauer H , Rahimi G , Hescheler J, Wartenberg M . 2000. Role of reactive oxygen species and phosphatidylinositol 3-kinase in cardiomyocyte differentiation of embryonic stem cells. F E B S Lett 476(3) :218-23. Shackleton M , Vai l lant F , Simpson K J , Stingl J, Smyth G K , Asselin-Labat M L , W u L , Lindeman GJ , Visvader JE. 2006. Generation of a functional mammary gland from a single stem cell. Nature 439(7072):84-8. Sharov A A , Piao Y , Matoba R, Dudekula D B , Qian Y , VanBuren V , Falco G , Martin PR, Stagg C A , Bassey U C , Wang Y , Carter M G , Hamatani T, A i b a K , Akutsu H , Sharova L , Tanaka TS, Kimber W L , Yoshikawa T, Jaradat S A , Pantano S, Nagaraja R, Boheler K R , Taub D , Hodes RJ , Longo D L , Schlessinger D , Keller J, Klo tz E , Kelsoe G , Umezawa A , Vescovi A L , Rossant J, Kunath T, Hogan B L , Curci A , D'Urso M , Kelso J, Hide W , K o M S . 2003. Transcriptome analysis of mouse stem cells and early embryos. PLoS B i o l 1(3):E74. Singh S K , Clarke ID, Terasaki M , Bonn V E , Hawkins C, Squire J, Dirks P B . 2003. Identification of a cancer stem cell in human brain tumors. Cancer Res 63(18):5821-8. Skottman H , Stromberg A M , Matilainen E , Inzunza J, Hovatta O, Lahesmaa R. 2005. Unique gene expression signature by human embryonic stem cells cultured under serum free conditions correlates with their enhanced and prolonged growth in an undifferentiated stage. Stem Cells. 32 Smith A G . 2001. Embryo-derived stem cells: of mice and men. Annu Rev Cel l Dev B i o l 17:435-62. Smith A G , Heath J K , Donaldson D D , Wong G G , Moreau J, Stahl M , Rogers D . 1988. Inhibition o f pluripotential embryonic stem cel l differentiation by purified polypeptides. Nature 336(6200):688-90. Solter D , Knowles B B . 1975. Immunosurgery of mouse blastocyst. Proc Nat l Acad Sci U S A 72(12):5099-102. Solter D , Knowles B B . 1978. Monoclonal antibody defining a stage-specific mouse embryonic antigen (SSEA-1) . Proc Natl Acad Sci U S A 75(11):5565-9. Stewart M H , Bosse M , Chadwick K , Menendez P, Bendall SC, Bhatia M . 2006. Clonal isolation of hESCs reveals heterogeneity within the pluripotent stem cell compartment. Nat Methods 3(10):807-15. Stingl J, Ei rew P, Ricketson I, Shackleton M , Vail lant F , Choi D , L i HI , Eaves C J . 2006. Purification and unique properties of mammary epithelial stem cells. Nature 439(7079):993-7. Sumi T, Fujimoto Y , Nakatsuji N , Suemori H . 2004. S T A T 3 is dispensable for maintenance of self-renewal in nonhuman primate embryonic stem cells. Stem Cells 22(5):861-72. Swain A , Narvaez V , Burgoyne P, Camerino G , Lovell-Badge R. 1998. D a x l antagonizes Sry action in mammalian sex determination. Nature 391 (6669):761-7. Swijnenburg R J , Tanaka M , Vogel H , Baker J, Kofidis T, Gunawan F, Lebl D R , Caffarelli A D , de Bruin JL , Fedoseyeva E V , Robbins R C . 2005. Embryonic stem cell immunogenicity increases upon differentiation after transplantation into ischemic myocardium. Circulat ion 112(9 Suppl):1166-72. Takahashi K , Mitsui K , Yamanaka S. 2003 a. Role of ERas in promoting tumour-like properties in mouse embryonic stem cells. Nature 423(6939):541-5. Takahashi K , Yamanaka S. 2006. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cel l 126(4):663-76. Takahashi T, Lord B , Schulze P C , Fryer R M , Sarang SS, Gullans SR, Lee R T . 2003b. Ascorbic acid enhances differentiation of embryonic stem cells into cardiac myocytes. Circulation 107(14):1912-6. Tanaka S, Kunath T, Hadjantonakis A K , Nagy A , Rossant J. 1998. Promotion of trophoblast stem cell proliferation by F G F 4 . Science 282(5396):2072-5. 33 Thomson J A , Itskovitz-Eldor J, Shapiro SS, Waknitz M A , Swiergiel JJ, Marshall V S , Jones J M . 1998. Embryonic stem cell lines derived from human blastocysts. Science 282(5391):1145-7. Tian L , Catt JW, O 'Nei l l C , K i n g N J . 1997. Expression of immunoglobulin superfamily cell adhesion molecules on murine embryonic stem cells. B i o l Reprod 57(3):561-8. T i l l JE, McCul loch E A . 1961. A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Radiat Res 14:213-22. Toyooka Y , Tsunekawa N , Akasu R, Noce T. 2003. Embryonic stem cells can form germ cells in vitro 10.1073/pnas. 1932826100. P N A S 100(20): 11457-11462. Tropepe V , Coles B L , Chiasson B J , Horsford D J , E l i a A J , Mclnnes R R , van der Kooy D . 2000. Retinal stem cells in the adult mammalian eye. Science 287(5460):2032-6. Tsuneto M , Yamazaki H , Yoshino M , Yamada T, Hayashi S. 2005. Ascorbic acid promotes osteoclastogenesis from embryonic stem cells. Biochem Biophys Res Comrnun 335(4): 1239-46. Vall ier L , Reynolds D , Pedersen R A . 2004. Nodal inhibits differentiation of human embryonic stem cells along the neuroectodermal default pathway. Dev B i o l 275(2):403-21. Viswanathan S, Benatar T, Mile ikovsky M , Lauffenburger D A , Nagy A , Zandstra P W . 2003a. Supplementation-dependent differences in the rates of embryonic stem cel l self-renewal, differentiation, and apoptosis. Biotechnol Bioeng 84(5):505-17. Viswanathan S, Benatar T, M i l e i k o v s k y M , Nagy A , Zandstra P W . 2003b. In vitro characterization of Oct4 promoter driven Green Fluorscent Protein expressing embryonic stem cells, in preparation. Wang J, Rao S, Chu J, Shen X , Levasseur D N , Theunissen T W , Orkin S H . 2006a. A protein interaction network for pluripotency of embryonic stem cells. Nature 444(7117):364-8. Wang L , Cerdan C , Menendez P, Bhatia M . 2006b. Derivation and characterization of hematopoietic cells from human embryonic stem cells. Methods M o l B i o l 331:179-200. Ward C M , Barrow K , Woods A M , Stern P L . 2003. The 5T4 oncofoetal antigen is an early differentiation marker of mouse ES cells and its absence is a useful means to assess pluripotency. J C e l l Sci 116(Pt22):4533-42. Warren L , Bryder D , Weissman IL , Quake SR. 2006. Transcription factor profiling in individual hematopoietic progenitors by digital R T - P C R . Proc Natl Acad Sci U S A 103(47):17807-12. 34 Wei C L , Miura T, Robson P, L i m S K , X u X Q , Lee M Y , Gupta S, Stanton L , Luo Y , Schmitt J, Thies S, Wang W , Khrebtukova I, Zhou D , L i u E T , Ruan Y J , Rao M , L i m B . 2005. Transcriptome profiling of human and murine E S C s identifies divergent paths required to maintain the stem cell state. Stem Cells 23(2):166-85. Wiles M V , Keller G . 1991. Multiple hematopoietic lineages develop from embryonic stem (ES) cells in culture. Development 111(2):259-67. W u S M , Fujiwara Y , Cibulsky S M , Clapham D E , Lien C L , Schultheiss T M , Orkin S H . 2006. Developmental Origin of a Bipotential Myocardial and Smooth Muscle Ce l l Precursor in the Mammalian Heart. Cel l . X u C, Inokuma M S , Denham J, Golds K , Kundu P, Gold JD, Carpenter M K . 2001. Feeder-free growth of undifferentiated human embryonic stem cells. Nat Biotechnol 19(10):971-4. X u R H , Peck R M , L i DS , Feng X , Ludwig T, Thomson JA. 2005. Basic F G F and suppression of B M P signaling sustain undifferentiated proliferation of human ES cells. Nat Methods 2(3): 185-90. Y i n g Q L , Nichols J, Chambers I, Smith A . 2003. B M P induction of Id proteins suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with S T A T 3 . Cel l 115(3):281-92. Zandstra P W , Bauwens C, Y i n T, L i u Q, Schiller H , Zweigerdt R, Pasumarthi K B , Fie ld L J . 2003. Scalable production of embryonic stem cell-derived cardiomyocytes. Tissue Eng 9(4):767-78. Zandstra P W , Le H V , Daley G Q , Griffith L G , Lauffenburger D A . 2000. Leukemia inhibitory factor (LIF) concentration modulates embryonic stem cell self-renewal and differentiation independently o f proliferation. Biotechnol Bioeng 69(6):607-17. • Zandstra P W , Nagy A . 2001. Stem cell bioengineering. Annu Rev Biomed Eng 3:275-305. Zhang J, Tarn W L , Tong G Q , W u Q, Chan H Y , Soh B S , L o u Y , Yang J, M a Y , Chai L , N g H H , Lufkin T, Robson P, L i m B . 2006. Sall4 modulates embryonic stem cell pluripotency and early embryonic development by the transcriptional regulation of Pou5f l . Nat Ce l l B i o l 8(10): 1114-23. zur Nieden NI , Kempka G , Ahr H J . 2003. In vitro differentiation of embryonic stem cells into mineralized osteoblasts. Differentiation 71(1): 18-27. 35 2. Correlation of Mouse Embryonic Stem Cell Gene Expression Profiles with Functional Measures of Pluripotency1 2.1. Introduction Embryonic stem cells (ESC) are characterized by their ability to both self-renew, as well as differentiate (Smith 2001; Suda et al. 1987). However, the molecular mechanisms, which regulate the decision between these two processes, are poorly understood. Mouse E S C were originally isolated from the inner cell mass ( ICM) of pre-implantation blastocysts (Evans and Kaufman 1981; Mart in 1981) and can be maintained in cell culture indefinitely without loss of their broad pluripotent differentiation capacity as determined by their ability to give rise to all three germ layers both in vitro and in vivo (Suda et al. 1987). The more recent establishment of human E S C lines (Thomson et al. 1998) has further increased interest in E S C , since they raise hope for an unlimited source of cells for tissue engineering and cell therapies in the future. However, realization of this potential requires an increased knowledge of the molecular mechanisms governing self-renewal and pluripotency to guide the development of processes that control expansion and differentiation of stem cells ex vivo. Unlike hematopoietic stem cells (HSC) for which quantitative assays of stem cell potential have been defined and validated, no such assays currently exist for E S C . In the murine system self-1 A version of this chapter has been published as Palmqvist, L , Glover, C H , Hsu, L , L u , M , Bossen, B , Piret, J M , Humphries, R K , Helgason, C D . Correlation of murine embryonic stem cell gene expression profiles with functional measures of pluripotency. Stem Cells 23(5): 663 -680,2005. 36 renewal is measured by the ability of mouse E S C to continuously proliferate in culture while maintaining an undifferentiated colony morphology (Abbondanzo et al . 1993). The most rigorous in vivo assay to establish functionality of cultured mouse E S C is blastocyst injection and measurement of their ability to give rise to chimeric mice, since it requires E S C contribution to all adult tissue, including germ cells (Smith 2001). However, injection of 10 - 15 E S C into a single blastocyst does not provide a quantitative measure^of stem cell potential at the single cell level. Two in vitro assays have been used extensively as surrogates for chimera formation when testing culture reagents or examining the consequences of genetic manipulation. The colony forming cell (CFC) assay is used to determine the plating efficiency of E S C populations under various conditions and thus may be considered indicative of self-renewal potential. Formation of embryoid bodies (EB) can be performed at a clonal level in vitro and reflects multi-lineage differentiation potential (Keller et al. 1993). The correlation between these in vitro assays and chimera generation has not been determined. Assessment of pluripotency has also relied on the expression of selected molecular markers. For murine E S C these have included alkaline phosphatase (AP), S S E A - 1 , and Oct4 (Kirschstein and Skirboll 2001). However, the correlation between marker expression and the various functional assays has not been extensively studied. Knowledge of the intricate mechanisms regulating E S C pluripotency and differentiation potential is currently limited to a few signaling pathways (i.e. leukemia inhibitory factor - LIF) and regulatory factors (i.e. Oct4 and Nanog). Thus very little is known about the tolerance limits of different culture condition for maintaining stem cell function during expansion or how these relate to altered gene expression patterns in E S C . Identification o f molecular markers that 37 correlate with pluripotency would be invaluable to enrich for the desired cells, as well as monitor their maintenance during expansion protocols. Achieving the goal o f defining the core stem cell regulatory network requires more precise characterization of the functional capacities of the cells for which the transcriptional profile is described. In this study we established gene expression profiles during early differentiation of the well-defined R l E S C line (Nagy et al. 1993) and correlated gene expression changes with both phenotypic and functional assessment of the same cells. Cellular markers used included SSEA-1 and Oct4, while blastocyst injections for chimeric mouse formation, E B assays and C F C counts were used to determine functional capacity. Undifferentiated E S C and E S C cultured without L IF for 18 hours and 72 hours were chosen for array analysis. We identified 473 unique genes as significantly differentially expressed during early E S C differentiation, and approximately one third of these have unknown biological function. Among the 275 genes whose expression decreased with E S C differentiation were several factors previously identified as important for, or as markers of, E S C pluripotency including statS, zfp42, sox2, gbx2, and bmp4. A significant number of the decreased genes also overlapped with previously published mouse and human E S C data. Reverse-transcription-polymerase chain reaction ( R T - P C R ) validation showed high correlation with the array data, and several genes were also shown to have similar changes after L I F removal in two other murine E S C lines. Expression of the commonly used E S C markers Oct4 and S S E A - 1 was also examined in parallel with the functional assays. However, a close correlation was not observed. Interestingly, among a 38 subset of 48 decreased genes that showed the closest correlation with the functional assays was the stem cell factor (SCF) receptor c-kit that can be a useful marker of undifferentiated E S C . 2.2. Materials and Methods 2.2.1. Culture of Mouse Embryonic Stem Cells and Embryonic Fibroblasts R l (Nagy et al. 1993), J l ( L i et al. 1992) and E F C (Nichols et al. 1998) E S C were routinely maintained at 37°C humidified air with 5% CO2 on a-layer of irradiated primary embryonic fibroblasts (PEF) and fed daily with a complete change of E S C maintenance media consisting of high glucose D M E M (all reagents obtained from Stem Cel l Technologies Inc., Vancouver, B C unless otherwise indicated) supplemented with 15% ESC-tested fetal bovine serum (FBS) , 0.1 m M non-essential amino acids, 2 m M glutamine, lOOOU/ml LIF , 100 U / m l penicillin, 100 pg/ml streptomycin, and 100 p M monothioglycerol ( M T G ; Sigma, Oakville, ON) . For gene expression profiling R l E S C were from passage 14 and had been frozen at 10 6 cells per vial . Cells were passaged every second day in maintenance cultures. To passage cells, a single cell suspension was generated by treatment with 0.25% Trypsin - 1 m M E D T A (T /E ; Invitrogen Life Technologies, Burlington, ON) for 5 minutes until cells detached from the culture vessel surface. T / E activity was then quenched with D M E M supplemented with 10% F B S . Cells were centrifuged at 1200 rpm for 7 minutes and resuspended in E S C maintenance media. Viable cells were plated at 1 x 10 6/100 mm dish on irradiated P E F and cultured for a further 48 hours at 37°C, 5% CO2 prior to harvest for R N A isolation, differentiation, and/or functional assessment. In subsequent experiments all cells used were within 5 passages of initial thawing. 39 P E F were maintained at 37°C humidified air with 5% CO2 in D M E M supplemented with 10% F B S and 100 u.M M T G . Cells to be irradiated were trypsinized, resuspended in 2 m L of media, and exposed to 60 G y from an X-ray source prior to replating for use as feeder cells. Alternatively, P E F were pelleted and placed in Trizol reagent for R N A isolation (Invitrogen). 2.2.2. Preparation of Embryonic Stem Cells for Differentiation E S C were thawed and cultured for two passages over 96 hours as described above. To prepare cells for differentiation cells were harvested and washed as described above, resuspended in ES cell maintenance medium, and pre-plated on tissue culture plates (Falcon) for one hour at 37°C, 5% CO2 to deplete contaminating P E F . A t the end of this pre-plating step, non-adherent E S C were discarded and the loosely adherent E S C were collected by gently washing the surface of the tissue culture plate. Cells were pelleted by centrifugation and viable cel l numbers were determined. Contaminating P E F in the undifferentiated (day 0) E S C samples was estimated to be less than 0.2% based on cell size during counting. A portion of pre-plated E S C were suspended at a density of 1-2 x 10 7 cells per 50 ml in liquid differentiation medium consisting of Iscove's modified Dulbecco's medium ( I M D M ) , 15% F B S selected for its ability to support E S C differentiation, 2 m M glutamine, 150 u.M M T G , and 40 ng/ml murine S C F and plated into 4 x 100 mm Petri style culture dishes (Falcon). Cells were cultured overnight (18 hours) at 37°C, 5% CO2. The following day, embryoid bodies (EB), both in suspension and loosely attached, were harvested and allowed to settle to the bottom of a 50 ml conical tube for approximately 10 minutes. The supernatant, containing mainly single cells, was 40 removed and the spontaneously pelleting E B fraction was collected by centrifugation at 1200 rpm for 7 minutes. E B were disrupted by incubation in T /E for 3 minutes at room temperature followed by passage through a 21-gauge needle to obtain single cell suspensions. Cells were washed with 10% D M E M and suspended in 2 ml of I M D M to count. The remainder of the pre-plated E S C were plated at a density of 10 4 cells per 35 mm low-adherence Petri dish in IMDM-based E S C differentiation methylcellulose consisting of 0.9% methylcellulose, 15% F B S , 2 m M glutamine, 150 p M M T G , and 40 ng/ml murine S C F . Cultures were grown for 3 days and all E B in each dish were harvested by carefully flooding the dish with I M D M and collecting the methylcellulose/EB solution. E B were washed twice in D M E M + 10% F B S to remove the residual methylcellulose and were then pooled and disrupted as described above. 2.2.3. Blastocyst Injection C57B1/6J mice (used as blastocyst donors) and B 6 C 3 F l females (used as pseudopregnant blastocyst recipients) were purchased from the in-house breeding program at the B C Cancer Agency An ima l Resource Center. A l l mice were maintained with sterilized food, water and bedding. A l l protocols were conducted according to guidelines set forth by the Canadian Counci l for A n i m a l Care and approved by the An ima l Care Committee at. the University of British Columbia. R l E S C were thawed and maintained for two passages on irradiated P E F with daily feeding of maintenance media. After the second passage, cells were harvested and subjected to 1 hour of pre-plating on plastic to deplete remaining P E F . Cells were then plated onto gelatin coated tissue culture dishes and fed with maintenance media without L IF for the 41 indicated lengths of time. To produce chimeras, 15 test cells (either E S C or differentiated) were injected into 3.5 day blastocysts from C57B1/6 mice as described previously (Helgason et al. 1998) and implanted back into pseudopregnant recipient females and allowed to gestate normally. Coat color was used to identify chimerism of the resulting pups. 2.2.4. Embryoid Body Formation Assays Single cell suspensions were collected on day 0 or prepared, as outlined above, during differentiation. Defined numbers of cells (500 - 20,000 depending on time following L I F removal) were plated in 35 mm petri-style dishes in the E S C differentiation methylcellulose medium described above to determine the efficiency of E B formation. E B numbers were determined microscopically following 5 - 6 days of culture and colonies were qualitatively scored as large or small. E B formation efficiency was calculated by dividing the total number of E B formed by the number of cells plated multiplied by 100. 2.2.5. Colony Forming Cell Assay Single cell suspensions collected oh day 0 or during differentiation were plated at various. \ . . . densities (500 - 20,000 cells per gelatinized 60 mm gridded tissue culture dish) to determine E S C C F C plating efficiency. Colonies were microscopically enumerated following 5 - 6 days of growth. To enable differential assessment of the colonies the protocols outlined in the A P detection kit (Sigma; 86-R) were modified for staining in 60 mm dishes. In brief, the media was removed from the colonies and 1 m L of room temperature fixative (prepared as per the Sigma protocol) was added for 30 seconds. The fixative was removed and colonies were washed with 2 m L phosphate buffered saline (PBS). Next 1.5 m L of alkaline dye mixture (prepared as per the 42 Sigma protocol) was added; the dish was incubated in the dark at room temperature for 15 minutes. Finally, the dye mixture was removed and the colonies were covered with 2 m L P B S for microscopic evaluation. The numbers o f stained (undifferentiated) vs. unstained (differentiated) colonies were determined. C F C plating efficiency was calculated by dividing the total number of A P positive colonies by the number of cells plated multiplied by 100. 2.2.6. RNA Extraction and Array Hybridization Single cell suspensions of test cells were prepared as described above and resuspended in Trizol (Invitrogen, Burlington, ON) at a density of 10 7 cells/mL. R N A was extracted following the manufacturer's instructions. Standard Affymetrix amplification protocols were used to prepare probe R N A for Affymetrix arrays with 5 pg of starting total R N A . Biotin-labeled amplified R N A was fragmented and hybridization cocktails were prepared according to the Affymetrix protocol. The mouse GeneChip M G _ U 7 4 v 2 chips were hybridized on a GeneChip System (Affymetrix) at the Genome Science Centre, B C Cancer Agency (Vancouver, Canada), according to the manufacturer's instructions. A l l experiments were performed in triplicate with the exception of P E F , which were analyzed in duplicate. The M I A M E (minimal information about a mieroarray experiment) guidelines were followed for the presentation of the data (Brazma et al. 2001). 1 2.2.7. Data Analysis The Affymetrix Mic ro Array Suite ( M A S ) 5.0 software was used to generate absolute expression estimates (Absence/Presence calls). Software default thresholds were used to determine the present or absent calls (cti=0.04, ci2 = 0.06 and x=0.015). Raw data (CEL-fi les) obtained from 43 M A S was then normalized and analyzed in GeneSpring (Silicon Genetics, Redwood City, C A ) . Data were normalized as follows: values below 0.01 were set to 0.01 and then each measurement was divided by the 50th percentile of all measurements in that sample. Each gene was divided by the median of its measurements in all samples. If the median of the raw values was below 10 then each measurement for that gene was divided by 10. We judged genes to be differentially expressed during E S C differentiation only when 1) the difference in expression between two time points was at least 2-fold; 2) the gene was identified as present in two out of three replicates or present or. marginal in al l three replicates by M A S in at least one of the time points. Addit ionally, genes scored as decreasing were removed that were scored as absent in the denominator of the fold change, as well as genes scored as increasing but absent in the numerator of the fold change; and 3) the extent of difference in expression was statistically significant (p<0.05, in parametric Welsh t-test). Classification of genes into functional categories was done by collecting annotations and keywords with the Onto-Express Tool (Draghici et al. 2003), Affymetrix NetAffx and the Simplified Gene Ontology.Tool in GeneSpring. The GenMapp 2.0 software tool was used to analyze signaling pathways (Dahlquist et al. 2002). 2.2.8. Quantitative RT-PCR R N A was isolated using Trizol and the samples were then treated with DNase I (amplification grade) prior to the R T - P C R according to the manufacturer's instructions (Invitrogen, Burlington, O N ) . c D N A was generated by reverse transcription (RT) with random primers and the Superscript II enzyme and RNase inhibitor (Invitrogen Life Technologies, Burlington, ON) . The R T reaction was incubated at 42°C for 50 min followed by 15 min at 70°C. c D N A was stored at -20°C for subsequent Q - P C R analysis. Gene transcripts were quantified by real-time P C R using 44 the iCycler apparatus (Bio-Rad Inc., Hercules, C A ) and were detected.with S Y B R Green as flurochrome (IQ S Y B R Green Supermix, BioRad Inc.). Gene sequences for primer design were obtained from the N C B I Reference Sequences database. Primers were chosen using the Primer3 software (Rozen and Skaletsky 2000). Primer sequences are provided in the Appendix A . Synthesis of the primers for this study was performed at Invitrogen. B L A S T searches were conducted to confirm gene specificity of the P C R . Relative expression changes were determined with the 2 " a a C T method (Livak and Schmittgen 2001) and the gapdh gene transcript was used to normalize the results. P C R efficiency was tested for each primer pair by dilution series of c D N A to make sure that the efficiency was appropriate for the 2" A A C T method (i.e. 95% or above). To identify amplification of any contaminating genomic D N A and ensure the specificity and the integrity of the P C R product, melt curve analyses were performed on all P C R products. N o products were obtained with real-time P C R from R N A samples when reverse transcription was omitted. Samples without template were included for each primer pair to identify contamination. Pearson correlation and Deming regression analysis were used to determine correlation and agreement respectively between the microarray and Q - R T - P C R results. 2.2.9. Flow Cytometry Antibodies used for phenotype analysis included phycoerythrin (PE)-conjugated anti-CD 117 (c-kit, clone A C K 4 5 ; B D Pharmingen, San Diego, C A ) , as wel l as purified anti-SSEA-1 (Clone M C - 4 8 0 ; Chemicon International Inc., Temecula, C A ) that was detected using a F I T C -conjugated anti-mouse I g M antibody ( B D Pharmingen). Single cell suspensions collected on day 0 or prepared from differentiating E S C were blocked for 10 minutes on ice with 5 pg/ml anti 45 mouse CD16/CD32 (Fc Block, B D Pharmingen) in P B S + 2% Fetal Bovine Serum (PBSFBS) . Cells were washed once with P B S F B S and then incubated on ice for 20 minutes with the primary antibody. Cells were then washed once, incubated with the secondary antibody i f needed, washed again, and then analysed by flow cytometry using a F A C S C a l i b u r flow cytometer and CELLQues t software ( B D Pharmingen). The forward scatter (FSC) versus side scatter (SSC) profile was used to gate on viable cells and an unstained sample was used to determine appropriate gating for quantification o f expression. Cel ls to be stained for Oct4 were resuspended in 100 pi of Hanks Buffered Saline Solution + 2% F B S (HF) and fixed with 100 p i of IntraPrep Permeabilization Reagent 1 (Immunotech, Westbrook, M E ) for 15 minutes at room temperature. Ce l l s were then washed wi th H F and permeabil ized wi th IntraPrep Permeabilization Reagent 2 for 5 minutes before incubation with a 1:100 dilution of mouse anti-mouse Oct3/4 monoclonal antibody (Transduction Laboratories, Lexington, K Y ) for 15 minutes at room temperature. Cells were washed with H F before staining with A P C labeled anti-mouse IgGi (BD Pharmingen). Samples were analyzed by flow cytometry as outlined above. 2.2.10. Comparison with Previously Published Gene Expression Array Data Genes in Table B l (decreased between 0 and 18 hours), B3 (decreased between 0 and 72 hours) and B5 (decreased between 18 and 72 hours) (Tables in Appendix B) were compared with the following data tables: (Bhattacharya et al. 2004), supplementary online Table 2; (Brandenberger et al. 2004), supplementary online Table 2; (Ivanova et al. 2002), supplementary online Table 1 (genes marked as either I or D); (Ramalho-Santos et al. 2002), database S I ; (Sato et al. 2003), database 1; (Sharov et al. 2003), dataset S7; (Sperger et al. 2003), supporting supplementary online Table 6; (Kel ly and Rizzino 2000), supplementary online Table 1; and (Ginis et al. 2004), 46 supplementary online Table 3 A . Comparison between (Ivanova et al. 2002) and (Ramalho-Santos et al. 2002) was performed on the basis of Affymetrix gene IDs. In (Ramalho-Santos et al. 2002), supplementary online Table 1 was refiltered as outlined in the original publication. Comparisons with human ES data sets (Bhattacharya et al. 2004; Brandenberger et al. 2004; Sato et al. 2003; Sperger et al. 2003) was done by converting public ID references from the human d a t a se ts i n t o m u r i n e A f f y m e t r i x c o d e s u s i n g E n s M a r t (http://www.ensembl.org/Multi/martview). For (Sharov et al. 2003), custom U codes were c o n v e r t e d to U n i g e n e codes u s i n g the t r a n s l a t i o n t o o l a v a i l a b l e (http://lgsun.grc.nia.nih.gov/geneindex/). Groups F , I, L , and O were chosen because these were the lists that contained genes expressed in E S C grown in L I F but not expressed in E S C grown without L IF for 4 or 18 hours. For (Kel ly and Rizzino 2000), Genbank codes were converted to Affymetrix codes using EnsMart. Fol lowing automated comparisons performed as described above, manual comparison between known genes was performed using gene names to ensure accurate comparison. 2.3. Results 2.3.1. Pluripotency During Early Differentiation We hypothesized that loss of E S C pluripotency as defined by measurable functional readouts correlates with significant alterations in gene expression. Identification of these gene expression changes would thus provide important insights into the genetic regulation of E S C pluripotency and facilitate the identification of new molecular markers of the undifferentiated E S C state. We relied on comparisons of three measures of E S C potential (chimera formation, E B formation, and 47 C F C assay) to select time points for analysis of gene expression profiles. Undifferentiated E S C were cultured on P E F which supply a number of as yet unidentified factors that enhance the plating efficiency of the E S C , as well as assist in the maintenance of the undifferentiated state. Pre-plating of E S C removes the P E F and enriches for E S C with the capacity to contribute to the developing blastocyst. It has been suggested that those E S C capable of loosely attaching to the feeder layer in a short (i.e. one hour) period of time have the highest likelihood of forming self-renewing colonies and thus may also be the most competent at contributing to the germ-line following blastocyst injection (Stewart 1993). We thus elected to use only pre-plated, loosely adherent mouse E S C for our analyses. The baseline activities for undifferentiated E S C in the three different assays were first determined. 100% of the pups resulting from injection of undifferentiated, pre-plated R l E S C into recipient blastocysts were chimeric (27 blastocysts injected; 6 born and analyzed), 6.90 ± 0.02 % of plated cells differentiated into embryoid bodies in the E B formation assay and 12.50 ± 0.04% of plated cells gave rise to A P positive E S C colonies in the C F C assay. These results were within the normal range for R l E S C . E S C differentiation was then initiated by L I F removal and re-plating without P E F . The morphology of the E S C before and after L I F removal is shown in Figure 2.1. A t 18 and 24 hours after P E F and L I F removal, the E S C looked very similar to one another and did not exhibit any clear signs of differentiation. Morphological differentiation was first, apparent at 48 hours and E B could be seen after 72 hours. Despite the lack of appreciable phenotypic differentiation within the first 24 hours, there were significant changes in the functional properties of the E S C . Cultured cells were harvested at subsequent 48 time points and tested in the three different assays. The blastocyst injection assay showed a rapid decrease in the number of chimeras obtained after initiation of E S C differentiation (Figure 2.2A). Only 28% of born pups (84 injected, 25 born and analyzed) were chimeras when E S C had differentiated for 24 hours and less than 5% after 72 hours of differentiation (66 injected, 29 born, 3% chimeras). The E B formation assay (Figure 2.2B) showed approximately 5.5% ± 1.0%, 3.7% ± 0.3% and 0.3% ± 0.1 % readout at 18, 24 and 72 hours after L I F removal respectively. In contrast, the frequency of cells replating in the C F C assay increased slightly during the first 24 hours but declined rapidly such that only 1.3% ± 0.3%> retained activity at 72 hours (Figure 2.2C). In summary, al l three assays showed clearly that most differentiation potential and self-renewing capacity was gone after 72 hours of differentiation. However, there was a high degree of variation during the first 24 hours amongst the in vitro and in vivo assays, with the E B formation assay correlating most closely with chimeric mouse formation. For gene expression profiling, because both the E B formation and chimera assays showed a pronounced decrease within the first 18 - 24 hours, we elected to use the earlier time point, along with 72 hours of differentiation, to compare against undifferentiated R l E S C . In parallel with functional assay analysis, the expression patterns of two markers commonly used to identify undifferentiated E S C , Oct4 and S S E A - 1 , were also analyzed to establish the level of correspondence in the readouts using each method. Although Oct4 and S S E A - 1 have been extensively used in E S C research, their expression patterns during E S C differentiation have not been studied in detail. Oct4 is used as a marker for E S C because of its requirement in E S C self-renewal (Nichols et al. 1998). The precise expression levels of oct4 are important for 49 determining E S C fate, and repression of oct4 induces loss of pluripotency and differentiation into trophoectoderm (Niwa et al. 2000). However, forced, constitutive expression of oct4 cannot prevent E S C differentiation, and a less than two-fold increase in expression causes differentiation into primitive endoderm and mesoderm. Thus, a critical amount of oct4 is required to sustain stem cell self-renewal but is not sufficient to prevent differentiation. SSEA-1 is a glycoprotein expressed during early embryonic development and by undifferentiated E S C . However, the precise role of S S E A - 1 in pluripotency and self-renewal has not been defined. E S C selected for expression of SSEA-1 and P E C A M 1 are enriched for cells that differentiate predominantly into epiblast cells in chimeric embryos (Furusawa et al. 2004). In the present study, protein expression of both SSEA-1 and Oct4 remained relatively unchanged throughout the first 72 hours of differentiation (Figure 2.2D). A t 120 hours following L IF removal, 18.5 ± . 1.1% of the cells still retained expression of S S E A - 1 , while 40.6 ± 6.6% of the cells continued to express Oct4. Thus there was no clear correlation between expression of these two markers and the E S C functional assays. 2.3.2. Array Analysis R l E S C from matched passage numbers in three separate experiments were collected at 0, 18, and 72 hours after removal of L IF and analyzed using Affymetrix GeneChip M G _ U 7 4 v 2 arrays containing 36,767 probe sets. Duplicate samples of P E F were also collected and analyzed to assess possible contamination in the undifferentiated E S C gene expression, samples. Hybridization, scanning, and production of raw data files were performed according to standard protocols. M A S software was used for initial scaling and expression analysis and data was then normalized and further analyzed in GeneSpring. To validate array reproducibility and the overall 50 variation of the data, hierarchical clustering analysis was performed. The data was first filtered for genes expressed in at least one of the three time points (present or marginal in all three replicates) resulting in 13,002 different probe sets. The average number of expressed genes for each time point was 12,818 (coefficient of variation, CoVar, 6%) in undifferentiated E S , 11,806 (CoVar 4%) at 18 hours and 12,297 (CoVar 1%) at 72 hours after L IF removal. The number of expressed genes at the different time points was not statistically significantly different. Hierarchical clustering was applied to the reduced gene set on individual array samples (three replicates for each time point) using Pearson correlation and average linkage clustering as implemented in GeneSpring. Individual replicates clustered tightly together according to their respective time points as seen in Figure 2.3 indicating that overall inter-experimental variation was low. Clustering results also reflected the temporal progression of the E S C differentiation and, as expected, the first two time points (0 hours and 18 hours) clustered more closely together while the 72 hour time point showed a more distinct expression pattern as reflected in the larger distance from the other two time points. These observations are also consistent with functional data (Figure 2.2). Genes that were differentially expressed during differentiation were determined using the following three criteria: 1) the difference in expression between two time points was at least 2-fold; 2) the gene was expressed in all three replicates and in at least one of the time points as determined by the detection algorithm in M A S 5.0; and 3) the extent of difference in expression was statistically significant (p<0.05 in parametric Welsh t-test). Additionally, decreased genes were removed that were scored as absent in the denominator of the fold change, as wel l as 51 •J increased genes scored as absent in the numerator of the fold change. Wi th this approach 473 unique genes were identified as significantly differentially expressed (275 decreased, 194 increased and 4 both decreased and increased) during early E S C differentiation (Tables B l - B6 , Appendix B) . Unigene and RefSeq IDs were used in the analysis to exclude redundant genes included in the array probe sets. To exclude the possibility that contaminating P E F cells could distort the data, duplicate samples of P E F were analyzed for gene expression on the M G _ U 7 4 v 2 GeneChip. Data was scaled in M A S and normalized in GeneSpring (as for E S C data) before 1% of the raw value (five times more than the maximum estimated contamination) obtained for the gene in P E F was subtracted from the value of the same gene in undifferentiated R l E S C . The analysis steps to find differentially expressed genes were then performed as described above. Only genes detected as decreased during differentiation in the original analysis were re-analyzed. Genes whose expression increased during differentiation would not be affected by P E F contamination in a way that, would give false positive results. This indicated that none of the genes that showed a significant 2-fold or greater decrease in the initial analysis lost their significance after P E F subtraction (data not shown) and suggests that the levels of P E F contamination were not sufficient to distort our E S C gene expression data. 2.3.3. Gene Expression Validation The expression patterns of several genes previously implicated in maintaining E S C pluripotency and self-renewal were analyzed to validate our data and the approach used to identify differentially expressed genes (Table 2.1). L I F binds to the gp l30 receptor that leads to 52 activation of the transcription factor S T A T 3 (Smith 2001). Stat3 was clearly detected in undifferentiated R l E S C and L I F removal decreased stat3 expression with the most pronounced effect during the first 18 hours (49% reduction). This was also seen for two other genes involved in the L I F / g p l 3 0 pathway, the Oncostatin M receptor, osmr, and the Interleukin 6 signal transducer, U6st. A known S T A T 3 target gene, piml (Shirogane et al. 1999), was also decreased significantly at the transcriptional level. Taken together, these observations confirm that the L I F / g p l 3 0 / S T A T 3 pathway was rapidly shut down following P E F and L I F removal. The bone morphogenetic proteins can act in combination with L I F to sustain self-renewal and preserve multi-lineage differentiation, chimera colonization, and germline transmission properties (Ying et al . 2003). Bmp4 transcript levels were significantly decreased during differentiation. However, the onset o f the decrease was later than for stat3 (between 18 and 72 hours, 80% reduction). The same was seen for the zfp42 gene, known to be highly expressed in undifferentiated E S C and down regulated after retinoic acid induced differentiation (Hosier et al. 1989; Rogers et al. 1991). A 90% reduction in zfp42 expression was seen between 18 and 72 hours after LIF removal. Transcript abundance of the akp2 gene coding for A P , commonly used as a marker for undifferentiated E S C , was also decreased significantly (55%) reduction between 18 and 72 hours). Oct4 was highly expressed in undifferentiated E S C but was not changed significantly during the first 72 hours after L IF removal consistent with the protein expression data (Figure 2.2D). Transcript levels o f the embryonal stem cell-specific gene 1 (Esgl or developmental pluripotency associated gene 5, dppaS) did not change during differentiation. This is consistent with the observation that esgl is a downstream target of oct4 and is downregulated slowly after oct4 suppression (Tanaka et al. 2002). The abundance offoxd3, a 53 gene expressed early in mouse embryonic development, also remained unchanged during the first 72 hours after L I F removal. The SRY-box containing transcription factor sox2 may act to maintain E S C pluripotency and is expressed in the I C M , epiblast and germ cells just like oct4 (Avi l ion et al. 2003). Sox2 was significant decreased (60% reduction between 18 and 72 hours, p=0.0136). Sox2, together with oct4, is involved in the regulation of Fibroblast growth factor 4 (fgf4), another factor confined to the I C M of the blastocyst (Avi l ion et al. 2003). Although fgf4 expression was reduced, (52% reduction between 18h and 72h) the change was not statistically significant (p=0.09). T h i s was also the case for the newly discovered marker o f E S C pluripotency dppa3 (Bowles et al. 2003) which decreased 78% between 0-and 72 hours (p = 0.073). Similarly, the recently discovered regulatory factor Nanog (Chambers et al. 2003; Mitsui et al. 2003), which can rescue E S C from L I F / S T A T 3 dependence and maintain oct4 expression, did not show a significant 2-fold change during the first 72 hours o f differentiation (40% reduction, p=0.07). Several genes indicative of E S C differentiation were also increased during differentiation. For example, the mesoderm marker brachyury (Wilkinson et al. 1990) increased 15-fold between 18 and 72 hours (p=0.003) and the ectoderm marker nestin (Wiese et al. 2004), increased approximately 3-fold between 18 and 72 hours (p=0.04). The epithelial cell marker promininl (Weigmann et al. 1997) increased 5-fold between 18 and 72 hours (p = 0.001). In conclusion most of these changes are consistent with the loss of E S C pluripotency as measured in all three assays and indicate that the thresholds used to define differentially expressed genes in this study are reasonable. 54 2.3.4. Quantitative RT-PCR validation To further confirm the fidelity of the microarray data, a set of 28 genes was selected and their transcript levels were tested using Q - R T - P C R . The genes selected included some of the genes mentioned above, as well as other genes increased, decreased, or unchanged after L I F removal according to the array results. The fold change was calculated between 0 and 18 hours, 18 and 72 hours, and 0 and 72 hours respectively. The analysis revealed a strong correlation between the array data and the Q - R T - P C R results (Pearson's correlation, r=0.87, Figure 2.4A), with a tendency for Q - R T - P C R results to show greater changes than the array (proportional bias 1.4 in Deming regression analysis). To determine i f the gene expression changes observed in the R l E S C are general and more broadly observed in multiple E S C lines, the expression patterns of 15 genes were followed for 72 hours after L IF removal in two other E S C lines, J l and E F C , and compared with the R l E S C line using Q - R T - P C R (Table 2.2). Overall, there was strong correlation between all three E S C lines (Figure 2.4 B-D) . Specifically, Q - R T - P C R analysis was able to verify the expression changes of genes such as lox, ankrdl and c-kit that showed significantly decreased transcript levels after 18 hours in all three E S C lines tested. Similarly, zfp42, sox2, leftb and mtf2 were all changed in a similar fashion in the R l , J l and E F C cell lines. Q - R T - P C R results also confirmed that oct4 levels were not decreased significantly during the first 72 hours of differentiation in any of the cell lines, in agreement with array results and observed protein levels (Figure 2.2D). In contrast, although the decrease in nanog was not significant in R l E S C by array, the Q - R T - P C R results indicated a consistent reduction in all cell lines. Only two of the tested genes showed different 55 expression patterns amongst the cell lines. The mesoderm marker brachyury and the homeobox transcript factor pbx3 were both increased during differentiation in the R l and J l E S C but not in the E F C cell line. This observation suggests that the various E S C lines might follow slightly altered differentiation pathways upon LIF removal. These validation analyses demonstrated that the early changes during differentiation uncovered by the array analysis could be verified with an independent method and could in most cases be observed in multiple E S C lines. Taken together, they provide confidence in our approach to identify differentially expressed genes with a high likelihood of exhibiting true expression level changes during the loss of E S C pluripotency. 2.3.5. Functional Classification of Differentially Expressed Genes We hypothesized that genes showing similar functional properties and expression patterns form interacting networks that contribute to the phenotypic and functional characteristics of the cells of interest. Functional classification of differentially expressed genes into appropriate biological processes was thus performed using NetAffx and GeneOntology Express as well as a simplified gene ontology tool in GeneSpring. The differentially expressed genes were separated into 13 main categories (Figure 2.5A). Many differentially expressed genes were, as expected, classified as being involved in development and differentiation (61 genes, 10%) or directly or indirectly involved in cell cycle control and cell proliferation (32 genes, 5%). Furthermore, at least 81 genes (13%) were classified as being involved in intracellular signal transduction, cell-cell signaling and/or response to external stimuli. 56 A closer look at the signaling pathways affected in E S C during differentiation revealed that the Yamaguchi sarcoma viral oncogene gene (yes) was decreased 60% between 0 and 72 hours. Yes has been shown to be regulated by LIF and to be important for E S C self-renewal (Anneren et al. 2004). Yes is a gene coding for a Src tyrosine kinase expressed in both mouse and human E S C and is down-regulated when these cells differentiate (Anneren et al. 2004). The Notch ligand, jagged-1, previously implicated in H S C self-renewal (Vas et al. 2004) exhibited a 55% decrease in expression between 0 and 18 hours. Self-renewal of E S C is influenced by the M A P K pathway where expression of E R K and SHP-2 counteract the proliferative effects of S T A T 3 and promote differentiation (Burdon et al. 1999). The Erk gene (mapkl or mapk3) and most other components in this pathway were expressed in undifferentiated E S C according to array results and their expression levels did not change significantly at the transcription level during the first 72 hours of differentiation (data not shown). Furthermore, two genes involved in the SHP-2 pathway, kras2 and mapkl2, were increased significantly during differentiation indicating an activation of this pathway. The gene coding for Grb2-associated binder 1 (gabl) was significantly decreased (58% reduction between 18 and 72 hours). Gab l binds to SHP-2 and is believed to suppress the M A P K pathway in E S C . Increased synthesis of Gab l together with Oct4 may suppress induction of differentiation (Smith 2001). Although most of the factors involved in the M A P K pathway are primarily regulated at the post-transcriptional level, the fact that most of the factors were detected in undifferentiated E S C , at least at the transcriptional level, suggests that undifferentiated E S C already harbour the required components to quickly respond to signals that promote differentiation. 57 The Wnt Signaling pathway is important for the maintenance of pluripotency in E S C and a recent study also showed that activation of the Wnt pathway by pharmacological inhibition of G S K - 3 was able to maintain pluripotency in both human and mouse E S C (Sato et al. 2004). Most components of the Wnt signaling pathway could be detected in both undifferentiated and differentiated E S C and some factors were significantly changed during differentiation (e.g.fzd5, wnt3, cyclind3). However, overall there was no consistent change at the transcription level o f several factors belonging to this pathway during the first 72 hours of differentiation (e.g. fi-catenin, gsk-3, axin, data not shown). The Hedgehog signaling pathway plays a critical role during development and has been extensively studied in multiple species and tissues (Cohen 2003; Ogden et al. 2004). Four components in this pathway, two receptors, ptchl and ptch2, and two transcription factors, glil and gli2, were decreased significantly during E S C differentiation. Furthermore, bmp4, a putative downstream target of this pathway was significantly decreased (discussed above) providing further evidence that this pathway might be involved in maintaining E S C self-renewal and pluripotency. Importantly, out of the 473 differentially expressed genes, the largest group of 173 genes had no known biological function. Further analysis of these genes, especially those whose expression changes closely correlate with loss of pluripotency might reveal novel mechanisms involved in E S C maintenance. 58 O f the 275 genes significant decreased during E S C differentiation, 48 genes showed a more than twofold decrease in transcript levels within the first 18 hour and continued to decrease or remained at a low level of expression until 72 hours (Table 2.3). Changes in the transcript levels for these genes correlated well with the functional assays (Figure 2.2). A t least half of them have been identified as ESC-enriched or downregulated during differentiation in previous studies (Brandenberger et al. 2004; Ivanova et al. 2002; Ke l ly and Rizzino 2000; Ramalho-Santos et al. 2002; Sato et al. 2003; Sharov et al. 2003; Sperger et al. 2003) (see below), but several of these genes have unknown biological function or have not previously been implicated in E S C maintenance (e.g. lox, tnc and jagl; Table 2.5). 2.3.6. Molecular Markers of Pluripotency We hypothesized that classification of differentially expressed genes according to cellular component, in combination with functional analyses, may lead to the identification of new markers of E S C potential (Figure 2.5B). A t least 60 gene products (11%) were considered localized to the plasma membrane and might therefore be good candidates as markers for E S C pluripotency. Both genes either increased or decreased during differentiation were included in this list o f possible E S C markers since the former might be used to detect the onset of differentiation and/or for negative selection strategies for isolating undifferentiated E S C . Protein expression of some of the genes has already been examined in relation to E S C differentiation (i.e. C D 9 , CD44, Osmr) (Haegel et al. 1994; Oka et al. 2002; Rose et al. 1994; Wheatley and Isacke 1995). Additional gene products of interest include the adhesion molecular V c a m l and the hedgehog signaling pathway receptors P tch l and Ptch2. One gene o f special interest 59 encodes the SCF receptor, c-kit. Array results indicated that expression of the c-kit gene was one of the most significant changes during differentiation (54% reduction in expression between 0 and 18 hours and 75% reduction between 0 and 72 hours, p=0.0016) and this was confirmed by Q - R T - P C R in multiple cell lines (Table 2.2). SCF was not added to the cell culture media used in these experiments and S C F was not expressed in undifferentiated E S C according to array results. However, SCF was expressed at high levels in P E F (data not shown). F low cytometry revealed that expression of c-kit on the cell surface of differentiating E S C closely paralleled the loss of functional potential as measured using the E B formation assay (Figure 2.6, Figure 2.7). A t later time points the correlation was not as close (i.e. E B formation capacity continued to decrease while protein expression remained unchanged) suggesting the emergence of differentiated cell types expressing c-kit protein. 2.3.7. Comparison with Previously Published Gene Expression Array Data Gene expression profiling to determine regulatory factors and signaling pathways present in E S C has been performed extensively in recent years (Bhattacharya et al. 2004; Brandenberger et al. 2004; Ginis et al. 2004; Ivanova et al. 2002; K e l l y and Rizzino 2000; Ramalho-Santos et al. 2002; Sato et al. 2003; Sharov et al. 2003; Sperger et al. 2003; Tanaka et al. 2002). It has been hypothesized that the undifferentiated state is conferred on various stem cell populations through the use of similar molecular mechanisms. Initial support for this hypothesis came from experiments comparing the gene expression profiles of multiple stem cell populations (Ivanova et al . 2002; Ramalho-Santos et al. 2002; Tanaka et al . 2002). However,, further analysis of available data indicated minimal overlap between different published stem cell-associated gene sets (Evsikov and Solter 2003; Fortunel et al. 2003). The discrepancy observed amongst these 60 studies likely arises in large part from significant differences in the strategies used to identify the stem cell profile. However, true differences in stem cell biology probably also exist among stem cells taken from different tissues (e.g., E S C , neural stem cells, H S C ) . Some studies have relied on comparison of E S C to terminally differentiated tissues (Kel ly and Rizzino 2000; Sato et al. 2003; Tanaka et al. 2002) to establish a list of genes enriched in E S C . This strategy is likely to miss genes involved in pluripotency that are transiently turned off early during the differentiation process but not uniquely expressed in the stem cell population. Another strategy used has been to compare stem cell populations arising from the same tissue source across species. A n example is the comparison of mouse with human E S C . Although mouse and human E S C are both derived from preimplantation blastocysts, they differ in responsiveness to extrinsic signals and in expression of surface markers; for example, L IF cannot sustain self-renewal of human E S C , even in the presence of serum, suggesting the existence of other signaling pathways essential for self-renewal in human E S C (Thomson et al. 1998). The evidence that human and mouse E S C share a common core molecular program is also somewhat conflicting (Bhattacharya et al. 2004; Ginis et al. 2004; Sato et al. 2003). However, the molecular mechanisms that confer pluripotency are likely evolutionary conserved, and thus comparison of mouse and human E S C array data might still be informative. The gene expression data reported here was compared with nine different array data sets, comprising four studies on murine E S C (Ivanova et al. 2002; Ke l ly and Rizzino 2000; Ramalho-Santos et al. 2002; Sharov et al. 2003), four studies on human E S C (Bhattacharya et al. 2004; Brandenberger et al. 2004; Sato et al . 2003; Sperger et al. 2003), and one study comparing 61 human and mouse E S C (Ginis et al . 2004) (summarized in Table 2.5) . Only genes downregulated during differentiation after L I F removal were used in the comparisons because most other studies only report ESC-enriched or ESC-specific genes. Complete gene lists derived from these comparisons can been found in Table B7 , Appendix B . Overall , there was a large overlap between the downregulated genes from our data and genes identified as ESC-enriched in the other data sets, despite the wide variety o f experimental designs and cell lines used. O f the genes identified as significantly decreased after L I F removal in our study, 60% (164 of 275) were found in at least one other data set, and 28% (78 of 275) were found in at least two other data sets. There was greater similarity with the murine data (129 of 275 or 47%o were found in at least one other murine data set) than with the human data (75 of 275 or 27%) were found in at least one other human data set). Genes that we identified as differentially expressed between 0 and 18 hours were the least represented in other published data (41%o between 0 and 18 hours compared with 67% between 18 and 72 hours and 63% between 0 and 72 hours), possibly because our approach to use early time points during differentiation for defining differential expression has not been widely used. Only two other studies included comparable early time points in their study (Kelly and Rizzino 2000; Sharov et al. 2003). Within individual data sets, the degree of overlap depended primarily on the number of genes in the starting list (Table 2.5). The data set generated by Ke l ly et al. (Kelly and Rizzino 2000) stood out because of its high degree of overlap (35%>), which can potentially be explained by the similarity in the two studies in terms of experimental design and the similarity of the genetic background of the E S C used. However, this encompasses relatively few genes (6 out of 62 17 genes). Furthermore, many genes that we report as decreased during the first 72 hours were reported as also being present in human E S C by Sato et al. (Sato et al. 2003). However, they, were not identified as differentially expressed in that study. This discrepancy could possibly be due to the later time point (3 weeks) used for determining differential expression in that study (data not shown). N o single gene was found enriched in every human and mouse E S C data set. Jadel, leftb, and smarcadl were the most commonly identified genes in other data sets (i.e., in six of nine other data sets). Leftb is a TGF(3 family member, which is expressed on the left side of developing mouse embryos and is implicated in left-right determination (Meno et al. 1996). Jadel was identified as a gene involved in anteroposterior axis development (Tzouanacou et al. 2003). Smarcadl has previously been identified as a marker of preimplantation embryos (Schoor et al. 1993). The gene coding for tenascin was downregulated within the first 18 hours after L I F removal (Table 2.3) and was also commonly observed as differentially expressed in other data sets (i.e., in four of eight data sets). Tenascin, also known as hexabrachion and cytotactin3, is an extracellular matrix protein with a spatially and temporally restricted tissue distribution that is tightly regulated during embryonic development and in adult tissue remodeling (Jones and Jones 2000). Other commonly identified markers of undifferentiated E S C were also found in this comparison. Sox2, zfp42, bmp4, and gbx2 were all observed in at least three other data sets. This also included the gene lox, coding for lysyl oxidase, found in three other murine data sets (Ivanova et al. 2002; Ramalho-Santos et al. 2002; Sharov et al. 2003). It was in fact one of the most pronounced and rapidly downregulated genes detected (Table 2.3). A t least one study 63 identified c-kit as enriched in human E S C (Sperger et al. 2003). Overall the comparison with previously published data indicated some intriguing similarities but also showed that most genes do not overlap in pairwise comparisons. This underlines the many differences that might arise when different sources of cells, culture conditions, microarray technique platforms, and data analysis tools are used and emphasizes the need to correlate expression changes with rigorous measures of E S C competence and differentiation potential. 2.4. Discussion The present study is distinct from previous studies in two important aspects. First and most important, only a selected population of germ line-competent E S C , grown under carefully controlled, optimized culture conditions, was used to establish the gene expression profiles. It is l ikely that substantial variations in gene expression arise in response to culture variables. Considering the multiplicity of culture variables that can be important for the biological heterogeneity of cell populations and their gene expression profiles, remarkably little information has been provided about the conditions used to generate the cells for the gene expression profiles reported thus far (Bhattacharya et al. 2004; Brandenberger et al. 2004; Ginis et al. 2004; Ivanova et al. 2002; Ke l ly and Rizzino 2000; Ramalho-Santos et al. 2002; Sato et al. 2003; Sharov et al. 2003; Sperger et al. 2003; Tanaka et al. 2002). Second, few of the previous studies addressing questions about a shared or common stem cell gene expression signature among different types of stem cells or different E S C lines, have involved correlative functional assays to assess the pluripotency and self-renewal capacity of the cells o f interest. B y combining the gene expression profiling data with assays measuring E S C pluripotency and self-renewal, it should be 6 4 possible to more precisely define the genes critical for specifying these properties. We combined defined culture and differentiation conditions with various measures of E S C pluripotency to determine optimal times for gene expression analysis using high-density oligonucleotide mieroarray. The results showed that the functional capacity of E S C declines rapidly under differentiation conditions, with the most pronounced changes in function occurring during the first 18 - 24 hours of differentiation. Array results were subsequently used to find genes whose expression correlated with loss of pluripotency and that could explain loss of this capacity during differentiation or serve as markers for pluripotent E S C . Analysis of the array data revealed that 473 genes were differentially expressed during the first 72 hours of E S C differentiation, suggesting they are potentially important for the maintenance o f E S C pluripotency and self-renewal. Important roles in the maintenance of undifferentiated E S C have previously been demonstrated for several of the downregulated factors such as stat3, zfp42, sox2, gbx2, and bmp4. Intriguingly, one third o f the differentially expressed genes have not been characterized in terms o f involvement in biological processes. More importantly, a refined list o f 48 genes whose transcript levels closely correlated with the functional assays (i.e., with early and persistent decreases in transcript levels after L IF removal) contains several genes with unknown function or genes that have not previously been suggested to play a role in E S C maintenance (Table 2.3). These genes may be novel candidates that play crit ical roles in the regulation of E S C pluripotency and self-renewal. Several of the genes in this list have also been found in previous 65 gene expression studies in human or mouse E S C (e.g., tnc and lox; Table 2.3). Furthermore, several promising candidate markers for pluripotent E S C were identified from the array analysis (Table 2.4). The changes in transcript levels observed for one gene, c-kit, were also verified at the protein level and showed good correlation with functional measures of E S C pluripotency (Figure 2.7). This finding was verified in two other murine E S C lines (Table 2.2). Two other commonly used E S C markers, Oct4 and S S E A - 1 , were also analyzed in parallel with the functional assays used in this study but showed poor correlation with the outcome in these (Figure 2.2D). Further studies are needed to determine i f c-kit is a more reliable marker of undifferentiated E S C and, more importantly, whether it can be used to enrich for these cells. Additional work is also required to determine the functional significance of these observations for E S C maintenance. It is also important to note in the context of these and other genes of interest that global gene expression profiling does not discriminate between changes arising at the level of transcription versus m R N A stability. Also , changes in transcript levels in a subset of cells may go undetected, and likewise subtle changes might arise from changes in just a small proportion of the cells. Transcriptional profiling of various stem cell populations has been used to determine the types of regulatory factors and signaling pathways present in pluripotent cells, including E S C (Bhattacharya et al. 2004; Brandenberger et al. 2004; Ginis et al. 2004; Ivanova et al. 2002; K e l l y and Rizzino 2000; Ramalho-Santos et al . 2002; Sato et al . 2003; Sharov et al. 2003; Sperger et al. 2003; Tanaka et al. 2002). Comparing genes that were significantly decreased in our analysis with both murine and human data sets (Bhattacharya et al. 2004; Brandenberger et 66 al. 2004; Sato et al. 2003; Sperger et al. 2003) gave a large overlap with 176 of 298 genes identified in at least one of these other studies. O f note, 139 of these have been identified in other murine E S C studies. This analysis was not able to distinguish biological differences between human and murine E S C because of the way it was performed. The level of overlap between our data and other murine studies was higher than the overlap with other human studies, but the difference could possibly be accounted for by the increased difficulty of gene comparison between species. However, additional E S C gene expression profiling carried out in an similarly stringent manner (i.e., correlating measurable functional properties with transcript levels) is needed to accurately assess and determine the confidence in the degree of overlap in differentially expressed genes after E S C differentiation. Induction of E S C differentiation by alternative methods or using E S C derived from mice of different genetic backgrounds w i l l also be important to properly assess the generality of these findings. Such analyses should be able to discriminate between treatment-specific or genetic-specific gene expression changes and those commonly observed under various conditions. These latter genes are more likely to play an important role in regulating E S C self-renewal and pluripotency. 67 Figure 2.1 Embryonic stem cell morphology during leukemia inhibitory factor removal Morphology of the R l E S C grown on P E F with L IF is shown. E S C were followed for 72 hours after P E F and LIF removal, and pictures were taken at the indicated times. ! ^* x200 > i o . ° x200 , >• •;. ' ( ,- -Q tf100 Abbreviations: E S C - embryonic stem cells, P E F - primary embryonic fibroblasts, L I F -leukemia inhibitory factor. 68 Figure 2.2, Assays of embryonic stem cell pluripotency fol lowing removal o f leukemia inhibitory factor Comparison of in vitro and in vivo functional measures of E S C potential with expression of the commonly used E S C markers Oct4 and S S E A - 1 . R l E S C were thawed, cultured and assayed as outlined in Materials and Methods. The frequency of cells capable of (A) generating chimeric mice (two independent experiments), (B) giving rise to embryoid bodies (three independent experiments), or (C) giving rise to colonies in the colony-forming cell assay (three independent experiments) is shown as a function of time after L I F removal. (D) F low cytometry (three independent experiments) was used to determine the percentage of cells expressing cell-surface SSEA-1 and intracellular Oct4. Data are the mean ± standard error of the mean for replicated experiments. Time (hours) Time (hours) Abbreviations: E S C - embryonic stem cells 69 Figure 2.3 Hierarchical clustering of array samples Pearson's correlation and average l ink ing clustering to determine reproducibili ty and interexperimental variability for array samples. Relative distance values are shown. 0.71 84 1.07 1.1 0.92 1.00 0.82 0.51 1 st experiment 0 hours 2nd experiment d j f f e r e n t i a t J o n 3rd experiment 3rd experiment 1 st experiment 2nd experiment 1st experiment 3rd experiment 2nd experiment 18 hours differentiation 72 hours differentiation 70 Figure 2.4 Correlation between array and quantitative R T - P C R results for multiple genes in multiple cells lines (A) Pearson's correlation analysis was done on the log ratio values obtained from the Q - R T - P C R and the array fold changes (n = 84). (B, C, D) Comparison of Q - R T - P C R results from the three different E S C lines are shown. Pearson's correlation analysis was done on the log ratios of the values obtained with the 2 " a a C T method (n = 45 in each comparison). B •2 0 2 Array log ratio ce u CL 10 8 6 4 2 0 -2 -4 •6 -8 -10 r = 0.94 -2 0 2 Rl - RT-PCR log ratio o a. 8 u u. '10 p 8 6 4 2 0 -2 -4 -6 -8 -10 r = 0.79 -2 0 2 Rl - RT-PCR log ratio I 10 8 -6 • 4' 2 0 -2 -4H -6 -8 -10 -8 r = 0.83 -4 -2 0 2 4 J1 - RT-PCR log ratio Abbreviations: Q - R T - P C R - quantitative R T - P C R , E S C - embryonic stem cells 71 Figure 2.5 Annotation of differentially expressed genes A l l genes differentially expressed during E S C differentiation were classified according to (A) biological process and (B) cellular component. Some genes are classified in more than one category, resulting in the total number of genes indicated in the figures being greater than the total number of differentially expressed genes. Development and differentiation, 61 10% Unknown, 173 27% Cell death, 14 2% Cell-cell signaling, 5 1% Signal transduction, 59 9% Response to external stimuli, 23 4% Cell growth and maintenance, 32 5% Transport, 32 5% Transcription, 84 13% Ceil proliferation, 32 5% Cell motility, 7 1% Cell adhesion, 17 3% Metabolism, 92 15% B Plasma membrane, 60 11% Unknown, 159 29% intracellular unspecified 35 6% Integral to membranes (other), 65 12% Extracellular, 42 8% Cytoplasm, 79 15% Nucleus, 103 19% 72 Abbreviations: E S C - embryonic stem cells 73 Figure 2.6 Flow cytometry profile to test for cell surface expression of c-kit following removal of leukemia inhibitory factor Undifferentiated R l E S C (0 hour) and E S C differentiated 18 and 72 hours after L I F removal were analyzed with flow cytometry using c-kit antibody. Negative control was unstained cells. 0 hour 18 hours 72 hours BsK'-'64% K;.. 29% p . ' • 1 8 % FSC Negative control c-kit Abbreviations: E S C - embryonic stem cells, L I F - leukemia inhibitory factor, S S C - side scatter, F S C - forward scatter. 74 Figure 2.7 Correlation between c-kit expression and embryoid body formation ability Expression of c-kit ( • ) was monitored at the indicated times during E S C differentiation using flow cytometry, and expression levels were plotted as percent gated cells versus time. EB-forming cell frequency ( • ) was determined in parallel. Time (hours) Abbreviations: E S C - embryonic stem cell, E B - embryoid body 75 Table 2.1 Expression change of genes previously reported as enriched in undifferentiated embryonic stem cells or to be markers of embryonic stem cell differentiation Gene Symbol Accession # Expression relative to 0 hours after LIT removal p-value 18 hours 72 hours STAT3 N M 011486 0.51 0.41 0.0128 . Osmr N M 011019 0.4 0.18 0.0321 Il6st N M 010560 0.62 0.4 0.0015 Piml N M 008842 1.08 0.36 0.0106 Bmp4 N M 007554 0.99 0.4 0.0051 Zfp42 N M 009556 1.08 0.12 0.0005 Akp2 N M 007431 1.24 0.55 0.0168 Oct4 N M 013633 1.09 1.01 0.4823 Sox2 N M 011443 0.9 0.41 0.0136 Fgf4 N M 010202 1.39 0.67 0.0902 Nanog X M 132755 0.82 0.59 0.069 FoxD3 N M 010425 1.74 0.95 0.549 Dppa3 N M 139218 1.06 0.24 0.0731 Dppa5 N M 025274 1.02 0.58 0.143 Brachyury N M 009309 0.48 7.3 0.0034 Nestin N M 016701 0.97 2.78 0.0384 Prominin 1 N M 008935 1.27 5.53 0.001 Relative expression levels compared with undifferentiated E S C are listed. The p values for the most significant fold change from Welsh t-tests are listed. Abbreviations: L I F - leukemia inhibitory factor. 76 Table 2.2 Comparison of relative expression levels obtained by quantitative R T - P C R in the J l , and E F C embryonic stem cell lines during differentiation R l , Gene symbol Accession # Expression relative to 0 Hrs after LIF removal R l ESC line Jl ESC line EFC ESC line 18 hours 72 hours 18 hours 72 hours 18 hours 72 hours Ankrdl N M 013468 0.09 0.07 0.02 0.05 0.03 0.002 Lox N M 010728 0.13 0.11 0.09 0.01 0.07 0.008 Kit N M 021099 0.35 0.17 0.22 0.12 0.36 0.12 Leftb N M 010094 0.58 0.41 ' 1.84 0.51 0.69 0.27 Piml N M 008842 0.81 0.48 0.66 0.39 0.48 0.22 Oct4 N M 013633 0.84 1.11 1.27 1.57 2 - 2.46 Nanog X M 132755 0.91 0.53 0.62 0.29 0.59 0.23 Piml N M 138606 0.87 4.14 1.87 9.19 1.32 17.15 Zfp42 N M 009556 0.9 0.05 1.07 0.01 0.97 0.005 Mtf2 N M 013827 1 0.3 0.93 0.13 0.68 0.1 Ptch N M 008957 1.15 0.19 2 0.13 1.68 0.23 Sox2 N M 011443 2.07 0.35 . 1.15 0.11 1.04 0.17 Hck N M 010407 2.3 0.05 1.11 0.01 1.19 0.03 Brachyury N M 009309 0.54 23.4 0.07 14.9 0.15 0.59 Pbx3 N M 016768 2.93 3.73 1.04 2.38 0.24 0.55 Abbreviations: L IF - leukemia inhibitory factor, E S C - embryonic stem cell.. 77 Table 2.3 Genes decreased during embryonic stem cell differentiation and most strongly correlated with the loss of pluripotency Expression at 18 and 72 hours is relative to 0 hours. Gene symbol Accession # 18 hours .72 Hours p value Decreased Vcaml N M 011693 0.08 0.08 0.0344 Cd44 N M 009851 0.08 0 .19 0.043 Cav N M 0 0 7 6 1 6 0.2 0.23 0 .0254 Empl N M 010128 0.22 0.09 0 .0302 Csfl N M 007778 0.36 0.49 0 .0337 Osmr N M 0 1 1 0 1 9 0.4 0 .18 0.0321 Abcal N M 013454 0.42 0.05 0.006 Jagl N M 0 1 3 8 2 2 0.45 0.47 0.0411 Kit N M 021099 0.46 0.23 0 .0016 Fzd5 N M 022721 0.47 0.31 0.0228 Il6st N M 010560 0.62 0.4 0 .0058 Ly75 N M 013825 0.8 0.31 0.035 Icaml N M 010493 0.81 0.31 0.0188 Gjb3 N M 008126 0.83 0.46 0.0423 Cd9 N M 0 0 7 6 5 7 0.99 0.45 ' 0 .0345 Slc2a3 N M 011401 1 0.44 0 .0027 9430079M16Rik N M 175414 1 0.49 0.031 Ppap2a N M 008903 1.05 0.49 0.0041 Gja7 N M 0 0 8 1 2 2 1.1 0.22 0 .0376 Abcbla N M 011076 1.1 ' 0.35 0.0201 Enah N M 010135 1.21 0.31 0 .0074 • Ptch N M 0 0 8 9 5 7 . 1.22 0.27 0 .0012 Akp2. N M 007431 1.24 0.55 0.0168 Slc29al N M 022880 1.37 - 0.44 0 .015 Trfr N M 011638 1.46 0.68 0 .039 Fgfr2 N M 010207 1.48 0.67 0 .0293 Utrn N M 011682 1.49 0 .59 0 .0013 Ptch2 N M 008958 1.53 0.38 0 .0054 Akl N M 0 2 1 5 1 5 1.53 0.66 0.0304 Cdldl N M 0 0 7 6 3 9 1.66 0.45 0.0291 Adaml9 N M 0 0 9 6 1 6 1.82 0.45 0 .002 1600025H15Rik N M 028064 1.88 0.83 0 .0062 Ltb N M 0 0 8 5 1 8 2 .29 0.85 0 .0202 Increased Enpp2 N M 015744 3.1 9.13 0 .0118 Dgka N M 016811 2.49 . 0.95 0 .0079 Cldn4 N M 009903 2 .36 1.6 0 .0472 Cldn6 N M 018777 2 .14 3.4 0.0081 Btla N M 177584 2.03 14.27 0 .0034 Stl4 N M 011176 1.73 2.45 0 .0319 78 Gene symbol.. Accession # 18 hours 72 hours p value If&9 N M 033608 1.55 2.22 0.005 Podxl N M 013723 1.54 2.54 0.0019 Itga8 N M 001001309 1.51 2.81 0.0298 Ddrl N M 007584 1.5 2.39 0.0264 Cldn7 N M 016887 1.48 5.97 0.00001 Spintl N M 016907 1.35 2.5 0.0138 Met N M 008591 1.35 2.18 0.0311 Rnfl28 N M 023270 1.34 2.51 0.0453 Slc39a8 N M 026228 1.33 2.32 0.0208 Tap! N M 013683 1.31 2.73 0.0074 Proml N M 008935 1.27 5.53 0.001 Calcr N M 007588 1.18 6.05 0.0338 Sfrp2 N M 009144 1.18 2.28 0.015 Tdpbp N M 009318 1.08 2.21 - 0.0202 Perp N M 022032 1.03 2.39 0.0269 Cd24a N M 009846 0.9 2.73 0 Dlkl N M 010052 0.82 1.77 0.0291 Cask N M 009806 0.79 1.91 0.0052 ' Gp38 N M 010329 0.79 1.63 0.0323 Amot N M 153319 0.76 4.1 0.0073 Lrp8 N M 053073 0.47 2.26 0.0255 p values for the most significant fold change from Welsh t-tests are listed. 79 Table 2.4 Differentially expressed genes with plasma membrane.localization Gene symbol Accession # Expression relative to 0 hours after LIF removal, 18 hours p value Expression relative to 0 hours after LIF removal, 72 hours p value Cell adhesion Collal N M 007742 0.05 0.00793 0.08 0.0278 Vcaml N M 011693 0.08 0.03805 0.08 0.0344 Thbsl N M 011580 0.11 0.01728 0.13 0.0203 Cell growth and maintenance •Acta2 N M 007392 0.13 0.0513 o.i 0.0472 Cav N M 007616 0.2 0.0288 0.23 0.0254 Empl N M 010128 0.22 0.1 0.09 0.0381 S100a6 N M 011313 0.38 0.113 . 0.16 0.0186 Akap2 N M 009649 0.45 0.131 0.28 0.0159 Metabolism Prss23 N M 029614 0.36 0.04408 0.46 0.174 Dnajb9 N M 013760 0.38 0.0682 0.31 0.0216 Abcal N M 013454 0.42 0.149 0.05 0.006 Eif2s2 N M 026030 0.43 0.00736 0.17' 0.017 Transcription Ankrdl N M 013468 0.04 0.01611 0.07 0.0657 Fos N M 010234 0.06 0.03985 0.11 0.0017 Fosb N M 008036 0.25 0.02309 0.33 0.0094 Fosl2 N M 008037 0.26 0.02917 0.31 0.283 A A 408868 N M 030612 0.27 . 0.01029 0.5 0.031 Klf4 N M 010637 0.32 0.00361 0.05 0.0021 Bcl3 N M 033601 0.38 0.00266 0.38 0.157 Tbx3 NM_011535 0.41 0.01023 0.3 0.0214 Aebp2 N M 009637 0.44 0.00747 0.43 0.0002 Egrl N M 007913 0.49 0.02355 0.44 0.0258 Mllt2h N M 133919 0.5 0.283 0.32 0.0277 Cell development and differentiation Cd44 N M 009851 0.08 0.00931 0.19 0.04301 Serpinel N M 008871 0.11 0.0007 0.05 0.06 lnhba N M 008380 0.32 0.03052 0.29 0.02966 Csfl N M 007778 0.36 0.03371 0.49 0.0711 Signal transduction Ccl2 N M 011333 0.04 0.0699 0.02 0.0279 Apbblip NM- 019456 0.18 0.04105 0.22 0.0664 Socs3 N M 007707 0.22 0.00055 0.15 0.0055 Osmr N M 011019 0.4 0.105 0.18 0.0321 Mras N M 008624 0.43 0.04385 0.07 0.00001 Jagl N M 013822 0.45 0.04943 0.47 0.0411 Kit N M 021099 0.46 0.0939 0.23 0.0016 Fzd5 N M 022721 0.47 0.0878 0.31 0.0228 Unknown biological process Lox N M 010728 0.02 0.00272 0.06 0.01 80 Gene symbol Accession # Expression relative to 0 hours after L I F /removal, 18 hours. p value Expression relative to 0 hours after L I F removal, 72 hours p Value Fbln2 N M 007992 0.14 0.01975 0.09 0.0054 Tnc N M 011607 0.14 0.02558 0.09 0.0134 BB120430 AI847445 0.14 0.01359 0.18 0.0224 Bgn N M 007542 0.15 0.01494 0.1 0.0071 C730049F20Rik AI840339 0.16 0.0534 0.22 0.0361 - Sdpr N M 138741 0.18 0.123 0.11 0.0278 Hnrphl N M 021510 0.33 0.04185 0.33 0.267 Timp2 N M 011594 0.41 0.03587 0.46 0.0298 AI504685 AI504685 0.44 0.01515 0.34 • 0.0004 C85523 C85523 0.45 0.0718 0.38 0.0449 All 15454 N M 175345 0.48 0.0596 0.31 0.0344 5730501N20Rik AI882080 0.49 0.01931 0.45 0.0068 Abbreviations: L IF - leukemia inhibitory factor. 81 Table 2.5 Summary of published gene expression studies in mouse and human embryonic stem cells used for comparison Kel'ereiiee experimental <lesi«n Cell line 1 ei liiinl<i»\ philfni in (if ne list used Number of «enes (herkip (Ivanova et al. 2002) Undifferentiated ESC grown on PEF in the presence ol LIF, cultured for one passage on gelatin before RNA extraction; expression compared with differentiated bone marrow and fetal liver cells CCE Affymetrix M G U74v2 Supplementary online Table 1 (refiltered according to author's criteria) 2,270 90 (97) (Kelly and Rizzino 2000) Undifferentiated ESC cultured on gelatin in the presence of LIF; expression compared with ESC differentiated for 96 hrs by treatment with 1 u M R A D3 Atlas Mouse cDNA expression arrays Table 1 17 6 (Ramalho-Santos et al. 2002) Undifferentiated ESC grown on PEF for two passages before R N A isolation; expression compared with differentiated bone marrow and tissue from the lateral ventricles on the brain C57/B16 Affymetrix M G U74v2 Database SI 1,787 82 (85) (Sharov et al. 2003) Expression profiles of 36 tissues from early development and adult stem cells, including undifferentiated ESC and ESC differentiated for 4 or 18 hrs in the absence of PEF and LIF R l cDNA library construction Data set S7, groups F, I, L, and O, with comparable annotation 260 16(18) (Bhattacharya et al. 2004) Expression profiles of five ESC lines grown on PEF was compared with 8-day-old EB outgrowths GE01,GE09, BG01,BG02, TE06 Custom oligonucleotide glass arrays Supplementary online Table 2 92 5 (Brandenberger et al. 2004) Expression profiles of three undifferentiated ESC lines compared with EB (8 days of differentiation), pre-hepatocytes (induced by addition of DMSO), and pre-neural cells (induced by addition of retinoic acid) H1,H7,H9 Massively parallel signature sequencing Supplementary online Table 2 532 19(20) (Sato et al. 2003) Undifferentiated ESC compared with ESC induced to differentiated neurons for 3 wks HI Affymetrix HGU133A Database 1 918 39 (40) (Sperger et al. 2003) Undifferentiated ESC compared with multiple cell lines, including germ cell tumor and other differentiated cell lines H1,H7, H9, H13, H14 Custom cDNA arrays Supplementary online Table 6 1,760 45 (51) (Ginis et al. 2004) Comparison between human and mouse undifferentiated ESC Human HI, mouse D3 RT-PCR/focused functional microarrays Table 3A 13 2 2.5. References Abbondanzo SJ, Gadi I, Stewart C L . 1993. Guide to techniques in mouse development. In: Wasserman P,DePamphilis M , editors. Toronto: Academic Press Inc. p 803-823. Anneren C, Cowan C A , Melton D A . 2004. The Src family of tyrosine kinases is important for embryonic stem cell self-renewal. J B i o l Chem. A v i l i o n A A , Nicol is S K , Pevny L H , Perez L , V iv i an N , Lovell-Badge R. 2003. Multipotent cell lineages in early mouse development depend on S O X 2 function. Genes Dev 17(1): 126-40. Bhattacharya B , Miura T, Brandenberger R, Mejido J, Luo Y , Yang A X , Joshi B H , Ginis I, Thies R S , Ami t M , Lyons I, Condie B G , Itskovitz-Eldor J, Rao M S , Puri R K . 2004. Gene expression in human embryonic stem cell lines: unique molecular signature. Blood 103(8):2956-64. Bowles J, Teasdale R P , James K , Koopman P. 2003. Dppa3 is a marker of pluripotency and has a human homologue that is expressed in germ cell tumours. Cytogenet Genome Res 101(3-4):261-5. Brandenberger R, Wei H , Zhang S, L e i S, Murage J, Fisk GJ , L i Y , X u C, Fang R, Guegler K , Rao M S , Mandalam R, Lebkowski J, Stanton L W . 2004. Transcriptome characterization elucidates signaling networks that control human E S cell growth and differentiation. Nat Biotechnol 22(6):707-16. Brazma A , Hingamp P, Quackenbush J, Sherlock G , Spellman P, Stoeckert C , Aach J, Ansorge W , B a l l C A , Causton H C , Gaasterland T, Glenisson P, Holstege F C , K i m IF, Markowitz V , Matese JC, Parkinson H , Robinson A , Sarkans U , Schulze-Kremer S, Stewart J, Taylor R, V i l o J, Vingron M . 2001. M i n i m u m information about a mieroarray experiment (MIAME)- toward standards for mieroarray data. Nat Genet 29(4):365-71. Burdon T, Chambers I, Stracey C, N i w a H , Smith A . 1999. Signaling mechanisms regulating self-renewal and differentiation of pluripotent embryonic stem cells. Cells Tissues Organs 165(3-4):131-43. Chambers I, Colby D , Robertson M , Nichols J, Lee S, Tweedie S, Smith A . 2003. Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. Ce l l 113(5):643-55. Cohen M M , Jr. 2003. The hedgehog signaling network. A m J M e d Genet 123A(l):5-28. Dahlquist K D , Salomonis N , Vranizan K , Lawlor S C , Conkl in B R . 2002. G e n M A P P , a new tool for viewing and analyzing mieroarray data on biological pathways. Nat Genet 31(l):19-20. 83 Draghici S, Khatri P , Bhavsar P, Shah.A, Krawetz S A , Tainsky M A . 2003. Onto-Tools, the toolkit of the modern biologist: Onto-Express, Onto-Compare, Onto-Design and Onto-Translate. Nucleic Acids Res 31(13):3775-81. Evans M J , Kaufman M H . 1981. Establishment in culture of pluripotential cells from mouse embryos. Nature 292(5819):154-6. Evsikov A V , Solter D . 2003. Comment on " 'Sternness': transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature". Science 302(5644):393; author reply 393.. Fortunel N O , Otu H H , N g H H , Chen J, M u X , Chevassut T, L i X , Joseph M , Bailey C, Hatzfeld J A , Hatzfeld A , Usta F , Vega V B , Long P M , Libermann T A , L i m B . 2003. Comment on " 'Sternness': transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature". Science 302(5644):393; author reply 393. Furusawa T, Ohkoshi K , Honda C, Takahashi S, Tokunaga T. 2004. Embryonic stem cells expressing both platelet endothelial cell adhesion molecule-1 and stage-specific embryonic antigen-1 differentiate predominantly into epiblast cells in a chimeric embryo. B i o l Reprod 70(5): 1452-7. Ginis I, Luo Y , Miura T, Thies S, Brandenberger R, Gerecht-Nir S, Ami t M , Hoke A , Carpenter M K , Itskovitz-Eldor J, Rao M S . 2004. Differences between human and mouse embryonic stem cells. Dev B i o l 269(2):360-80. Haegel H , Dierich A , Ceredig R. 1994. CD44 in differentiated embryonic stem cells: surface expression and transcripts encoding multiple variants. Dev Immunol 3(4):239-46. Helgason C D , Damen JE, Rosten P, Grewal R, Sorensen P, Chappel S M , Borowski A , Jirik F , Krysta l G , Humphries R K . 1998. Targeted disruption o f SHIP leads to hemopoietic perturbations, lung pathology, and a shortened life span. Genes Dev 12(11):1610-20. Hosier B A , LaRosa GJ , Grippo JF, Gudas L J . 1989. Expression of R E X - 1 , a gene containing zinc finger motifs, is rapidly reduced by retinoic acid in F9 teratocarcinoma cells. M o l Cel l B i o l 9(12):5623-9. Ivanova N B , Dimos JT, Schaniel C , Hackney J A , Moore K A , Lemischka IR. 2002. A stem cell molecular signature. Science 298(5593):601-4. Jones P L , Jones FS . 2000. Tenascin-C in development and disease: gene regulation and cell function. Matrix B i o l 19(7):581-96. 84 Keller G , Kennedy M , Papayannopoulou T, Wiles M V . 1993. Hematopoietic commitment during embryonic stem cell differentiation in culture. M o l Cel l B i o l 13(l):473-86. K e l l y D L , Rizz ino A . 2000. D N A mieroarray analyses of genes regulated during the differentiation of embryonic stem cells. M o l Reprod Dev 56(2):113-23. Kirschstein R, Skirbol l L R . 2001. Stem Cells: Scientific Progress and Future Research Directions. National Institutes of Health, Department of Health and Human Services. Appendix E , 1-11 p. L i E , Bestor T H , Jaenisch R. 1992. Targeted mutation of the D N A methyltransferase gene results in embryonic lethality. Cel l 69(6):915-26. L ivak K J , Schmittgen T D . 2001. Analysis of relative gene expression data using real-time quantitative P C R and the 2(-Delta Delta C(T)) Method. Methods 25(4):402-8. Mart in G R . 1981. Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc Natl Acad Sci U S A 78(12):7634-8. Meno C, Saijoh Y , Fujii H , Ikeda M , Yokoyama T, Yokoyama M , Toyoda Y , Hamada H . 1996. Left-right asymmetric expression of the T G F beta-family member lefty in mouse embryos. Nature 381(6578): 151-5. Mitsui K , Tokuzawa Y , Itoh H , Segawa K , Murakami M , Takahashi K , Maruyama M , Maeda M , Yamanaka S. 2003. The homeoprotein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells. Cel l 113(5):631-42. Nagy A , Rossant J, Nagy R, Abramow-Newerly W , Roder JC. 1993. Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc Nat l Acad Sci U S A 90(18):8424-8. Nichols J, Zevnik B , Anastassiadis K , N i w a H , Klewe-Nebenius D , Chambers I, Scholer H , Smith A . 1998. Formation of pluripotent stem cells in the mammalian embryo depends on the P O U transcription factor Oct4. Cel l 95(3):379-91. N i w a H , M i y a z a k i J, Smith A G . 2000. Quantitative expression o f Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells. Nat Genet 24(4):372-6. Ogden S K , Ascano M , Jr., Stegman M A , Rqbbins DJ . 2004. Regulation of Hedgehog signaling: a complex story. Biochem Pharmacol 67(5):805-14. Oka M , Tagoku K , Russell T L , Nakano Y , Hamazaki T, Meyer E M , Yokota T, Terada N . 2002. C D 9 is associated with leukemia inhibitory factor-mediated maintenance of embryonic stem cells. M o l B i o l Cel l 13(4):1274-81. 85 Ramalho-Santos M , Y o o n S, Matsuzaki Y , Mul l igan R C , Mel ton D A . 2002. "Sternness": transcriptional profiling of embryonic and adult stem cells. Science 298(5593):597-600. Rogers M B , Hosier B A , Gudas L J . 1991. Specific expression of a retinoic acid-regulated, zinc-finger gene, Rex-1, in preimplantation embryos, trophoblast and spermatocytes. Development 113(3):815-24. Rose T M , Weiford D M , Gunderson N L , Bruce A G . 1994. Oncostatin M (OSM) inhibits the differentiation of pluripotent embryonic stem cells in vitro. Cytokine 6(l):48-54. Rozen S, Skaletsky H . 2000. Primer3 on the W W W for general users and for biologist programmers. Methods M o l B i o l 132:365-86. Sato N , Meijer L , Skaltsounis L , Greengard P, Br ivanlou A H . 2004. Maintenance of pluripotency in human and mouse embryonic stem cells through activation of Wnt signaling by a pharmacological GSK-3-specific inhibitor. Nat M e d 10(l):55-63. Sato N , Sanjuan I M , Heke M , Uchida M , Naef F , Brivanlou A H . 2003. Molecular signature of human embryonic stem cells and its comparison with the mouse. Dev B i o l 260(2):404-13. Schoor M , Schuster-Gossler K , Gossler A . 1993. The Etl-1 gene encodes a nuclear protein differentially expressed during early mouse development. Dev D y n 197(3):227-37. Sharov A A , Piao Y , Matoba R, Dudekula D B , Qian Y , VanBuren V , Falco G , Martin PR, Stagg C A , Bassey U C , Wang Y , Carter M G , Hamatani T, A i b a K , Akutsu H , Sharova L , Tanaka TS, Kimber W L , Yoshikawa T, Jaradat S A , Pantano S, Nagaraja R, Boheler K R , Taub D , Hodes RJ , Longo D L , Schlessinger D , Keller J, Klo tz E , Kelsoe G , Umezawa A , Vescovi A L , Rossant J, Kunath T, Hogan B L , Curci A , D'Urso M , Kelso J, Hide W , K o M S . 2003. Transcriptome analysis of mouse stem cells and early embryos. PLoS B i o l 1(3):E74. Shirogane T, Fukada T, Muller J M , Shima D T , H ib i M , Hirano T. 1999. Synergistic roles for Pim-1 and c -Myc in STAT3-mediated cell cycle progression and antiapoptosis. Immunity 11(6):709-19. Smith A G . 2001. Embryo-derived stem cells: of mice and men. Annu Rev Ce l l Dev B i o l 17:435-62. Sperger J M , Chen X , Draper JS, Antosiewicz JE, Chon C H , Jones SB, Brooks JD, Andrews P W , Brown PO, Thomson J A . 2003. Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors. Proc Natl Acad Sci U S A 100(23): 13350-5. Stewart C L . 1993. Production of chimeras between embryonic stem cells and embryos. Methods Enzymol 225:823-55. 86 Stewart C L , Kaspar P, Brunet L J , Bhatt H , Gadi I, Kontgen F , Abbondanzo SJ. 1992. Blastocyst implantation depends on maternal expression of leukaemia inhibitory factor. Nature 359(6390):76-9. Suda Y , Suzuki M , Ikawa Y , Aizawa S. 1987. Mouse embryonic stem cells exhibit indefinite proliferative potential. J Cel l Physiol 13 3 (1): 197-201. Tanaka TS, Kunath T, Kimber W L , Jaradat S A , Stagg C A , Usuda M , Yokota T, N i w a H , Rossant J, K o M S . 2002. Gene expression profiling of embryo-derived stem cells reveals candidate genes associated with pluripotency and lineage specificity. Genome Res 12(12): 1921-8. Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz M A , Swiergiel JJ, Marshall V S , Jones J M . 1998. Embryonic stem cell lines derived from human blastocysts. Science 282(5391):1145-7. Tzouanacou E , Tweedie S, Wilson V : 2003. Identification of Jadel , a gene encoding a P H D zinc finger protein, in a gene trap mutagenesis screen for genes involved in anteroposterior axis development. M o l Cel l B i o l 23(23):8553-2. Vas V , Szilagyi L , Paloczi K , Uher F. 2004. Soluble Jagged-1 is able to inhibit the function of its multivalent form to induce hematopoietic stem cell self-renewal in a surrogate in vitro assay. J L e u k o c B i o l 75(4):714-20. Weigmann A , Corbeil D , Hel lwig A , Huttner W B . 1997. Prominin, a novel microvilli-specific polytopic membrane protein of the apical surface of epithelial cells, is targeted to plasmalemmal protrusions of non-epithelial cells. Proc Natl Acad Sci U S A 94(23): 12425-30. Wheatley SC, Isacke C M . 1995. Induction of a hyaluronan receptor, CD44, during embryonal carcinoma and embryonic stem cell differentiation. Cel l Adhes Commun 3(3):217-30. Wiese C, Rolletschek A , Kania G , Blyszczuk P, Tarasov K V , Tarasova Y , Wersto R P , Boheler K R , Wobus A M . 2004. Nestin expression—a property of multi-lineage progenitor cells? Ce l l M o l Life Sci 61(19-20):2510-22. -Wilkinson D G , Bhatt S, Herrmann B G . 1990. Expression pattern of the mouse T gene and its role in mesoderm formation. Nature 343(6259):657-9. Will iams R L , Hil ton D J , Pease S, Wil lson T A , Stewart C L , Gearing DP , Wagner E F , Metcalf D , N i c o l a N A , Gough N M . 1988. M y e l o i d leukaemia inhibitory factor maintains the developmental potential of embryonic stem cells. Nature 336(6200):684-7. 87 Xue Y , Gao X , Lindsel l C E , Norton C R , Chang B , Hicks C, Gendron-Maguire M , Rand E B , Weinmaster G , Gridley T. 1999. Embryonic lethality and vascular defects in mice lacking the Notch ligand Jaggedl. Hum M o l Genet 8(5):723-30. Y i n g Q L , Nichols J, Chambers I, Smith A . 2003. B M P induction of Id proteins suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with S T A T 3 . Cel l 115(3):281-92. 88 3. Increased Capacity of Mouse Embryonic Stem Cells to Form Embryoid Bodies Following Optimization of Ascorbic Acid, PD98059 And Chondroitin Sulphate Levels2 3.1. Introduction The ability of embryonic stem cells (ESC) to differentiate into all mature cell types represented has stimulated investigations to explore their use to replace tissues damaged through injury, disease or aging. However, clinical scale application of many stem cell-based therapies w i l l require the development of protocols for their expansion as an undifferentiated population prior, to the initiation of differentiation. Since regulation o f stem cell survival, proliferation and differentiation are likely modulated by the integration of multiple signaling pathways (Davey and Zandstra 2004), development of these protocols requires systematic methods to investigate the influence of multiple environmental cues simultaneously. Mouse E S C (Evans and Kaufman 1981; Martin 1981) provide a useful model system to develop methods for the optimization of stem cell cultures because of their abundance, relative ease of propagation and purity compared to adult stem cell populations. In addition to leukemia inhibitory factor (LIF), a number of soluble and insoluble factors have been shown to influence E S C fate when added to the culture environment indicating the importance of understanding the 2 A version of this chapter wi l l be submitted for publication. Glover, C H , Eaves, C J , Piret, J M . Increased capacity of mouse embryonic stem cells to form embryoid bodies following the optimization of ascorbic acid, PD98059 and chondroitin sulphate levels. 91 influence of multiple signalling pathways in directing cell fate (Prudhomme et al. 2004). For example, inhibition of E R K 1 / 2 activation by the specific M A P K inhibitor PD98059 (PD) in the presence of gp l30 signaling leads to an increase in E S C self-renewal (Burdon et al. 1999). Similarly, we have shown that ascorbic acid ( A A ) in the presence of L I F increases the proportion of EB-forming cells in culture (Chapter 4). It appears that components of the extracellular matrix also influence cell fate since E S C are maintained with higher efficiency when cultured on a layer of inactivated feeders than when this is replaced with gelatin (Glover et al . 2006) although the specific components have not been identified. Chondroitin sulphate (CS) is a component of the extracellular matrix that is able to bind to growth factors and modulate their activity as wel l as affect development and cell division (Sugahara et al. 2003). Here, we hypothesized that CS would have an influence oh cell fate and investigated its influence in conjunction with P D and A A . A s new factors with beneficial effects on stem cell maintenance and expansion are identified, their incorporation into existing culture medium can involve interactions, either negative or positive. The large number of factors present in cell culture medium makes single factor approaches impractical due to the number of experiments that must be performed. In addition, single factor approaches are not well suited to uncover complex interactions that are seen when changes in multiple factors are made. Statistical design of experiments, specifically factorial design (FD) and central composite design (CCD) experiments, provide an efficient and rigorous statistical approach to efficiently determine individual and multiple factor responses simultaneously. They make use of geometric principles for statistical sampling and minimize the 92 variance of estimated coefficients obtained by regression. In addition, these designs can minimize the impact of unimportant factors (e.g. reagent batch, errors in cell counts, volume addition etc.) by grouping experiments into homogeneous blocks and by inducing replication. This is particularly useful when the magnitude but not the trend of assay readouts is highly variable. Statistical design of experiments have previously been used to enhance stem cell expansion and differentiation for a variety of stem cell types. For example, a F D experiment was used to screen multiple factors for their influence on hematopoietic stem cell expansion, and this was followed by a C C D to optimize the level of 3 of these factors for maximal expansion (Audet et al. 2002). Statistical design of experiments has also been used in the context of differentiation. Chang et al., investigated the effect of 5 factors on endoderm differentiation from E S C (Chang and Zandstra 2004). Experiments of this nature w i l l become increasingly important as clinical needs for stem cell based therapies develop. In particular, clinical application of these therapies w i l l require the development of xeno-free culture methods to avoid subsequent tissue rejection and immunological reaction (Martin et al. 2005). A s replacements for xeno-derived components of cell culture medium are discovered, significant effort w i l l have to be made to incorporate these into cell culture medium and to ensure that there are no negative consequences of addition as a result of factor interactions. In the present study, we have determined the influence of A A , CS and P D on E S C as measured by the subsequent ability of treated E S C to form clonal E B . EB-forming cell frequencies were measured because this assay provides a relatively sensitive measure which correlates closely with the ability o f ESC-derived cells to contribute to the formation of embryos following 93 blastocyst injection (Chapter 2). Furthermore, the majority of differentiation protocols require the formation of E B prior to differentiation towards specific cell types. A statistical approach was used to determine factor interactions where a positive or negative interaction represents joint effects that are more or less than additive. Results from the interaction studies suggested specific molecular targets for A A and also allowed the determination of a cocktail for the three factors that maximised E B output. 3.2. Materials and Methods 3.2.1. Embryonic Stem Cell Maintenance Cultures R l (passage 17) E S C and primary embryonic fibroblasts (PEF, StemCell Technologies) were maintained as described in Chapter 2. 3.2.2. Embryonic Stem Cell Experimental Cultures E S C were thawed and maintained and prepared for experiments as described in Chapter 2. Experimental cultures were performed on tissue culture dishes (Sarstedt) coated with 0.1% porcine gelatin (Sigma) with cells plated at a density of 100,000 cells per 60 mm tissue culture dish. Cells were cultured for 72 hours with daily medium change before harvesting and testing in the EB-formation assay. Cells were counted using trypan-blue exclusion to test for viability. Differentiation media were based on maintenance medium with the following differences (a) L IF removal - maintenance medium minus L I F , (b) D M S O - maintenance medium minus L I F plus 1%> D M S O (Sigma), (c) R A - maintenance medium plus 2 u.M retinoic acid ( R A , Sigma) (d) A A + L I F - maintenance medium plus 50 u.g/mL A A (Sigma) and (e) A A - L I F - maintenance 94 medium without L IF plus 50 u.g/mL A A . Concentrated R A was prepared at a concentration of 10 m M by dissolving 30 mg powder in 10 m L 100% ethanol and stored at 4°C in the dark. Media was prepared by adding 10 p L of thawed R A stock solution to 50 m L maintenance medium. Aliquots of stock A A were prepared by dissolving in dFkO to a final concentration of 10 mg/mL. P D was dissolved in D M S O to a final concentration of 2 m M . Aliquots were frozen at -20°C for subsequent use. CS was dissolved in D M E M to a concentration of 100 mg/mL and stored at 4°C for subsequent use. Cells were harvested daily for functional assay analysis ( C F C and E B assays - performed as described in Chapter 2) or for R N A extraction. 3.2.3. Design of Experiments A two-level F D was used to screen initial factor levels and interactions (Table 3.1). High levels used in the F D were chosen based on the optimal level observed in the single factor dose responses. The F D experiment was repeated 3 times with no centre points included. Results from these experiments were used to estimate parameter values for the following model: Y-K + PAA{AA) + Pcs(CS) + Pm{PD) +PAAICs{AA • CS) + PMIPD{AA • PD) + PCSIPD{CS • PD) where Y is either the net growth rate, the frequency of EB-forming cells or the yield of E B -forming cells; K the intercept term; A A , CS and P D the coded factor terms (-1 or +1); p\ the main effect parameters and |3j/j the second-order interaction terms (i,j = A A , C S , P D ; i*j). Analysis of variance was used to estimate the statistical significance of the effects. Factor concentrations are varied on a log scale to satisfy the assumption that data are normally 95 distributed. Determination of coefficients and statistical analysis was performed using J M P I N (SAS Institute). Further optimization of factor levels was performed using 2-factor C C D to construct response surface maps of net expansion, EB-forming cell frequency and yield of EB-forming cells (Table 3.2, Table 3.3). Orthogonal, rotatable designs were used including 4 replicated centre points, 4 factorial points and 4 axial points. The data was analyzed to estimate parameter values for the following model: Y = K + fi1{xl) + p2{x2) +A,2(*i • *2) + Aj(*i2) + ^ 2,2(4) where x; are the log 10 (concentration of factor i); Bi the main effect parameters, By the second-order interaction terms and By the quadratic coefficients. Analysis of variance was used to determine the significance of the parameters. Determination of coefficients and statistical analysis was performed using J M P I N (SAS Institute). 3.2.4. Flow Cytometry Single cell suspensions were prepared and blocked for 10 minutes on ice with 5 pg/ml anti-mouse C D 1 6 / C D 3 2 (Fc Block, B D Pharmingen) in phosphate buffered saline plus 2% F B S (PBSFBS) . Cells were washed once with P B S F B S and then incubated on ice for 20 minutes with a phycoerythrin (PE)-conjugated anti-SSEA-1 antibody (clone M C - 4 8 0 , R & D Systems). Cells were then washed twice before subsequent preparation for annexin-V analysis. Annexin-V staining was performed using instructions provided in the B D kit using 7 - A A D to exclude 96 nonviable cells. Samples were analyzed by flow cytometry using a FACSCal ibu r flow cytometer running CELLQues t software (BD). Data was analyzed using Flojo (Treestar). 3.2.5. RNA extraction and Quantitative RT-PCR These methods are described in Chapter 2. Primer sequences are shown in Appendix A . 3.3 . Results 3.3.1. Dose Response of Ascorbic Acid, Chondroitin Sulphate or PD98059 To define the dose response effects of A A , C S and P D on the expansion of E S C , R l E S C were cultured for 72 hours in adherent cultures on gelatin with daily medium exchanges in the presence of several concentrations of A A , CS or P D . Addi t ion of A A and P D significantly increased the yield of EB-forming cells (p<0.05 for both) with 75 pg/mL A A and 20 p,M P D yielding the greatest expansion at 3.6 ± 0.1 and 1.9 ± 0.1-fold, respectively, over untreated cultures (Figure 3.1 A , Figure 3.2A). However, there was no significant difference between cells cultured at 50 or 75 u.g/mL A A . Both factors decreased the average net cell growth rate (the difference between cell proliferation and death) in a dose-dependent fashion (p<0.05, Figure 3 . IB, Figure 3.2B). The decreased net growth rate was compensated by an increased frequency of cells that were able to read out in the E B assay (p<0.05, Figure 3.1C, Figure 3.2C) leading to an overall expansion of the number of cells able to read out in the E B assay. Addit ion of CS between 0 and 2000 pg/mL led to a slight increase in the number of EB-forming cells, but this change was not statistically significant (data not shown). This increase was due to 97 an increase in the frequency of EB-forming cells and there was no apparent trend in the net growth rate. 3.3.2. Interactions between Ascorbic Acid, Chondroitin Sulphate and PD98059 A 2-level, 3-way F D (Table 3.1, Table 3.3) was used to investigate the influence of A A , P D and CS , individually and in combination, on the yield of EB-forming cells, the net cell growth rate and the frequency of EB-forming cells (Figure 3.3). In 8 experiments, a combination of CS and P D consistently produced the largest number of EB-forming cells, with a total yield 2.13 ± 0.27 times greater than untreated cultures. However, this was not significantly different from any of the other combination treatments which yielded EB-forming cell numbers as listed when compared to untreated cultures: 1.95 ± 0.08 (CS + A A ) , 1.65 ± 0.14 (PD + A A ) and 1.73 ± 0.17 (PD + C S + A A ) (numbers are fold-increase over untreated cultures). The combination treatments all yielded higher numbers of EB-forming cells than the single factor experiments (1.30 ± 0.05 (CS), 1.34 ± 0.04 (PD) and 1.41 ± 0.05 (AA)) . Using analysis of variance, these experiments were analyzed to estimate parameter values for equation 1. The magnitude of each parameter and the p-value associated with them represent the influence of each factor individually and in combination on the particular culture output being modeled (yield of EB-forming cells, net cell growth, and frequency of EB-forming cells). Each individual component exhibited a significant positive effect on the yield of EB-forming cells (p<0.05) including C S , which influenced yield by increasing the frequency of EB-forming cells (p<0.05), but had no significant influence on the net cell growth. The F D revealed a significant negative interaction between P D and A A on the yield of EB-forming cells (p<0.05) through a 98 reduction in the frequency of EB-forming cells (p<0.1). Interestingly, there was a significant, positive interaction between these two factors with respect to total cell growth (p<0.1) but this was insufficient to counteract the negative effect on EB-forming cell potential. 3.3.3. Central Composite Design Experiments to Determine Non-Linear Responses Two C C D experiments were used to confirm factor interactions observed in the F D , to determine non-linear responses to these factors and to optimize concentrations for addition of these factors in combination. The first C C D experiment (Table 3.2, Table 3.3, Figure 3.4A) was performed with P D and CS in the absence of A A since this was the combination from the F D that maximized EB-forming cell output. Since a negative interaction was observed between A A and P D at the concentrations tested in the F D , lower concentrations of each of these two factors were investigated in the C C D to explore this effect further. The coded value of +1 for C S and P D were fixed at the levels tested in the F D experiment with CS being varied between 2 and 1250 pg/mL and P D between 0.08 and 50 p M around these points. The positive influence of both CS and P D on the expansion of EB-forming cells was confirmed (p<0.05 and p<0.1 respectively) and a negative interaction between CS and P D revealed (p<0.1). Furthermore, there was a large, negative, quadratic term for P D (LogPD 2 ) on the yield o f EB-forming cells (p<0.05) indicating that P D should be added at lower concentrations to avoid high dose inhibition. This was consistent with the results seen in the single-factor dose response of P D (Figure 3.2). 99 Since C S exhibited a large positive effect on EB-forming cell yield individually, its level was fixed at 500 pg/mL and the interaction between P D and A A was explored further in a second C C D experiment (Table 3.2, Table 3.3, Figure 3.4B). The coded value of +1 for A A was at 50 pg/mL as for the F D experiment with the corresponding range from 0.2 to 130 pg/mL. The range for P D was reduced because of the negative interaction between it and the other two factors and the high dose inhibition that was observed in the previous experiments. Thus, for P D the coded value of +1 was at 10 p M with the corresponding range from 0.04 to 26 p M . Interestingly at the lower levels of P D tested in this experiment (and contrary to the result observed in the F D experiment, Figure 3.3), a positive interaction between A A and P D was observed (p<0.1). However, the high dose inhibition of P D was observed again (p<0.05). Furthermore, A A exhibited a significant, positive quadratic term (p<0.1). 3.3.4. Verification of Cocktail to Maximize Expansion of Embryoid Body-Forming Cells A s a result of the above experiments, a more favourable cocktail for the expansion of E B -forming cells was proposed to be: A A - 50 p.g/mL, C S - 500 p,g/mL and P D - 2 piM. To confirm that this would promote the expansion of EB-forming cells as expected this cocktail was compared to (1) the expansion of untreated cells, (2) a combination which consisted of all factors at concentrations determined from their isolated dose responses ( A A - 50 pg/mL, CS - 500 pg/mL and P D - 20 p M ) , (3) a two-way combination of A A and C S : A A - 50 pg/mL, CS - 500 pg/mL, and (4) a two-way combination of P D and C S : P D - 20 p M , CS - 500 pg/mL (Figure 3.5). 100 A s expected, the optimized cocktail expanded EB-forming cells 3.1 ± 0.2-fold more than untreated cells (p<0.05). The cocktail derived from the optimal concentrations during individual dose responses expanded EB-forming cells almost 1.8 ± 0.2-fold over untreated cultures (p<0.05) but was less than both two way combinations as a result of the high-dose inhibition of P D and negative interactions between P D and A A . 3.3.5. Influence of Ascorbic Acid, Chondroitin Sulphate and PD98059 on Self-Renewal and Apoptosis A s indicated by the F D and C C D results, there was a significant negative interaction between A A and P D on the yield of EB-forming cells at their optimal single factor concentrations but a positive interaction when P D was tested at lower concentrations. To gain further insight into the cellular kinetics responsible for this interaction, we investigated the activity of these factors, as well as C S , on self-renewal and apoptosis. Cells were cultured in maintenance medium in the presence of each of the single factors alone ( A A , P D or CS) or in a C S / P D combination for 72 hours. After this time, the C F C assay was used to determine the influence of these components on self-renewal over this period. P D alone increased the yield of C F C 2.1 ± 0.1-fold (p<0.05) while addition of A A showed a marginal increase in C F C (1.2 ± 0.1-fold increase, p>0.05, Figure 3.6A). Addit ion of CS in combination with P D increased the yield of C F C 4.8 ± 1.1-fold compared to untreated controls (p<0.05) while C S alone increased C F C yield 1.4 ± 0.4-fold (p>0.05, Figure 3.6B). 101 These results suggested that the addition of P D either increased the proliferation rate of undifferentiated cells or decreased their differentiation rate. To distinguish between these two possibilities, the proportion of colonies containing undifferentiated E S C was determined in C F C assays using A P staining. Addit ion of P D to C F C assays did not lead to a statistically significant increase in the proproportion of undifferentiated colonies in culture (Figure 3.6C). A n enhancement of the self-renewal ability of cells combined with the presence of similar numbers of differentiated and undifferentiated colonies suggested that addition of P D increased the proliferation rate of undifferentiated cells. The proportion of undifferentiated colonies increased from 62% to 83%> when CS and P D were added in combination and there was a concordant decrease in the proportion of differentiated colonies. This was in contrast to the addition of CS or P D alone where there was no significant change in the frequency of both types of colonies (Figure 3.6C). The inability of A A to increase the yield of C F C implied that its primary action was through a change in the relative apoptosis rates of differentiated and undifferentiated cells. Consistent with this, the proportion of AP-positive colonies in C F C was indistinguishable from untreated cultures while the proportion of AP-negative colonies was decreased (Figure 3.6C). To test specifically whether A A increased the apoptosis rate of differentiated cells, cells were cultured for 48 hours in the presence of A A with no medium change. Cells were harvested and stained for SSEA-1 (to distinguish between differentiated and undifferentiated cells), Annex in -V (to identify cells undergoing apoptosis) and 7 - A A D (to identify nonviable cells). There was no significant change in the apoptosis rate of SSEA-1-posit ive cells under any of the treatments. In contrast, A A 102 significantly increased the proportion of SSEA-1-negative, 7-AAD-negative cells undergoing apoptosis relative to untreated cultures (3.4 ± 1.2-fold, p<0.05, Figure 3.6D). 3.3.6. Differentiation Ability of Treated Cells To verify that the ability of E S C to differentiate to specific lineages was not compromised by addition of A A , CS and P D , cells were grown in medium with all three added at the optimal concentration ( A A - 50 pg/mL, CS - 500 pg/mL and P D - 2 p M ) for 48 hours with daily media exchange and then, to induce differentiation, replated into medium without L I F or into medium containing L I F and R A . A s a control, parallel cultures were maintained in standard conditions for the same length of time. The gene expression level of 5 lineage marker genes was compared to untreated cells using Q - R T - P C R (Figure 3.7). Removal of L IF induced the increased expression of genes associated with ectoderm, neural and mesoderm differentiation (fgf5, nestin, and brachyury respectively), but had little effect on the expression of genes associated with endodermal differentiation (foxa2 and soxl 7). In contrast, treatment with R A strongly induced the markers of neural and endodermal differentiation, but had no significant effect on the expression of brachyury (mesodermal differentiation) and fgf5 expression (ectodermal differentiation). Although the specific levels of each of these makers was different, in all cases differentiation towards the specific lineages appeared to be enhanced indicating that differentiation ability of A A , CS or P D treated cells was not compromised. 103 3.4. Discussion This work has demonstrated that A A , CS and P D all have positive influences on the absolute y ie ld of EB-forming cells. Furthermore, statistical design of experiments allowed the optimization o f the levels of individual factors to maximize EB-forming cell yields. These experiments also revealed an unexpected negative interaction between P D and A A , suggesting that ascorbic acid acts through suppression of the M A P K pathway. A A , CS and P D each led to an increase in the expansion of EB-forming cells relative to untreated controls although the specific mechanism of this enhancement appears to be different for each factor. While all three factors increased the frequency of EB-forming cells in culture, both A A and P D decreased the net growth rate whereas C S had no effect on the net growth rate. Nevertheless, P D increases the overall yield of C F C indicating an enhancement of E S C self-renewal not obtained with A A . The fact that there was no change in the proportion of undifferentiated colonies obtained from cells in PD-treated cultures indicated that the increased yield of C F C was due to an increased rate of proliferation of undifferentiated relative to differentiated cells ones. The effect of CS on culture output only became significant in the presence of P D and A A , yet no factor interactions involving CS were revealed in the F D experiment. This suggests that CS acted either through the stabilization of P D and A A , both of which are known to be labile, or by binding these compounds and making them more available to be taken up by the cells. However, when P D was tested at a lower concentration, as in the first C C D , a negative interaction between 104 CS and P D was revealed. This suggests that there is a critical level of P D below which CS has an inhibitory effect suggesting some kind of binding mechanism. This observation is also consistent with the observation in the second C C D experiment, that P D alone has no significant effect since in this experiment a constant level of CS was maintained throughout. CS has previously been shown to stabilize growth factors and increase their functional availability (Sugahara et al. 2003). The combination of P D and CS led to a significant reduction in the rate of spontaneous differentiation in culture, as demonstrated by the increased proportion of undifferentiated colonies in culture. CS is a glycosaminoglycan and is distributed widely throughout the body primarily as a modification on many extracellular matrix proteins. CS has previously been shown to have functions in C N S development, wound repair, growth factor signaling and cell division (Sugahara et al. 2003). Since optimal expansion of E S C takes place in contact with irradiated P E F , addition of CS to cultures that do not contain P E F may be necessary to replace this lost source. In development, an essential role for CS has been described in C. elegans where disruption of a gene required for production of C S resulted in defects in cytokinesis and early embryonic death (Mizuguchi et al. 2003). The influence of A A was minimal on the yield of C F C indicating that it had little effect on self-renewal. Rather, A A increased the apoptosis rate of differentiated cells as demonstrated by an increased apoptosis rate among SSEA-1-negative cells as wel l as a decreased number of differentiated colonies in AA-treated cultures. There have been no previous reports of an enhancement on E S C culture output by A A in the presence of L I F . Some clues as to its mechanism of action are given by the negative interaction that was observed between A A and 105 P D in the F D experiment. Since it is known that P D inhibits the activation of E R K 1 / 2 , it could be hypothesized that addition of A A similarly inhibits this signaling pathway. However, given the differential response of cells to the two factors in terms of self-renewal, the mode of action could also be through an effect on another step in the M A P K signal transduction pathway, or through a parallel pathway which influences the same molecular targets. Surprisingly, a positive interaction was observed between A A and P D in the second C C D experiment. A possible explanation for this finding could be that there is a critical level of M A P K suppression that maximizes E S C output and above this, E S C output is decreased. Thus, when P D and A A are both present at lower levels as is the case in the second C C D experiment, a positive interaction is observed, but when they are at higher levels as seen in the F D , a negative interaction is observed. Support for this hypothesis is provided by the high dose inhibition observed for P D (Figure 3.2, Figure 3.4) and the observation that a combination of all three compounds, albeit with P D at a lower level, outperforms any tested two-way combination of the factors (Figure 3.5). This is similar to the high dose inhibition seen in many growth factor dose responses (Audet et al. 2002) It is known that A A acts as an antioxidant in many cellular systems including E S C ( C H G , unpublished observations) where it reduces the levels of toxic reactive oxygen species such as hydrogen peroxide and the superoxide ion (McEligot et al. 2005). Hydrogen peroxide has been shown to cause the activation of the M A P K pathway (Guyton et al . 1996), suggesting that 106 addition of A A leads to a suppression of the M A P K pathway through a reduction in intracellular hydrogen peroxide. Qualitatively there is no difference in the ability of the cells to differentiate towards specific lineages when treated with C S , P D or A A . However, it appears that treatment increases the proportion of cells differentiating towards the mesodermal and endodermal lineages. It would be interesting to explore this further to see whether treatment in specific ways can enhance differentiation towards specific lineages as has been previously observed for A A . A A can be used, in the absence of L I F , to drive differentiation to specific differentiated cell types including cardiomyocytes (Takahashi et al. 2003), neurons (Lee et al. 2000) and osteoblasts (Buttery et al. 2001). It is unclear in these protocols whether A A is in fact directing cell fate or merely enhancing the activity of other compounds present in the differentiation medium. A n inherent problem with the functional assays used here is the shifting magnitude of readout that is observed from experiment to experiment despite consistency within an experiment. It is important to note that despite this shift, trends remain consistent. The cause of this change is unclear but it can provide a challenge in process development and has caused us to propose surrogate assays for E S C developmental potential (Chapter 4). In situations such as these, statistical design of experiments such as F D and C C D can be very useful tools to minimize the impact of these unimportant sources of variance. This is done by inducing replication within blocks of experiments. In this way statistically significant results can be obtained from a minimal number of experiments. This advantage is in addition to providing specific information 107 regarding factor interactions that would not be uncovered by single factor dose response type experiments. In summary, we demonstrate here that A A and P D both increased E S C culture output when added in the presence of L I F . Furthermore, CS when added in combination with either of these factors also had a positive impact on E S C output. Statistical design of experiments allowed the optimization o f E S C culture output with respect to these three factors and uncovered unanticipated factor interactions. Further investigation into the mechanism by which each of these factors act revealed that P D and CS each reduce the spontaneous differentiation observed in standard E S C culture while A A increase the loss of differentiating cells. Despite the optimization methods shown here, the yields of E S C are still not comparable to cultures with P E F , suggesting that identification of more molecules is required to fully mimic these conditions. While signalling through LIF/gpl30/Stat3 and hence the action of P D on human E S C is not relevant, experiments should be performed with A A and CS to determine their activity in these cultures. These experiments should provide a useful framework for making further improvements in the output of undifferentiated E S C in culture. 108 Figure 3.1 Dose response of embryonic stem cells to ascorbic acid (A) Y i e l d of EB-forming cells (B) net growth rate and (C) frequency o f EB-forming cells following treatment of E S C cultures for 72 hours. B u o 2 o.oo o 0 20 40 60 80 100 Ascorbic Acid concentration ((ig/mL) a. 0 20 40 60 80 100 Ascorbic Acid concentration (ng/mL) 18 o e 16 £8 CL X U • O 14 "o 12 o O 0 20 40 60 80 100 Ascorbic Acid concentration (ng/mL) Abbreviations: E B - embryoid body, E S C - embryonic stem cell. 109 Figure 3.2 Dose response of embryonic stem cells to PD98059 (A) Y i e l d of EB-forming cells (B) net growth rate and (C) frequency of EB-forming cells following treatment of E S C cultures for 72 hours. B 10 20 30 40 PD98059 concentration (uM) 10 20 30 40 PD98059 concentration <uM) 10 20 30 40 a, PD98059 concentration (\xM) Abbreviations: E B - embryoid body, E S C - embryonic stem cell. 110 Figure 3.3 Parameter estimates derived from factorial design experiments Parameter estimates for the response of (A) yield of EB-forming cells, (B) net growth rate and (C) frequency of EB-forming cells. Parameters derived from factorial design experiment. ** indicates p < 0.05, * indicated p < 0.1. A B PD x AA CSxAA CSxPD AA PD CS -0.2 -0.1 0.0 0.1 0.2 Parameter Estimates 0.3 PDx AA CSxAA CS x PD AA PI) CS — Parameter Estimates Abbreviations: P D - PD98059, A A - ascorbic acid, CS - chondroitin sulphate - 2 - 1 0 I Parameter Estimates 111 Figure 3.4 Parameter estimates describing yield of embryoid body-forming cells in two central composite design experiments (A) Experiment varying P D and CS (no A A ) . (B) Experiment varying A A and P D with CS held constant. See Table 3.3 for levels of factors in these two experiments. ** indicates p < 0.05, * indicated p < 0.1. A ^ ^ ^ = ^ ^ B PD*PD > * CS'CS — CS'PD — — i * PD * CS ** 1 1 i 1 1 1 1 i ' 1 1 1 i 1 ' 1 1 r^- --0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 -0.04 -0.02 0.00 0.02 0.04 0.06 Parameter Estimates Parameter Estimates Abbreviations: P D - PD98059, A A - ascorbic acid, CS - chondroitin sulphate 112 Figure 3.5 Verification of factor combination that maximizes embryoid body-forming cell output compared to other factor combinations and untreated cells Numbers in column labels indicate concentrations. Units are \xg/mL for A A and CS , p M for P D . ** indicates p < 0.05. 1.6—1 untreated 500CS+20PD 500CS+50AA 500CS+20PD+50AA 500CS+50AA+2PD Treatment Abbreviations: P D - PD98059, A A - ascorbic acid, CS - chondroitin sulphate 113 Figure 3.6 Effect of ascorbic acid, chondroitin sulphate and PD98059 on self-renewal and apoptosis (A) Y ie ld of CFC-forming cells following treatment with A A and P D (B) Y i e l d of C F C forming cells following treatment with C S and P D , either individually or in combination (C) Proportion of A P positive and A P negative cultures in C F C containing A A , C S , P D and P D + C S . (D) Influence of A A on apoptosis rates of SSEA-1 positive and negative cells. B 100000 80000 y 3 60000 C M o 2 40000 '>> 20000 untreated Ascorbic Acid Treatment PD98059 untreated CS PD CS-PI) Treatment D 3 10 > 6 B 2 a. o untreated AA CS PD CS+PD untreated Ascorbic acid PD98059 Treatment Treatment Abbreviations: C F C - colony forming cell, P D - PD98059, A A - ascorbic acid, C S chondroitin sulphate, A P - alkaline phosphatase. 114 Figure 3.7 Affect of differentiation by either leukemia inhibitory factor removal or retinoic acid addition on lineage marker genes of ascorbic acid, chondroitin sulphate and PD98059-treated cells Brachyury -LIF RA Differentiation Treatment -LIF R A -LIF RA Differentiation Treatment Differentiation Treatment Abbreviations: L IF - leukemia inhibitory factor, R A - retinoic acid 115 Table 3.1 Design matrix for a two-level, three-factor factorial design experiment Specific concentrations for each factor are shown in Table 3.3. Run .. A A C S PP 1 -1 - l -1 2 1 - l -1 3 -1 l -1 4 -1 - l 1 5 1 l -1 6 1 - l 1 7 -1 l 1 8 1 l 1 Abbreviations: P D - PD98059, A A - ascorbic acid, CS - chondroitin sulphate 116 Table 3.2 Design matrix for a two-factor central composite design experiment Specific concentrations for each factor are shown in Table 3.3. Run Factor 1 Factor 2 1 -1.63 0 2 -1 -1 3 -1 +1 4 0 -1.63 5 0 +1.63 6 +1 -1 7 +1 +1 8 +1.63 0 9 0 0 10 0 0 11 0 0 12 0 0 117 Table 3.3 Factor concentrations used in the factorial design and central composite design experiments Units for A A and CS are pg/mL and units for P D are p M . Level Factorial Design Experiment 1 s t Cen Desig tral Composite n Experiment 2 n d Central Composite Design Experiment A A CS PD A A ' ' CS PD ' A A • CS PD -1.63 - - - - 2 0.08 0.2 500 0.04 -1 0 0 0 - 5 0.2 0.5 500 0.1 0 - - - - 50 2 5 500 1 +1 50 500 20 - 500 20 50 500 10 +1.63 - - - - 1250 50 130 500 26 Abbreviations: P D - PD98059, A A - ascorbic acid, CS - chondroitin sulphate 118 3.5. References Audet J, Mi l l e r C L , Eaves CJ , Piret J M . 2002. Common and distinct features of cytokine effects on hematopoietic stem and progenitor cells revealed by dose-response surface analysis. Biotechnol Bioeng 80(4):393-404. Burdon T, Stracey C , Chambers I, Nichols J, Smith A . 1999. Suppression of SHP-2 and E R K signalling promotes self-renewal of mouse embryonic stem cells. Dev B i o l 210(l):30-43. Buttery L D , Bourne S, Xynos JD, Wood H , Hughes FJ , Hughes SP, Episkopou V , Polak J M . 2001. Differentiation of osteoblasts and in vitro bone formation from murine embryonic stem cells. Tissue Eng 7(l):89-99. Chang K H , Zandstra P W . 2004. Quantitative screening of embryonic stem cell differentiation: endoderm formation as a model. Biotechnol Bioeng 88(3):287-98. Davey R E , Zandstra P W . 2004. Signal processing underlying extrinsic control of stem cell fate. Curr Opin Hematol 11(2):95-101. Evans M J , Kaufman M H . 1981. Establishment in culture of pluripotential cells from mouse embryos. Nature 292(5819):154-6. Glover C H , Mar in M , Eaves C J , Helgason C D , Piret J M , Bryan J. 2006. Meta-Analysis o f Differentiating Mouse Embryonic Stem Cel l Gene Expression Kinetics Reveals Early Change of a Small Gene Set. PLoS Comput B i o l 2(1 l ) :e l58. Guyton K Z , L i u Y , Gorospe M , X u Q, Holbrook N J . 1996. Activation of mitogen-activated protein kinase by H 2 0 2 . Role in cell survival fol lowing oxidant injury. J B i o l Chem 271(8):4138-42. Laabs T, Carull i D , Geller H M , Fawcett JW. 2005. Chondroitin sulfate proteoglycans in neural development and regeneration. Current Opinion in Neurobiology 15(1):116-120. Lee S H , Lumelsky N , Studer L , Auerbach J M , M c K a y R D . 2000. Efficient generation of midbrain and hindbrain neurons from mouse embryonic stem cells. Nat Biotechnol 18(6):675-9. Mart in G R . 1981. Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc Natl Acad Sci U S A 78(12):7634-8. Matsuda T, Nakamura T, Nakao K , A r a i T, Katsuki M , Heike T, Yokota T. 1999. S T A T 3 activation is sufficient to maintain an undifferentiated state of mouse embryonic stem cells. Embo J 18(15):4261-9. 119 M c E l i g o t A J , Yang S, Meyskens F L , Jr. 2005. Redox regulation by intrinsic species and extrinsic nutrients in normal and cancer cells. Annu Rev Nutr 25:261-95. Mizuguchi S, Uyama T, Kitagawa H , Nomura K H , Dejima K , Gengyo-Ando K , Mi tan i S, Sugahara K , Nomura K . 2003. Chondroitin proteoglycans are involved in cell division of Caenorhabditis elegans. Nature 423(6938):443-8. N i w a H , Burdon T, Chambers I, Smith A . 1998. Self-renewal of pluripotent embryonic stem cells is mediated via activation of S T A T 3 . Genes Dev 12(13):2048-60. Prudhomme W , Daley G Q , Zandstra P, Lauffenburger D A . 2004. Multivariate proteomic analysis of murine embryonic stem cell self-renewal versus differentiation signaling. Proc Natl Acad Sci U S A 101(9):2900-5. Sato N , Meijer L , Skaltsounis L , Greengard P, Br ivan lou A H . 2004. Maintenance of pluripotency in human and mouse embryonic stem cells through activation of Wnt signaling by a pharmacological GSK-3-specific inhibitor. Nat M e d 10(l):55-63. Smith A G , Heath JK , Donaldson D D , Wong G G , Moreau J, Stahl M , Rogers D . 1988. Inhibition of pluripotential embryonic stem cel l differentiation by purified polypeptides. Nature 336(6200):688-90. Sugahara K , M i k a m i T, Uyama T, Mizuguchi S, Nomura K , Kitagawa H . 2003. Recent advances in the structural biology of chondroitin sulfate and dermatan sulfate. Curr Opin Struct B i o l 13(5):612-20. Takahashi K , Murakami M , Yamanaka S. 2005. Role of the phosphoinositide 3-kinase pathway in mouse embryonic stem (ES) cells. Biochem Soc Trans 33(Pt 6): 1522-5. Takahashi T, Lord B , Schulze P C , Fryer R M , Sarang SS, Gullans SR, Lee RT . 2003. Ascorbic acid enhances differentiation of embryonic stem cells into cardiac myocytes. Circulation 107(14):1912-6. Y i n g Q L , Nichols J, Chambers I, Smith A . 2003. B M P induction of Id proteins suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with S T A T 3 . Cel l 115(3):281-92. Youn B S , Sen A , Behie L A , Girgis-Gabardo A , Hassell J A . 2006. Scale-up of breast cancer stem cell aggregate cultures to suspension bioreactors. Biotechnol Prog 22(3):801-10. 120 4. Meta-Analysis of Differentiating Mouse Embryonic Stem Cell Gene Expression Kinetics Reveals Early Change of a Small Gene Set3 4 .1 . Introduction Various types of stem cells are now recognized as responsible both for the generation of tissues and organs during embryonic development and also for the subsequent maintenance and repair of these tissues and organs throughout adult life. This has led to considerable interest in the potential of these stem cell populations to be exploited as cellular therapies for medical conditions where tissue damage or malfunction is severe and irreversible. The cl inical realization of stem cell-based therapies w i l l , however, rely on the development of robust, scalable methods for the ex vivo expansion and controlled manipulation of these cel l populations. Development of such protocols requires extensive testing of multiple factors and culture conditions (Zandstra and Nagy 2001) but this is currently inhibited by the lack of rapid endpoints of stem cell frequency that can be used in high throughput assays. The specific identification of most stem cell types currently relies on the use of functional assays to detect their developmental potential, either in vitro or in vivo (van Os et al. 2004). Such assays are thus, by their very nature, retrospective, protracted, cumbersome and labor-intensive. 3 A version of this chapter has been published as Glover, C H , Mar in , M , Eaves, C J , Helgason, C D , Piret, J M , Bryan, J . Meta-analysis of differentiating mouse embryonic stem cell gene expression kinetics reveals early change of a small gene set. P L o S Computational Biology 2(11): e l58,2006. 121 These features make such assays impractical for large-scale studies and/or rapid screening methodologies. Monitoring critical changes in gene expression using either microarray or high-throughput quantitative R T - P C R (Q-RT-PCR) platforms offers a potentially attractive solution but depends on the identification of a set of genes whose expression changes predict decreased stem cell frequency with adequate precision and specificity. Recently, several groups have described differences in the gene expression profiles of several types of stem cells and their differentiating progeny (Brandenberger et al. 2004; Ivanova et al. 2002; Ramalho-Santos et al. 2002; Sato et al. 2003; Sharov et al. 2003; Tanaka et al. 2002). Most of these investigations have resulted in lists of genes that are too large for comprehensive assessment of their functional significance or specificity. Moreover, many have focused on the detection of altered patterns of gene expression that are more likely to be indicative of emerging differentiated lineages than of altered transcription of genes responsible for sustaining the stem cell state. In many cases, the actual stem cell content of the population was insufficient to infer any changes in the stem cell subset. Mouse embryonic stem cells (ESC) are less problematic in this regard because of their ease of propagation in culture as a predominantly undifferentiated population (Evans and Kaufman 1981; Mar t in 1981) and the availability o f wel l defined protocols for inducing their rapid differentiation into particular lineages. A few genes that are important to the maintenance of the pluripotent status of mouse E S C , like oct4 (Niwa et al. 2000) and nanog (Chambers et al. 2003), have been identified. However, recent studies of the rate at which functional measures of stem cell frequency of mouse E S C are lost indicated that these occur well before changes in oct4 expression are initiated (Chapter 2). The goal of this study 122 was to identify a robust set of early gene expression changes that would be reliable indicators of decreased pluripotent cell content in mouse E S C cultures, regardless of the differentiation stimulus applied. In the fol lowing, the E S C signature change is defined as a set of gene expression changes that are indicative of E S C loss from a population. To identify candidates for inclusion in the E S C signature change, we sought genes that exhibited a pattern of expression consistent with functional assay output in 3 independently generated data sets from ESC-derived cell populations that had been treated for up to 96 hours in several ways to induce differentiation. This strategy required innovation in statistical methodology since the E S C signature change is more complicated than simple differential expression. Here, we present a statistically rigorous approach where the probability that a gene exhibits a predetermined expression pattern is estimated using a semi-parametric bootstrap. The definition of the E S C signature change was specific to each experimental context and, therefore, we obtained from each data set an objective summary of the evidence that a gene is part of the E S C signature change. Genes that exhibited the strongest evidence across all 3 data sets were then tested in other maintenance or differentiating conditions and shown to successfully predict functional assay readout, indicating their potential to be used as an assay in high-throughput screening experiments. 123 4.2 . Materials and Methods 4.2.1. Embryonic Stem Cell Maintenance Cultures J l ( L i et al. 1992) (passage 14) and R l (Nagy et al. 1993) (passage 17) E S C lines and primary embryonic fibroblasts (PEF) were maintained as described in Chapter 2 4.2.2. Embryonic Stem Cell Experimental Cultures Experiments were performed as described in Chapter 3. 4.2.3. Embryoid Body and Colony Forming Cell assays Assays were performed as described in Chapter 2. 4.2.4. RNA Extraction and Array Hybridization Cytoplasmic R N A was extracted using the RNeasy mini kit (Qiagen, Mississauga, ON) . Standard Affymetrix protocols (Affymetrix, Santa Clara, C A ) were used to generate R N A probes from 5 pg of extracted R N A . Samples were hybridized to M O E 4 3 0 A & B chips on a Genechip system (Affymetrix) at the Ottawa Genome Sciences Centre (Ottawa, ON) according to the manufacturer's instructions. 4.2.5. Gene Expression Data Sets Gene expression data was obtained at several time points in 3 independent experiments in which various differentiation induction protocols were applied to 3 mouse E S C lines. Summaries of each of the experiments are shown in Table 4.1. 124 Dimethylsulfoxide ( D M S O ) and retinoic acid (RA)- induced differentiation o f R l cells ( D M S O / R A data set): Independent, duplicate cultures of R l cells (Nagy et al. 1993) were cultured for 96 hours with leukemia inhibitory factor (LIF) ± 2 p M R A or without L I F + 1% D M S O . Samples were hybridized to Affymetrix M O E 4 3 0 A and B Genechips (Table 4.1, ArrayExpress (Parkinson et al. 2005) accession number: E - M E X P - 4 1 2 ) . Induction of R l E S C differentiation by L I F removal ( R l - L R data set): These data have been described in detail in Chapter 2. Briefly, R l cells were cultured in the absence of L I F for 0, 18 and 72 hours. R N A for the 18-hour sample was generated from cells cultured in suspension, while R N A for the 72 hour sample was generated from cells cultured in mefhylcellulose-containing medium. Samples were generated independently in triplicate and hybridized to Affymetrix M G _ U 7 4 v 2 A , B and C Genechips (Table 4.1, ArrayExpress accession number: E -MEXP-414) . Induction of R l , J l and V6.5 differentiation by L I F removal ( M - L R data set): This data set is available from StemBase (www.scgp.ca:8080/StemBase/) (Perez-Iratxeta et al. 2005). R l , J l ( L i et al. 1992) and V6.5 (Eggan et al. 2001) E S C were transferred onto 0.1% gelatin-coated dishes for 48 hours with L IF prior to inoculation in Petri dishes in the absence of L I F and R N A extracts obtained from 0 to 96 hours later. R N A samples were generated independently in triplicate and hybridized to M O E 4 3 0 A and B Affymetrix Genechips. (Table 4.1, Gene Expression Omnibus (Edgar et al. 2002) accession numbers: GSE2972 ( R l data), GSE3231 (V6.5 data) and GSE3749 (J l data)). 125 4.2.6. Gene Expression Analysis In this analysis, as is common practice, hybridization data for each probeset was considered independently, although we recognize that transcripts for many genes would be captured by multiple probesets. Furthermore, although the correspondence between probeset and gene is not, as a rule, one-to-one, we refer to the expression from each probeset as i f it reflected the expression of one gene, unless otherwise stated. A l l pre-processing, including background correction, normalization, probeset summarization, and log2 transformation, was carried out with the R M A algorithm (Irizarry et al. 2003) in the affy package (Gautier et al. 2004) from Bioconductor (Gentleman et al. 2004) and processed data returned by R M A are referred to as expressions. In the equations below we follow these conventions: • Observed intensities were denoted by Yt c o n d , where i indexed the biological replicate, i.e. i€E | l , K , « } , and cond denoted the corresponding condition, i.e. treatment or time. • A l l models are gene specific although, for the sake of simplicity, an explicit index for gene was avoided. • Wi th in the observations for one gene, the random errors e were assumed to be independent and identically distributed and to have expectation zero and a finite, gene specific variance o2 , where exp\s D M S O / R A , R l - L R , or M - L R . 126 To summarize gene expression changes, we fit linear models to the R M A processed data. For the D M S O / R A data we used the following model: ^i,cond = ^+UF + Pcond + £i,cond 0 ) where /J,+UF was the expected intensity in the +LIF control condition, B+UF = 0 by definition, and P D M S 0 (PRA) was the effect of D M S O (RA) treatment, relative to +LIF. For the R l - L R data we used the following model: Yit = ft,* + PISH * A.ISA + P™ * ^TZH + £u (2) where (x0h was the expected intensity at time 0, j8 1 8 A ((372h) was the effect of 18 hours vs. 0 hours (72 hours vs. 18 hours), and Istatement was 1 i f the statement was true and Istatement was 0 otherwise. For the M - L R data we used the following model: Yit = Mo + PiJ + P q u J 2 + (3) where /J,0 was the expected intensity at time 0, t was log2 transformed and centered time where 0 hours was changed to 3 hours to avoid undefined values, and /3 / i n and f}quad gave the linear and quadratic effects of time, respectively. 4.2.7. Defining the Embryonic Stem Cell Signature Change in Terms of Gene Expression For the D M S O / R A data, a gene had to fulfill the following requirements to be included in the E S C signature change: 127 • Absolute change: ficond > Cahs, or ficond < -Cabs for cond = D M S O and cond = R A • Change relative to variability: > Crel for cond = D M S O and cond = R A o • Consistency: PDMSO >0&pRA>0, or PDMSO < 0 & pM < 0 For the R l - L R data, expression requirements were as follows: • Absolute change: Pl8h + p12h > Cahs or, /3 1 8 A + p12h < -Cahs • Change relative to variability: l^ 1 8 * + & 2 *l > Crel o • Consistency: /3{Sh s 0 & fi12h a 0, or finh =s 0 & fi12h £ 0 For the M - L R data, the definition of an interesting expression pattern was more complicated. We required a large absolute difference in the expected expression intensity between the start time, tmin, of the study and the end, f m a x . Specifically, we required that Given equation 3, it can be shown that this is equivalent to the following requirement: C TD -C TD TD, TD, q m d TD, TD, where TDk = - ?* i n. The relative expression change requirement that E[Y, )-E(Y,. ) \ 'max / \ *rrun / •rd o > Cre! (5) is equivalent to the following condition: 128 C TD -C TD R > _i£L(j_ _ Z _ L ft o r ft < Q-- 1 ft PG U A, > ^ ^ ^ P/,„ P , ^ < r A o t ^ P m Consistency was built into the E S C signature change definition by noting the location in the time course of the vertex of the quadratic fit (equation 3). The requirement for strictly increasing or decreasing expression patterns was relaxed by allowing genes whose vertex fell before C m i n or after Cm„ to be retained, where C • and C v are specified relative to the standardized, iTiaX ' in in rritiA A transformed study time. This requirement is captured in the following condition: / B ^ PM t min „ \ r-quad I „ < C- or > C TD2 m n TD2 In all experiments, the specific values of the user-specified thresholds are given in Table 4.2. Note that the final results of this analysis are highly robust to modest differences in these thresholds. 4.2.8. Confidence Values One thousand simulated data sets were generated for each of the original data sets by adding the original fitted averages and a randomly sampled residual (with replacement) from the residuals associated with that gene within the data set. Note that for the M - L R data set, the bootstrap data was derived from time and cell line specific averages and residuals, not from quadratic fits. In order to maintain the covariance between genes, the same random selection of residuals were used for all genes in a simulated data set. The simulated data was then assessed relative to the E S C signature change definitions given above. The proportion of times that a gene fulfilled the definition, i.e. the confidence value ( C V ) , was calculated. A s a practical measure, within each 129 experiment, we used several values for Cabs and averaged the resulting C V to obtain the C V used for Pareto Front Analysis (PFA) . This was an expedient method for reducing the frequency of C V equal to 0 or 1, which are highly undesirable for P F A . The use of multiple thresholds makes the final results of our analysis robust to modest changes. The most stringent values of Cabs were chosen such that the number of genes retained in each data set was approximately equal and a range of less stringent thresholds was applied to create further distinction among the C V s . 4.2.9. Comparison of MOE430 and MG_U74v2 Chips To compare gene samples hybridized to the M O E 4 3 0 and M G U 7 4 chips, 6 different comparisons were used. Two comparisons were generated from the Affymetrix-defined "good" or "best" comparisons (for more information see www.affymetrix.com). Resourcerer (Tsai et al. 2001) was used to generate lists of comparable probesets based on E G O , Unigene, Locuslink and Refseq comparisons. A list of 2-way comparable probesets was generated by ordering evidence for comparison as follows: Affymetrix Best > Affymetrix Good > E G O > Unigene = Locuslink = Refseq. In this way 21, 271 one-to-one mappings between the 2 chips were made. 4.2.10. Gene Ontology Functional information about differentially expressed genes was obtained by loading Affymetrix identifiers into the Database for Annotation, Visualization, and Integrated Discovery 2.0 (Dennis et al. 2003). Gene ontologies were determined and compiled at several different levels of the ontology. 130 4.2.11. Quantitative RT-PCR Quantitative R T - P C R ( Q - R T - P C R ) was performed as described in Chapter 2 and primer sequences are shown in Appendix A . 4.3. Results 4.3.1. Functional Assay Analysis To define the time course of changes in the biological properties of E S C subjected to the differentiation protocols used for gene expression analyses, R l E S C were cultured on 0.1% gelatin coated tissue culture dishes with L I F ± R A or without L I F ± D M S O and aliquots then sequentially tested in 2 colony assays for undifferentiated cell activity - the colony forming cell (CFC) assay performed in liquid cultures containing L I F and the embryoid body (EB)-forming cell assay performed in a semi-solid medium without L I F (Figure 4.1 A , B) . The loss of these activities more closely parallels the loss of stem cell activity measured by contribution to chimeric mice than the loss of expression of S S E A - 1 or Oct4 (Chapter 2). In the starting population 20.9 ± 0.2% of the R l cells were detectable as C F C and 11.2 ± 0.4% gave rise to E B . Exposure to R A had the fastest effect causing a reduction of both these values ~20-fold within 24 hours, whereas simply removing L I F (with or without D M S O addition) caused a corresponding reduction of these values ~6 and 12-fold in the same time frame. After 96 hours, C F C and E B -forming cell frequencies were less than 1% in al l treated cultures. In control cultures the frequencies of both C F C and EB-forming cells were sustained at half o f the starting value as noted previously when R l cells are transferred from cultures containing feeders to gelatin-coated dishes (Chapter 2). 131 To verify that each treatment induced cells to differentiate towards different lineages, we used Q-R T - P C R to monitor changes in transcript levels for 5 differentiation markers in samples obtained after 96 hours of treatment with the 3 differentiation protocols (Figure 4.1 C) . A l l fold-changes were statistically significant relative to the +LIF condition unless otherwise stated. A s expected, removal of L I F , with or without D M S O treatment induced the increased expression of genes associated with ectoderm, neural and mesoderm differentiation: fgf5 (Haub and Goldfarb 1991), nestin (Dahlstrand et al. 1995), and brachyury (Wilkinson et al. 1990), respectively. However, there was little effect on the expression of genes associated with endodermal differentiation: foxa2 and soxl 7 (Kubo et al. 2004). In contrast, treatment with R A strongly induced the markers of neural and endodermal differentiation, but decreased the expression o f brachyury (mesodermal differentiation) and had no significant effect on fgf5 expression (ectodermal differentiation). Overall, all treatments generated a mixed populations of cells. 4.3.2. Gene Expression Analysis For the data from each experiment we applied a gene specific linear model to separate the observed expression into a level for that gene under a reference condition (e.g. +LIF or time 0) plus effects due to treatment and random fluctuation due to biological variabili ty and measurement error. For example, to analyze the D M S O / R A data set, the following model was used: expression = expression in "+LIF" + change due to treatment ( D M S O or R A ) + noise In this case, the three model parameters of primary interest were the change attributable to the effect o f D M S O , the change attributable to the effect of R A , and the typical magnitude of the 132 noise. These changes can be visual ized easily by plotting parameter estimates on a "transcriptome plot" where each gene is represented by a single point (Figure 4.2). For the data from the D M S O / R A experiments, most of the genes in such a plot were found to lie close to the origin (Figure 4.2 A ) indicative of their unaltered expression fol lowing either treatment. However, it is interesting to note that for those genes whose observed expression change was greater than 2 in either treatment (463 genes), both treatments appeared to have similar effects, as indicated by the fact that 98% of these were either increased (209 genes, top right quadrant) or decreased (243 genes, bottom left quadrant) by both treatments. A similar model was fit to the R l - L R data set to obtain estimates of expression changes 18 and 72 hours after the removal of L I F , as wel l as a measure of the noise when this differentiation induction protocol was used (Figure 4.2 C) . For the M - L R data set, use of an analysis of variance-type model, such as those described above, would have resulted in a large number of model parameters. Since interpretability of the parameters is so important in our context, we preferred a smaller, smoother quadratic model based on time. This model was able to capture the temporal trends of expression changes and principal component analysis strongly suggested that a linear combination of constant, linear and quadratic terms explained almost all of the data variability. 4.3.3. Identification of a Robust Set of Early Gene Expression Changes that Indicate Decreased Frequency of Undifferentiated Embryonic Stem Cells We defined the E S C signature change as a set of gene expression changes that were associated with decreased frequencies of E S C as indicated by functional assay readouts during E S C culture. 133 To identify genes that exhibit patterns consistent with the E S C signature change, we imposed three requirements on each data set. When customized to a specific experimental context, these requirements constitute an expression-based definition of the E S C signature change, a prerequisite for developing an appropriate statistical procedure. The 3 requirements used to select expression changes for inclusion were: (1) large change in absolute magnitude, (2) consistent change for all treatments and cell lines, and (3) large change relative to gene specific variability. The first two requirements can be visualized as retaining expression changes falling in certain regions of the transcriptome plots shown in Figure 4.2 (namely, those regions containing black points). The third requirement cannot be visualized directly in a transcriptome plot, but its effect is revealed by the fact that some expression changes in the highlighted regions are not retained, due to large gene specific variance. Applications of these requirements are shown in Figure 4.2 with full explanation detailed in Section 4.2.7. Specific values of the thresholds used for each data set are shown in Table 4.2. 4.3.4. Confidence Values One way to detect E S C signature change genes is to identify those genes whose observed expression patterns fall in regions of interest in transcriptome plots (Figure 4.2). However, this approach ignores the biological and technical noise contained in the observed data. Furthermore, it fails to distinguish between genes whose observed expression changes barely fulfi l l our requirements from those that substantially exceed the specified thresholds. For genes of the latter type, we have more confidence that their true, long-run expression patterns are compatible with our definition of the E S C signature change. We therefore decided to define and exploit a 134 probabilistic quantity that measured our confidence, given the observed data, that a gene exhibits an expression pattern consistent with the E S C signature change (Bryan et al. 2002; van der Laan and Bryan 2001). With in each experiment, we defined a quantity p g for each gene g: the probability that, in a hypothetical repeat of the experiment, the observed expression change of this gene would fulfill our requirements. Genes with true expression changes that substantially exceed all relevant thresholds have a p g greater than those that barely fulfill the requirements. If two genes shared common expression changes but differed with respect to their background variability, the p g o f the gene with less variability would be greater. A l s o , as biological replication increases, the p g of true E S C signature change genes approach 1 and those of all other genes approach 0. Just as p-values are used to rank genes with respect to differential expression, we used p g to rank genes with respect to their consistency with the E S C signature change. Note that genes of primary interest have a p g value near 1, not near 0, as is the case with p-values. B y definition, knowing p g requires knowledge of the true change in expression following induction of differentiation, which is not available. Therefore, we estimated p g by calculating the proportion of bootstrap data sets in which gene g exhibited data that fulfilled our requirements (van der Laan and Bryan 2001) and referred to this quantity as the confidence value (CV) (Efron and Tibshirani 1998). 4.3.5. Meta-Analysis via Pareto Optimization After calculating C V for all genes in the 3 experiments, we conducted a meta-analysis to identify the gene expression changes across all data sets most compatible with the E S C signature change. Genes with expression changes most correlated with decreased frequency of E S C pluripotency 135 would have a C V near 1 in all 3 experiments. If we were working with only one data set, we could rank the genes by C V in decreasing order. However, with C V arising from 2 or more data sets, the task of ranking becomes considerably more difficult. In fact, it is only possible to partially order the genes and we accomplished this with P F A (Fleury et al. 2002). Briefly, in P F A , a comparison is made between all pairs of genes and gene g is said to dominate gene k if, in all experiments, the C V s of gene g are greater than or equal to those of gene k, with strict inequality in at least one experiment. The set of genes not dominated by any others is called the 1 s t Pareto front (PF) and contains the most promising candidates for the E S C signature change. This set is removed from the analysis and the same principle of non-domination is then used to derive successive PFs. A more detailed explanation of P F A can be found in (Fleury et al. 2002). The first 5 PFs identified changes in expression of 89 probesets representing 88 genes (10, 7, 17, 27 and 28 probesets on PFs 1-5, respectively). The genes on PFs 1 to 5 are shown in Table 4.3. 4.3.6. Q - R T - P C R Verif icat ion of A r r a y Results Experiments were undertaken to test, by an independent strategy, the consistency o f the candidates identified from the analysis with the definition of the E S C signature change. Accordingly, R N A extracts were prepared from R l and J l cells cultured for 0, 24, 72 and 96 hours with L IF ± R A or without L I F ± D M S O . Q - R T - P C R was used to measure the changes in levels of 22 selected transcripts (relative to the cells cultured in the presence of LIF) . Nine of these were for genes in the 1 s t PF {103728_at, esrrb, nrObl, tell, hck, gbx2, kl/2, fbxol5 and sppl), 4 for genes in the 2 n d PF (tcfcp2ll, 8430410A17Rik, zfp42, klf4), 5 for genes in the 3 r d PF (sox2,jam2, morc,podxl, sodl), 2 in the 4 t h PF (nrld2, kit) and 2 in the 5 t h PF (mt/2, nmycl). These genes were purposefully chosen to have both high and moderate confidence in their E S C 136 signature change membership (i.e. from the I s to the 5 PFs). ' This tested the breadth of the correlation between the Q - R T - P C R and array results across the complete set of genes contained in the first 5 PFs. Q - R T - P C R results were compared to their corresponding array data, except in the R l - L R data set where the 24-hour Q - R T - P C R results were compared to the 18-hour array data. A l l comparisons of the data sets for matched treatments are plotted in Figure 4.3. Results from the Q - R T - P C R measurements and the microarray analyses were strongly correlated in both cell lines ( R l cell line: Figure 4.3 A , r=0.76; J l cell line: Figure 4.3 B , r=0.82) although the individual changes in gene expression measured by Q - R T - P C R were generally larger than those apparent from the microarray data. There was also a strong correlation between the data obtained for the 2 different cell lines tested (r = 0.86, Figure 4.3 C) . Overall , o f the 22 genes tested, 18 demonstrated the expected kinetics when assessed using Q - R T - P C R . In particular, 7 genes evaluated (i.e. from PF 1: 103728_at, klf2, nrObl and tell; from PF 2: 8430410A17Rik and zfp42; and sox2 from PF 3) showed rapid (within 24 hours) and sustained changes in expression in both E S C lines in all differentiation induction protocols (Figure 4.4). These 7 genes were tested for their ability to predict the time course of functional changes in populations of E S C treated with another differentiation protocol, i.e. exposure to 50 [xg/mL ascorbic acid ( A A ) in the absence of L I F , a treatment reported to promote the generation of cardiac myocytes from undifferentiated E S C (Takahashi et al. 2003). Accordingly, R l E S C were cultured for 5 days on 0.1% gelatin coated tissue culture dishes in standard maintenance 137 conditions and with A A ± L I F and changes in gene expression compared to the loss of E B -forming potential. A s expected, E B potential decreased to 40% of its starting value in the first 24 hours after transferring the cells to the control (+LIF) conditions (without feeders) and then stayed constant over the remaining 5 days of the experiment (Figure 4.5 A ) . Cells cultured without L I F with added A A showed a rapid decrease in E B potential to almost undetectable levels by day 3. In the presence of A A and L I F , there was an enhanced yield of EB-forming cells (with a doubling of the proportion of EB-forming cells when compared to the control +LIF conditions, Chapter 3). Figure 4.5 C shows the time course of changes in transcript levels for the 7 genes that had previously been identified as showing rapid changes in expression in all tested differentiation conditions. It can be seen that all were reduced in the cells exposed to A A in the absence of L IF , consistent with the concordant rapid loss in E B potential. Moreover, when A A was added in the presence of L I F , the level of expression of these genes increased relative to the +LIF control cells, consistent with their opposite biological response to the A A treated cells in the absence of L I F . Zfp42 showed the most rapid increase in expression in the AA+LIF-treated cells and the increase in expression of 103728_at and sox2 was the most delayed. Nevertheless, significantly increased expression of all 7 genes relative to the +LIF controls was seen by 96 hours. Together, these results indicate that changes in expression of these 7 genes can be used to infer concordant functional changes in populations of E S C in culture. 138 4.4. Discussion In this work, we have identified a small set of genes that exhibit the E S C signature change, i.e. whose altered expression is consistently and temporally correlated with an altered frequency of functionally-defined, undifferentiated cells in E S C cultures. This result is important because we had previously found that significant changes of more established molecular markers of undifferentiated mouse E S C (Oct4 and SSEA-1) may not occur until well after the biological hallmarks of these cells have been lost (Chapter 2). B y undertaking an integrated analysis of gene expression changes induced by exposure of E S C to multiple differentiation stimuli and the use of objective statistical methods, we identified 7 genes whose altered expression correctly predicted concomitant functional changes induced by other treatments. Importantly, the expression changes of these genes reflected both decreased and increased frequency of E S C . It is interesting to note that while we did not specifically require a high expression level, the genes identified all exhibited expression above up to two fold higher than the average intensity of expression seen on the microarrays. This was due to our requirement that genes exhibit a large fold change relative to the estimated error. Since genes that are expressed at low levels in cells tend to have a high error estimate, these are effectively eliminated. Four of these 7 genes have been shown previously to be involved in the maintenance of mouse E S C or during early development. These include nrObl (Swain et al. 1998), sox2 (Avi l ion et al. 2003), tell (Narducci et al. 2002) and zfp42 (Rogers et al. 1991). Among the remaining 81 genes on the first 5 PFs, an additional 11 have been reported to be involved in some aspect of development (see Table C I , Appendix C) . Most notably hck (Ernst et al . 1994), fbx015 139 (Tokuzawa et al. 2003), dppa3/stella (Bortvin et al. 2003) and klf4 ( L i et al. 2005) have all been specifically implicated in maintenance or differentiation of E S C , while expression of eed (Mor in -Kens i ck i et al . 2001) is required for embryonic viabi l i ty before implantation. Interestingly, oct4 (Niwa et al. 2000) and nanog (Chambers et al. 2003) were not included in the first 5 PFs. While oct4 was differentially expressed in both the D M S O / R A and M - L R data sets, it was not changed in the R l - L R data set as shown previously (Chapter 2). Nanog was differentially expressed in both the R l - L R and D M S O / R A data sets but did not show any change in the M - L R data set. Several previous mieroarray studies have been performed to uncover signalling pathways and regulatory factors required for the maintenance of human and mouse E S C (Bhattacharya et al. 2004; Brandenberger et al. 2004; Ivanova et al. 2002; Ramalho-Santos et al. 2002; Sato et al. 2003; Sharov et al. 2003; Sperger et al. 2003). These have each uncovered large numbers of genes whose expression was increased or absent in undifferentiated cells and, in some cases, little overlap has been found between the genes thus affected (Evsikov and Solter 2003; Fortunel et al. 2003). A s a further step towards assessing the validity of our identified genes, we compared our data to 2 previously published data sets that sought to identify genes uniquely expressed in mouse E S C (Ivanova et al. 2002; Ramalho-Santos et al. 2002). O f the 88 genes highlighted here, 75 were also changed in the other data sets (Table C2 , Appendix C). This high degree of correspondence supported the validity of our very different approach to a similar biological question. Interestingly, a comparison of our results to genes whose expression has been reported to accompany the differentiation of human E S C (Bhattacharya et al . 2004; 140 Brandenberger et al. 2004; Sato et al. 2003; Sperger et al. 2003), revealed far fewer similarities, as reported by others (Ginis et al. 2004). We found that only 26 of the 88 genes, exhibited similar expression changes in at least one of the 4 human E S C studies. Moreover, in some cases, the gene expression level changed in the opposite direction. For example, in this work, podxl was found to be strongly increased as mouse E S C differentiate, whereas the human homolog was found to be decreased in 3 studies of human E S C differentiation (Bhattacharya et al. 2004; Brandenberger et al. 2004; Sato et al. 2003). Overall, only 6 of the 88 genes were not identified as being altered in any other published data sets of differentiating E S C . The goal of this work was to identify a small number of genes, suitable for the development of an expression-based assay to estimate the frequency of E S C in culture. To achieve this, we have taken care to seek gene expression changes that fulfill several criteria beyond simple differential expression, including large, rapid and consistent changes in more than one cell line following the induction of differentiation using multiple methods. In terms of statistical analysis, we required (1) a quantitative index that reflected each gene's compliance with the predefined E S C signature change (for use within each data set); and (2) a meta-analytic procedure for ranking genes based on their compliance with the E S C signature change definition in multiple data sets. A s our quantitative index, we chose the probability that a gene would exhibit the E S C signature change in a hypothetical repeat of the experiment, instead of the more conventional p-value, the probability that a gene would exhibit data as, or more "extreme' than, the observed change i f its true expression were unchanged in the study. This choice was necessitated because in our 141 application, the E S C signature change is defined by more than a requirement for differential expression. In the majority of microarray applications, the genes of interest are characterized by differential expression, where the complementarity of the null hypothesis (no differential expression) and the biologically interesting state (differential expression) permits the p-value to serve as an index of biological interest. Here we employed an index that originates with an explicit definition of the biologically interesting target profile. This is an application of methods previously described for identifying genes with biologically specific expression patterns (Bryan et al. 2002; van der Laan and Bryan 2001). The task of identifying E S C signature change genes can be viewed also as an instance of the so-called "problem of regions" (Efron and Tibshirani 1998), in which the term "confidence value" is first established. In certain settings, C V s can be shown to be approximations of Bayesian a posteriori probabilities. Although not formally established here, the C V can be interpreted heuristically as the posterior probability that a gene truly exhibits the E S C signature change. A s our method of meta-analysis, we used Pareto Front Analysis (PFA) (Fleury et al. 2002) to (partially) rank genes based on three independent measures of E S C signature change compliance, as opposed to the more prevalent practice of integrating experiment-specific fold-changes (Stevens and Doerge 2005; Wang et al. 2004), p-values (Levesque et al. 2006; Rhodes et al. 2002), test statistics (Estrada et al. 2006; Ghosh et al. 2003), or effect size estimates (Choi et al. 2003; H u et al. 2005). In these works, the common goal is a unified list o f differentially expressed genes that is accompanied by an estimated error rate, usually the false discovery rate (Benjamini and Hochberg 1995; Storey and Tibshirani 2003). The methodological choices and 142 innovations of this work are motivated by departures from this common goal and our techniques may prove useful in other studies where biological interest is not synonymous with differential expression. P F A was first applied to gene expression data by Fleury, et al (Fleury et al. 2002). In that work, P F A was used to optimize multiple indices within one study as opposed to our use, which is the optimization of a comparable index, the C V , across distinct but related studies. Yang et al present another compelling technique for the synthesis of competing measures of differential expression within a single experiment (Yang et al. 2005) and it may be possible to extend their methodology for use in meta-analysis. Meta-analysis of mieroarray data is an increasingly common technique to capitalize on the combined power of biologically related but distinct data sets (Choi et al. 2003; Estrada et al. 2006; Ghosh et al. 2003; H u et al. 2005; Levesque et al. 2006; Rhodes et al. 2002; Stevens and Doerge 2005; Wang et al. 2004). In addition to the usual advantage of increasing the effective sample size, the primary benefit o f meta-analysis in our application is to insulate our biological findings from confounding experimental and biological effects (Estrada et al. 2006; Levesque et al. 2006). For example, in the R l - L R data set, changes induced by differentiation could have been confounded with changes caused by the removal of feeders from the culture. However, this effect was not present in the other data sets, therefore any common gene expression changes cannot be attributed to the removal of feeders. A n example of profound differences in gene expression caused by small changes in culture conditions was reported by Skottman et al., who demonstrated effects on 1417 genes in human E S C cultured in serum containing vs. serum-free 143 conditions, despite comparable levels of expression of other markers of their undifferentiated state (Skottman et al. 2006). In summary, meta-analysis of multiple gene expression data sets from populations of mouse E S C induced to differentiate has revealed multiple genes whose altered expression provides a robust and timely indication of changes in pluripotency. These findings suggest the importance of the products of these genes in the molecular regulation of the undifferentiated state of E S C and provide a useful basis for developing high throughput approaches for the bio-monitoring of E S C cultures. 144 Figure 4.1 The effect of leukemia inhibitory factor removal with or without addition of D M S O or retinoic acid on the maintenance and differentiation of embryonic stem cells (A) Colony-forming cell ( C F C ) frequency and (B) embryoid body (EB)-forming ability o f cells from the E S C cultures assessed at varying times after initiation of the treatment (+LIF controls = • , -LIF = or - L I F + D M S O = • , +LIF+RA = T ) . * denotes the data for the +LIF sample are significantly different from all other treatments. (C) Gene expression of differentiation markers was monitored by Q - R T - P C R after 96 hours of treatment. Results shown are relative to the +LIF control cells. Time (hours) Abbreviations: E S C - embryonic stem cell, L I F - leukemia inhibitory factor, R A - retinoic acid, Q - R T - P C R - quantitative R T - P C R Figure 4.2 Transcriptome plots of estimated expression changes, based on fitting models to each data set A l l plots have density shading to demonstrate the number of points (genes) in different regions. Lines illustrate examples of some of the requirements that make up the definition of the E S C signature change and observed gene expression patterns that fulfilled all requirements are marked as #. Experiment-specific implementations of requirements are explained below. A ) D M S O / R A data set. The requirement for large absolute changes is illustrated by the solid lines. Consistency across conditions implied that genes must exhibit a change in the same direction in both treatments (bottom left or top right quadrant). (B) R l - L R data set. Note that the y axis is the change seen at 72 hours relative to that seen at 18 hours. The requirement for large absolute changes is illustrated by the solid lines. The criterion for consistency was applied by requiring that the change 18 hours after LIF removal be in the same direction as that after 72 hours (i.e., in the lower left or upper right quadrants), regardless of its magnitude. (C) M - L R data set. The requirement for large absolute changes is illustrated by the dashed lines. To meet the consistency criterion, we required that a temporal gene expression trend either increase or decrease continuously over the duration of the experiment (solid lines). This requirement was relaxed slightly to retain trends with a direction change occurring either very early or very late. 1:8 1:4 1:2 1:1 2:1 4:1 8:1 - 1:2 1:1 2:1 Figure 4.3 Comparison of differently measured changes in gene expression within and between two embryonic stem cell lines grown for 0, 24, 72 and 96 hours in +LIF±RA or - L I F ± D M S O (A) Comparison of Q - R T - P C R and mieroarray results for the R l E S C line (r = 0.76). (B) Comparison of Q - R T - P C R and mieroarray results for the J l E S C line (r = 0.82). (C) Comparison of results between the R l and J l E S C lines (r = 0.87). Q - R T - P C R T Q - R T - P C R " B • * • * * * . * • * • * • « - T 1 ^ • • t 1 1:64 1:16 I 4 1:1 A m i v 4:1 16:1 1:64 1:16 1:4 1:1 4:1 Array Jl 1:4 - c • • • • • • » • * • i i i i 1:256 1:64 1:16 1:4 1:1 R l Abbreviations: L IF - leukemia inhibitory factor, R A - retinoic acid, Q - R T - P C R - quantitative R T - P C R , E S C - embryonic stem cell. Figure 4.4 Quantitative R T - P C R profiles of transcript levels for seven genes that showed rapid decrease in embryonic stem cells subjected to several differentiating conditions A l l levels of expression are relative to the +LIF control. +LIF control - L I F + D M S O = JI, +LIF+RA = 6. 1.5-, _ 1.5T Klf2 1.04 -LIF = ® , or t 1 103728_at 1 M i l • 1 1 1.0 0.5 • 1 8430410A17Rik i i A N 1 ¥ ¥ 20 40 60 80 100 0 1.5 NrObl -¥ ¥ i-o-20 40 60 80 100 20 40 60 80 100 Tcl1 t m 1 II 20 40 60 80 100 Zfp42 • i n II 1 1 ¥ 1 1 0 20 40 60 80 100 Time (h) Abbreviations: L I F - leukemia inhibitory factor, R A - retinoic acid. 148 Figure 4.5 Comparison o f biological and molecular changes in embryonic stem cells stimulated to differentiate by exposure to ascorbic acid (A) Changes in embryoid body (EB) frequency. (B) Changes in the levels of transcripts for seven E S C signature change genes measured by Q-RT-PCR. +LIF ( • ) , +LIF+AA ( • ) and - L I F + A A (A) A 10n 40 60 80 Time (h) B + 0 o ~-' 4 T3 03 3 co § 2 o c 1 o 103728 at a A i 8430410A17Rik o Klf2 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 Nr0b1 -4 k o Sox2 Tcl1 A 03 0 20 40 60 80 100 6 20 40 60 80 100 0 20 40 60 80 100 I-X 0) 2| 1* / Zfp42 0 20 40 60 80 1°° T ime (h) Abbreviations: E S C - embryonic stem cell, Q - R T - P C R - quantitive R T - P C R , LIF - leukemia inhibitory factor, A A - ascorbic acid Table 4.1 Summary of all microarray experiments used in this study Note that all replicates are biological. ...v .'/ DMSO/RA Rl-LR M-LR Condition l%DMSO/no LIF 2 \iM retinoic acid/10 ng/mL LIF 10 ng/mL LIF (positive control) LIF removal LIF removal Cell line R l R l R l , J l , V6.5 Timepoint (hours) 96 0, 18, 72 0, 6, 12, 18,24, 36, 48, 96 Number of replicates per condition (total number of chips) 2(6) 3(9) 3(72) Affymetrix Chip MOE430 M G U74v2 MOE430 Abbreviations: L IF - leukemia inhibitory factor 150 Table 4.2 Thresholds used in the definition of the embryonic stem cell signature chang Parameter D M S O / R A R l - L R M - L R Cabs. 2.0,2.3,3.0 1.5, 1.7, 2.0 1.4, 1.8, 2.2 Cmin •"* 0.4 (12 h) Cmax 0.9 (68 h) C r e ] 1.84(DMSO), 2.72 (RA) 1.22 1.57 Number of genes, CV>0.5 265 269 240 Abbreviations: R A - retinoic acid, C V - confidence value 151 Table 4.3 Genes identified on the first five Pareto fronts MG_U74v2 MOE430 Confidence Values Fold change at 96 hours Fold change following LIF removal (R1LR) Fold change following L I F removal for 96 hours (MLR) Gene Name Gene Symbol D M S O / R A R 1 L R M L R D M S O R A 18 h 72 h J l R l V6.5 Pareto Front 1 103728 at 1456521 at 1.00 1.00 0.07 0.18 0.19 0.65 0.24 0.58 0.64 0.29 Transcribed locus — 168508 at 1436926 at 1.00 0.86 0.35 0.09 0.09 0.63 0.20 0.36 0.61 0.24 estrogen related receptor, beta Esrrb 93141 at 1417760 at 1.00 0.81 1.00 0.06 0.08 0.84 0.10 0.09 0.12 0.08 nuclear receptor subfamily 0, group B, member 1 NrObl 93296 at 1422458 at 1.00 0.98 0.10 0.19 0.19 0.78 0.22 0.48 0.73 0.42 T-cell lymphoma breakpoint 1 Tell 93483 at 1449455 at 1.00 0.85 0.67 0.18 0.17 0.89 0.15 0.36 0.44 0.50 hemopoietic cell kinase lick 94200 at 1420337 at 0.92 0.98 1.00 0.23 0.30 0.69 0.22 0.26 0.29 0.25 gastrulation brain homeobox 2 Gbx2 96109 at 1448890 at 1.00 1.00 0.01 0.11 0.11 0.74 0.41 0.49 0.78 0.31 Kruppel-like factor 2 (lung) Klf2 96162 at 1427238 at 0.83 1.00 0.53 0.11 0.33 0.63 0.10 0.53 0.51 0.18 F-box protein 15 FbxolS 97519 at 1449254 at 0.02 0.99 0.55 0.08 0.77 0.54 0.19 0.54 0.44 0.10 secreted phosphoprotein 1 Sppl 99561 f at 1448393 at 0.00 1.00 1.00 5.23 0.67 1.34 4.15 4.67 7.25 3.34 claudin 7 Cldnl Pareto Front 2 103761 at 1418091 at 1.00 1.00 0.07 0.19 0.26 0.64 0.13 0.60 0.66 0.22 transcription factor CP2-like 1 TcfcplU 108097 at 1450626 at 0.93 0.80 0.10 0.15 0.29 0.74 0.27 0.69 0.71 0.25 mannosidase, beta A, lysosomal Manba 108712 at 1434917 at 0.98 0.77 1.00 0.26 0.29 0.73 0.32 0.36 0.38 0.28 cordon-bleu Cobl 160684 at 1423786 at 1.00 0.73 1.00 0.27 0.23 0.97 0.46 0.29 0.24 0.17 RIKEN cDNA 843041 OA 17 gene 843041 OA 17 Rik 95033 at 1426810 at 0.50 0.97 0.97 0.42 0.30 0.84 0.41 0.31 0.33 0.39 jumonji domain containing 1A Jmjdia 98414 at 1418362 at 1.00 0.73 0.97 0.09 0.10 0.88 0.14 0.28 0.40 0.11 zinc finger protein 42 Zfp42 99622 at 1417394 at 0.70 1.00 0.00 0.09 0.33 0.26 0.07 0.30 0.14 0.11 Kruppel-like factor 4 (gut) Kl/4 Pareto Front 3 100009 r at 1416967 at 1.00 0.67 0.00 0.16 0.07 0.89 0.36 0.59 1.11 0.39 SRY-box containing gene 2 Sox2 101560 at 1415856 at 0.00 0.60 1.00 1.22 0.43 1.09 1.85 3.37 4.71 3.22 em big in Emb 102012 at 1418895 at 0.00 0.87 0.30 0.61 2.18 0.74 0.48 0.54 0.66 0.52 src family associated phosphoprotein 2 Scap2 102332 at 1448370 at 0.38 0.99 0.00 0.45 0.42 0.79 0.38 1.02 0.95 0.78 Unc-51 like kinase 1 (C. elegans) Ulkl Fold change Fold change MG_U74v2 MOE430 Confidence Values Fold change at 96 hours following L I F removal (R1LR) following L IF removal for 96 hours (MLR) Gene Name Gene Symbol D M S O / R A R 1 L R M L R D M S O R A 18 h 72 h J l R l V6.5 108784 at 1455604 at 0.91 0.92 0.02 0.32 0.25 0.51 0.36 0.46 0.63 0.58 expressed sequence AI427138 AN27138 112828 at 1448688 at 0.87 0.11 1.00 3.07 5.89 1.18 1.45 10.45 4.96 4.45 podocalyxin4ike Podxl 115445 at 1435374 at 0.84 0.88 0.03 0.22 0.32 0.64 0.34 0.75 0.56 0.38 Transcribed locus — 116872 at 1435437 at 0.82 0.72 0.96 3.08 3.01 1.27 2.07 2.78 2.49 2.84 SET domain-containing protein 7 MGl: 192050 1 117246 at 1448845 at 0.69 0.27 1.00 0.23 0.31 0.77 0.65 0.29 0.26 0.32 ribonuclease P 25 subunit (human) Rpp25 133365 at 1436568 at 0.41 0.86 0.87 0.23 0.44 0.74 0.42 0.41 0.31 0.44 junction adhesion molecule 2 Jam2 133819 at 1419418 a at 1.00 0.66 0.99 0.08 0.09 0.79 0.24 0.38 0.33 0.09 microrchidia More 163005 s at 1429366 at 0.99 0.76 0.00 0.19 0.22 0.65 0.28 0.49 0.91 0.34 leucine rich repeat containing 34 Lrrc34 92770 at 1421375 a at 0.38 0.98 0.00 2.1 1 5.38 0.33 0.16 0.36 0.35 0.05 SI00 calcium binding protein A6 (calcyclin) S100a6 93864 s at 1421624 a at 0.04 0.48 1.00 0.30 0.65 1.03 0.36 0.38 0.35 0.34 enabled homolog (Drosophila) Enah 94745 f at 1427479 at 0.01 1.00 0.00 0.73 5.61 0.49 0.26 0.89 0.35 1.00 eukaryotic translation initiation factor IA Eifla 96752 at 1424067 at 0.00 0.99 0.06 0.63 0.80 0.69 0.41 0.67 0.75 0.61 intercellular adhesion molecule lea ml 97890 at 1416041 at 0.83 0.88 0.06 0.21 0.32 0.87 0.38 0.57 0.73 0.23 serum/glucocorticoid regulated kinase Sgk Pareto Front 4 100030 at 1448562 at 0.34 0.91 0.01 0.47 0.17 0.83 0.36 0.25 0.39 0.85 uridine phosphorylase 1 Uppl 100301 at 1422986 at 1.00 0.57 0.44 0.21 0.21 0.92 0.54 0.40 0.57 0.37 estrogen related receptor, beta Esrrb 103342 at 1448653 at 0.11 0.86 0.61 0.45 0.55 0.72 0.49 0.41 0.33 0.25 embryonic ectoderm development Eed 103653 at 1449590 a at 0.00 0.95 0.03 0.38 1.16 0.73 0.47 0.39 0.39 0.41 muscle and microspikes RAS Mras 104139 at 1452094 at 0.00 0.45 1.00 1.50 1.88 0.91 0.57 5.73 3.69 2.91 procollagen-proline, 2-oxoglutarate 4-dioxygenase (proline 4-hydroxylase), alpha 1 polypeptide P4hal 104544 at 1423327 at 0.53 0.90 0.00 0.16 0.41 0.75 0.43 0.88 0.36 0.42 RIKEN cDNA 4930517K11 gene 4930517K11 Rik 107103 at 1416958 at 0.00 0.96 0.00 0.51 0.73 0.60 0.43 0.80 0.91 0.72 nuclear receptor subfamily 1, group D, member 2 Nrld2 M G U74v2 MOE430 Confidence Values Fold change at 96 hours Fold change following LIF removal (R1LR) Fold change following L IF removal for 96 hours (MLR) Gene Name Gene Symbol D M S O / R A R 1 L R M L R D M S O R A 18 h 72 h J l R l V6.5 108010 at 1418318 at 0.87 0.54 0.99 3.13 16.53 1.17 1.87 4.61 3.73 2.54 ring finger protein 128 Rnfl28 108048 at 1454788 at 0.67 0.08 1.00 2.62 5.18 1.00 1.37 3.02 3.47 3.13 ADP-ribosylation factor-like 7 Aril 113673 at 1423508 at 0.56 0.72 0.02 0.28 0.42 0.69 0.38 0.57 0.59 0.36 MYST histone acetyltransferase monocytic leukemia 4 Myst4 115804 at 1438237 at 1.00 0.57 0.68 0.16 0.14 0.88 0.50 0.49 0.22 0.12 RIKEN cDNA C230088H06 gene — 133204 at 1455425 at 1.00 0.47 0.99 0.14 0.08 0.96 0.43 0.28 0.33 0.26 Expressed sequence BB001228 BB001228 160370 at 1416552 at 0.99 0.70 0.00 0.17 0.25 0.91 0.44 0.93 1.08 0.44 developmental pluripotency associated 5 Dppa5 160828 at 1426858 at 0.76 0.62 0.25 0.31 0.33 0.64 0.53 0.38 0.38 0.31 inhibin beta-B Inhbb 93063 at 1427442 a at 0.71 0.00 1.00 2.70 9.60 0.50 0.85 6.52 3.84 3.02 amyloid beta (A4) precursor protein App 93104 at 1426083 a at 0.00 0.79 0.62 0.92 1.83 0.65 0.48 2.74 3.04 1.93 B-cell translocation gene 1, anti-proliferative BtgJ 94270 at 1448169 at 0.85 0.86 0.00 4.12 12.01 1.27 2.57 12.20 18.17 1.76 keratin complex 1, acidic, gene 18 Krtl-18 94354_at 1421840 at 0.00 0.98 0.00 0.37 0.83 0.53 0.36 1.12 0.67 0.75 ATP-binding cassette, sub-family A (ABC1), member 1 Abcal 95518 at 1424683 at 0.00 0.13 1.00 1.74 3.14 1.22 1.43 7.05 5.20 3.40 RIKEN cDNA 1810015C04 gene 1810015C04 Rik 95531 at 1454890 at 0.35 0.08 1.00 3.90 2.22 0.79 3.31 20.76 7.86 5.62 angiomotin Amot 96042 at 1448610 a at 0.53 0.72 0.94 0.43 0.32 0.75 0.52 0.30 0.42 0.28 superoxide dismutase 2, mitochondrial Sod2 96841 at 1451069 at 0.00 1.00 0.00 0.71 0.46 0.73 0.41 1.39 1.26 1.09 proviral integration site 3 Pim3 96900 at 1433720 s at 1.00 0.62 0.00 0.18 0.12 0.92 0.54 0.45 0.85 0.48 Nur77 downstream gene 2 MGL214355 8 97317 at 1415894 at 0.00 1.00 0.00 1.05 0.79 2.00 5.20 1.05 1.00 1.21 Ectonucleotide pyrophosphatase/phosphodiesterase 2 Enppl 97426 at 1416529 at 0.00 0.96 0.00 0.67 6.50 0.28 0.15 0.75 0.63 0.57 epithelial membrane protein 1 Empl 97520 s at 1423506 a at 0.08 0.97 0.00 1.73 3.84 1.21 2.33 0.66 1.13 1.39 neuronatin Nnat 99956 at 1452514 a at 0.00 1.00 0.00 0.67 0.96 0.49 0.25 2.31 1.57 1.04 kit oncogene Kit MG_U74v2 MOE430 Confidence Values Fold change at 96 hours Fold change following L I F removal (R1LR) Fold change following LIF removal for 96 hours ( M L R ) Gene Name Gene Symbol D M S O / R A R 1 L R M L R D M S O R A 18 h 72 h J l R l V6.5 Pareto Front 5 103048 at 1417155 at 0.98 0.39 0.64 0.30 0.25 0.96 0.58 0.35 0.50 0.35 neuroblastoma myc-related oncogene 1 Nmycl 103234 at 1424847 at 0.92 0.01 0.78 0.29 0.25 1.02 0.94 0.43 0.35 0.45 neurofilament, heavy polypeptide Nefh 103737 at 1418753 at 0.53 0.69 0.33 0.39 0.43 0.95 0.51 0.40 0.63 0.65 — — 108279 at 1455300 at 0.50 0.52 0.63 0.35 0.42 0.93 0.35 0.37 0.51 0.29 RIKEN cDNA E130014J05 gene E130014J05 Rik 110429 at 1455333 at 0.01 0.80 0.08 0.28 0.66 0.72 0.36 0.39 0.32 0.25 cDNA sequence BC023928 BC023928 111970 at 1460711 at 0.76 0.39 0.46 0.36 0.24 0.92 0.57 0.54 0.54 0.40 RIKEN cDNA 4930461P20 gene 4930461P20 Rik 115058 at 1434362 at 0.41 0.68 0.91 2.65 2.23 1.25 2.04 2.47 2.28 2.83 expressed sequence AW550831 AW550831 116214_at 1456329 at 0.58 0.37 0.72 2.44 3.11 1.19 1.91 2.29 2.11 2.45 RIKEN cDNA A230098A12 gene A230098A12 Rik 116435 at 1418076 at 0.00 0.93 0.01 1.90 0.78 1.65 2.22 1.70 2.03 1.24 suppression of tumorigenicity 14 (colon carcinoma) Stl4 160253 at 1423754 at 0.06 0.74 0.31 0.34 0.56 0.82 0.51 0.53 0.66 0.34 interferon induced transmembrane protein 3 lfitm3 161042 at 1427912 at 0.75 0.75 0.00 0.35 0.29 1.27 1.92 0.80 0.75 0.77 carbonyl reductase 3 Cbr3 161106 r at 1443892 at 1.00 0.00 0.97 0.21 0.16 1.09 0.78 0.36 0.21 0.20 . . . — 162522 f at 1437015_x_a t 0.17 0.00 0.98 0.51 0.22 0.95 0.78 0.29 0.42 0.39 phospholipase A2, group IB, pancreas Pla2glb 163288 at 1460471 at 0.98 0.70 0.00 0.23 0.26 0.82 0.36 0.78 0.93 0.40 RIKEN cDNA 2410146L05 gene 2410146L05 Rik 163489 at 1418488 s at 0.00 0.85 0.42 1.52 1.06 1.35 2.24 3.03 2.75 1.72 receptor-interacting serine-threonine kinase 4 Ripk4 163715 at 1429399 at 1.00 0.50 0.32 0.19 0.17 0.92 0.47 0.34 0.49 0.59 ring finger protein 125 Rnfl25 165699 r at 1453299 a at 0.56 0.00 1.00 0.42 0.37 0.98 1.02 0.30 0.23 0.29 purine-nucleoside phosphorylase Pup 166142 r at 1436799 at 1.00 0.00 0.97 0.13 0.13 1.16 0.96 0.39 0.27 0.17 RIKEN cDNA D230005D02 gene D230005D0 2Rik 167088 r at 1456242 at 0.41 0.47 0.78 0.34 0.45 0.98 0.13 0.27 0.47 0.24 LOC433110 LOC433110 Fold change following L IF removal Fold change following LIF removal for 96 hours Fold change at 96 hours MG_U74v2 MOE430 Confidence Values (R1LR) (MLR) Gene Symbol D M S O / R A R 1 L R M L R D M S O R A 18 h 72 h J l R l V6.5 Gene Name 92275 at 1418147 at 0.00 0.95 0.00 j 0.56 0.45 0.77 0.42 1.25 1.48 0.92 transcription factor AP-2, gamma Tcfap2c 92476 at 1449288 at 1.00 0.33 0.75 0.23 0.16 0.85 0.62 0.43 0.47 0.29 growth differentiation factor 3 Gdft 92550 at 1417156 at 0.98 0.63 0.00 5.12 5.63 1.22 2.04 1.90 1.70 1.22 keratin complex 1, acidic, gene 19 Krtl-19 93271 s at 1450186 s at 0.58 0.01 0.88 2.44 3.17 0.82 1.38 2.24 2.90 2.98 GNAS (guanine nucleotide binding protein, alpha stimulating) complex locus Gnas 95584 at 1453223_s_at 0.03 0.08 1.00 0.25 0.64 1.08 0.63 0.34 0.23 0.11 developmental pluripotency associated 2 Dppa2 96203 at 1424713 at 0.06 0.27 0.92 0.50 0.59 1.06 0.54 0.23 0.43 0.11 calmodulin-like 4 Calml4 97083 at 1441023 at 0.00 0.99 0.00 1.00 0.92 0.51 0.35 1.10 0.93 1.07 eukaryotic translation initiation factor 2, subunit 2 (beta) Eif2s2 97283 at 1424295 at 0.68 0.83 0.00 0.37 0.32 0.84 0.36 1.14 2.48 0.44 developmental pluripotency-associated 3 Dppa3 97442 at 1416832 at 0.00 0.14 1.00 1.20 1.05 1.08 1.47 7.27 3.55 4.90 solute carrier family 39 (metal ion transporter), member 8 Slc39a8 4.5. References A v i l i o n A A , Nicol is S K , Pevny L H , Perez L , V iv i an N , Lovell-Badge R. 2003. Multipotent cell lineages in early mouse development depend on S O X 2 function. Genes Dev 17(1): 126-40. Benjamini Y , Hochberg Y . 1995. Controll ing the False Discovery Rate; A Practical and Powerful Approach to Mult ip le Testing. Journal of the Royal Statistical Society Series B 57(1):289 - 300. Bhattacharya B , Miura T, Brandenberger R, Mejido J, Luo Y , Yang A X , Joshi B H , Ginis I, Thies RS , Ami t M , Lyons I, Condie B G , Itskovitz-Eldor J, Rao M S , Puri R K . 2004. Gene expression in human embryonic stem cell lines: unique molecular signature. Blood 103(8):2956-64. Bortvin A , Eggan K , Skaletsky H , Akutsu H , Berry D L , Yanagimachi R, Page D C , Jaenisch R. 2003. Incomplete reactivation of Oct4-related genes in mouse embryos cloned from somatic nuclei. Development 130(8): 1673-80. Brandenberger R, Wei H , Zhang S, L e i S, Murage J, Fisk GJ , L i Y , X u C, Fang R, Guegler K , Rao M S , Mandalam R, Lebkowski J, Stanton L W . 2004. Transcriptome characterization elucidates signaling networks that control human E S cel l growth and differentiation. Nat Biotechnol 22(6):707-16. Bryan J, Pollard K S , van der Laan M J . 2002. Paired and unpaired comparison and clustering with gene expression data. Stat Sinica 12(1):87-110. Chambers I, Colby D , Robertson M , Nichols J, Lee S, Tweedie S, Smith A . 2003. Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. Ce l l 113(5):643-55. Choi J K , Y u U , K i m S, Y o o OJ . 2003. Combining multiple mieroarray studies and modeling interstudy variation. Bioinformatics 19 Suppl l:i84-90. Dahlstrand J, Lardelli M , Lendahl U . 1995. Nestin m R N A expression correlates with the central nervous system progenitor cell state in many, but not all , regions of developing central nervous system. Brain Res Dev Brain Res 84(1): 109-29. Dennis G , Jr., Sherman B T , Hosack D A , Yang J, Gao W , Lane H C , Lempick i R A . 2003. D A V I D : Database for Annotation, Visualizat ion, and Integrated Discovery. Genome B i o l 4(5):P3. Edgar R, Domrachev M , Lash A E . 2002. Gene Expression Omnibus: N C B I gene expression and hybridization array data repository. Nucleic Acids Res 30(1):207-10. 157 Efron B , Tibshirani R. 1998. The problem of regions. A n n Stat 26(5): 1687-1718. Eggan K , Akutsu H , Loring J, Jackson-Grusby L , K lemm M , Rideout W M , 3rd, Yanagimachi R, Jaenisch R. 2001. Hybr id vigor, fetal overgrowth, and viabili ty of mice derived by nuclear cloning and tetraploid embryo complementation. Proc Natl Acad Sci U S A 98(11):6209-14. Ernst M , Gearing DP , Dunn A R . 1994. Functional and biochemical association of Hck with the LIF/ IL-6 receptor signal transducing subunit gpl30 in embryonic stem cells. Embo J 13(7): 1574-84. Estrada B , Choe SE, Gisselbrecht SS, Michaud S, Raj L , Busser B W , Halfon M S , Church G M , Michelson A M . 2006. A n integrated strategy for analyzing the unique developmental programs of different myoblast subtypes. PLoS Genet 2(2):el6. Evans M J , Kaufman M H . 1981. Establishment in culture of pluripotential cells from mouse embryos. Nature 292(5819): 154-6. Evsikov A V , Solter D . 2003. Comment on " 'Sternness': transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature". Science 302(5644):393; author reply 393. Fleury G , Hero A O , Yosida S, Carter T, Barlow C, Swaroop A . Pareto analysis for gene filtering in mieroarray experiments. Proc X I European Signal Processing Conference; 2002; Toulouse, France, p 165-168. Fortunel N O , Otu H H , N g H H , Chen J, M u X , Chevassut T, L i X , Joseph M , Bailey C, Hatzfeld J A , Hatzfeld A , Usta F, Vega V B , Long P M , Libermann T A , L i m B . 2003. Comment on " 'Sternness': transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature". Science 302(5644):393; author reply 393. Gautier L , Cope L , Bolstad B M , Irizarry R A . 2004. affy—analysis of Affymetrix GeneChip data at the probe level. Bioinformatics 20(3):307-15. Gentleman R C , Carey V J , Bates D M , Bolstad B , Dettling M , Dudoit S, El l i s B , Gautier L , Ge Y , Gentry J, Hornik K , Hothorn T, Huber W , Iacus S, Irizarry R, Leisch F, L i C, Maechler M , Rossini A J , Sawitzki G , Smith C, Smyth G , Tierney L , Yang J Y , Zhang J. 2004. Bioconductor: open software development for computational biology and bioinformatics. Genome B i o l 5(10):R80. Ghosh D , Barette TR, Rhodes D , Chinnaiyan A M . 2003. Statistical issues and methods for meta-analysis of mieroarray data: a case study in prostate cancer. Funct Integr Genomics 3(4): 180-8. 158 Ginis I, Luo Y , Miura T, Thies S, Brandenberger R, Gerecht-Nir S, Ami t M , Hoke A , Carpenter M K , Itskovitz-Eldor J, Rao M S . 2004. Differences between human and mouse embryonic stem cells. Dev B i o l 269(2):360-80. Haub O, Goldfarb M . 1991. Expression of the fibroblast growth factor-5 gene in the mouse embryo. Development 112(2):397-406. H u P, Greenwood C M , Beyene J. 2005. Integrative analysis of multiple gene expression profiles with quality-adjusted effect size models. B M C Bioinformatics 6:128. Ihaka R, Gentleman R. 1996. R: A Language for Data Analysis and Graphics. Journal of Computational and Graphical Statistics 5(3):299-314. Irizarry R A , Bolstad B M , C o l l i n F, Cope L M , Hobbs B , Speed TP . 2003. Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 31(4):el5. Ivanova N B , Dimos JT, Schaniel C , Hackney J A , Moore K A , Lemischka IR. 2002. A stem cell molecular signature. Science 298(5593):601-4. Kubo A , Shinozaki K , Shannon J M , Kouskoff V , Kennedy M , Woo S, Fehling H J , Kel ler G . 2004. Development of definitive endoderm from embryonic stem cells in culture. Development 131(7):1651-62. Levesque M P , Vernoux T, Busch W , C u i H , Wang J Y , B l i l o u I, Hassan H , Nakajima K , Matsumoto N , Lohmann J U , Scheres B , Benfey P N . 2006. Whole-genome analysis o f the S H O R T - R O O T developmental pathway in Arabidopsis. PLoS B i o l 4(5):el43. L i E , Bestor T H , Jaenisch R. 1992. Targeted mutation of the D N A mefhyltransferase gene results in embryonic lethality. Cel l 69(6):915-26. L i Y , McCl in t ick J, Zhong L , Edenberg H J , Yoder M C , Chan RJ . 2005. Murine embryonic stem cell differentiation is promoted by SOCS-3 and inhibited by the zinc finger transcription factor Klf4 . Blood 105(2):635-7. L ivak K J , Schmittgen T D . 2001. Analysis of relative gene expression data using real-time quantitative P C R and the 2(-Delta Delta C(T)) Method. Methods 25(4):402-8. Mart in G R . 1981. Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc Natl Acad Sci U S A 78(12):7634-8. Morin-Kensicki E M , Faust C, LaMantia C, Magnuson T. 2001. Cel l and tissue requirements for the gene eed during mouse gastrulation and organogenesis. Genesis 31(4): 142-6. 159 Nagy A , Rossant J, Nagy R, Abramow-Newerly W , Roder JC. 1993. Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc Natl Acad Sci U S A 90(18):8424-8. Narducci M G , Fiorenza M T , Kang S M , Bevilacqua A , D i Giacomo M , Remotti D , Picchio M C , Fidanza V , Cooper M D , Croce C M , Mangia F , Russo G . 2002. T C L 1 participates in early embryonic development and is overexpressed in human seminomas. Proc Natl Acad Sci U S A 99(18):11712-7. N i w a H , Miyazaki J, Smith A G . 2000. Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells. Nat Genet 24(4):372-6. Palmqvist L , Glover C H , Hsu L , L u M , Bossen B , Piret J M , Humphries R K , Helgason C D . 2005. Correlation of murine embryonic stem cell gene expression profiles with functional measures of pluripotency. Stem Cells 23(5):663-80. Parkinson H , Sarkans U , Shojatalab M , Abeygunawardena N , Contrino S, Coulson R, Fame A , Lara G G , Holloway E , Kapushesky M , L i l j a P, Mukherjee G , Oezcimen A , Rayner T, Rocca-Serra P, Sharma A , Sansone S, Brazma A . 2005. ArrayExpress—a public repository for microarray gene expression data at the E B I . Nucleic Acids Res 33(Database issue):D553-5. Perez-Iratxeta C , Palidwor G , Porter C J , Sanche N A , Huska M R , Suomela B P , Muro E M , Krzyzanowski P M , Hughes E , Campbell P A , Rudnicki M A , Andrade M A . 2005. Study of stem cell function using microarray experiments. F E B S Lett 579(8): 1795-801. Ramalho-Santos M , Y o o n S, Matsuzaki Y , Mul l igan R C , Mel ton D A . 2002. "Sternness": transcriptional profiling of embryonic and adult stem cells. Science 298(5593):597-600. Rhodes D R , Barrette T R , Rubin M A , Ghosh D , Chinnaiyan A M . 2002. Meta-analysis of microarrays: interstudy validation of gene expression profiles reveals pathway dysregulation in prostate cancer. Cancer Res 62(15):4427-33. Rogers M B , Hosier B A , Gudas L J . 1991. Specific expression of a retinoic acid-regulated, zinc-finger gene, Rex-1, in preimplantation embryos, trophoblast and spermatocytes. Development 113(3):815-24. Sato N , Sanjuan I M , Heke M , Uchida M , Naef F, Brivanlou A H . 2003. Molecular signature of human embryonic stem cells and its comparison with the mouse. Dev B i o l 260(2):404-13. Sharov A A , Piao Y , Matoba R, Dudekula D B , Qian Y , VanBuren V , Falco G , Martin PR, Stagg C A , Bassey U C , Wang Y , Carter M G , Hamatani T, A i b a K , Akutsu H , Sharova L , Tanaka TS, Kimber W L , Yoshikawa T, Jaradat S A , Pantano S, Nagaraja R, Boheler K R , Taub D , Hodes RJ , Longo D L , Schlessinger D , Keller J, Klo tz E , Kelsoe G , Umezawa A , Vescovi A L , Rossant J, 160 Kunath T, Hogan B L , Curci A , D'Urso M , Kelso J, Hide W , K o M S . 2003. Transcriptome analysis of mouse stem cells and early embryos. PLoS B i o l 1(3):E74. Skottman H , Stromberg A M , Matilainen E , Inzunza J, Hovatta O, Lahesmaa R. 2006. Unique gene expression signature by human embryonic stem cells cultured under serum-free conditions correlates with their enhanced and prolonged growth in an undifferentiated stage. Stem Cells 24(l):151-67. Sperger J M , Chen X , Draper JS, Antosiewicz JE, Chon C H , Jones SB, Brooks JD, Andrews P W , Brown P O , Thomson J A . 2003. Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors. Proc Natl Acad Sci U S A 100(23): 13350-5. Stevens JR, Doerge R W . 2005. Combining Affymetrix mieroarray results. B M C Bioinformatics 6:57. Storey JD, Tibshirani R. 2003. Statistical significance for genomewide studies. Proc Nat l Acad Sci U S A 100(16):9440-5. Swain A , Narvaez V , Burgoyne P, Camerino G , Lovell-Badge R. 1998. D a x l antagonizes Sry action in mammalian sex determination. Nature 391(6669):761-7. Takahashi T, Lord B , Schulze P C , Fryer R M , Sarang SS, Gullans SR, Lee R T . 2003. Ascorbic acid enhances differentiation of embryonic stem cells into cardiac myocytes. Circulation 107(14):1912-6. Tanaka TS , Kunath T, Kimber W L , Jaradat S A , Stagg C A , Usuda M , Yokota T, N i w a H , Rossant J, K o M S . 2002. Gene expression profiling of embryo-derived stem cells reveals candidate genes associated with pluripotency and lineage specificity. Genome Res 12(12): 1921-8. Tokuzawa Y , Kaiho E , Maruyama M , Takahashi K , Mitsui K , Maeda M , N i w a H , Yamanaka S. 2003. F b x l 5 is a novel target of Oct3/4 but is dispensable for embryonic stem cell self-renewal and mouse development. M o l Cel l B i o l 23(8):2699-708. Tsai J, Sultana R, Lee Y , Pertea G , Karamycheva K , Antonescu V , Cho J, Parvizi B , Cheung F, Quackenbush J. 2001. R E S O U R C E R E R : a database for annotating and l inking mieroarray resources within and across species. Genome Biol:2:software0002.1-0002.4. van der Laan M J , Bryan J. 2001. Gene expression analysis with the parametric bootstrap. Biostatistics 2(4):445-61. van Os R, Kamminga L M , de Haan G . 2004. Stem cell assays: something old, something new, something borrowed. Stem Cells 22(7): 1181-90. 161 Wang J, Coombes K R , Highsmith W E , Keating M J , Abruzzo L V . 2004. Differences in gene expression between B-cel l chronic lymphocytic leukemia and normal B cells: a meta-analysis of three mieroarray studies. Bioinformatics 20(17):3166-78. Wilkinson D G , Bhatt S, Herrmann B G . 1990. Expression pattern of the mouse T gene and its role in mesoderm formation. Nature 343(6259):657-9. Yang Y H , Xiao Y , Segal M R . 2005. Identifying differentially expressed genes from mieroarray experiments via statistic synthesis. Bioinformatics 21(7): 1084-93. Zandstra P W , Nagy A . 2001. Stem cell bioengineering. Annu Rev Biomed Eng 3:275-305. 162 5. Conclusions and Future Directions 5.1. Conclusions The full clinical potential of embryonic stem cell (ESC) based therapies w i l l only be realized with the development of methods for the large-scale expansion of undifferentiated cells in culture. A t present, this is hindered by incomplete knowledge of intra- and extra-cellular determinants of stem cell fate. This thesis reports the development and exploitation of novel endpoints for the detection of loss of pluripotency within a population, the discovery of new reagents that w i l l enable enhanced expansion of undifferentiated E S C , the use of statistical design of experiments to optimize the expansion of E S C and the identification of candidate genes that may be involved in determining E S C fate. Integral to determining effectors of E S C fate decisions are adequate methods to demonstrate changes in the developmental potential of cells following a particular treatment. Conventional methods so rely on the detection of expression of markers such as Oct4 or S S E A - 1 . While assessment of these is relatively easy and lends itself to many high-throughput technologies, such as fluorescent-activated cell sorting, work reported in Chapter 2 demonstrates that loss of these markers is a lagging indicator of loss of pluripotency as measured by the ability of cells to contribute to chimeric mice when injected back into recipient blastocysts. This result suggests that while cells not expressing Oct4 or S S E A - 1 have indeed lost pluripotency, there is developmental heterogeneity among those cells that are still expressing these markers. Two in vitro assays, the embryoid body (EB) and colony-forming cell (CFC) assays, show a high degree 163 of correlation with the chimeric mouse assay following L I F removal, thus providing more reliable indicators of the frequency of pluripotent cells in a population. B y tracking gene expression changes in parallel with loss of readout in these assays, several early indicators of loss of pluripotency were identified. One of these, c-kit, is a cell surface marker and thus lends itself to easy detection by antibody staining. It was shown that loss of c-kit expression closely parallels loss of ability to read out in the E B assay and this provides a more reliable marker of stem cell state than Oct4 or S S E A - 1 . Subsequent experiments by our lab and others have demonstrated that c-kit is involved in the survival of E S C as they differentiate (Bashamboo et al. 2006; L u et al. 2006), demonstrating the utility, o f using functional assays to determine novel markers of pluripotency. Chapter 3 reports the discovery of two new factors that enhance E S C expansion - ascorbic acid ( A A ) and chondroitin sulphate (CS) - and the use of statistical design of experiments to integrate and optimize these reagents along with PD98059 (PD), a specific inhibitor of M A P K signalling which has previously been shown to increase E S C expansion (Burdon et al. 1999), into existing protocols for the expansion of E S C . For this work, the E B assay was used to measure E S C potential. Several advantages exist to using statistical design of experiments. In particular, factorial (FD) and central composite designs ( C C D ) enable a small number of experiments to be run which w i l l optimally cover a large amount of state space and identify potential negative interactions or synergies between tested factors. Here, a 3-fold greater expansion of E S C was demonstrated relative to untreated controls in an optimized combination of these three factors determined from a series of F D and C C D experiments. Since a negative interaction was 164 demonstrated between A A and P D , optimal expansion resulted from a reduction of the concentration of P D that was added to the cultures. Furthermore, since the molecular target of P D is known, the discovery of the negative interaction between this factor and A A , suggests that ascorbic acid enhances E S C maintenance by suppressing the M A P K pathway, although this remains to be proved directly. Thus, in addition to allowing the optimization of the culture environment, the design of experiments approach also allowed the discovery of unanticipated interactions between media components and suggests future directions for research to understand the specific effect of A A . The success of the optimization efforts demonstrated in Chapter 3 hinged on the sensitivity of the E B assay that was used to determine E S C potential. Unfortunately this assay does not lend itself to the high-throughput experiments that would be required for full optimization studies due to its labour-intensive nature. Thus there is a need for the development of more sensitive, high-throughput assays. The goal of the study described in Chapter 4 was to discover genes that are changed concurrent with a loss of pluripotency with the intention that these could be used as the basis of a high-throughput assay for detection of E S C in culture. In addition, it was hypothesized that these genes would be enriched with those that had a biological role in conferring the undifferentiated state on E S C . Several studies with similar goals have been performed but all have resulted in large lists of candidate genes that are too numerous for the subsequent low-throughput studies that must be performed to determine their biological significance. In this study we were able to improve on this by using a more inclusive experimental design and by developing novel statistical tools that allowed a combined analysis of data sets that we generated 165 in conjunction with several data sets that had been generated by other, independent groups. Gene expression profiling by microarray was used to monitor the expression level of up to 45,000 transcripts simultaneously. L ike many large-scale data generating technologies, optimal value was obtained by careful consideration of the samples analyzed. Here, gene expression changes at short time intervals following the loss of pluripotency in mouse E S C following the induction of differentiation by multiple stimuli were chosen. This approach relied on the inclusion of previously generated data sets resulting in greater robustness to the findings since they were independent of the lab that the cells were grown in and the individual that had processed the gene expression samples, both factors having been shown to influence gene expression data (Irizarry et al. 2005). Using novel statistical techniques, we were able to determine a short list of 88 genes - considerably smaller than determined in other studies. Upon further investigation, we were able to propose a gene expression assay consisting of seven genes that we demonstrated was able to assess both increases and decreases in the frequency of stem cells within a population. Since assessment of a small number of gene expression changes can be done in a high-throughput way using a variety of technologies, this provides a rapid assay that can be used in subsequent experiments where high-throughput capacity is required. Since this work was performed, several genes that were reported have been shown to be functionally involved in maintenance and differentiation of E S C indicating the value of this approach in identifying relevant genes. Specifically Esrrb and T e l l have been shown to be required for self-renewal (Ivanova et al. 2006). Overexpression of Kl f4 , in combination with Oct4, Sox2 and c-M y c , was able to dedifferentiate terminally differentiated adult fibroblasts in to ESC- l ike cells (Takahashi and Yamanaka 2006). 166 5.2. Future Directions Work performed in this thesis is intended to guide future scale-up efforts of clinically relevant stem cell populations. Here, all work was performed on mouse E S C that have no clinical value. Human E S C are closely related to mouse E S C in terms of tissue of origin, and while differences between the two cell types means that direct application of these findings may not be possible, extension of many of the methods described here w i l l be of great value. Specifically, there is a great need for more assays in the human E S C system. For ethical reasons, chimera contribution can not be assessed and clonal assays are still impractical due to their inefficiency. Thus, an assay based on specific gene expression changes would not only allow more high-throughput discovery but may also allow discovery of new biology. Direct application to the human system of the findings that A A , CS and P D are beneficial for the expansion of E S C would be interesting. A t present, optimal expansion of h E S C takes place on a layer of inactivated feeders (Thomson et al. 1998) although this can be replaced with matrigel and conditioned media ( X u et al. 2001). This similarity with the mouse system, and the fact that CS is an integral part of the extracellular matrix, suggests that it may also have beneficial effects for hESC. The relevance of P D and A A to the human system may be more questionable though since it appears that very different signalling pathways are operational in the two different systems (Wei et al. 2005). This is based on the assumption that A A acts through suppression of the M A P K pathway. If the antioxidant power of A A also has other effects, then A A may have other, unanticipated effects on h E S C . Nevertheless, the statistical designs of experiments described here w i l l have great power to determine optimal protocols for the both the expansion 167 and differentiation of hESC. Indeed, F D experiments appear to have already applied to develop optimal differentiation protocols for the generation of pancreatic cells from h E S C (D'Amour et al. 2006). Chapter 3 uses F D experiments to determine a negative interaction between the A A and P D . From this data, I hypothesize that A A has its effect by indirectly suppressing the M A P K pathway by reducing the concentration of intracellular reactive oxygen species. This remains to be definitively demonstrated and a greater exploration of the role of reactive oxygen species in E S C maintenance and differentiation would be interesting. Increased intracellular R O S concentration has been shown to have a role in driving m E S C to differentiate towards cardiomyocytes (Sauer et al . 2000). Whi le this seems to contradict previous reports that A A also increases differentiation towards cardiomyocytes (Takahashi et al. 2003), it is clear that modulation of R O S does have an influence on cell fate decisions. The gene list reported in Chapter 4 provides many new candidates that require testing to understand their role in maintenance and differentiation of E S C . Genes with unknown function such as 103728_at and 8430410A17Rik would be particularly interesting to focus on. Podxl provides an interesting candidate to investigate because it is expressed on the cell surface and thus may provide a useful marker for m E S C . Subsequent experiments have demonstrated that podxl protein expression tracks gene expression (C. Glover , unpublished observations). Furthermore, podxl has been shown to be an anti-adhesive molecule (Takeda et al. 2000) thus 168 investigation of this molecule may provide further clues as to how cell fate decisions can be modulated by cell-cell contact. Final ly, the gene expression assay described in Chapter 4 is designed to be used in high-throughput settings where screenings of multiple compounds or optimization type experiments are being performed. Testing of this assay in such a setting would be very interesting. Wi th the recent advances made in combinatorial chemistry, studies of this type are becoming increasingly important to identify novel modulators of cell fate decisions as well as to identify new biological pathways that may be active in cells (Chen et al. 2006). Furthermore, the development of microfluidic systems for both culture ( K i m et al. 2006) and analytical purposes (Warren et al. 2006) mean that experiments of this type can be performed rapidly and inexpensively. 169 5.3. References Bashamboo A , Taylor A H , Samuel K , Panthier JJ, Whetton A D , Forrester L M . 2006. The survival of differentiating embryonic stem cells is dependent on the S C F - K I T pathway. J Ce l l Sci 119(Pt 15):3039-46. Burdon T, Stracey C, Chambers I, Nichols J, Smith A . 1999. Suppression of SHP-2 and E R K signalling promotes self-renewal of mouse embryonic stem cells. Dev B i o l 210(l):30-43. Chen S, Do JT, Zhang Q, Yao S, Yan F, Peters E C , Scholer H R , Schultz P G , Ding S. 2006. Self-renewal of embryonic stem cells by a small molecule 10.1073/pnas.0608156103. P N A S 103(46): 17266-17271. D A m o u r K A , Bang A G , Eliazer S, Ke l ly O G , Agulnick A D , Smart N G , Moorman M A , Kroon E , Carpenter M K , Baetge E E . 2006. Production of pancreatic hormone-expressing endocrine cells from human embryonic stem cells. Nat Biotechnol 24(11): 1392-401. Irizarry R A , Warren D , Spencer F , K i m IF, Biswal S, Frank B C , Gabrielson E , Garcia J G , Geoghegan J, Germino G , Griff in C , Hi lmer S C , Hoffman E , Jedlicka A E , Kawasaki E , Martinez-Murillo F, Morsberger L , Lee H , Petersen D , Quackenbush J, Scott A , Wilson M , Yang Y , Y e SQ, Y u W . 2005. Multiple-laboratory comparison of mieroarray platforms. Nat Methods 2(5):345-50. Ivanova N , Dobrin R, L u R, Kotenko I, Levorse J, DeCoste C, Schafer X , L u n Y , Lemischka IR. 2006. Dissecting self-renewal in stem cells with R N A interference. Nature 442(7102):533-8. K i m L , Vahey M D , Lee H Y , Voldman J. 2006. Microfluidic arrays for logarithmically perfused embryonic stem cell culture. Lab Chip 6(3):394-406. L u M , Glover C H , Humphries R K , Piret J M , Helgason C D . 2006. The role o f c-kit in maintaining murine embryonic stem cell pluripotency. in preparation. Sauer H , Rahimi G , Hescheler J, Wartenberg M . 2000. Role of reactive oxygen species and phosphatidylinositol 3-kinase in cardiomyocyte differentiation of embryonic stem cells. F E B S Lett 476(3):218-23. Takahashi K , Yamanaka S. 2006. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cel l 126(4):663-76. Takahashi T, Lord B , Schulze P C , Fryer R M , Sarang SS, Gullans SR, Lee R T . 2003. Ascorbic acid enhances differentiation of embryonic stem cells into cardiac myocytes. Circulation 107(14):1912-6. 170 Takeda T, Go W Y , Orlando R A , Farquhar M G . 2000. Expression of podocalyxin inhibits cell-cell adhesion and modifies junctional properties in Madin-Darby canine kidney cells. M o l B i o l Cel l ll(9):3219-32. Thomson J A , Itskovitz-Eldor J, Shapiro SS, Waknitz M A , Swiergiel JJ, Marshall V S , Jones J M . 1998. Embryonic stem cell lines derived from human blastocysts. Science 282(5391): 1145-7. Warren L , Bryder D , Weissman IL , Quake SR. 2006. Transcription factor profiling in individual hematopoietic progenitors by digital R T - P C R . Proc Natl Acad Sci U S A 103(47): 17807-12. Wei C L , Miura T, Robson P, L i m S K , X u X Q , Lee M Y , Gupta S, Stanton L , Luo Y , Schmitt J, Thies S, Wang W , Khrebtukova I, Zhou D , L i u E T , Ruan Y J , Rao M , L i m B . 2005. Transcriptome profiling of human and murine E S C s identifies divergent paths required to maintain the stem cell state. Stem Cells 23(2):166-85. X u C, Inokuma M S , Denham J, Golds K , Kundu P, Gold JD, Carpenter M K . 2001. Feeder-free growth of undifferentiated human embryonic stem cells. Nat Biotechnol 19(10):971-4. 171 Appendix A Table A l Primers used for quantitative R T - P C R throughout this study Product 5'-primers 3'-primers Length Gene Name C C C A A A C G T G A G G A A C A A C T A C T T T G G G C A G C A T G A C A A G 101 103728 at C T A C A A C A A G A G C C C G C A G T C C C A T C G C A T G G G A A T A A T 104 8430410A17Rik G C A C C T G C C T T A C C A A C T C T T T T C A G G G C A T T T T T C A A G G 95 Akp2 A A A G C G A G A A A C T G C G A G A G G G C T C C T T C A C A A C T G G A A C 148 Ankrdl T G A G G A G T T T C C A T C A C G A A C A C C T C A T T C T C T G G G A T G C 107 Bmp4 CCTGTCGCAGTTGGGTTC A T T G C C A T T G C T G A A G T G C 105 Bnip3 C T G C G C T T C A A G G A G C T A A C C C A G G C C T G A C A C A T T T A C C 91 Brachyury T A T G T G G C C C T G C T A C A A C A G G A G G T C T C T G A C A C C A G G A 105 Ebaf, Leftb T A T C A A G G C C C T G A C C A C T C C T C A T C T G G T C C C C A A G T G T 114 Esrrb A T T G G A T G T G T G A G C A G A C G A A T G C A G C T G C G T A C T T C C T 77 FbxolS T G T A C T G C A G A G T G G G C A T C A C A A T C C C C T G A G A C A C A G C 121 Fgf5 C A T C C G A C T G G A G C A G C T A T G T G T T C A T G C C A T T C A T C C 92 Foxa2 A C G T C G G G A G T C T G G A T C T C A T A G T C T C C T G C G G C C A T A 112 Fzd5 A A C T T T G G C A T T G T G G A A G G A T G C A G G G A T G A T G T T C T G G 130 Gapdh A G A C G G C A A A G C C T T C T T G TTGACTCGTCTTTCCCTTGC 113 Gbx2 T C G G A G A G C T C A T C T G G A A T T C T C T G G C G A G T G A A G A T C C 95 Gus C A G G A A C T C G T G C T C C A C T A C A T C T T T C T C C C A T G G C T T C 103 Hck G A T G T C A C T G C T C T G C A A G G A A T T G G C T G A G T T T G G C A A C 168 Inpp5d, s-SHIP G A G T G G A A G A A G G T G G G A C A C A C A G C G A T A C T C T C C A G C A 145 Jam2 GATCTGCTCTGCGTCCTGTT C T G A T T G T G C T G G A T G G A T G 109 Kit A G C C T A T C T T G C C G T C C T T T C C A T G G A G A G G A T G A A G T C C 138 Kl/2 CTCTGCTCCCGTCCTTCTC A G A G A G T T C C T C A C G C C A A C 104 Klf4 C A G A G G A G A G T G G C T G A A G G C C A G G A C T C A A T C C C T G T G T 116 Lox G C A G G G A G G A T T C A T G T T G T C C A T A T T G C C C G A T G A A C T T 129 More G C A G G C C A T A A A A C C A A T G T G C A G C A C T G C A G C A T C T T T A 84 Mtf2 A T G C C T G C A G T T T T T C A T C C G A G G C A G G T C T T C A G A G G A A 109 Nanog C T G C A G G C C A C T G A A A A G T T T T C C A G G A T C T G A G C G A T C T 90 nestin A G C A C C T C C G G A G A G G A T A C A A G T G G T T A C C G C C T T G T T G 133 Nmycl T A T C T G A A A G G G A C C G T G C T A T C T G C T G G G T T C T C C A C T G 92 NrObl C A G T G C T C C T G G C A T G A C T A T G G A A T C C T G A T G C C A C A T 99 Nrld2, Thra G T T G G A G A A G G T G G A A C C A A C C A A G G T G A T C C T C T T C T G C 107 Oct4, Pou5fl A G A G C C A A A T T G A C C C A G A T A T G G G A C G C G T T C T A C T C T G • 125 Pbx3 C A A G A G G A G A C C T C C A G C A G C C A A G T G A G A A A C G T G A G C A 90 Pcaf C G C T G C T C A A G G A C A C A G T G T A G C G A T G G T A G C G A A T C C 84 Piml T T C A C G A T G A G C C G T A C A C T C A G G G C A T G G T A C T G G T G T C 88 Pim2 T G A C A T G G A A G C C C A C T A C A A C T G T T G C A G T G T G G A G A C G 111 Podxl C T T C G C T C T G G A G C A G A T T T G C C G C A G T T C T T T T G A A T G T 121 Ptch T G G C T T C C C T G A C A G A T A C C C C T T C G A A C G T G C A C T G A T A 120 Rexl C C T G T C T G T C A G G A G G T G C T T C C A C G A C A T A G G G G A C A A T 123 Rnfl38 C C G A G G A G A A G T A C C A C G A G A A T A T G T C C C C C A C C A T T G A 95 Sod2 172 5'-primers 3'-primers Product Length Gene Name C A G A A C C C A G A T C T G C A C A A G C T T C T C T G C C A A G G T C A A C 79 Sox 17 G G G G T G C A A A A A G A G G A G A G C T A G T C G G C A T C A C G G T T T T 101 Sox2 G C T T G G C T T A T G G A C T G A G G A G G T C C T C A T C T G T G G C A T C 82 Sppl G G C A C C T T G G A T T G A G A G T C A C T C T T G C A G G A A T C G G C T A 122 Stat3 A C A A G C G G G G T A C A G A G A T G C C A G T C C A G A G C A C C T C A C T 81 T-box3 C C G C C C C T A C A G T A T G T G T T A G C C G G A T T T C A T A C G A C T G 104 Tcfcp2ll G A T C T G G G A G A A G C A C G T G T A C G C A A G A T C A C C T G G A A T T T 93 Tell 173 Appendix B Table B1 Genes decreased between 0 and 18 hours after leukemia inhibitory factor removal Gene Symbol Probe TD Accessions Ratio 0 to 18 hours p-value Lox 160095 at N M 010728 0.02 0.0027 Cd44 114697 at N M 009851 0.03 0.0311 Bgn 167023 f at N M 007542 0.03 0.0406 Ankrdl 102048 at N M 013468 0.04 0.0161 Collal 94305 at N M 007742 0.05 0.0079 Fos 160901 at N M 010234 0.06 0.0399 Cd44 103005 s at N M 009851 0.07 0.0095 Vcaml 92559 at N M 011693 0.08 0.0381 Cd44 109403 at N M 009851 0.08 0.0093 CxclU 160511 at N M 013655 0.1 0.0341 Socs3 92232 at N M 007707 0.11 0.006 Thbsl 160469 at N M 011580 0.11 0.0173 Serpinel 94147 at N M 008871 0.11 0.0007 BB120430 105858 at AI847445 0.14 0.0136 Fbln2 100928 at N M 007992 0.14 0.0198 Col3al 98331 at X52046 0.14 0.0352 Tnc 101993 at N M 011607 0.14 0.0256 Bgn 96049 at N M 007542 0.15 0.0149 Flrt3 164233 at N M 178382 0.17 0.0244 Apbblip 102710 at N M 019456 0.18 0.041 Cav 160280 at N M 007616 0.2 0.0288 Socs3 162206 f at AV374868 0.22 0.0006 Fosb 103990 at N M 008036 0.25 0.0231 Fosl2 109982 at N M 008037 0.26 0.0292 AA 408868 98988 at N M 030612 0.27 0.0103 Inhba 100277 at N M 008380 0.32 0.0305 Klf4 99622 at N M 010637 0.32 0.0036 Hnrphl 108850 at N M 021510 0.33 0.0418 Bdnf 102727_at N M 007540 0.35 0.0171 Csfl 101450 at N M 007778 0.36 0.0337 2310046G15Rik 94238 at N M 029614 0.36 0.0441 Caldl 109059_at N M 145575 0.37 0.0403 App 93063 at N M 007471 0.37. 0.0136 Bcl3 105489 at N M 033601 0.38 0.0027 Tbx3 103538 at N M 011535 0.41 0.0102 Timp2 93507 at N M 011594 0.41 0.0359 Eif2s2 97083 at N M 026030 0.43 0.0074 Mras 103653 at N M 008624 0.43 0.0439 Aebp2 105298 at N M 009637 0.44 0.0075 Ch25h 104509 at N M 009890 0.44 0.0065 174 Gene Symbol Probe ID , Accession??; Ratio 0 to 18 hours p-value AI504685 116239 at AI504685 0.44 0.0151 Jagl 116304 at N M 013822 0.45 0.0494 AI158842 137194 r at AI158842 0.45 0.0184 M32461 93022 i at N M 054045 0.45 0.0233 Stambp 112159 at N M 024239 0.46 0.0254 Atf3 104155 f at N M 007498 0.48 0.0193 Egrl 98579 at N M 007913 0.49 0.0236 5730501N20Rik 163583 at AI882080 0.49 0.0193 Nek6 109012 at N M 021606 0.49 0.0447 175 Table B2 Genes increased between 0 and 18 hours after leukemia inhibitory factor removal Gene Symbol Probe ID Accessions Ratio 0 to18 hours p-value 2410012C07Rik 166904 at AW046085 8.4 0.0464 AI461653 115158 at N M 178920 7.5 0.0194 2600010E01Rik 165640 at N M 175181 3.2 0.0051 Hebpl 103085 at N M 013546 2.9 0.0314 Fbxw5 96799 at N M 013908 2.7 0.019 9430093H08Rik 97391 at N M 145462 2.6 0.0082 Nedd8 162180 r at AV367714 2.5 0.0141 Six4 93000 g at N M 011382 2.5 0.0008 Myb 92644 s at N M 010848 2.5 0.038 Dgka 103596 at N M 016811 2.5 0.0079 Cldn4 101410 at N M 009903 2.4 0.0472 2310010J17Rik 161018 at AI661590 2.3 0.0137 Lamrl 96016 at AW045665 2.2 0.0001 C730015A04 105817 at N M 177788 2.2 0.0122 BC027231 109595 at N M 145972 2.2 0.0156 Dnmt3a 100479 at N M 007872 2.1 0.0282 Cldn6 163262 at N M 018777 2.1 0.0081 4933429DllRik 171343 at AV259517 2.1 0.0379 Casp9 100368 at N M 015733 2.1 0.0331 Slc40al 109069 at N M 016917 2.1 0.037 Grhl2 162724 at N M 026496 2.1 0.0036 E130303B06Rik 160997 at N M 198299 2.1 0.0025 Erbb3 96771 at AI006228 2.1 0.0352 Sixl 92722 f at N M 009189 2.1 0.0354 1810037117Rik 102234 at N M 024461 2.1 0.0381 Nqol 94351 r at N M 008706 2 0.0176 Lefl 103629_g_at N M 010703 2 0.0099 Fxfl5 97721 at N M 008003 2 0.0113 176 Table B3 Genes decreased between 18 and 72 hours after leukemia inhibitory factor removal Gene Symbol Probe ID Accessions Ratio 18 to 72 hours p-yalue Klf2 96109 at N M 008452 0.04 0.0016 Gbx2 94200 at N M 010262 0.07 0.0089 NrObl 93141 at N M 007430 0.09 0.0067 Esrrb 100301 at N M 011934 0.1 0.0021 A ass 103389 at N M 013930 0.11 0.0062 Zfp42; Rexl 98414 at N M 009556 0.11 0.0005 Tell 93296 at N M 009337 0.12 0.0001 Pfkp 97834 g at N M 019703 0.13 0.0008 Epasl 102698 at N M 010137 0.14 0.0048 0610013D04Rik 96786 s at N M 030697 0.14 0.021 D930018N21Rik 103761 at N M 023755 0.15 0.0067 Fbxol5 96162 at AA683807 0.16 0.0316 Klf4 99622 at N M 010637 0.17 0.0022 Pfkp 97833 at N M 019703 0.18 0.001 Gja7 104635 r at N M 008122 0.2 0.0376 Pcolce 93349 at N M 008788 0.21 0.0328 B830022L21Rik 95156 g at N M 133206 0.22 0.0207 A230106A15Rik 104030 at AI848841 0.22 0.0024 0610013D04Rik 96785 at N M 030697 0.22 0.0045 Gli 94097 at N M 010296 0.22 0.0044 MovlO 103025 at N M 008619 0.24 0.0101 AA 183875 103728 at A A 183875 0.24 0.0094 AU046074 110041 s at AU046074 0.24 0.0424 Ddc • 160074 at N M 016672 0.24 0.0012 Adaml9 103554 at N M 009616 0.24 0.0161 Ryr3 97126 at D38218 0.25 0.0315 Ptch2 96557 at N M 008958 0.25 0.0054 Enah 93864 s at N M 010135 0.25 0.0074 4930404NllRik 112858 at AI854616 0.26 0.0495 C77370 95963 at C77386 0.26 0.0099 Leftb 102345 at N M 010094 0.26 0.0124 Bmp4 93456 r at N M 007554 0.26 0.0409 Myolf 101708 at N M 008660 0.27 0.009 Bnip3 93836 at N M 009760 0.27 0.0045 GU2 100395 at X99104 0.27 0.0111 Cdldl 103422 at N M 007639 0.27 0.0291 Mybl2 100023 at N M 008652 0.27 0.0005 Texl9 102418 at N M 028602 0.28 0.0026 Ceacaml 102805 at N M 011926 0.28 0.0139 Ptk9l 94020 at N M 011876 0.28 0.0066 Pern 101368 at N M 008818 0.28 0.0245 Acas2l 160921 at N M 080575 0.28 0.0049 Trap la 102764 at N M 011635 0.29 0.0043 Aire 97159 at N M 009646 0.3 0.0004 177 Gene Symbol Probe ID Accession^ Ratio 18 to 72 hours p-value Tm7sf3 103402 at AI848522 0.31 0.0179 6720457D02 Rik 97129 at AA617494 0.31 0.0446 2810008P14Rik 104150 at N M 146164 0.31 0.0052 TcflS 97717 at N M 009328 0.31 0.0261 Slc29al 95733 at N M 022880 0.32 0.0081 Rex2 95346 at N M 009051 0.32 0.0096 D930018N21Rik 97193 at N M 023755 0.32 0.0019 Abcbla 102910 at N M 011076 0.32 0.0213 Tcea3 102344 s at N M 011542 0.32 0.0337 G2an 135386 r at AW228115 0.33 0.0402 Sppl 97519 at N M 009263 0.33 0.0486 Piml 99384 at N M 008842 0.34 0.0106 Foxpl 96183 at N M 053202 0.34 0.0381 AI451904 104938 at AI451904 0.34 0.0463 Aldo3 160546 at AW121134 0.34 0.0004 Rbpsuh 101902 at N M 009035 0.35 0.0061 Ncoa3 102024 at N M 008679 0.35 0.0051 Catnall 94174 at N M 018761 0.35 0.0138 Smarcad1 98991 at X69942 0.35 0.008 B3gntl 160137 at N M 016888 0.35 0.0263 AI847795 104266 at AI847795 0.35 0.0007 843041OA 17Rik 160684 at N M 173737 0.35 0.0057 Gabbrl 98011 at N M 019439 0.36 0.0005 A730024F05Rik 114567 at N M 198295 0.36 0.0098 Anxa9 97798 at N M 023628 0.36 0.0033 C230069C04 100321 f at N M 172929 0.36 0.0181 5830411El ORik 95655 at N M 028696 0.36 0.0184 Ptch 104031 at N M 008957 0.37 0.0135 2700038C09Rik 93838 at N M 025598 0.37 0.0024 Ltb 102940 at N M 008518 0.37 0.0202 Gadd45a 102292 at N M 007836 0.38 0.0082 Icaml 96752 at N M 010493 0.38 0.0236 Dgka 103596 at N M 016811 0.38 0.0128 C530030I18 161096 at AA209597 0.38 0.0037 Pcqf 104070 at N M 020005 0.38 0.0104 Dnajb6 104625 at AA874130 0.38 0.0083 Ulkl 102332 at N M 009469 0.38 0.0024 Acp6 96744 at N M 019800 0.39 0.0025 Empl 97426 at N M 010128 0.39 0.0302 Txnip 160547 s at AI839138 0.39 0.0288 Sec 1411 95664 at N M 028777 0.39 0.0399 Utrn 92507 at N M 011682 0.39 0.0013 Pltp 100927 at N M 011125 0.4 0.0179 Mtf2 103657 i at AI836552 0.4 0.0002 Bmp4 93455 s at N M 007554 0.4 0.0051 Epb4.1l4a 92231 at N M 013512 0.4 0.0204 178 Gene Symbol Probe ID Accessions Ratio 18 to 72 hours p-value Gfpt2 103737 at N M 013529 04 0.0059 DlPasl 93201 at N M 033077 0.41 0.0487 Rasgrp2 103282 at N M 011242 0.41 0.0243 Enah 100472 at N M 010135 0.41 0.0125 Rasa4 160965 at A A 163960 0.41 0.0023 1190005106Rik 93143 at N M 197988 0.41 0.0444 Btebl 93527 at N M 010638 0.41 0.0155 Bcap29 160876 at N M 007530 0.42 0.0378 Gabl 93902 at N M 021356 0.42 0.0035 BC004728 94423 at N M 174992 0.42 0.0109 Jmjdl 95033 at AW049513 0.42 0.0072 6430402H10Rik 97279 at N M 145567 0.42 0.0124 E2fl 102963 at N M 007891 0.43 0.0228 Pdelb 93382 at N M 008800 0.43 0.0228 Cox7al 102749 at N M 009944 0.43 0.0096 AW490810 168859 at AV304692 0.44 0.029 Akl 96801 at N M 021515 0.44 0.0304 1810037I17Rik 102234 at N M 024461 0.44 0.0252 Rnfl38 99529 f at N M 019706 0.44 0.0095 C77009 95028 r at C77009 0.44 0.017 1110036003Rik 100471 at N M 176830 0.44 0.0022 Ldh2 101990_at N M 008492 0.44 0.0436 Slc2a3 93804 at N M 011401 0.44 0.0124 1600025H15Rik 97320 at N M 028064 0.44 0.0062 D530020C15Rik 100977 at N M 172665 0.44 0.0255 Akp2 92796 at N M 007431 0.45 0.0168 Pim3 96841 at N M 145478 0.45 0.0196 Stk38 94460 at N M 134115 0.45 0.0113 Scdl 94056 at N M 009127 0.45 0.0105 Trim25 100475 at D63902 0.45 0.0407 Ii 162171 f at AV363696 0.45 0.0293 Sox2 100009 r at N M 011443 0.45 0.0136 D17H6S56E-5 104333 at N M 033075 0.46 0.0124 Dmclh 96498 at N M 010059 0.46 0.0157 Xlr3b 101883 s at N M 011727 0.46 0.0462 Cd9 95661 at N M 007657 0.46 0.0452 9430077C05Rik 97138 at N M 178059 0.46 0.0285 Ppap2a 98508 s at N M 008903 0.46 0.0033 Soxl5 96824 at N M 009235 0.46 0.0396 Xbpl 94821 at N M 013842 0.46 0.0154 Cdc5l 95610 at AA636547 0.46 0.0036 Eifla 94745 f at N M 010120 0.46 0.0269 Dmrtl 100374 at N M 015826 0.46 0.0052 Trfr 103957 at N M 011638 0.47 0.039 Elovl2 94393 r at N M 019423 0.47 0.0189 Jmj 94341 at N M 021878 0.47 0.046 179 Gene Symbol Probe ID Accessions Ratio 18 to 72 hours p-value Ung 102792 at N M 011677 0.47 0.0377 Zfp292 99001 at N M 013889 0.48 0.0148 Trim23 163698 at N M 030731 0.48 0.0041 Ppm lb 101836 at N M 011151 0.48 0.0199 Pipox 101844 at N M 008952 0.48 0.0486 Mtf2 102069 at N M 013827 0.48 0.0243 Ccnd3 160545 at N M 007632 0.48 0.0015 4930461P20Rik 93478 at AA612483 0.48 0.0297 Znrfi 104396 at N M 199143 0.48 0.0289 Ldbl 160498 at N M 010697 0.48 0.046 Scap2 102012 at N M 018773 0.48 0.0338 H2qfy 97452 at N M 012015 0.49 0.0059 AW046396 106872 at N M 177836 0.49 0.0246 6330406L22Rik 100511 at N M 027521 0.49 0.0306 Stx8 97977 at N M 018768 0.49 0.0346 4931417E21Rik 104556 at AI847766 0.49 0.0033 Trim 13 103691 at N M 023233 0.49 0.012 Mxil 160138 at N M 010847 0.49 0.0313 9230112O05Rik 102807 at N M 173347 0.49 0.0171 Brcal 102976 at N M 009764 0.49 0.0009 Sgk 97890 at N M 011361 0.5 0.0161 Acadm 92581 at N M 007382 0.5 0.0212 AA415014 161106 r at AA415014 0.5 0.0285 Ezhl 100486 at N M 007970 0.5 0.0096 C80913 104589 at N M 011274 0.5 0.0458 Jadel 95883 at N M 172303 0.5 0.0381 180 Table B4 Genes increased between 18 and 72 hours after leukemia inhibitory factor removal Gene Symbol Probe ID Accessions Ratio 18 to 72 hours p-value AI836367 100154 at AI836367 2 0.0255 MGC25977 103078 at AA881018 2.1 0.0034 BC023829 103756 at AA832886 2.2 0.0302 Tpm4 95542 at AI835858 2.4 0.038 Tpm4 95543 at AI843046 2.6 0.0367 3 U 0003Al 7Rik 96135 at AA833425 2.1 0.0003 Acadl 95425 at N M 007381 2.1 0.0497 App 93063 at N M 007471 2.3 0.0161 Atf3 104155 f at N M 007498 2.3 0.0103 Dpi 103886 at N M 007874 2.1 0.005 ldb3 92614 at N M 008321 2.7 0.0036 Itm2a 93511 at N M 008409 2.9 0.004 Modi 101082 at N M 008615 3 0.0013 Proml 93389 at N M 008935 3.5 0.0061 Proml 93390_g_at N M 008935 4.4 0.0023 Pxmp3 103047 at N M 008994 2.1 0.0113 Satbl 99665 at N M 009122 2.5 0.0022 Sox 11 101631 at N M 009234 2.1 0.0208 Sparc 97160 at N M 009242 2.2 0.0209 T 93941 at N M 009309 15.2 0.0034 Tpd52ll 101446 at N M 009413 2 0.0103 Cask 102248 f at N M 009806 2.4 0.0052 Ccngl 160127 at N M 009831 2.1 0.0186 Cd24a 100600 at N M 009846 3 0 Col4al 101093 at N M 009931 2 0.0212 DM 101975 at N M 010052 2.2 0.0291 Gng2 100418 at N M 010315 2 0.0491 Gp38 104469 at N M 010329 2.1 0.0323 Hmga2 99058 at N M 010441 2.7 0.0307 Idbl 100050 at N M 010495 2.8 0.0147 ldb2 93013 at N M 010496 5.2 0.0132 Krtl-18 94270 at N M 010664 2.5 0.0415 Lmo4 98122 at N M 010723 2.3 0.021 Psmb8 102791 at N M 010724 2.3 0.0021 Pitx2 102788 s at N M 011098 7.8 0.0013 Psme2 100588 at N M 011190 2.2 0.024 Tead2 96940 at N M 011565 2.3 0.0192 Copeb 98083 at N M 011803 2 0.0257 Cpe 99642 i at N M 013494 2.4 0.0313 Tap! 103035 at N M 013683 2.1 0.0077 Ptpla 166430 i at N M 013935 2.4 0.0369 Enpp2 97317 at N M 015744 2.9 0.0118 Nes 103549 at N M 016701 2.9 0.0431 Pbx3 93615 at N M 016768 2.5 0.0226 181 Gene Symbol Probe ID Accession^ Ratio 18 to 72 hours p-value . Vamp8 100345 f at N M 016794 2.2 0.0116 Cldnl 99561 f at N M 016887 4 0.0003 Ctps2 160652 at N M 018737 2.1 0.0004 Tm4sf2 93326 at N M 019634 2.2 0.0209 Ptpn9 100976 at N M 019651 2.9 0.0141 Tm4sf6 92555 at N M 019656 2 0.0095 Perp 97825 at N M 022032 2.3 0.0198 Ifi30 97444 at N M 023065 2.1 0.0458 Tpml 160532 at N M 024427 2.3 0.0066 2610301B20Rik 94893 at N M 026005 2.3 0.0067 1110003B01Rik 160367 at N M 026131 2.2 0.0246 3632413B07Rik 97958 at N M 027230 2.5 0.0002 2310058J06Rik 103100 at N M 133784 2 0.0157 Arhu 96747 at N M 133955 2.2 0.0309 Pim2 101926 at N M 138606 2.2 0.0237 Hip] 167859 at N M 146001 2.3 0.0026 Snk 92310 at N M 152804 3.7 0.0066 Amot 95531 at N M 153319 5.4 0.0093 1810057B09Rik 98958 at N M 175009 2.2 0.0093 5730453H04Rik 94247 at N M 175338 2.3 0.0405 B430218L07Rik 93467 at N M 198037 3.8 0.0003 182 Table B5 Genes decreased between 0 and 72 hours after leukemia inhibitory factor removal Gene Symbol Probe ID Accessions Ratio 0 to 72 hours p-value Ccl2 102736 at N M 011333 0.02 0.0279 Bgn 167023 f at N M 007542 0.03 0.0257 Tnc 162362 f at AV230686 0.03 0.0258 Klfl 96109 at N M 008452 0.03 0.0007 Socs3 92232 at N M 007707 0.04 0.0122 Abcal 94354 at N M 013454 0.05 0.006 Klf4 99622 at N M 010637 0.05 0.002 Gbx2 94200 at N M 010262 0.06 0.0072 Lox 160095 at N M 010728 0.06 0.01 Mras 103653 at N M 008624 0.07 0 Collal 94305 at N M 007742 0.08 0.0278 Ndp52ll 95562 at AA615127 0.08 0.0136 NrObl 93141 at N M 007430 0.08 0.0083 Vcaml 92559 at N M 011693 0.08 0.0344 Empl 97426 at N M 010128 0.09 0.0381 Fbln2 100928 at N M 007992 0.09 0.0054 Tnc 101993 at N M 011607 0.09 0.0134 D930018N21Rik 103761 at N M 023755 0.09 0.0014 Nfib 99440 at N M 008687 0.09 0.0047 Acta2 93100 at N M 007392 0.1 0.0472 Esrrb 100301 at N M 011934 0.1 0.0011 Bgn 96049 at N M 007542 0.1 0.0071 Fbxol5 96162 at AA683807 0.11 0.0212 Tell 93296 at N M 009337 0.11 0.0001 Fos 160901 at N M 010234 0.11 0.0017 Sdpr 167384 s at N M 138741 0.11 0.0278 Zfp42 98414 at N M 009556 0.12 0.0018 Thbsl 160469 at N M 011580 0.13 0.0203 A ass 103389 at N M 013930 0.13 0.007 Tbx3 166427 f at N M 011535 0.14 0.0329 Cish3 162206 f at AV374868 0.15 0.0055 S100a6 92770 at N M 011313 0.16 0.0186 Eif2s2 97083 at N M 026030 0.17 0.017 Pfkp 97834 g at N M 019703 0.17 0.0033 AA183875 103728 at AA183875 0.17 0.0029 Pcolce 93349 at N M 008788 0.17 0.0231 BB120430 105858 at AI847445 0.18 0.0224 Osmr 102255 at N M 011019 0.18 0.0321 Sppl 97519 at N M 009263 0.19 0.0126 Cd44 109403 at AW121933 0.19 0.043 Mras 164306 f at AV226634 0.2 0.0292 Gja7 104635 r at N M 008122 0.22 0.0473 C730049F20Rik 96038 at AI840339 0.22 0.0361 Cav 160280 at N M 007616 0.23 0.0254 183 Gene Symbol Probe ID Accessions Ratio 0 to 72 hours p-value Kit 99956 at N M 021099 0.23 0.0016 D6Ertd32e 95618 at AI843884 0.24 0.0496 D930018N21Rik 97193 at N M 023755 0.24 0.0113 Pfkp 97833 at N M 019703 0.24 0.0002 Ferll3 112941 f at AA960514 0.24 0.029 Trpsl 110414 at N M 032000 0.25 0.0369 Lamal 103729 at N M 008480 0.26 0.0346 Wdt2-pending 107952 i at AA606601 0.26 0.0439 Pern 101368 at N M 008818 0.27 0.0293 A230106A15Rik 104030 at AI848841 0.27 0.0012 Myolf 101708 at N M 008660 0.27 0.0083 Akap2 163655 s at N M 009649 0.28 0.0159 RM2 95617 at N M 011250 0.28 0.039 Eifla 94745 f at U28419 0.28 0.0091 My lip 106479 at N M 153789 0.28 0.0452 Gli 94097 at N M 010296 0.29 0.0496 Inhba 100277 at N M 008380 0.29 0.0297 Tbx3 103538 at N M 011535 0.3 0.0214 Trim2 163115 at N M 030706 0.3 0.0307 Dnajb9 96680 at N M 013760 0.31 . 0.0216 All 15454 115333 at N M 175345 0.31 0.0344 Icaml 96752 at N M 010493 0.31 0.0188 G2an 135386 r at AW228115 0.31 0.0253 6720457D02Rik 97129 at AA617494 0.31 0.0462 AA 175286 113283 at N M 010156 0.31 0.0402 AI427138 108784 at AI427138 0.31 0.0228 Ly75 103258 at N M 013825 0.31 0.035 Ulkl 102332 at N M 009469 0.31 0.0012 Nrld2 99076 at N M 011584 0.31 0.0194 Tcfap2c 92275 at N M 009335 0.32 0.0482 Mllt2h 115212 at N M 133919 0.32 0.0277 Trap la 102764 at N M 011635 0.33 0.0109 Dpys 113671 at N M 022722 0.33 0.0005 Fosb 103990 at N M 008036 0.33 0.0094 Mybl2 100023 at N M 008652 0.33 0.0013 Pim3 96841 at N M 145478 0.34 0.0107 AI504685 116239 at AI504685 0.34 0.0004 MovlO 103025 at N M 008619 0.34 0.0104 Abcbla 102910 at N M 011076 0.35 0.02 Fgfl7 98730 at N M 008004 0.36 0.0044 GH2 100395 at X99104 0.36 0.0122 C77370 95963 at C77386 0.36 0.0303 Scap2 102012 at N M 018773 0.36 0.0001 Piml 99384 at N M 008842 0.36 0.0154 Nidi 100120 at N M 010917 0.36 0.0471 8430410A17Rik 160684 at N M 173737 0.37 0.007 184 Gene Symbol Probe ID Accessions Ratio.0 to 72 hours p-value Gfpt2 103737 at N M 013529 0.37 0.0018 Leftb 102345 at N M 010094 0.37 0.0186 Nrld2 107103 at N M 011584 0.37 0.0474 AI85U55 117246 at N M 133982 0.38 0.0126 4930422J18Rik 161817 f at AV376312 0.38 0.0119 C85523 99849 at C85523 0.38 0.0449 Mapt 102742_g_at N M 010838 0.38 0.0203 D130027M04 112257 at AA958476 0.38 0.0427 Bnip3 93836 at N M 009760 0.39 0.0111 493043 lPHRik 108866 at N M 172723 0.39 0.0444 Tuba3 160420 r at N M 009446 0.39 0.0215 AW490810 168859 at AV304692 0.4 0.0134 AI837485 133552 at AI837485 0.4 0.0075 1110035O14Rik 94192 at N M 021524 0.4 0.0349 Jmjdl 95033 at AW049513 0.4 0.0209 Kif3a 161275 at AV324729 0.4 0.0223 Bmp4 93455 s at N M 007554 0.4 0.0052 Camk4 104455 at N M 009793 0.4 0.0329 Yes 100015 at N M 009535 0.4 0.0079 Il6st 135888 at N M 010560 0.4 0.0058 Stat3 99100 at N M 011486 0.41 0.0128 Sox2 100009 r at N M 011443 0.41 0.02 Ddc 160074 at N M 016672 0.41 0.0046 Catnall 94174 at N M 018761 0.42 0.017 AUO14924 109299 f at AU014924 0.42 0.0398 Epb4.1l4a 92231 at N M 013512 0.42 0.0173 0610013D04Rik 96785 at N M 030697 0.42 0.019 Cnot4 163709 i at N M 016877 0.42 0.013 Zfhxla 99052 at N M 011546 0.42 0.0096 Tm7sf3 103402 at AI848522 0.42 0.0287 Scmhl 103399 at N M 013883 0.42 0.0115 Rbpsuh 101902 at N M 009035 0.43 0.0004 493243 lF02Rik 95964 at N M 028475 0.43 0.0304 Msc 93234 at N M 010827 0.43 0.025 Texl9 102418 at N M 028602 0.43 0.0075 Aebp2 105298 at N M 009637 0.43 0.0002 9930116P15Rik 111970 at AI616223 0.43 0.0454 Ptk9l 94020 at N M 011876 0.43 0.0141 2810008P14Rik 104150 at N M 146164 0.44 0.0067 Uppl 100030 at N M 009477 0.44 0.0096 Slc29al 95733 at N M 022880 0.44 0.015 Aldo3 160546 at AW121134 0.44 0.0257 Epb4.1l4a 161603 r at AV308245 0.44 0.0005 Mtf2 103657 i at AI836552 0.44 0.0031 Zfp36ll 93324 at N M 007564 0.44 0.0123 Slc2a3 93804 at N M 011401 0.44 0.0027 185 Gene Symbol Probe ID Accessions Ratio 0 to 72 hours p-value Egrl 98579 at N M 007913 0.44 0.0258 Anxa9 97798 at N M 023628 0.44 0.0025 B3gntl 160137 at N M 016888 0.44 0.0041 Klf5 97937 at N M 009769 0.44 0.0236 Adaml9 103554 at N M 009616 0.45 0.002 2310005N03 Rik 160316 at N M 025511 0.45 0.023 AI851470 112406 at AI851470 0.45 0.0186 Echdcl 113214 at N M 025855 0.45 0.0469 5730501N20Rik 163583 at AI882080 0.45 0.0068 Gadd45a 102292 at N M 007836 0.45 0.0011 Cd9 95661 at N M 007657 0.45 0.0345 5830411ElORik 95655 at N M 028696 0.45 0.0384 H2afy 97452 at N M 012015 0.46 0.0115 C230069C04 100321 f at N M 172929 0.46 0.0231 Jadel 95883 at N M 172303 0.46 0.0267 Sgk 97890 at N M 011361 0.46 0.0009 C530030I18 161096 at AA209597 0.46 0.0208 Ac as 21 160921 at N M 080575 0.46 0.0283 D7Ertdl43e 95954 at AA437728 0.46 0.0038 Gjb3 104232 at N M 008126 0.46 0.0423 Timp2 93507 at N M 011594 0.46 0.0298 Xbpl 94821 at N M 013842 0.46 0.0045 Ddahl 115112 at N M 026993 0.47 0.0289 Nfatc2ip 96561 at N M 010900 0.47 0.0087 Stx8 97977 at N M 018768 0.47 0.0046 Cdc37l 115976 at N M 025950 0.47 0.032 2410043F08Rik 115028 at N M 133754 0.47 0.0161 Il6st 94345 at N M 010560 0.47 0.0119 Jagl 116304 at N M 013822 0.47 0.0411 Aard 160968 at N M 175503 0.47 0.0367 Arg2 98473 at N M 009705 0.47 0.0037 2310009N05Rik 160801 at N M 025861 0.47 0.0086 Rgsl6 94378 at N M 011267 0.48 0.0082 8430420F16Rik 111789 at AI642433 0.48 0.0444 Mtf2 102069 at N M 013827 0.48 0.0096 Rgsl6 161609 at AV349152 0.48 0.0026 3230402J05Rik 106144 at AI846635 0.49 0.0425 Btgl 93104 at N M 007569 0.49 0.0418 Ppap2a 98508 s at N M 008903 0.49 0.0041 9430079M16Rik 104091 at N M 175414 0.49 0.031 4930553M18Rik 104639 i at N M 029248 0.49 0.0253 Smarcadl 98991 at X69942 0.49 0.024 3110031O14Rik 116738 at N M 026081 0.49 0.0208 Ak4 99959 at N M 009647 0.49 0.01 2700017106Rik 162909 at AI846731 0.49 0.0234 Sod2 96042 at N M 013671 0.5 0.0018 186 Gene Symbol Probe ID Accessions Ratio 0 to 72 hours p-value 1810013L24Rik 131177 r at AA636427 0.5 0.0038 Fut9 105675 at N M 010243 0.5 0.0125 AI595304 113495 at AI595304 0.5 0.0489 U10008H02Rik 96156 at AA874329 0.5 0.0463 E130012A19Rik 160712 r at N M 175332 0.51 0.0217 187 Table B6 Genes increased between 0 and 72 hours after leukemia inhibitory factor removal Gene Symbol Probe ID Accessions ' Ratio 0 to 72 hours p-value Prphl 104751 at N M 013639 17.6 0.0122 Thbs4 113830 at N M 011582 16.2 0.0003 Evxl 98815 at N M 007966 14.7 0.0337 Btla 116608 at N M 177584 14.3 0.0033 A1461653 115158 at N M 178920 13.9 0.0058 Aloxl5 98758 at N M 009660 11.1 0.0353 Enpp2 97317 at N M 015744 9.1 0.0165 T 93941 at N M 009309 7.3 0.0126 Pitx2 102788 s at N M 011098 6.1 0.0002 Calcr 101135 at N M 007588 6 0.0338 Cldn7 99561 f at N M 016887 6 0 Nef3 117269 at N M 008691 5.6 0.0415 Proml 93390 g at N M 008935 5.5- 0.001 Galnt3 99011 at N M 015736 4.7 0.0391 Eomes 103532 at AW120579 4.7 0.0162 Proml 93389 at N M 008935 4.5 0.0312 Amot 95531 at N M 153319 4.1 0.0073 Fox hi 97789 at N M 007989 3.9 0.0376 B430218L07Rik 93467 at N M 198037 3.9 0.001 Krtl-18 94270 at N M 010664 3.9 0.0124 Nudtll 166849 at . N M 021431 3.9 0.0101 Cldn6 93703 at N M 018777 3.7 0.013 AU067722 135354 s at AU067722 3.6 0.005 AV022166 168363 i at AV022166 3.5 0.036 Cldn6 163262 at N M 018777 3.4 0.027 -Pbx3 93615 at N M 016768 3.4 0.0073 Nefl 103575 at N M 010910 3.4- 0.0494 Ssbp2 103504 at AI837107 3.4 0.0102 Cxcll2 100112 at N M 013655 3.3 0.0231 Wnt3 99325 at N M 009521 3.3 0.0027 AA407151 94663 at AA407151 3.3 0.004 Krt2-8 101009 at N M 031170 3.2 0.0451 Msxl 101526 at N M 010835 3.2 0.0275 Cnnl 99942 s at N M 009922 3.1 0.0228 Plcg2 103503 at N M 172285 3.1 0.0147 Gatm 96336 at N M 025961 3.1 0.0072 1110003B01Rik 160367 at N M 026131 3.1 0.0069 Hes6 97334 at N M 019479 3 0.0109 Cbr3 168153 i at AV217999 3 0.0253 Crlfl 161046 at N M 018827 3 0.0045 Cbr3 161042 at N M 173047 2.9 0.0048 Bhlhb5 129880 s at N M 021560 2.9 0.0227 Ifi30 97444 at N M 023065 2.9 0.0162 Dapkl 114748 at N M 029653 2.8 0.0197 Gene Symbol Probe ID Accessions _ Ratio 0 to 72 hours . p-value Itga8 166680 at AI447669 2.8 0.0298 2600010E01Rik 165640 at N M 175181 2.8 0.0054 Idb2 93013 at N M 010496 2.8 0.022 Tead2 96940 at N M 011565 2.8 0.0058 Modi 101082 at N M 008615 2.8 0.0016 Nes 103549 at N M 016701 2.8 0.0384 Idbl 100050 at N M 010495 2.8 0.0314 Mdk 160561 at N M 010784 2.7 0.0139 Ankrd3 163489 at N M 023663 2.7 0.04 Tdpl 103035 at N M 013683 2.7 0.0074 Cd24a 100600 at N M 009846 2.7 0.0001 5330410G16Rik 162573 at N M 182991 2.7 0.048 Gatm 109824 f at AI835703 2.7 0.0023 Erbb3 96771 at AI006228 . • 2.7 0.0146 Arhgef5 160977 at AA726063 2.6 0.0094 St6gall 94432 at N M 009175 2.6 0.0498 Arhu 96747 at N M 133955 2.6 0.0009 Nnat 97520 s at N M 010923 2.6 0.0092 MGC27770 97116 at N M 145986 2.6 0.0038 5730467H21 Rik 168550 at N M 175270 2.5 0.0013 Podxl 112828 at N M 013723 2.5 0.0019 Ctps2 160652 at N M 018737 2.5 0.0082 Tc/7 97994 at N M 009331 2.5 0.0322 Rnfl28 108010 at N M 023270 2.5 0.0453 Tor2a 97557 at AI841457 2.5 0.0016 Vamp8 100345 f at N M 016794 2.5 0.0024 Spintl 97206 at N M 016907 2.5 0.0138 Ptpla 166430 i at N M 013935 2.5 0.0019 Prphl 161482 f at AV068234 2.5 0.0028 2810432L12Rik 95440 at N M 025944 2.5 0.0038 2610319K07Rik 104744 at N M 028264 2.5 0.0002 Lamrl 96016 at AW045665 2.5 0.0284 Gclm 160335 at N M 008129 2.5 0.0315 Tpm4 95543 at AI843046 2.5 0.0049 Ptpn9 100976 at N M 019651 2.5 0 Stl4 116435 at N M 011176 2.5 0.0319 Facl5 97456 at AI838021 2.4 0.006 Nme3 94982 f at N M 019730 2.4 0.0085 Tm4sf6 92555 at N M 019656 2.4 0.0113 Cldn7 129928 f at AW214436 2.4 0.0147 Nudtll 163611 f at N M 021431 2.4 0.0207 Kiflc 160679 at N M 153103 2.4 0.0069 Edg4 114162 at N M 020028 2.4 0.0301 Gng3 94986 at NM010316 2.4 0.0321 Ddrl 100155 at N M 007584 2.4 0.0264 Perp 97825 at N M 022032 2.4 0.0268 189 Gene Symbol Probe ID Accessions , Ratio 0 to 72 hours p-value ; Sox 11 101631 at N M 009234 2.4 0.0434 Mapkl2 92323 at N M 013871 2.4 0.0047 Ssbp3 117147 at N M 023672 2.4 0.0416 Snk 92310 at N M 152804 2.3 0.0295 MGC25977 103078 at AA881018 2.3 0.0114 6030440G05Rik 108014 at N M 145148 2.3 0.0166 3632413B07Rik 97958 at N M 027230 2.3 0.0055 Slc39a8 97442 at N M 026228 2.3 0.0208 Zdhhc2 162858 at N M 178395 2.3 0.0158 Ptpnl3 98424 at N M 011204 2.3 0.0191 D10Ertd755e 130824 i at AUO19706 2.3 0.0428 Cyfip2 162711 at N M 133769 2.3 0.0268 Sfrp2 93503 at N M 009144 2.3 0.015 1810029F08Rik 104302 f at N M 029635 2.3 0.0145 Lrp8 138065 at N M 053073 2.3 0.0255 Unc5b 112955 at N M 029770 2.3 0.0367 Kras2 97991 at N M 021284 2.3 0.0467 AU024549 131137 at AU024549 2.3 0.004 Trip6 94948 at N M 011639 2.2 0.0044 Oxct 92845 at N M 024188 2.2 0.0073 SST3 109554 at N M 172463 2.2 0.0155 Igsf9 129147 r at N M 033608 2.2 0.005 2300002F06Rik 96672 at N M 175606 2.2 0.0081 Tapbp 100154 at AI836367 2.2 0.0202 Phfl5 105005 at N M 199299 2.2 0.0024 Cugbp2 97255 at N M 010160 2.2 0.018 Cyb5 98533 at N M 025797 - 2.2 0.0248 Pcbd 99056 at N M 025273 2.2 0.044 Hipl 167859 at N M 146001 2.2 0.0073 AW228646 135355 at AW228646 2.2 0.0184 Satbl 99665 at N M 009122 2.2 0.0285 AI838057 117151 at AI838057 2.2 0.0311 Mx2 102699 at N M 013606 2.2 0.0299 Clu 95286 at N M 013492 2.2 0.0018 Psme2 100588 at N M 011190 2.2 0.0149 Binl 92220 s at , N M 009668 2.2 0.0234 Them2 93780 at N M 025790 2.2 0.0318 Dpi 96115 at N M 007874 2.1 0.0455 2810428I15Rik 160278 at N M 025577 2.1 . 0.027 Rem2 163542 at N M 080726 2.1 0.0225 Hmga2 99058 at N M 010441 2.1 0.0439 Frda 101407 at N M 008044 2.1 0.0052 AW214234 129180 f at AW214234 2.1 0.0099 Lmo4 98122 at N M 01Q723 2.1 0.0243 AW046694 104445 at AW046694 2.1 0.0241 Laptm4b 100571 at N M 033521 2.1 0.0369 190 Gene Symbol Probe ID Accessions Ratio 0 to 72 Hours p-value 1600022A19Rik 113913 at N M 146062 2.1 0.0374 2810453I06Rik 104393 at N M 026050 2.1 0.0496 2610301B20Rik 94893 at N M 026005 2.1 0.0214 Vamp8 93305 f at N M 016794 2.1 0.0173 1200003O06Rik 94549 at N M 025813 2.1 0.0031 Tpd52U 101446 at N M 009413 2.1 0.0038 A hey 96026 at N M 016661 2.1 0.0068 Siat8c 99504 at N M 009182 2.1 0.0399 130001 lC24Rik 93327 at N M 029564 2.1 0.0054 AI586067 105461 at AI586067 , 2.1 0.0388 Purg 110145 at N M 152821 2.1 0.0468 All 73484 131796 at A l l 73484 2.1 0.0487 Ppplrl4c 115575 at N M 133485 2 0.0418 Nme3 94981 i at N M 019730 2 0.0003 Ube2d2 93069 at N M 019912 2 0.0034 PsmblO 101486 at N M 013640 . 2 0.0058 1110049F12Rik 160329 at N M 025411 2 0.0197 BC023829 103756 at AA832886 2 0.0415 Mab21l2 95379 at N M 011839 2 0.0108 AW547186 116943 at N M 177592 2 0.026 Rabl5 113199 at N M 134050 2 0.019 Ttc8 163404 at N M 029553 2 0.0435 0610013E23Rik 160507 at N M 029788 2 0.0037 D10Ertd755e 130825 f at AUO19706 2 0.0043 AW550831 115058 at AA756546 2 0.0084 BC026370 162607 i at N M 198167 2 0.011 Vapb 112389 at N M 019806 2 0.0419 AW555464 131216 f at AI463306 ' 2 0.0421 191 Table B7 Comparison of changed genes to previously published data sets to Bhattac Overlap Kelly Overlap harya Brandenbe Sato et Sperger with all Ivanova and Ramalho- (Sharov with all (Ginis Overlap et al. rger et al. al. etal. human etal. Rizzino Santos et et al. mouse et al. with all Gene Name Affy ID Accession 2004) 2004) 2003) 2003) data sets 2002) 2000) al. 2002) 2003) data sets 2004) data sets Decreased between 0 and 18 hours Tnc 101993 at N M 011607 0 0 0 1 1 1 0 1 1 3 0 4 Lox 160095 at N M 010728 0 0 0 0 0 1 0 1 1 3 0 3 Fbln2 100928 at N M 007992 0 0 0 0 0 1 0 1 0 2 0 2 Bdnf 102727 at N M 007540 0 0 0 0 0 1 0 1 0 2 0 2 Tbx3 103538 at N M 011535 0 0 0 0 0 1 0 1 0 2 0 2 2310()46G15Rik 94238 at N M 029614 0 0 0 0 0 1 0 1 0 2 0 2 Bgn 96049 at N M 007542 0 0 0 0 0 1 0 1 0 2 0 2 Cav 160280 at N M 007616 0 0 0 0 0 1 0 1 0 2 0 2 Serpinel 94147 at N M 008871 0 0 0 0 0 1 0 1 0 2 0 2 Ankrdl 102048 at N M 013468 0 0 0 0 0 0 0 1 0 1 0 1 Atf3 104155 f at N M 007498 0 0 1 0 1 0 0 0 0 0 0 1 Bgn 167023 f at N M 007542 0 0 0 0 0 1 0 0 0 1 0 1 Collal 94305 at N M 007742 0 0 0 0 0 1 0 0 0 1 0 1 Col3al 98331 at X52046 0 0 0 0 0 0 0 1 0 1 0 1 Cxcll2 160511 at N M 013655 0 0 0 0 0 0 0 1 0 1 0 1 Egrl 98579 at N M 007913 0 0 0 1 1 0 0 0 0 0 0 1 Klf4 99622 at N M 010637 0 0 0 0 0 0 1 0 0 1 0 1 Socs3 162206 f at N M 007707 0 0 0 1 1 0 0 0 0 0 0 1 Socs3 92232 at N M 007707 0 0 0 1 1 0 0 0 0 0 0 1 Thbsl 160469 at N M 011580 0 1 0 0 1 0 0 0 0 0 0 1 Decreased between 18 and 72 hours Leftb 102345 at D83921 1 1 1 1 4 1 0 1 0 2 0 6 Jadel 95883 at N M 172303 1 1 1 1 4 1 0 1 0 2 0 6 Smarcadl 98991 at X69942 0 1 1 1 3 1 0 1 1 3 0 6 Ung 102792 at N M 011677 0 1 1 0 2 1 1 1 0 3 0 5 100009_r_a Sox2 t N M 011443 1 1 0 0 2 I 0 0 0 1 1 4 Overlap with all data sets -i- -r - t -t- -r -t -t -r -1- m m r i CN O l CN r-1 .23 • S~ .E 5 5 o « • s o o o o O o o o o O o o o O o O o O o o o o O © © o S Overlap with all j mouse data sets m - oi - CN CN o l CN CN CN — CN o CN - f N — CN PI - CN CN - - r 1 (Sharov| etal. i 2003) o o o © O - o - i o — — — o o O o o o - o © o © O Ramalho-Santos et al. 2002) o o -H O - O - - — - o o o o - -Kelly and Rizzino 2000) — o o o O o o o o o o o o o o o o o o o o - o o © © © o Ivanova et al. 2002) — - - — - - - - o — o - o o - o - — — - © — — © © -Overlap with all human data sets - CN CN ol CN CN CN - o CN - m — CN o O l o r l o o o Sperger et al. 2003) - - — - — - - - o o - © — o - — O o - — © - © o - - o Sato et ai. 2003) o - - - o - - o o o - - - o - o o o o - o o © o o Brandenbe rger et al. 2004) o - o o O o - o o o o o o o o - o - o - o o © © o © © o as e o o - o o o o o o c c 6 o o O 0 o o © © © o o - © o Accession NM 008652 NM 008952 NM 009035 NM 008492 NM 013827 NM 013529 NM 011401 NM 021878 U28419 WI 009556 NM 019706 NM 010135 NM 172665 NM 013827 NM 011635 NM 010207 NM 013512 NM 007431 NM 009337 NM 007554 NM 021356 NM 010262 NM 022880 NM 015826 NM 008818 NM 173347 NM 009616 AA 183875 Gene Name Affy ID 100023 at 101844 at 101902 at 101990 at 103657_i_a t 103737 at 93804 at 94341 at 94745 f at 98414 at 99529 f at 100472 at 100977 at 102069 at 102764 at 162171_f_a t 92231 at 92796 at 93296 at 93456 r at 93902 at 94200 at 95733 at 100374 at 101368 at 102807 at 103554 at 103728 at Gene Name Affy ID 5 Pipox Rbpsnh <N =1 1 Gfpt2 <*i q Jmi Eifla CS 4 i Enah D530020C15Rik 1 Trap la 1 S-1 — 1 — i "* Akp2 Tell •» B. 5 03 Gabl CN Slc29al Dmrtl Pern 9230112005Rik Adaml9 AA183875 193 5 Gene Name Afly ID Accession Bhattac harya et al. 2004) Brandenbe rger et al. 2004) Sato et al. 2003) Sperger et al. 2003) Overlap with all human data sets Ivanova et al. 2002) Kelly and Rizzino 2000) Ramalho-Santos et al. 2002) (Sharov et al. 2003) Overlap with all mouse data sets (Ginis et al. 2004) Overlap with all data sets D930018N2IRik 103761 at N M 023755 0 0 0 0 0 1 0 1 0 2 0 2 C80913 104589 at N M 011274 0 0 0 0 0 1 0 1 0 2 0 2 Aldo3 160546 at AW121134 0 0 1 0 1 1 0 0 1 0 2 843041OA 17Rik 160684 at N M 173737 0 0 0 0 0 1 0 1 0 2 0 2 AA415014 161106_r_a t AA415014 0 0 0 0 0 1 0 1 0 2 0 2 Pcolce 93349 at N M 008788 0 0 0 0 0 1 0 1 0 2 0 2 Bmp4 93455 s at N M 007554 0 1 0 1 2 0 0 0 0 0 2 Btebl 93527 at N M 010638 0 0 0 0 0 1 0 1 0 2 0 2 Bnip3 93836 at N M 009760 0 0 1 0 1 0 0 1 0 1 0 2 Enah 93864 s at N M 010135 0 0 0 1 1 0 0 1 1 0 2 C atrial 1 94174 at N M 018761 0 0 0 1 1 0 0 1 0 1 0 2 C77009 95028 r at C77009 0 0 0 0 0 1 0 1 0 2 0 2 Fbxol5 96162 at AA683807 0 0 0 0 0 1 0 1 0 2 0 2 Ptchl 96557 at N M 008958 0 0 0 0 0 0 0 1 1 2 0 2 0610013D04Rik 96785 at N M 030697 0 0 0 0 0 1 0 1 0 2 0 2 6720457D02Rik 97129 at AA617494 0 0 0 0 0 1 0 1 0 2 0 2 Empl 97426 at N M 010128 0 0 0 0 0 1 0 1 0 2 0 2 Sppl 97519 at N M 009263 0 0 0 0 0 1 0 1 0 2 0 2 Sgk 97890 at N M 011361 0 0 0 0 0 1 0 1 0 2 0 2 Ppap2a 98508 s at N M 008903 0 0 1 0 1 0 0 1 0 1 0 2 Piml 99384 at N M 008842 0 0 1 1 2 0 0 0 0 0 0 2 GH2 100395 at X99104 0 0 0 1 1 0 0 0 0 0 0 1 1110036O03Rik 100471 at N M 176830 0 0 0 1 1 0 0 0 0 0 0 1 Pltp 100927 at N M 011125 0 0 1 0 1 0 0 0 0 0 0 1 Ppm lb 101836 at N M 011151 0 0 1 0 1 0 0 0 0 0 0 1 Xlr3b 101883_s_ at N M 011727 0 0 0 0 0 1 0 0 0 1 0 , Ncoa3 102024 at N M 008679 0 0 0 0 0 0 0 0 1 1 0 1 Gadd45a 102292 at N M 007836 0 0 0 0 0 0 0 1 0 1 0 1 Overlap with all data sets (Ginis etal. 2004) o © © © © © o © © © © © © © © © © © o © © © © © © Overlap with all mouse data sets - - O - © © © — © — © © - - — © © © - © © - -(Sharov et al. 2003) o © o © o © © © © © © © o © o © © o © © © © © © © Ramalho-Santos et \ al "M)02^ o - o o — © o o - © — o © © © © © © © — © © © — -Kelly and Rizzino 2000) j - o © o o © © o o © © © o © © © o © © © © © © © © Ivanova etal. 2002) o o o - o o © o © © © © o - - — © © © © - © © © © Overlap with all human data sets o o - o © © - © - © © © - © © - © © Sperger et al. o o - o © © o © © © © - © © © © © © o - © © Sato et al. 2003) o o o © © - — - o © © o © © © © © - © © o © © © © Brandenbe rger et al. 2004) o © © © © o © © © - © © © © © © © © © © © © © © Bhattac harya etal. 2004) o o o o © © © o © © © © © © © © © © © © © © © © © Accession NM 011542 NM 028602 NM 019456 NM 011076 NM 009764 NM 011242 NM 013930 NM 016811 NM 023233 NM 011638 NM 008957 AA874130 NM 008122 NM 177836 NM 008060 NM 016672 NM 016888 NM 010697 AI839138 NM 007530 AA209597 NM 007707 NM 007707 NM 011682 NM 007382 Gene Name Any ID 102344_s_ at 102418 at 102710 at 102910 at 102976 at 103282 at 103389 at 103596 at 103691 at 103957 at 104030 at 104625 at 104635_r_a t 106872 at 135386_r_a 160074 at 160137 at 160498 at 160547_s_ at 160876 at 161096 at 162206_f_a t 92232 at 92507 at 92581 at Gene Name Any ID Tcea3 Apbblip Abcbla Brcal & B> &; Aass Dgka Trim 13 Trfr Ptch so § g Gjal AW046396 e a CN Ddc I 03 Ldbl Txnip a, CN & rtt •o U 05 o CO <"n "3 o Utrn Acadm 195 sO O N Bhattac Overlap Kelly Overlap harya Brandenbe Sato et Sperger with all Ivanova and Ramalho- (Sharov with all (Ginis Overlap etal. rger et al. al. et al. human etal. Rizzino Santos et et al. mouse et al. with all Gene Name Affy ID Accession 2004) 2004) 2003) 2003) data sets 2002) 2000) al. 2002) 2003) data sets 2004) data sets App 93063 at N M 007471 0 0 1 0 1 0 0 0 0 0 0 1 NrObl 93141 at N M 007430 0 0 0 0 0 0 0 1 0 1 0 1 U90005106Rik 93143 at N M 197988 0 0 0 0 0 1 0 0 0 1 0 1 C330022M23 93478 at AA612483 0 0 0 0 0 0 0 1 0 1 0 1 Scdl 94056 at N M 009127 0 0 1 0 1 0 0 0 0 0 1 Gli 94097 at N M 010296 0 0 0 1 1 0 0 0 0 0 1 Jmjdl 95033 at AW049513 0 0 0 0 0 0 0 1 0 1 0 1 Rex2 95346 at N M 009051 0 0 0 0 0 0 0 1 0 1 0 1 Cdc5l 95610 at AA636547 0 0 0 0 0 1 0 0 0 1 0 1 5830411E10Rik 95655 at N M 028696 0 0 0 0 0 0 0 1 0 1 0 1 C77370 95963 at C77386 0 0 0 0 0 0 0 1 0 1 0 1 Kip 96109 at N M 008452 0 0 0 0 0 0 1 0 0 1 0 1 Dmclh 96498 at N M 010059 0 0 1 0 1 0 0 0 0 0 1 Acp6 96744_at N M 019800 0 0 0 0 0 1 0 0 0 1 0 1 lcam I 96752 at N M 010493 0 0 0 0 0 1 0 0 0 1 0 1 9430077C05Rik 97138 at N M 178059 0 0 0 0 0 1 0 0 0 1 0 1 D930018N21Rik 97193 at N M 023755 0 0 0 0 0 1 0 0 0 1 0 1 1600025H15Rik 97320 at N M 028064 0 0 0 0 0 1 0 0 0 1 0 1 H2afy 97452 at N M 012015 0 0 1 0 1 0 0 0 0 0 1 TcflS 97717 at N M 009328 0 0 0 0 0 1 0 0 0 1 0 1 Pfkp 97833 at N M 019703 0 0 0 0 0 1 0 0 0 1 0 1 Pfkp 97834 g at N M 019703 0 0 0 0 0 1 0 0 0 1 0 1 Klf4 99622 at N M 010637 0 0 0 0 0 0 1 0 0 1 0 1 Decreased between 0 and 72 hours Leftb 102345 at D83921 1 1 1 1 4 1 0 1 0 2 0 6 Jadel 95883 at N M 172303 1 1 1 1 4 1 0 1 0 2 0 6 Smarcad I 98991 at X69942 0 1 1 1 3 1 0 1 1 3 0 6 Sox2 100009 r at N M 011443 1 1 0 1 3 1 0 0 0 1 1 5 Ak4 99959 at N M 009647 0 1 1 1 3 1 0 1 0 2 0 5 Mybl2 100023 at N M 008652 0 0 0 1 1 1 1 1 0 3 0 4 Overlap with all data sets -t -t -T -r -r rn m rn rn m rn rn rn rn cn cn O l CN O l O l O l O l CN r i CN CN (Ginis et al. 2004) o © © © O © © - © © o © o O o © o © © © © © © © © © © © © Overlap with all mouse data sets CN on CN CN rn CN CN CN CN cn CN CN CN rn cn CN O l CN O l CN CN — - © (Sharov et al. 2003) o — O - o © - O © o © O © © - © - © o © © © © © © © © Ramalho-Santos et al 2002^ — © © © Kelly and Rizzino 2000) o o o o © © © O © © o o o o o © © - © © © o © © © © © © © Ivanova et al. 2002) © Overlap with all human data sets CN - CN CN CN CN CN - CN - © - - © © o i © © © © - © © - — CN Sperger et al. — — — - - — © - - —1 © o - - © © © - © © © © - © © © © © Sato et al. 2003) © - o o © o - - o o o - o © - © © o O © o O — — -Brandenbe rger et al. 2004) o © o o © - o © © o O o o o © © © © © © © © © © © o © © -Bhattac harya etal. 2004) © © o o © © © - O o o o o o o © © o © o o © © © © © © © O Accession NM 009035 NM 011607 NM 013827 NM 013529 NM 011607 NM 011401 U28419 NM 009556 NM 013827 NM 011333 NM 010838 NM 011635 NM 010728 NM 013512 NM 013512 NM 007392 NM 009337 NM 010262 NM 022880 NM 009535 NM 007992 NM 008818 NM 011535 NM 009616 AAl 83875 NM 023755 109909VV NM 022721 NM 022722 Affy ID 101902 at 101993 at 103657 i at 103737 at 162362 f at 93804 at 94745 f at 98414 at 102069 at 102736 at 102742 g at 102764 at 160095 at 161603 r at 92231 at 93100 at 93296 at 94200 at 95733 at 100015 at 100928 at 101368 at 103538 at 103554 at 103728 at 103761 at 107952 i at 108784 at 1 13671 at cu s Z CU a O Rbpsuh Tnc Mtf2 1 0 Tnc i*i Q <N "o &a Eifla Zfp42 1 Ccl2 Mapt Trapla Lox Epb4.1l4a > -> 5-ylcta2 Tell <N ' O Sic 2 9a 1 Yes Fblnl Pern Adaml9 AAl83875 D93()0l8N2lRi k n r- I FzdS Dpys 197 Bhattac harya Brandenbe Sato et Sperger Overlap with all Ivanova Kelly and Ramalho- (Sharov Overlap with all (Ginis Overlap etal. rger et al. al. et al. human et al. Rizzino Santos et et al. mouse et al. with all Gene Name Affy ID Accession 2004) 2004) 2003) 2003) data sets 2002) 2000) al. 2002) 2003) data sets 2004) data sets Cav 160280 at N M 007616 0 0 0 0 0 1 0 1 0 2 0 2 Tuba3 160420 r at N M 009446 0 0 1 0 1 1 0 0 0 1 0 2 Aldo3 160546 at AW121134 0 0 1 0 1 1 0 0 0 1 0 2 8430410AllRik 160684 at N M 173737 0 0 0 0 0 1 0 1 0 2 0 2 E130012A19Rik 160712 r at N M 175332 0 0 0 0 0 1 0 1 0 2 0 2 Trim! 163115 at N M 030706 0 1 0 1 2 0 0 0 0 0 0 2 Akap2 163655 s at N M 009649 0 0 0 1 1 1 0 0 0 1 0 2 Cnol4 163709 i at N M 016877 0 1 0 1 2 0 0 0 0 0 0 2 Tcfap2c 92275 at N M 009335 0 0 0 1 1 0 0 1 0 1 0 2 Msc 93234 at N M 010827 0 0 0 0 0 1 0 1 0 2 0 2 Pcolce 93349 at N M 008788 0 0 0 0 0 1 0 1 0 2 0 2 Bmp4 93455 s at N M 007554 0 1 0 1 2 0 0 0 0 0 2 Bnip3 93836 at N M 009760 0 0 1 0 1 0 0 1 0 1 0 2 Catnal 1 94174 at N M 018761 0 0 0 1 1 0 0 1 0 1 0 2 D7Ertdl43e 95954 at AA437728 0 0 0 0 0 1 0 1 0 2 0 2 Bgn 96049 at N M 007542 0 0 0 0 0 1 0 1 0 2 0 2 Fbxol5 96162 at AA683807 0 0 0 0 0 1 0 1 0 2 0 2 0610013D04Rik 96785 at N M 030697 0 0 0 0 0 1 0 1 0 2 0 2 6720457D02Rik 97129 at AA617494 0 0 0 0 0 1 0 1 0 2 0 2 Empl 97426 at N M 010128 0 0 0 0 0 1 0 1 0 2 0 2 Sppl 97519 at N M 009263 0 0 0 0 0 1 0 1 0 2 0 2 Sgk 97890 at N M 011361 0 0 0 0 0 1 0 1 0 2 0 2 Klf5 97937 at N M 009769 0 0 0 0 0 1 0 1 0 2 0 2 Ppap2a 98508 s at N M 008903 0 0 1 0 1 0 0 1 0 1 0 2 Nrld2 99076 at N M 011584 0 0 1 0 1 1 0 0 1 0 2 Piml 99384 at N M 008842 0 0 1 1 2 0 0 0 0 0 2 Uppl 100030 at N M 009477 0 0 0 0 0 0 0 1 0 1 0 1 Nidi 100120 at N M 010917 0 0 0 0 0 0 0 1 0 1 0 1 GU2 100395 at X99104 0 0 0 1 1 0 0 0 0 0 0 1 Osmr 102255 at N M 011019 0 0 0 1 1 0 0 0 0 0 o 1 sC sC Gene Name Affy ID Accession Bhattac harya etai. 2004) Brandenbe rger et al. 2004) Sato et al. 2003) Sperger etal. 2003) Overlap with all human data sets Ivanova et al. 2002) Kelly and Rizzino 2000) Ramalho-Santos et al. 2002) (Sharov etal. 2003) Overlap with all mouse data sets (Ginis et al. 2004) Overlap with all data sets Gadd45a 102292 at N M 007836 0 0 0 0 0 0 0 1 0 1 0 1 TexJ9 102418 at N M 028602 0 0 0 0 0 0 0 1 0 1 0 1 Abcbla 102910 at N M 011076 0 0 0 0 0 1 0 0 0 1 0 1 Aass 103389 at N M 013930 0 0 1 0 1 0 0 0 0 0 1 Lama I 103729 at N M 008480 0 0 0 1 1 0 0 0 0 0 1 Ptch 104030 at N M 008957 0 0 0 0 0 0 0 1 0 1 0 1 G/b3 104232 at N M 008126 0 0 0 0 0 0 0 1 0 1 0 1 104635 r at N M 008122 0 0 0 1 1 0 0 0 0 0 1 4930553MI8Rik 104639 i at N M 029248 0 0 0 0 0 0 0 1 0 1 0 1 Fut9 105675 at N M 010243 0 0 0 0 0 1 0 0 0 1 0 1 My lip 106479 at N M 153789 0 0 0 1 1 0 0 0 0 1 Nrld2 107103 at N M 011584 0 0 0 0 0 1 0 0 0 1 0 1 Trpsl 110414 at N M 032000 0 0 0 0 0 0 0 1 1 0 1 C330022M23 111970 at A1616223 0 0 0 0 0 1 0 0 0 1 0 1 Dl30027M04 112257 at AA958476 0 0 0 0 0 1 0 0 0 1 0 1 Ferll3 112941 f at AA960514 0 0 0 0 0 1 0 0 0 1 0 1 A1595304 113495 at AI595304 0 0 0 0 0 1 0 0 0 1 0 1 2410043F08Rik 115028 at N M 133754 0 0 0 0 0 1 0 0 0 1 0 1 Ddahl 115112 at N M 026993 0 0 0 0 0 1 0 0 0 1 0 1 3110031O14Rik 116738 at N M 026081 0 0 0 0 0 1 0 0 0 1 0 1 G2an 135386 r at N M 008060 0 0 0 0 0 1 0 0 0 1 0 1 Ddc 160074 at N M 016672 0 0 0 0 0 1 0 0 0 1 0 1 B3gntl 160137 at N M 016888 0 0 0 1 1 0 0 0 0 1 Thbsl 160469 at N M 011580 0 1 0 0 1 0 0 0 0 1 Aard 160968 at N M 175503 0 0 0 0 0 1 0 0 0 1 0 1 C530030118 161096 at AA209597 0 0 0 0 0 1 0 0 0 1 0 1 4930422J18Rik 161817 f at AV376312 0 0 0 0 0 1 0 0 0 1 0 1 Mras 164306 f at N M 008624 0 0 0 0 0 1 0 0 0 1 0 1 Tbx3 166427 f at N M 011535 0 0 0 0 0 1 0 0 0 1 0 1 Bgn 167023 f at N M 007542 0 0 0 0 0 1 0 0 0 1 0 Bhattac Overlap Kelly Overlap harya Brandenbe Sato et Sperger with all Ivanova and Ramalho- (Sharov with all (Ginis Overlap et al. rger et al. al. etal. human et al. Rizzino Santos et et al. mouse et al. with all Gene Name Affy ID Accession 2004) 2004) 2003) 2003) data sets 2002) 2000) al. 2002) 2003) data sets 2004) data sets NrObl 93141 at N M 007430 0 0 0 0 0 0 0 1 0 1 0 1 Zfp36ll 93324 at N M 007564 0 1 0 0 1 0 0 0 0 0 0 1 Gli 94097 at N M 010296 0 0 0 1 1 0 0 0 0 0 0 1 Collal 94305 at N M 007742 0 0 0 0 0 1 0 0 0 1 0 1 H6st 94345 at N M 010560 0 0 0 0 0 1 0 0 0 1 0 1 Abcal 94354 at N M 013454 0 0 0 1 1 0 0 0 0 0 1 Jmjdl 95033 at AW049513 0 0 0 0 0 0 0 1 0 1 0 1 583041 IE WRik 95655 at N M 028696 0 0 0 0 0 0 0 1 0 1 0 1 C77370 95963 at C77386 0 0 0 0 0 0 0 1 0 1 0 1 Sod2 96042 at N M 013671 0 0 0 0 0 0 0 1 0 1 0 1 Klf2 96109 at N M 008452 0 0 0 0 0 0 1 0 0 1 0 1 1110008H02 Rik 96156 at AA874329 0 0 0 0 0 0 0 1 0 1 0 1 Nfatc2ip 96561 at N M 010900 0 0 0 0 0 0 0 1 0 1 0 1 Dnajb9 96680 at N M 013760 0 1 0 0 1 0 0 0 0 0 1 learn 1 96752 at N M 010493 0 0 0 0 0 1 0 0 0 1 0 1 D930018N21Ri k 97193 at N M 023755 0 0 0 0 0 1 0 0 0 1 0 1 H2afy 97452 at N M 012015 0 0 1 0 1 0 0 0 0 0 1 Pfkp 97833 at N M 019703 0 0 0 0 0 1 0 0 0 1 0 1 Pfkp 97834 g at N M 019703 0 0 0 0 0 1 0 0 0 1 0 1 Egrl 98579 at N M 007913 0 0 0 1 1 0 0 0 0 0 1 Nfib 99440 at N M 008687 0 0 0 0 0 1 0 0 0 1 0 1 Klf4 99622 at N M 010637 0 0 0 0 0 0 1 0 0 1 0 1 Kit 99956 at N M 021099 0 0 0 1 1 0 0 0 0 0 0 1 Appendix C Table C I Function of the 88 genes on the first five Pareto fronts Gene Ontology Term Count Percentage Physiological Process 56 62.2 Metabolism 40 44.4 Cellular Physiological Process 28 31.1 Cell Growth And/Or Maintenance 24 26.7 Regulation Of Biological Process 23 25.6 Macromolecule Metabolism 22 24.4 Protein Metabolism 20 22.2 Nucleobase, Nucleoside, Nucleotide And Nucleic Acid Metabolism 18 20 Regulation Of Metabolism 16 17.8 Regulation Of Physiological Process 16 17.8 Development 15 16.7 Regulation Of Transcription 14 15.6 Regulation Of Nucleobase, Nucleoside, Nucleotide And Nucleic Acid Metabolism 14 15.6 Transcription 14 15.6 Regulation Of Transcription, Dna-Dependent 13 14.4 Transcription, Dna-Dependent 13 14.4 Protein Modification 11 12.2 Morphogenesis 10 11.1 Phosphate Metabolism 9 10 Phosphorus Metabolism 9 10 Protein Amino Acid Phosphorylation 8 8.9 Phosphorylation 8 8.9 Organogenesis 8 8.9 Axonogenesis 4 4.4 Cell Migration 4 4.4 Regulation Of Cell Proliferation 4 4.4 Cell Motility 4 4.4 Response To Chemical Substance 4 4.4 Chemotaxis 3 3.3 Taxis 3 3.3 Negative Regulation Of Transcription 3 3.3 Negative Regulation Of Nucleobase, Nucleoside, Nucleotide And Nucleic Acid Metabolism 3 3.3 Growth 3 3.3 Negative Regulation Of Cytokine Biosynthesis 2 2.2 Negative Regulation Of Cytokine Production 2 2.2 Negative Regulation Of Protein Biosynthesis 2 2.2 Negative Regulation Of Biosynthesis 2 2.2 201 Table C2 Comparison of genes on the first five Pareto Fronts with previously published data sets M G U74v2 MOE430 Gene Name Gene Symbol Bhatta charya Branden burger Sato Sperger Overlap with human Ivanova Ramalho -Santos Overlap with mouse Overlap with all 100009 r at 1416967 at SRY-box containing gene 2 Sox2 1 1 0 1 3 1 0 1 4 100030 at 1448562 at uridine phosphorylase 1 Uppl 0 0 0 0 0 1 1 2 2 101560 at 1415856 at embigin Emb 0 0 0 0 0 0 0 0 0 102012 at 1418895 at src family associated phosphoprotein 2 Scap2 0 0 0 0 0 1 0 1 1 102332 at 1448370 at Unc-51 like kinase 1 (C. elegans) Ulkl 0 0 0 0 0 0 0 0 0 103048 at 1417155 at neuroblastoma myc-related oncogene 1 Nmycl 0 0 0 0 0 1 1 2 2 103234 at 1424847 at neurofilament, heavy polypeptide Nefh 0 0 1 1 2 0 1 1 3 103342 at 1448653 at embryonic ectoderm development Eed 0 0 0 0 0 1 0 1 1 103653 at 1449590 a at muscle and microspikes RAS Mras 0 0 0 0 0 1 0 1 1 103728^at 1456521 at Transcribed locus 0 0 0 0 0 1 1 2 2 103737 at 1418753 at — — 0 0 0 1 1 1 0 1 2 103761 at 1418091 at transcription factor CP2-like 1 Tcfcp2ll 0 0 0 0 0 1 1 2 2 104139 at 1452094 at procollagen-proline, 2-oxoglutarate 4-dioxygenase (proline 4-hydroxylase), alpha 1 polypeptide P4hal 0 0 0 0 0 0 0 0 0 104544 at 1423327 at RIKEN cDNA 4930517K11 gene 4930517 KllRik 0 0 0 0 0 1 1 2 2 107103 at 1416958 at nuclear receptor subfamily 1, group D, member 2 Nrld2 0 0 0 0 0 1 0 1 1 108010 at 1418318 at ring finger protein 128 Rnfl28 0 0 0 0 0 0 0 0 0 108048 at 1454788 at ADP-ribosylation factor-like 7 Arl7 0 0 0 1 1 0 0 0 1 108097 at 1450626 at mannosidase, beta A , lysosomal Manba 0 0 0 0 0 1 0 1 1 108279 at 1455300 at RIKEN cDNA E130014J05 gene E130014 JOSRik 0 0 0 0 0 1 0 1 1 M G U74v2 MOE430 Gene Name Gene Symbol Bhatta charya Branden burger Sato Sperger Overlap with human Ivanova Ramalho -Santos Overlap with mouse Overlap with all 108712 at 1434917 at cordon-bleu Gobi 0 0 0 1 1 1 0 1 2 108784 at 1455604 at expressed sequence AI427138 AI42713 8 0 0 0 0 0 , 0 1 110429 at 1455333 at tensin 3 Tns3 0 0 0 1 1 1 0 1 2 111970 at 1460711 at RIKEN cDNA 4930461P20 gene 4930461 PlORik 0 0 0 0 0 0 1 1 12828 at 1448688 at podocalyxin-like Podxl 1 1 1 0 3 0 3 113673 at 1423508 at M Y S T histone acetyltransferase monocytic leukemia 4 Myst4 0 0 0 1 1 0 2 1 15058 at 1434362 at expressed sequence AW550831 AW5508 31 0 0 0 0 0 0 1 115445 at 1435374 at Transcribed locus — 0 0 0 0 0 1 0 1 1 115804 at 1438237 at RIKEN cDNA C230088H06 gene 0 0 0 0 0 0 1 116214_at 1456329 at R I K E N cDNA A230098A12 gene A230098 A12Rik 0 0 0 0 0 0 1 116435 at 1418076 at suppression of tumorigenicity 14 (colon carcinoma) Stl4 0 0 0 0 0 0 1 116872 at 1435437 at SET domain-containing protein 7 MG1.192 0501 0 0 0 0 0 0 1 117246 at 1448845 at ribonuclease P 25 subunit (human) Rpp25 0 0 0 0 0 0 1 133204 at 1455425 at Expressed sequence BB001228 BB00122 8 0 0 0 0 0 1 2 133365 at 1436568 at junction adhesion molecule 2 Jam2 0 1 0 0 1 1 0 1 2 133819 at 1419418 a at microrchidia More 0 0 0 0 0 1 0 1 1 160253 at 1423754 at interferon induced transmembrane protein 3 lfitm3 0 0 0 0 0 0 0 0 0 160370 at 1416552 at developmental pluripotency associated 5 Dppa5 0 0 0 0 0 1 1 2 2 M G U74v2 MOE430 Gene Name Gene Symbol Bhatta charya Branden burger Sato Sperger Overlap with human Ivanova Ramalho -Santos Overlap with mouse Overlap with all 160684 at 1423786 at RIKEN cDNA 8430410A17 gene 8430410 A17Rik 0 0 0 0 0 1 2 2 160828 at 1426858 at inhibin beta-B Inhbb 0 0 0 0 0 1 1 2 2 161042 at 1427912 at carbonyl reductase 3 Cbr3 0 0 0 0 0 1 1 1 161106 r at 1443892 at 0 0 0 0 0 1 1 2 2 162522 f at 1437015 x at phospholipase A2, group IB, pancreas Pla2glb 0 1 0 1 2 0 1 3 163005 s at 1429366 at leucine rich repeat containing 34 Lrrc34 0 0 0 0 0 1 0 1 1 163288 at 1460471 at RIKEN cDNA 2410146L05 gene 2410146 L05Rik 0 0 0 0 0 0 1 1 163489 at 1418488 s at receptor-interacting serine-threonine kinase 4 Ripk4 0 0 0 0 0 0 1 1 163715 at 1429399 at ring finger protein 125 Rnfl25 0 0 0 0 0 1 0 1 1 165699 r at 1453299 a at purine-nucleoside phosphorylase Pnp 0 0 0 0 0 1 0 1 1 166142_r_at 1436799 at RIKEN cDNA D230005D02 gene D230005 D02Rik 0 0 0 0 0 0 1 1 167088 r at 1456242 at LOC433110 LOC433 110 0 0 0 0 0 0 1 1 168508 at 1436926 at estrogen related receptor, beta Esrrb 0 0 0 0 0 1 0 1 1 92275 at 1418147 at transcription factor AP-2, gamma Tcfap2c 0 0 0 1 1 1 1 2 3 92476 at 1449288 at growth differentiation factor 3 Gdf3 1 0 1 0 2 1 1 2 4 92550 at 1417156 at keratin complex 1, acidic, gene 19 Krtl-19 0 1 1 1 -> i 0 0 3 92770 at 1421375 a at SI00 calcium binding protein A6 (calcyclin) SW0a6 0 0 0 0 0 1 0 1 1 93063 at 1427442 a at amyloid beta (A4) precursor protein App 0 0 1 1 2 0 1 3 93104 at 1426083 a at B-cell translocation gene 1, anti-proliferative BM 0 0 0 0 0 0 1 1 M G U74v2 MOE430 Gene Name Gene Symbol Bhatta charya Branden burger Sato Sperger Overlap with human Ivanova Ramalho -Santos Overlap with mouse Overlap with all 93141 at 1417760 at nuclear receptor subfamily 0, group B, member 1 NrObl 0 0 0 0 0 1 1 2 2 93271 s at 1450186 s at GNAS (guanine nucleotide binding protein, alpha stimulating) complex locus Gnas 0 1 0 1 2 0 0 0 2 93296 at 1422458 at T-cell lymphoma breakpoint 1 Tell 0 0 0 0 0 1 1 2 2 93483 at I449455 at hemopoietic cell kinase Hck 0 0 0 0 0 1 0 1 1 93864 s at 1421624 a at enabled homolog (Drosophila) Enah 0 0 0 1 1 0 0 0 1 94200 at 1420337 at gastrulation brain homeobox 2 Gbx2 0 0 0 0 0 1 1 2 2 94270 at 1448169 at keratin complex 1, acidic, gene 18 Krtl-18 1 1 1 0 3 0 0 0 3 94354 at 1421840 at ATP-binding cassette, sub-family A (ABC1), member 1 Abcal 0 0 0 0 0 1 0 1 1 94745_fat 1427479 at similar to Eukaryotic translation initiation factor 1A (elF-lA) (eIF-4C) MGC107 533 0 0 0 1 1 1 1 2 3 95033 at 1426810 at jumonji domain containing 1A Jmjdia 0 0 1 1 2 1 1 2 4 95518 at 1424683 at RIKEN cDNA 1810015C04 gene 1810015 C04Rik 0 0 0 0 0 1 0 1 1 95531 at 1454890 at angiomotin Amot 0 0 0 1 1 1 1 2 3 95584 at 1453223 s at developmental pluripotency associated 2 Dppa2 0 0 0 0 0 1 1 2 2 96042 at 1448610 a at superoxide dismutase 2, mitochondrial Sod2 0 0 1 0 1 1 1 2 3 96109 at 1448890 at Kruppel-like factor 2 (lung) Kip 0 0 0 0 0 0 1 1 1 96162 at 1427238 at F-box protein 15 Fbxol5 0 0 0 0 0 1 1 2 2 96203 at 1424713 at calmodulin-like 4 Calml4 0 0 0 0 0 1 0 1 1 96752 at 1424067 at intercellular adhesion molecule learn 1 0 0 0 0 0 1 0 1 1 96841 at 1451069 at proviral integration site 3 Pim3 0 0 0 0 0 1 0 1 1 M G U74v2 MOE430 Gene Name Gene Symbol Bhatta charya Branden burger Sato Sperger Overlap with human Ivanova Ramalho -Santos Overlap with mouse Overlap with all 96900 at 1433720 s at Nur77 downstream gene 2 MG1.214 3558 0 0 0 0 0 1 0 1 1 97083 at 1441023 at eukaryotic translation initiation factor 2, subunit 2 (beta) Eif2s2 0 0 0 1 1 1 0 1 2 97283 at 1424295 at developmental pluripotency-associated 3 Dppa3 0 0 0 0 0 0 1 1 1 97317 at 1415894 at ectonucleotide pyrophosphatase/phosphodiestera se 2 Enpp2 0 0 0 0 0 0 0 0 0 97426 at 1416529 at epithelial membrane protein 1 Empl 0 0 0 0 0 1 1 2 2 97442 at 1416832 at solute carrier family 39 (metal ion transporter), member 8 Slc39a8 0 0 0 0 0 1 0 1 1 97519 at 1449254 at secreted phosphoprotein 1 Sppl 0 0 0 0 0 1 1 2 2 97520 s at 1423506 a at neuronatin Nnat 0 1 0 1 2 0 0 0 2 97890 at 1416041 at serum/glucocorticoid regulated kinase Sxk 0 0 0 0 0 1 1 2 2 98414 at 1418362 at zinc finger protein 42 Zfp42 1 0 0 0 1 1 1 2 3 99561 f at 1448393 at claudin 7 Cldn7 0 0 0 1 1 1 0 1 2 99622 at 1417394 at Kruppel-like factor 4 (gut) Klf4 0 0 0 0 0 1 0 1 1 99956 at 1452514 a at kit oncogene Kit 0 0 0 1 1 1 0 1 2 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0100378/manifest

Comment

Related Items