EPAS1 REGULATION AND FUNCTION

IN THE HUMAN TROPHOBLAST

NORHAM ERLYANI BTE ABDUL HAMID

B.Sc. (Hons.), NUS

A THESIS SUBMITTED

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

NUS GRADUATE SCHOOL FOR INTEGRATIVE SCIENCES AND

ENGINEERING

NATIONAL UNIVERSITY OF SINGAPORE

2013

ACKNOWLEDGEMENTS

Writing this doctoral dissertation has been most challenging and it would not have been possible without the support of many people.

First of all, my gratitude goes to A*STAR, NGS, and GIS for giving me the opportunity to pursue the Doctor of Philosophy. I would like to thank A/P Paul

Robson for being a tremendous mentor, and for his invaluable guidance. Under his care, I was able to develop and grow into an independent researcher. His unyielding encouragement gave me confidence to pursue the work I have achieved today. For always pointing me in the right direction and having the student’s best interest at heart, I must thank the panel of advisors at GIS as well as my thesis advisory committee; A/P Neil Clarke, Asst. Prof. Tara Huber, Asst. Prof. Thilo Hagen, and Dr.

Shyam Prabhakar.

I owe special thanks to my co-workers, Dr. Wishva Herath and Dr. Li Juntao. Their indispensable expertise in Bioinformatics is deeply appreciated, without which much of the data would have been rendered unintelligible to me. To my peers Mehran

Rahmani, Nani Djunaidi, and Tapan Mistri, I thank you all for the sharing of ideas as well as laboratory material. Together, we can conquer the PhD! Additionally, I would also like to extend my thanks to Sun Lili, Jameelah Sheikh, Woon Chow Thai, and all members of the lab for their support and cooperation, ensuring research in the Robson lab proceeds as smoothly as possible.

To all colleagues at GIS, past and present, thank you for making this experience a memorable one for me. To Rani Ettikan, our dearest lab manager, we owe it to you for translating funding into laboratory reagents and consumables with great

i efficiency. To Li Pin, my surrogate mother at work, thank you for the stimulating chats and discussions, and for always keeping me and Heather on track.

To Heather Ang, my confidante and partner in Science, I thank you for the numerous discussions and exchange of ideas, as well as the long nights of thesis-writing and editing, staying up well into the witching hour. Our lively chats and inextinguishable discourse has not only challenged my outlook, but has most definitely fueled my development as a mature (maybe), thinking individual. Also to the creators of Fringe, thank you for validating my love for Science and the geek in me.

I would like to thank Dr. Elisabeth Zielonka, Nadine Richter, and Syahirah Abd

Karim for reading and editing my thesis, and for their valuable input and suggestions.

To my closest friends; Norishikin, thank you for always having faith in me and for always pushing me to do better; Fadilah, Iskandar, and little Hayder, thank you for always being there for me and for keeping me sane. Specially to little Hayder, your toothy smiles and little steps have kept me grounded and looking to brighter days ahead.

To Zulkarnain, my dearest husband, I thank you for your undying love and incredible support, no matter the nautical miles; I would not have made it without these wings.

Most importantly, I would like to thank my parents for their endless support throughout my life. Mama and Daddy, you raised me, you guided me, and you supported me every step of the way. For that, I can never thank you enough. I hope the person I am today is a daughter you can be proud of. To them, I dedicate this thesis.

ii TABLE OF CONTENTS

SUMMARY vii

LIST OF TABLES ix

LIST OF FIGURES x

LIST OF SYMBOLS xii

CHAPTER 1: INTRODUCTION 1 1.1. Understanding placental development in humans. 2 1.1.1. From zygote to blastocyst. 3 • Human embryonic stem cells. 4 • The early trophoblasts; progenitor to the placenta. 5 1.1.2. Extraembryonic membranes and the placenta. 8 1.1.3. Implantation and placentation. 12 1.1.4. Molecular basis of trophoblast formation. 17 1.1.5. Identifying trophoblast cell types. 22 1.1.6. Modeling human placentation. 28 1.1.7. Preeclampsia: A multi-factorial pregnancy complication. 30 1.2. EPAS1 and the HIF family. 34 1.2.1. The canonical HIF pathway: Oxygen-dependent regulation. 38 1.2.2. HIF target genes. 42 1.2.3. HIF1α vs. HIF2α: Are they redundant? 45 1.3. Role of hypoxia and HIF in development. 48 1.3.1. Placentation begins in an environment of physiological hypoxia. 49 • EPAS1 is expressed most abundantly in the human placenta. 50 1.3.2. HIF activity impacts placental development. 54 1.3.3. EPAS1 as a candidate gene that confers reproductive success. 58 • Impact of SNP on gene expression. 61 1.4. Experimental hypothesis and approach. 67

iii CHAPTER 2: MATERIALS & METHODS 69 2.1. Cell culture. 70 2.2. Gene expression assay. 71 2.2.1. RNA extraction. 71 2.2.2. Complementary DNA synthesis. 72 2.2.3. Quantitative real-time polymerase chain reaction. 72 2.3. Assay for alternative transcripts. 73 2.4. Screen of human cell lines for presence of SNP. 75 2.5. Microarray. 76 2.6. Immunocytochemistry. 76 2.7. ChIP-Seq. 77 2.7.1. Chromatin immunoprecipitation. 77 2.7.2. Quality check of ChIP-DNA: Western blot and qPCR. 80 2.7.3. Sequencing. 82 2.7.4. Bioinformatics analysis of ChIP-Seq targets. 84 2.8. RNA-Seq. 84

CHAPTER 3: FUNCTIONAL CHARACTERIZATION OF THE TARGET EPAS1 SNP 86 3.1. Introduction. 87 3.2. Characterizing the locus of the EPAS1 SNP. 89 3.3. Alternative splicing between exon 2 and 7. 93 3.4. Screening of human cell lines for the target EPAS1 SNP. 99 3.5. Discussion. 101 3.5.1. Significance of the locus of the SNP. 101 3.5.2. Impact of cryptic splice sites in exon 6. 106 3.5.3. The SNP screening process. 110 3.6. Conclusion. 111

CHAPTER 4: IDENTIFICATION OF EPAS1 TARGETS 113 4.1. Introduction. 114 4.2. In vitro differentiation into trophoblast lineage at 20% and 3% oxygen. 117

iv 4.3. EPAS1 targets by ChIP-Seq. 123 4.4. In vitro & in vivo expression of EPAS1 binding targets. 133 4.5. Discussion. 137 4.5.1. Establishing the trophoblast lineage in 3% oxygen. 137 4.5.2. Verification of ChIP. 140 4.5.3. Assessment of EPAS1/HIF2α targets in the human trophoblast. 142 4.5.4. EPAS1 regulation and preeclampsia. 145 4.5. Concluding remarks. 147

CHAPTER 5: GLOBAL TRANSCRIPTOMIC ANALYSIS OF THE PRE- AND POST-FLOW PLACENTA 150 5.1. Introduction. 151 5.2 mRNA-seq data generation. 154 5.2.1. Human chorionic gonadotropin genes. 155 5.3. Differential analysis and gene ontology. 159 5.3.1. Hemoglobin genes. 163 5.3.2. Blood coagulation. 166 5.3.3. Complement system. 169 5.4. Genes most abundantly expressed in the placenta. 171 5.4.1. Chorionic somatomammotropin hormone 1. 171 5.4.2. Growth differentiation factor 15. 171 5.4.3. Peptidylprolyl isomerase B. 172 5.4.4. Epstein-Barr virus induced 3. 172 5.4.5. Ferritin light polypeptide. 173 5.5. Genes associated with preeclampsia and other pregnancy complications. 177 5.5.1. Leptin. 177 5.5.2. Inhibin genes. 177 5.5.3. Endoglin and fms-related tyrosine kinase 1. 178 5.5.4. Secreted phosphoprotein 1. 179 5.5.5. Hepatocyte growth factor. 179 5.6. Discussion. 181 5.6.1. Hemoglobin switching by the placenta. 181

v 5.6.2. Changing state of the placenta: Anti-coagulatory to pro-coagulatory. 184 5.6.3. Complement system and pregnancy. 186 5.6.4. Recognizing important placental genes. 188 5.6.5. Influence of the uterine environment on preeclampsia pathogenesis. 192 5.7. Concluding remarks. 199

CHAPTER 6: CLOSING REMARKS 203

BIBLIOGRAPHY 206

APPENDIX 241

vi SUMMARY

Early development in humans occur in an environment of low oxygen. During the first trimester, maternal blood flow into the placenta is blocked by trophoblastic plugs; a phase we term ‘pre-flow’. Placental perfusion only begins upon the removal of these plugs, between 10 to 12 weeks of gestation. The sudden presence of blood at the end of the first trimester (i.e. post-flow) brings with it oxygenation, drastically modifying the uterine environment.

We identified EPAS1 (also known as HIF2α), a transcription factor responding in low oxygen and which has been implicated in placental development and function, to be pivotal to the placental response to the changing oxygen conditions of its environment. Moreover, the EPAS1 gene recently emerged as the gene with the strongest signal of natural selection. A SNP was found in native Tibetan highlanders, a geographically-isolated population that has survived generations at reduced oxygen levels, which appeared to confer not only survivability, but also reproductive fitness.

Given the rapid enrichment of this particular allele,, we reasoned that natural selection could likely be acting in utero, and that EPAS1 was not only instrumental in placental development, but that this SNP also effected a profound change in its role.

Using a hESC-derived trophoblast culture maintained at 3% oxygen, I took a three- pronged approach to this study. I first characterized the SNP found in Tibetan people, assessing the impact of a nucleotide change in its position. Molecular analysis of

EPAS1 transcripts uncovered two cryptic splice sites in close proximity to the SNP.

Utilization of either cryptic splice site would lead to modification of the adjacent exon, likely resulting in attenuation of EPAS1 activity.

vii Through ChIP-Seq and preliminary analysis, I identified 35 potential EPAS1 targets.

Twelve were found by global microarray analysis to be upregulated under hypoxia in vitro, while eight were confirmed through gene expression analysis of human placental samples. From our findings, seven genes had previously been identified as

HIF targets in a different cell type, including EGFR and ELF3. Strikingly, FLT1 and

LEP, two genes frequently implicated in preeclampsia were also associated with our results. Preeclampsia, a condition of the placenta characterized by inadequate perfusion of the tissue, which likely leads to dysregulation of oxygen-sensitive gene expression, while intensively studied, remains largely not well understood. Given the oxygen-sensitive nature of EPAS1 and its capacity in mediating gene expression in the trophoblast, there is a likely role for EPAS1 in the gene dysregulation that occurs in preeclampsia.

Analysis of human placental samples were achieved through RNA-Seq, and by comparison between pre-flow and post-flow gene expression. Through this analysis, I was additionally able to explore the global transcriptomic landscape of the early placenta, and contrast changes in gene expression across the changing uterine environment. Our analysis demonstrated the potential to identify biological processes activated upon placental perfusion, as well as to further our understanding of the influence of this developmental milestone on pregnancy outcomes.

viii LIST OF TABLES

Table 1.1 (A) Frequently used definitions describing low oxygen. Table 1.1 (B) Physiological oxygen tensions in a selection of normal human tissue types. Table 1.2. (A) Members of the HIF family. Table 1.2. (B) Associations as well as transcriptional activity of the HIF complexes. Table 1.3. Selected HIF target genes grouped according to their function. Table 1.4. Placental defects in mice lacking in components of the HIF transcription complex. Table 1.5. SNPs/Mutations in splicing regulatory elements and resulting outcome. Table 4.1. Preliminary analysis of EPAS1 binding sites. Table 5.1. mRNA-Seq samples and library information. Table 5.2. Pre-flow expression of genes Table 5.3. Post-flow expression of genes. Supplementary Table 1. 681 genes differentially expressed between pre- and post- flow placentae. Supplementary Table 2. List of pathways and genes from gene ontology analysis.

ix LIST OF FIGURES

Figure 1.1. Development of the human embryo. Figure 1.2. Development of the extraembryonic membranes during early pregnancy. Figure 1.3. Progression of implantation and invasion of the endometrial lining by trophoblasts. Figure 1.4. Trophoblast lineage cell types and proposed markers. Figure 1.5. Oxygen tension measurements in various stem cell niches. Figure 1.6. HIF regulation under aerobic and hypoxic conditions. Figure 1.7. Expression pattern for EPAS1 and HIF1A across different human tissues. Figure 1.8. Expression of EPAS1 by mRNA-Seq in hESC-derived trophoblast culture. Figure 3.1. Locus of EPAS1 SNP shows some sequence conservation. Figure 3.2. Locus of EPAS1 SNP has specific base preferences. Figure 3.3. Screening of exon-exon junctions in the EPAS1 mature transcript showed under-representation of exon 4-5 junction and exon 5-6 junction. Figure 3.4. Alternative splicing between exons 2 and 7 of the EPAS1 gene. Figure 3.5. Novel splice acceptor sites in exon 6. Figure 3.6. Representative sequencing chromatograms from screened samples. Figure 3.7. Exon 6 as the spacer/linker between two functional domains of EPAS1. Figure 4.1. Culture of hES and trophoblast-induced cells are comparable between 20% and 3% oxygen conditions. Figure 4.2. Venn analysis of upregulated genes. Figure 4.3. Expression of pluripotency markers as well as trophoblast markers and placental genes in BMP4/SU treatment at 20% and 3% oxygen conditions. Figure 4.4. Identification of EPAS1/HIF2α in vitro. Figure 4.5. Strongest EPAS1 binding site is ~50kb downstream of SLC2A1. Figure 4.6. 2nd-strongest EPAS1 binding site is just downstream of HIF1A.

Figure 4.7. Peak #4 is nearest to ARRDC3.

x Figure 4.8. Peak #14 is found within exon 1 of EGLN1, in the 5’ UTR. Figure 4.9. Peak #28 is found ~40kb downstream of FLT1. Figure 4.10. In vitro expression of EPAS1 targets. Figure 4.11. In vivo expression of EPAS1 targets by RNA-Seq. Figure 4.12. 35 genes associated with EPAS1 binding sites. Figure 5.1. Expression of hCG genes. Figure 5.2. Heatmap of differentially expressed genes. Figure 5.3. Gene ontology analysis. Figure 5.4. Expression of hemoglobin genes indicates hemoglobin switching. Figure 5.5. Expression of genes involved in the blood coagulation cascade. Figure 5.6. Expression of genes involved in the complement cascade. Figure 5.7. Genes abundantly expressed in the placenta. Figure 5.8. Expression of genes implicated in preeclamptic pregnancies.

xi LIST OF SYMBOLS

(-)BID17 negative binding site, >1kbp upstream of BID 17 (NF)-κB nuclear factor of kappa light polypeptide gene enhancer in B- cells 0.1%RSB 0.1% Reduced SDS Buffer 5’ UTR 5’ untranslated region A, T, C, G nucleotide bases adenine, thymine, cytosine, and guanine A2M alpha-2-macroglobulin ABCC8 ATP-binding cassette, sub-family C (CFTR/MRP), member 8 ABCG2 ATP-binding cassette, sub-family G (WHITE), member 2 ACVR1B activin A receptor, type IB ACVRL1 activin A receptor type II-like 1 ADAM-19 ADAM metallopeptidase domain 19 ADM adrenomedullin ALDOA aldolase A, fructose-bisphosphate ALPPL2 alkaline phosphatase, placental-like 2 AMOTL2 angiomotin like 2 Ang-1 (mouse) angiopoietin 1 ARID3A AT rich interactive domain 3A (BRIGHT-like) ARNT aryl hydrocarbon receptor nuclear translocator ARRDC3 arrestin domain containing 3 ATP adenosine triphosphate BAC bacterial artificial chromosome bFGF basic fibroblast growth factor bHLH basic helix-loop-helix bHLH-PAS basic helix-loop-helix Per-Arnt-Sim BHLHE40 basic helix-loop-helix family, member e40

BID17 novel HIF-binding site

xii BMP4 bone morphogenetic protein 4 C1QA complement component 1, q subcomponent, A chain C1QB complement component 1, q subcomponent, B chain C1QC complement component 1, q subcomponent, C chain C1R complement component 1, r subcomponent C3AR1 complement component 3a receptor 1 C4ORF3 chromosome 4 open reading frame 3 C7 complement component 7 CA9 carbonic anhydrase 9 CBP/p300 Creb-binding protein CCL14 chemokine (C-C motif) ligand 14 CCL8 chemokine (C-C motif) ligand 8 CCR1 chemokine (C-C motif) receptor 1 CD55 CD55 molecule, decay accelerating factor for complement CD59 CD59 molecule, complement regulatory protein cDNA complementary DNA CDS coding sequence Cdx2 (mouse) caudal type homeobox 2 CDX2 caudal type homeobox 2 CEACAM1 carcinoembryonic antigen-related cell adhesion molecule 1 CEBPA CCAAT enhancer binding protein alpha CERES composite exonic regulatory element of splicing CFB complement factor B CFD complement factor D CFTR cystic fibrosis transmembrane-conductance receptor CGA chorionic gonadotropin, alpha subunit CGB chorionic gonadotropin, beta subunit

ChIP chromatin immunoprecipitation

xiii CITED-2 Cbp/p300-interacting transactivator, with Glu/Asp-rich carboxy-terminal domain, 2 CK7 cytokeratin 7 CLK3 CDC-like kinase 3 COX cytochrome c oxidase COX4I2 cytochrome c oxidase subunit IV isoform 2 cRNA synthesized RNA CSH1 chorionic somatomammotropin hormone 1 Ct threshold cycle CTGF connective tissue growth factor CVS chorionic villus sampling CX3CL1 chemokine (C-X3-C motif) ligand 1 CXCL16 chemokine (C-X-C motif) ligand 16 CXCR6 chemokine (C-X-C motif) receptor 6 CypB cyclophilin B DAF delay accelerating factor DAPI 4',6-diamidino-2-phenylindole DEC-1, -2 deleted in esophageal cancer 1, 2 DLX3 distal-less homeobox 3 DMEM/F12 Dulbecco's Modified Eagle Medium: Nutrient Mixture F-12 Media DNA deoxyribonucleic acid dsDNA double stranded DNA DTT dithiothreitol EBI3 Epstein-Barr virus induced 3 ECM extracellular matrix EDTA ethylenediaminetetraacetic acid EGFR epidermal growth factor receptor

EGLN1 egl nine homolog 1 (C. elegans); PHD1 gene

xiv EGLN3 egl nine homolog 3 (C. elegans); PHD3 gene ELF3 E74-like factor 3 (ets domain transcription factor, epithelial- specific) Elf5 (mouse) E74-like factor 5 ELF5 E74-like factor 5 (ets domain transcription factor) ELK-1 member of ETS oncogene family ENG endoglin Eomes (mouse) eomesodermin EPAS1 endothelial PAD domain 1 EPCR endothelial protein C receptor EPO erythropoietin ER endoplasmic reticulum ERVWE1 endogenous retrovirus group W, member 1 ESAM endothelial cell adhesion molecule ESE exonic splicing enhancer ESS exonic splicing silencer ETS E-twenty six family of transcription factors ETS-1 v-ets erythroblastosis virus E26 oncogene homolog 1 (avian) Ets2 (mouse) E26 avian leukemia oncogene 2, 3' domain EVTs extravillous trophoblasts F13A1 coagulation factor XIII, A subunit (A1 polypeptide) F3 coagulation factor III, tissue factor FDR false discovery rate FGF1 fibroblast growth factor 1 FGF2 fibroblast growth factor 2 FGF4 fibroblast growth factor 4 FGFR fibroblast growth factor receptor FIH factor inhibiting HIF

FLK-1 protein-tyrosine kinase receptor

xv FLT1 fms-related tyrosine kinase 1 FPKM fragments per kilobase of transcript per million mapped fragments FSH follicle stimulating hormone FTL ferritin light polypeptide GAPDH glyceraldehyde 3-phosphate dehydrogenase GATA2 GATA binding protein 2 Gata3 (mouse) GATA binding protein 3 GATA3 GATA binding protein 3 GATA4 GATA binding protein 4 GCM1 glial cells missing homolog 1 (Drosophila) GDF15 growth differentiation factor 15 GFAP glial fibrillary acidic protein GKN1 gastrokine 1 GLUT1 glucose transporter 1 HADH hydroxyacyl-CoA dehydrogenase HAND1 heart and neural crest derivatives expressed 1

HbA, α2β2 adult hemoglobin, 2 alpha and 2 beta chains HBA1 hemoglobin alpha subunit 1 HBA2 hemoglobin alpha subunit 2 HBB hemoglobin beta subunit HBD hemoglobin delta subunit HBE1 hemoglobin epsilon subunit 1

HbF, α2γ2 fetal hemoglobin, 2 alpha and 2 gamma chains HBG1 hemoglobin gamma subunit 1 HBG2 hemoglobin gamma subunit 2 HBS HIF-binding sequence HBZ hemoglobin zeta subunit hCG human chorionic gonadotropin hormone

xvi HEPES-KOH 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, potassium hydroxide-buffered HERV-FRD endogenous retrovirus group FRD, member 1 hESC human embryonic stem cell HGF hepatocyte growth factor HIF hypoxia inducible factor HIF-1 hypoxia inducible factor, transcription complex 1 HIF-2 hypoxia inducible factor, transcription complex 2 HIF-α hypoxia inducible factor, alpha subunit HIF-β hypoxia inducible factor, beta subunit HIF1A gene for hypoxia inducible factor 1, alpha subunit HIF1α hypoxia inducible factor 1, alpha subunit HIF2α hypoxia inducible factor 2, or EPAS1, alpha subunit HIF3A gene for hypoxia inducible factor 3, alpha subunit HIF3α hypoxia inducible factor 3, alpha subunit HILPDA hypoxia inducible lipid droplet-associated HIST1H1C histone cluster 1, H1c HLA-G human leukocyte antigen HO-1 heme oxygenase (decycling) 1 HRE hypoxia-responsive elements hTSC human trophoblast stem cell ICM inner cell mass ID2 inhibitor of DNA binding 2 IDH2 isocitrate dehydrogenase 2 (NADP+), mitochondrial IFN-γ interferon-gamma IgG immunoglobulin G, antibody isotype IL10 interleukin-10 IL12 interleukin-12

IL1R2 interleukin 1 receptor, type II

xvii IL1RAP interleukin 1 receptor accessory protein IL27 interleukin-27 IL2RB interleukin 2 receptor, beta IL35 interleukin-35 IL6 interleukin-6 INHA inhibin, alpha INHBA inhibin, beta A IP S/N supernatant from immunoprecipitation reaction iPSCs induced pluripotent stem cells

IPTG isopropyl β-D-1-thiogalactopyranoside ISE intronic splicing enhancer ISS intronic splicing silencer IUGR intrauterine growth restriction IVF in vitro fertilization IVS intervillus space JAM2 junctional adhesion molecule 2 KISS1R KISS1 receptor KLF5 Kruppel-like factor 5 (intestinal) KLK12 kallikrein 12 KRT18 keratin 18 KRT19 keratin 19 L3MBTL1 l(3)mbt-like 1 (Drosophila) LARGE like-glycosyltransferase LDHA lactate dehydrogenase A LEP leptin LH luteinizing hormone LiCl lithium chloride

LIF leukemia inhibitory factor

xviii LTR long terminal repeat M-CSF macrophage colony stimulating factor MAC membrane attack complex MAFK v-maf musculoaponeurotic fibrosarcoma oncogene homolog K (avian) MAPK4 mitogen-activated protein kinase 4 MCAD medium-chain acyl-CoA dehydrogenase MID1 midline 1 (Opitz/BBB syndrome) MIR17HG miR-17-92 cluster host gene (non-protein coding) miRNA micro RNA mRNA messenger RNA MSX2 msh homeobox 2 mTOR mammalian target of rapamycin MXD1 MAX dimerization protein 1 MXI1 MAX interactor 1, dimerization protein MYF5 myogenic factor 5 MYH mutY homologue base excision repair MYOD1 myogenic differentiation 1 NaCl sodium chloride NaDOC sodium deoxycholate NANOG homeobox transcription factor NANS N-acetylneuraminic acid synthase NAT10 N-acetyltransferase 10 (GCN5-related) NEUROD1 neuronal differentiation 1 NP40 nonyl phenoxypolyethoxylethanol OAS1 2'-5'-oligoadenylate synthetase 1, 40/46kDa OCT4 POU class 5 homeobox 1 ODD oxygen degradation domain

OPN osteopontin

xix P2X(7) purinergic receptor P2X, ligand-gated ion channel, 7 PAI-1 plasminogen activator inhibitor-1 PAPPA pregnancy-associated plasma protein A, pappalysin 1 PAS Per(period circadian protein)-ARNT-Sim(single-minded protein) PAX6 paired box 6 PBS phosphate buffered saline PCR polymerase chain reaction PDGF platelet derived growth factor PDK1 pyruvate dehydrogenase kinase, isozyme 1 PE primitive endoderm PECAM1 platelet/endothelial cell adhesion molecule 1 PGF placenta growth factor PHD3 prolyl hydroxylase 3 PHDs prolyl hydroxylase PKM2 pyruvate kinase, muscle PP5 placental protein 5 PPARD peroxisome proliferator-activated receptor delta PPARG peroxisome proliferator-activated receptor gamma PPIB peptidylprolyl isomerase B PPP1R13L protein phosphatase 1, regulatory subunit 13 like PROCR gene for the receptor for activated protein C PVDF polyvinylidene fluoride pVHL von Hippel–Lindau tumor suppressor PVR poliovirus receptor, CD155 qPCR quantitative polymerase chain reaction RBP RNA binding protein rhBMP4 recombinant human bone morphogenetic protein 4

RIN RNA integrity number

xx RNA ribonucleic acid RPKM reads per kilobase per million mapped reads SALL4 sal-like 4 (Drosophila) SDS sodium dodecyl sulphate SEMA4C sema domain, immunoglobulin domain (Ig), transmembrane domain (TM) and short cytoplasmic domain, (semaphorin) 4C SERPINE2 serpin peptidase inhibitor, clade E (nexin, plasminogen activator inhibitor type 1), member 2 sFlt1 soluble fms-like tyrosine kinase-1 SGA small-for-gestational age SHOX short stature homeobox SLC28A1 solute carrier family 28 (sodium-coupled nucleoside transporter), member 1 SLC2A1 solute carrier family 2, facilitated glucose transporter membrane 1; GLUT1 SLC2A3 solute carrier family 2 (facilitated glucose transporter), member 3; GLUT3 SNP single nucleotide polymorphism snRNP small nuclear ribonucleoprotein SOD2 superoxide dismutase 2, mitochondrial SOX1 SRY (sex determining region Y)-box 1 SOX17 SRY (sex determining region Y)-box 17 SOX2 SRY (sex determining region Y)-box 2 SPP1 secreted phosprotein 1 SSEA4 stage-specific embryonic antigen-4 SU5402 fibroblast growth factor receptor 2 inhibitor SYCP3 synaptonemal complex protein 3 T/BRACHYURY T, brachyury homolog (mouse) TAD transactivation domain

TAE Tris-acetate buffer

xxi TAP1 transporter 1, ATP-binding cassette, sub-family B (MDR/TAP) TAP2 transporter 2, ATP-binding cassette, sub-family B (MDR/TAP) TBS Tris-buffered saline Tcfap2c (mouse) transcription factor AP-2, gamma TE trophectoderm Tead4 (mouse) TEA domain family member 4 TF tissue factor TFAP2A transcription factor AP-2 alpha (activating enhancer binding protein 2 alpha) TFAP2C transcription factor AP-2 gamma (activating enhancer binding protein 2 gamma) TFPI tissue factor pathway inhibitor TFPI2 tissue factor pathway inhibitor 2 TGF-β1 transforming growth factor, beta 1 TGFβ transforming growth factor, beta Tie-2 (mouse) endothelial-specific receptor tyrosine kinase TIMP2 TIMP metallopeptidase inhibitor 2 TNNT2 troponin T type 2 (cardiac) TP63 tumor protein p63 Tris-HCl trisaminomethane, hydrochloride-buffered TSH thyroid stimulating hormone TSS transcription start site U2AF65 U2 small nuclear ribonucleoprotein auxiliary factor 65 kDa VEGF vascular endothelial growth factor VEGF-A vascular endothelial growth factor A VEGFR vascular endothelial growth factor receptor Vegfr2 (mouse) vascular endothelial growth factor receptor 2 VGF VGF nerve growth factor inducible

VHL von Hippel-Lindau tumor suppressor, E3 ubiquitin protein ligase

xxii X-gal 5-bromo-4-chloro-indolyl-β-D-galactopyranoside ZO-1 tight junction protein 1

ζ2ε2 embryonic hemoglobin, 2 epsilon chains and 2 zeta chains

xxiii CHAPTER 1: INTRODUCTION

1 1.1. Understanding placental development in humans.

The fertilized egg, otherwise known as the zygote, undergoes numerous asymmetric cell divisions before implantation in the mother’s uterus, during a phase known as preimplantation. The zygote forms a spherical structure, the blastocyst, comprising two distinct cell populations - the inner cell mass (ICM), and the trophectoderm (TE).

The TE is the precursor to the placenta, the formation of which marks the first differentiation event that takes place in the embryo (Tarkowski and Wroblewska,

1967; Johnson and Ziomek, 1981; Arnold and Robertson, 2009; Rossant and Tam,

2009). The TE develops into the extraembryonic tissue which supports the growth and development of the embryo, while cells in the ICM give rise to the embryo proper.

Much of what is understood about early human development is inferred from studies in model mammals, primarily the mouse. Very little has been observed in the first weeks of embryonic development in humans; given the limitations of accessing human material, very few specimens have been studied and documented. In some cases, information about a particular stage of development originates from a single specimen (Boyd and Hamilton, 1960). Relevance of information gleaned from animal studies remains unclear since cellular interactions that mediate implantation and development of the placenta vary greatly even amongst primates (Martin, 2008).

Nevertheless, comparing across species may be a helpful beginning to elucidating the early stages of human embryo development, implantation, as well as placental development.

2 1.1.1. From zygote to blastocyst.

Fertilization of the human oocyte takes place in the oviduct, adjacent to the ovary

(Fig. 1.1). The zygote then makes its way to the uterus, over a period of five to seven days. During this time, the zygote divides continuously, with the first cleavage giving rise to two identical cells. Upon reaching the 8-cell stage, the early embryo begins to compact, and cells come together in a tight array interconnected by gap junctions. By the 16-cell stage, the compacted embryo, now known as the morula, sees the first segregation of cells along the basolateral cleavage plane, with smaller inner cells forming the ICM, and larger outer epithelial cells forming the TE. The ICM will eventually give rise to all the tissues of the embryo’s body, as well as to the non- trophoblastic extraembryonic tissues, such as the yolk sac, allantois, and amnion, all of which support and protect the developing embryo throughout gestation. The TE on the other hand, will generate the trophoblast cells of the chorion, its contribution to the embryo’s extraembryonic tissue, ultimately forming the placenta (Sutherland et al., 1990; Gilbert, 2000).

The cell segregation observed in the morula is accompanied by the formation of a fluid-filled cavity called the blastocoel, filled with fluid secreted by TE cells. As a result of cavitation and the accompanying physical separation of the ICM from the

TE, the morula now becomes a blastocyst, characterized by an outer sphere of flattened TE cells around a cluster of small round cells of the ICM in the fluid-filled blastocoel (Gilbert, 2000). The ICM is positioned asymmetrically within the blastocoel, attached to TE cells at one end. The outermost layer of cells of the ICM

3 that face the blastocoel is known as the hypoblast, and the remainder of the ICM is known as the epiblast.

This segregation is followed by the next differentiation event, in which the hypoblast form the beginnings of the extraembryonic endoderm, later developing into the outer layer of the yolk sac tissue. The epiblast meanwhile harbors pluripotent stem cells, which will give rise to the embryo proper (Gilbert, 2000). By definition, pluripotent stem cells are self-renewing, and have the ability to generate all three primary germ layers (Evans and Kaufman, 1981; Martin, 1981).

At this point, three differentiated cell types exist in the embryo - the trophoblast, epiblast, and hypoblast, also known as the primitive endoderm (PE). By this time, the preimplantation blastocyst is ready to implant in the uterine wall. During the next major phase of development known as gastrulation, the epiblast will give rise to the three primary germ layers - the endoderm, or the innermost layer that gives rise to the epithelial lining of, for example, the gastrointestinal and respiratory tract; the mesoderm, which gives rise to muscles; and the ectoderm, the outermost layer that forms the skin and the nervous system.

Human embryonic stem cells.

The pluripotent stem cells harbored within the ICM can and have been, as technology has allowed us, harvested for laboratory use. These cells, now known as human embryonic stem cells (hESCs), were derived from donated human embryos that had been produced by in vitro fertilization (IVF) for clinical purposes (Thomson et al.,

1998). hESCs are self-renewing and have the potential to be expanded limitlessly in

4 vitro. Their innate ability to bear all three germ layers has been replicated both in vitro, via directed differentiation or embryoid body formation (Itskovitz-Eldor et al.,

2000), and in vivo, as evidenced by the formation of teratomas in immunodeficient mice (Thomson et al., 1998; Stojkovic et al., 2004). They also have the capacity to give rise to germ cells, extra-embryonic tissue, and trophoblast cells, as demonstrated by previous studies (Xu et al., 2002; Hyslop et al., 2005; Babaie et al., 2007).

In hESCs, the transforming growth factor β (TGFβ) superfamily signaling plays a prominent role in the early cell fate decisions of embryonic development. hESC pluripotency is maintained via the TGFβ or Activin/Nodal signaling pathways, as well as the fibroblast growth factor 2 (FGF2)-mediated pathways (Vallier et al., 2004;

Beattie et al., 2005; Vallier et al., 2005; James et al., 2005; Greber et al., 2007; Greber et al., 2008). Both TGFβ and FGF signals synergize to inhibit bone morphogenetic protein (BMP) signaling, sustain expression of pluripotency-associated genes such as

NANOG, OCT4, and SOX2, and promote long-term undifferentiated proliferation of hESCs (Xu et al., 2008).

The early trophoblasts; progenitor to the placenta.

Trophoblast or TE formation represents the very first differentiation event and lineage restriction that a sub-population of the totipotent cells of the early embryo undergoes.

The TE gives rise to two main cell types - the cytotrophoblast and the syncytiotrophoblast. Cytotrophoblast cells make up the inner layer, and give rise to the outer layer of the implanting blastocyst, known as the primitive syncytium, which is made up of multinucleated syncytiotrophoblasts. The primitive syncytium appears to secrete enzymes that digest the lining of the maternal endometrium (Fisher et al.,

5 1989), allowing expansion of the primitive syncytium into the resultant fluid-filled spaces, or lacunae. Processes of the primitive syncytium, referred to as trabeculae, advance into the lacunae and as they do so expand the lacunae, which may go on to breach maternal sinusoids (Boyd and Hamilton, 1970; Hertig et al., 1956; Enders,

1989; Herzog, 1909).

The trophoblast cells go on to form the extraembryonic membrane called the chorion, ultimately developing into a large part of the placenta, which is crucial for healthy fetal development and protection. Placental development is therefore primarily dependent on trophoblast stem cell differentiation (Jurisicova et al., 2005).

6 Figure 1.1. Development of the human embryo. The formation of the human embryo begins with fertilization in the oviduct. The zygote divides continuously, and at the 8-cell stage, the early embryo begins to compact. The first segregation of cells occurs at the 16-cell stage, along the basolateral cleavage plane. The smaller inner cells give rise to the inner cell mass, and the larger outer epithelial cells form a layer known as the trophectoderm. © Terese Winslow, 2001.

7 1.1.2. Extraembryonic membranes and the placenta.

While the embryonic epiblast is undergoing gastrulation to form the early fetal structures, the extraembryonic cells proceed to form the distinctly mammalian tissues that enable the fetus to survive within the maternal uterus. There are four extraembryonic membranes associated with fetal development, namely the amnion, allantois, yolk sac, and chorion. The amnion and allantois arise from the epiblast, the yolk sac is derived from the PE, and the chorion is formed from the TE (Gilbert,

2000).

The amnion is the innermost membranous sac enclosing the embryo. The thin amniotic membrane comprises a single layer of extraembryonic ectodermal cells lined by a non-vascularized layer of extraembryonic mesoderm. The amniotic membrane surrounds the body of the embryo, creating a fluid-filled cavity known as the amniotic sac, in which the fetus is suspended for the duration of the pregnancy

(Fig. 1.2). The amniotic sac not only serves as a buffer against mechanical injury to the fetus, it also accommodates growth and allows fetal movement. Generally, amniotic fluid is thought to be a dilute transudate of maternal plasma, but the origins and exchange dynamics of amniotic fluid are complex and not completely understood

(Carlson, 2007). While the amniotic membrane itself secretes fluids, components of maternal serum are also able to pass through the amniotic membrane. As pregnancy advances, there are increasing contributions from fetal urine, filtration from maternal blood vessels, and possibly filtration from fetal vessels in the umbilical cord and chorionic plate.

8 Anatomically, while the amnion appears dorsal to the embryonic disk, which later gives rise to the organs and body of the fetus, the yolk sac is formed ventral to the embryo. A narrow stalk connects the yolk sac to the forming embryonic gut. As pregnancy progresses, the yolk stalk becomes very long and attenuated as it is incorporated into the body of the umbilical cord, while the yolk sac itself moves nearer the chorionic plate of the placenta (Fig. 1.2).

The allantois is also embedded in the umbilical cord. The mesodermal wall of the allantois gives rise to the blood vessels that form the umbilical circulatory arc, consisting of the arteries and veins that supply the chorion, or placenta. Later in development, the proximal part of the allantois becomes continuous with the forming urinary bladder.

The chorion too is connected to the embryo via the umbilical cord. It is the outermost protective membrane encapsulating the fetus, and harbors the trophoblast cells. These trophoblast cells eventually give rise to numerous, small, finger-like extensions called villi that penetrate deeply into the uterine tissue. Although the entire surface of the chorion is initially covered by villi, the villi across the cavity from the embryonic pole gradually degenerate, producing the smooth chorion (chorion laeve; Fig. 1.2). At the embryonic pole meanwhile, the villi increase in number and branch, forming the villous chorion (chorion frondosum), which becomes highly vascular. As described earlier, the fetal vascular compartment that comes in contact with the chorion arises from the allantois, and during the course of development, blood vessels grow out of the allantois into the chorion.

9 It is the chorionic villi which brings maternal and fetal blood into proximity, allowing the exchange of nutrients and waste between mother and fetus, yet preventing maternal and fetal blood from mixing. Thus, the chorion frondosum ultimately makes up the principal embryonic part of the placenta, the structure through which materials are exchanged between mother and fetus.

The placenta is the extraembryonic tissue essential for healthy development of the fetus during gestation; its malformation can trigger a spontaneous abortion (Trounson et al., 2000). The placenta anchors the developing embryo to the uterine wall, and is not only responsible for delivering oxygen and nutrition from the mother, but also to secrete hormones supporting fetal growth in the uterus, and to produce regulators of immune response that prevent rejection of the fetus.

The placenta is a vascular structure with feto-maternal compartments; the chorionic villi is fetal, and the maternal portion consists of the area of the endometrium called the decidua basalis, into which the villi penetrate. Implantation of the developing blastocyst in the maternal endometrium signifies the beginning of the formation of the placenta, requiring a cooperative effort between the extraembryonic tissues of the embryo and the endometrial tissues of the mother. While the placenta is morphologically established early on, maternal blood supply to the placenta does not start until the end of the first trimester of pregnancy (at approximately 12 to 13 weeks). As a result, the placenta forms and exists in an environment that is low in oxygen during the first trimester. Throughout the rest of pregnancy, the placenta continues to grow to keep up with the demands of the developing fetus.

10 Figure 1.2. Development of the extraembryonic membranes during early pregnancy. (A) The amnion appears dorsal to the embryo, while the yolk sac is formed ventral to the embryo. The trophoblast forms the beginnings of the chorion. (B) The trophoblasts give rise to villi. (C) The villi at the embryonic pole increase in number and branch. (D) The villi across from the embryonic pole atrophy, producing the smooth chorion, or chorion laeve. The chorion frondosum, or placental villi, make up the embryonic portion of the placenta. (E) As pregnancy progresses, the yolk stalk becomes very long and attenuated, while the yolk sac moves nearer the chorion. The yolk stalk becomes embedded in the umbilical cord. The fetal blood vessels grow out of the allantois, and is carried through the umbilical cord, to supply the chorion.

11 1.1.3. Implantation and placentation.

Approximately one week after fertilization, trophoblasts participate in a complex dialogue with maternal cells that enables implantation, a process that quickly sequesters the human embryo within the uterine wall. This begins with a phase known as apposition, where the blastocyst adheres to the uterine epithelium via TE cells (termed “polar”) that are immediately adjacent to the ICM at the embryonic pole of the blastocyst (Fig. 1.3A; Hertig et al., 1959). These “polar” cells comprise approximately one-third of the TE just prior to implantation, and are morphologically different from the remainder of the TE distal to the ICM (termed “mural” trophectoderm; Herzog, 1909). The polar TE cells, in the first hours of implantation, proliferate and invade, whereas mural TE cells do not (Fig. 1.3B; Enders, 1976).

In the following hours, the trophoblastic cells proliferate to form a double layer, as they progressively invade the endometrial epithelium. The inner layer consists of cytotrophoblasts, while the outer layer, in contact with maternal tissue, comprises of the syncytiotrophoblast, supported by proliferation and fusion of underlying cytotrophoblasts. The syncytiotrophoblast is a continuous system, not interrupted by intercellular spaces, that forms branching, finger-like extensions which deeply invade the endometrium. Micro-protrusions from the apical surface of the uterine epithelium, known as pinopodes, inter-digitate with microvilli on the apical surface of the syncytiotrophoblast (Lopata et al., 2002). This first stage, lasting from day 7 to day 8 post-fertilization, is defined as the prelacunar period (Enders, 1989).

The next stage is the lacunar or pre-villous stage, characterized by the formation of connected syncytial trophoblast-lined lacunae. Small vacuoles appear within the

12 syncytiotrophoblastic mass, and enlarge rapidly, becoming confluent and forming a system of lacunae. The separating lamellae and pillars of syncytiotrophoblast are called the trabeculae. This stage lasts from day 8 to day 13 post-fertilization (Fig,

1.3C and D). Meanwhile, by day 12, the blastocyst becomes completely embedded as the uterine epithelium regrows to cover the site of implantation (Hertig et al., 1956).

At about the same time, the cytotrophoblasts beneath the syncytial layer proliferate and stream out of the trophoblast layer (Fig. 1.3C; Pijnenborg et al., 1981a). They appear as trophoblast cell columns penetrating through the trabeculae, reaching the tip, where they extend through the syncytiotrophoblast and spread laterally, coalescing to form a cytrotrophoblastic shell layer, between the syncytium and the endometrium.

The cytotrophoblast cell columns grow into the lacunae, marking the initiation of the early villous stage. The cell columns then begin to branch, representing primary villi, which are composed of a cytotrophoblast core with a covering of syncytiotrophoblast.

Their presence marks the beginning of the villous stages of placentation. Further proliferative activity, with branching of the primary villi, initiates the development of primitive villous trees. By the third week of gestation, these villi contain fetal blood vessels and occasional macrophages surrounding a core of mesenchymal connective tissue (Hertig, 1968).

Villi that maintain contact with the trophoblastic shell are known as anchoring villi.

Meanwhile, the lacunar system transforms into the intervillous space.

Cytotrophoblasts at the base of the anchoring villi proliferate outwards from the underlying basement membrane, creating the path of migration into the uterine

13 epithelium and ultimately invasion of the maternal spiral arteries. Migrating cytotrophoblasts are referred to as extravillous trophoblasts (EVTs). Those that infiltrate the entire endometrium and inner third of the myometrium, through a process termed interstitial invasion, are known as interstitial EVTs. Endovascular trophoblasts are EVTs that penetrate the uterine vasculature, in a process known as endovascular invasion (Pijnenborg et al., 1981a; 1981b).

Endovascular invasion places trophoblasts in direct contact with maternal blood, establishing utero-placental circulation. While the basic structure of the definitive placenta is complete by the fourth week of gestation, communication with the terminal branches of the spiral arteries have not yet been established, and maternal blood remains excluded from the intervillous space (Boyd and Hamilton, 1970). At eight-weeks gestation, direct channels connecting the maternal spiral arteries with the intervillous space are observed, but the path of blood flow remains blocked by trophoblast plugs that occlude the spiral arteries (Hustin and Schaaps, 1987; Burton et al., 1999). Full blood flow into the intervillous space only takes place upon the removal of the plugs, between 10 to 12 weeks of gestation (Hustin and Schaaps,

1987; Jauniaux et al., 1992; Jauniaux et al., 1995; Jaffe et al., 1997; Burton et al.,

1999).

Therefore, during the first trimester, the placenta forms and develops in a low oxygen environment. At this time, the conceptus (or the early embryo), is also undergoing growth and organogenesis, and therefore derives its required nutrition from secretions of the uterine glands, in what is known as histiotrophic nutrition (Burton et al., 2002).

When the trophoblastic plugs are broken down, the sudden presence of blood

14 drastically modifies the uterine environment; the blood brings with it oxygenation, as well as hemotrophic nutrition, which is the exchange of blood-borne materials between the maternal and fetal circulations (Wooding and Flint, 1994).

15 Figure 1.3. Progression of implantation and invasion of the endometrial lining by trophoblasts. Images are from the Carnegie embryo collection hosted by The Virtual Human Embryo Project (http://www.ehd.org/virtual-human-embryo/). A-B) Sections of a rhesus monkey (Macaca mulatta) embryo (specimen #C 610) initiating implantation at Carnegie Stage 4. C-D) Sections through human embryos at Carnegie Stage 5a-1 (Panel C: Carnegie embryo #8020) and Carnegie Stage 5b (Panel D: Carnegie embryo #8004) representing approximately 0.5 and 2.0 days post-implantation, respectively.

16 1.1.4. Molecular basis of trophoblast formation.

Pre-implantation development encompasses a series of important developmental events, beginning with fertilization of the oocyte and ending with the formation of the blastocyst. The specification of the TE and ICM lineages during blastocyst formation represents the first developmental decision during human embryo development, paving the way for all subsequent developmental events. The earliest stages of trophoblast differentiation are critical for mediating implantation and fostering normal placental growth and function during gestation. Moreover, there is growing evidence that aberrant trophoblast development is associated with serious pregnancy complications, including recurrent miscarriages, preeclampsia, and restricted fetal growth. Therefore, a comprehensive understanding of pre-implantation development is fundamental.

In the mouse, identification of the determinants of trophoblast cell fate mainly stem from the analysis of targeted mutations (Rossant and Cross, 2001; Senner and

Hemberger, 2010). The initiation of TE development in mouse begins with Tead4, which controls expression of Cdx2, Eomes, and Tcfap2c. Further downstream, other transcription factors essential to normal TE development include Gata3, Elf5 and

Ets2. However, the relevance of these findings to human TE development are limited.

The similarities between mouse and human TE development end at the mutually exclusive expression of key transcription factors in ICM (OCT4, SOX2, and NANOG) and TE (CDX2; Kimber et al., 2008).

The molecular basis of human pre-implantation is not well known, owing to a scarcity of material, as well as ethical and legal limitations. Most available

17 knowledge is based on mouse (Hamatani et al., 2004; Wang et al., 2004) or bovine

(Ushizawa et al., 2004; Mamo et al., 2006), and limited data coming from non-human primates. Previously, characterization of early pre/post-implantation developmental stages of the human placental trophoblast lineage were mainly based on morphological studies (Hertig et al., 1956; Enders, 1989). Recently however, technological advances have permitted the exploration of global gene expression profiles from limited amounts of material, leading to renewed interest and exponential growth of information within the field (Zhang et al., 2009; Vassena et al.,

2011; Bai et al., 2012).

Generally, pre-implantation development consists of two main phases, governed by either maternal or embryonic genes (Zhang et al., 2009; Vassena et al., 2011). The timing of the switch from maternal to embryonic gene expression remains unclear.

Vassena et al. (2011) demonstrated the first wave of transcriptional activation of the embryonic genome at the 2-cell stage, while Zhang et al. (2009) reported expression of mainly maternal genes up to the 4-cell stage, and from 8-cell embryo to blastocyst, maternal genes are downregulated while embryonic genes are upregulated.

The trophoblast cell lineage is specified early in the blastocyst stage, leading to the emergence of the TE surrounding the ICM. Expression microarray analysis of mechanically dissected human TE samples revealed a core transcriptional regulatory circuitry of 13 tightly interconnected transcription factors that are expressed during

TE development, and maintained later in mature placenta (Bai et al., 2012). These include CEBPA, GATA2, GATA3, GCM1, KLF5, MAFK, MSX2, MXD1, PPARD,

PPARG, PPP1R13L, TFAP2C and TP63. The expression profile of these factors

18 suggests a critical role in not only establishing, but also maintaining the trophoblast lineage.

Moreover, some of these transcription factors have already been reported to be important for trophoblast differentiation or placenta formation. For instance, GCM1 is a known placenta-specific transcription factor, responsible for inducing ERVWE1 expression in trophoblast, which cause cell fusion and contribute to the formation of the syncytium (Yamada et al., 2000; Rawn and Cross, 2008; Black et al., 2010).

GATA2 and GATA3 are known to be expressed in mouse trophoblast, and in placenta, for regulation of placenta-specific genes (Ma et al., 1997). In the mouse, GATA3 is selectively expressed in the TE of peri-implantation embryo and directly regulates expression of trophoblast transcription factor, CDX2 (Home et al., 2009). PPARG, a nuclear hormone receptor implicated in the differentiation of various cell types, is essential for trophoblast and placental differentiation in mice, as evidenced by embryonic lethality in PPARG-null mice, due to defects in these tissues (Barak et al.,

1999).

Bai et al. (2012) also demonstrated the TE-associated expression of genes known for their role in placenta, namely CGA, PGF, ALPPL2, and ABCG2. An earlier study had also confirmed the upregulation of TE-associated genes CDX2 and GATA2, as well as pregnancy-associated hormone human chorionic gonadotropin (hCG; encoded by genes CGA and CGB) in human blastocysts (Kimber et al., 2008). Studies of human embryonic stem cell differentiation also demonstrate that BMP4, GATA2, GCM1 and

ID2, as well as deficiencies in the polycomb group protein L3MBTL1, are all

19 associated with differentiation into a TE phenotype (Schulz et al., 2008; Hoya-Arias et al., 2011).

Trophoblast invasion and migration, along with formation of the corresponding trophoblast cell subtype, are likely controlled by components of the trophoblast itself as well as exogenously by the maternal microenvironment, through molecular and cellular interactions. Despite these advances in understanding the genetic regulation of the TE lineage, the exogenous regulation of this differentiation process has not been well defined. Studies from the mouse have identified a number of exogenous factors with the potential to regulate TE lineage formation. Collagen IV has been shown to induce TE differentiation from mouse embryonic stem cells, which can be further increased by stimulation with basic fibroblast growth factor (bFGF) or FGF4

(Schenke-Layland et al., 2007). In humans, numerous factors in the extracellular microenvironment are thought to be capable of directing TE differentiation, such as soluble factors - TGF, platelet derived growth factor (PDGF), leukemia inhibitory factor (LIF), macrophage colony stimulating factor (M-CSF), and the chemokines

CX3CL1, CCL8, and CCL14. Environmental factors include cell–cell contacts, extracellular matrix (ECM) composition, oxygen tension, as well as other environmental stressors (Suzuki et al., 1999; Hannan and Salamonsen, 2008;

Dimitriadis et al., 2010). Indeed, many of these factors have been demonstrated to play an important role in trophoblast differentiation later in the first trimester of human pregnancy (Bischof et al., 2000).

Whilst the knowledge of the events underlying cell fate determination in the human blastocyst is limited compared to the corresponding events in mouse implantation,

20 even less is known about the events that occur after the initial commitment to the TE.

Trophoblast lineage formation in the mouse has been mapped from TE to placenta, due in large part to the readily available blastocysts and trophoblast stem cells.

However, this information gleaned from mouse is not entirely transferable to humans as the labyrinthine structure of the murine placenta is quite unlike the villous structure of the human placenta. Additionally, maternal exchange in mouse occurs across three trophoblast layers, as opposed to the single syncytiotrophoblast layer in humans (Carter, 2007).

It is evident that there is much yet to be understood about human implantation during early development. The time between formation of the blastocyst and formation of the placenta still remains largely unknown given the limitations of accessing material for study. The progression of human trophoblast cell types, arising from TE during the earliest days of pregnancy, to the mature trophoblast cell types that constitute the placenta and exist throughout pregnancy still remain largely a “black box” in our understanding of early development in humans. Unlike the blastocyst itself, which is usually received from IVF clinics as donations to research, the development of the trophoblast and its lineages that follow is mandated by its biological environment and ability to invade and migrate into the uterine epithelium. Moreover, in the earliest weeks that follow conception, most women are barely even aware of the forming blastocyst and their pregnancy. As a result, the earliest that human trophoblast cells are normally accessed is at six weeks post-conception. To that end, much work has been focused on identifying human trophoblast cell types, and growing these cells in the laboratory, in order to understand the development of the placenta.

21 1.1.5. Identifying trophoblast cell types.

The human TE differentiates to give rise to the main cell lineages of the villous placenta - the cytotrophoblast, syncytiotrophoblast and extravillous trophoblast (EVT;

Fig. 1.4). As ongoing studies pursue the understanding of how these lineages develop, what factors stimulate and regulate this differentiation, as well as when differentiation occurs, studies to date have been able to isolate trophoblast cell populations and provide markers that identify the trophoblast cell subtypes.

The human trophoblast stem cell (hTSC) however remains elusive, and hence so does its understanding and characterization. Isolation and characterization of the hTSC will not only expound the differentiation of the human trophectoderm into trophoblast, but will also enable the generation of differentiated trophoblast populations in vitro; thereby allowing a more accurate and detailed study into what separates one trophoblast population from another (Hyman and Simons, 2011). On the other hand, the stem cell population of the TE lineage has been isolated from mouse blastocysts and maintained in vitro in the presence of FGF4, heparin and a feeder layer of mouse embryonic fibroblast cells (Tanaka et al., 1998). These cells are found to be immortal in culture and have the potential to differentiate into multiple mouse trophoblast cell lineages. However, whilst hESCs can be differentiated to TE

(see 1.1.6. Modeling human placentation), to date, efforts to isolate true hTSCs from human embryos have not yielded an in vitro trophoblast stem cell line comparable to those from mice, likely due to differing growth factor requirements and regulatory pathways (Douglas et al., 2009).

22 While cytotrophoblasts undergo self-renewal and have the potential to differentiate down one of two mutually exclusive differentiation pathways, either the non-invasive villous syncytiotrophoblast or the invasive extravillous cytotrophoblast (EVT), they are not considered to be the true hTSC. Previous studies have suggested that by the eighth week of gestation, cytotrophoblasts are divided into subpopulations that may be committed to differentiation into either the syncytiotrophoblast or the EVT

(Aboagye-Mathiesen et al., 1996; James et al., 2005; James et al., 2007). This pre- commitment thus casts a doubt on their previously recognized bi-potent character.

Nevertheless, cytotrophoblasts have been recognized as specialized cells that actively invade the uterus, forming the infrastructure that anchors the conceptus to the maternal endometrium and taps the supply of maternal blood. This requires the cytotrophoblast to assume highly specialized characteristics; some commonly associated with tumor cells, while others are typical of endothelial cells. Expression at the mRNA level revealed upregulated levels of EPAS1, a gene primarily expressed in endothelial cells and functions in adaptive responses to hypoxia such as in cancer

(Janatpour et al., 1999).

Interestingly, the isolation of trophoblasts with stem cell-like properties from first trimester placentae has been recently reported (Spitalieri et al., 2009; Genbacev et al.,

2011). Spitalieri et al. (2009) claimed to have identified a sub-population of cytotrophoblasts in chorionic villus samples that expressed typical stem cell markers, and were capable of differentiating into multiple lineages in vitro as well as in vivo.

OCT4, SSEA4, and alkaline phosphatase, all of which are indicators for pluripotency, were expressed in approximately 20% of the cytotrophoblast population, with

NANOG and SOX2 expression observed in a further subset of these cells. However,

23 the cells isolated in this study could be induced to differentiate into cells with features of neuronal, myogenic or pancreatic cells, suggesting they were not true trophoblast stem cells.

In a separate study, Genbacev et al. (2011) identified human trophoblast progenitors that predominantly reside in the chorion of first trimester placenta. In isolated colonies, a proportion of these cells expressed markers of trophoblast progenitor cells

- cytokeratin 7 (CK7), OCT4, tight junction protein ZO-1, and GATA4. Away from the original attachment site, these cells expressed markers of trophoblast differentiation, like GCM1 (syncytiotrophoblast) and HLA-G (EVT). Single cell suspensions isolated from these colonies lost polarity, and did not differentiate in culture, suggesting that pure populations of trophoblast progenitors had been isolated.

These progenitor cells could also be induced to differentiate into cytotrophoblast and syncytiotrophoblast lineages, suggesting they could indeed be true trophoblast stem cells. However, in order to confirm that these cells are true trophoblast stem cells, it would be useful to further characterize for common markers of stem cells, as well as those for human trophectoderm and the human equivalents to markers of mouse trophoblast stem cells (James et al., 2012).

From second trimester onwards, villous trophoblasts are in contact with maternal blood, and mediates nutrient exchange, while EVTs are responsible for invasion of the decidua and remodeling of the uterine arteries. In an earlier study which co- related trophoblast cell type with ability to invade (villous trophoblasts with poor invasiveness, and EVTs with high invasiveness), microarray analysis yielded 991 invasion/migration related transcripts (Bilban et al., 2009). EVTs were found to be

24 significantly enriched in pathways regulating ECM synthesis and remodeling, cell adhesion, as well as carbohydrate metabolism. Several genes involved in physiological and pathological cell invasion, including A disintegrin and metalloprotease-12, -19, -28, as well as spondin-2, were upregulated in EVTs. EVTs also demonstrated enhanced expression of several collagen or laminin isoforms, the predominant matrix proteins in the endometrium (Das et al., 2002), as well as several isoforms of lactate dehydrogenase, aldehyde dehydrogenase, and phosphofructokinase, all of which are enzymes regulating carbohydrate metabolism.

Altogether, these observations suggest that EVTs may secrete substrates relevant to cell invasion as well as their degrading enzymes.

Additionally, in a more recent whole genome microarray study examining transcriptional changes associated with differentiation from villous to EVT (Apps et al., 2011), EVTs were found to be associated with genes involved in migration (ECM proteases such as PAPPA, ADAM-19, laeverin, TIMP2, and SERPINE2), immune modulation (HLA-G, TAP1 and -2, CFB, PVR), and cytokine or angiogenic factor secretion (CTGF, PGF, FGF1, GKN1, VGF, and CXCL16, and its receptor CXCR6).

Cytokine/chemokine receptors such as CCR1, IL1R2, IL1RAP, IL2RB, ACVRL1 and

ACVR1B are all increased, suggesting EVT acquires an enhanced ability to respond to the wide range of cytokines secreted in the decidua. Adhesion molecules regulating cell migration were also upregulated, including integrins α-1, -2 and -5, as well as

ESAM, and JAM2, both of which are involved in regulating the adhesion of leukocytes to endothelium. Their up-regulation by EVT suggests trophoblast may utilize these molecules to interact with decidual leukocytes during invasion

(Wegmann et al., 2006).

25 Meanwhile, villous trophoblasts (or cytotrophoblast) were significantly enriched in genes implicated in stress response, signal transduction, and gene regulation, among others (Bilban et al., 2009), which is unsurprising given their vulnerability to maternal blood supply. Genes regulating response to heavy metals, oxygen species, and DNA damaging agents, including SOD2, HO-1, and several heat shock proteins were upregulated in these cells. Furthermore, several signal transduction pathways are modulated, including glucocorticoid-, cAMP/Ca2+-, and G protein-coupled receptor signaling. Interestingly, there was also enhanced expression of a large group of gene sets regulating the nuclear import/export and translation factor activity, which may be tied to the potential of the cytotrophoblast to progress down the differentiation pathway. Additionally, in demonstration of the validity of their findings, Bilban et al. (2009) showed an inverse relationship of HO-1 expression with trophoblast migration, acting via upregulation of PPARG.

The other study by Apps et al. (2011) reported upregulation of human endogenous retroviruses HERV-FRD and ERVWE1, as well as transcription factor ELF5, in villous trophoblasts, indicating its potential to generate the syncytiotrophoblast cell type.

Overall, these studies cumulatively present a yet unfinished picture of trophoblast cell type markers. While all-encompassing, the contents of these preliminary lists must be validated before a definitive list can be issued as the gold standard for identifying the specific trophoblast cell type. Nevertheless, these studies provide a good starting point for the search of defining trophoblast traits.

26 Figure 1.4. Trophoblast lineage cell types and proposed markers. Formation of the trophectoderm (TE) marks the first differentiation event in the development of the embryo. The self-renewing human trophoblast stem cells (hTSCs) have yet to be definitively isolated and characterized. Several genes identified in mouse models to have limited expression in the TE and found to be essential for proper trophoblast formation are proposed for the identification of hTSCs (adapted from James et al., 2012).

27 1.1.6. Modeling human placentation.

The human implantation process is unique, as exemplified by the placenta representing the most structurally and morphologically diverse organ found in mammals (Mossman, 1991). As such, animal models are subpar estimates at best.

While knock-out studies in mice have provided insight on the possible roles of various factors in the implantation process, differences in reproductive physiology between species limit the relevance. Non-human primates, representing our closest cousins on the evolutionary scale, provide a more physiologically relevant model, thus in vivo and in vitro studies in baboons have contributed to the current understanding.

Ethical restrictions and limited availability of human placental tissue prohibit human implantation studies. Where practical and ethical considerations prevent the use of human blastocysts, cells that closely represent the trophectoderm have been used as a model in its stead. Studies have shown that hESCs can be induced to differentiate into the trophoblast lineage via three different mechanisms: 1) The transgenic knockdown of pluripotency genes OCT4, NANOG, and SOX2 result in the expression of trophoblast markers (Matin et al., 2004; Babaie et al., 2007); 2) Spontaneously differentiating trophoblasts have been collected from embryoid bodies formed in the absence of growth factors that support pluripotency. The trophoblasts are purified from these embryoid bodies either by clonal selection for hCG production (Harun et al., 2006) or by their adhesive properties (Peiffer et al., 2007); 3) Feeder-free hESCs can be cultured in the presence of BMP4, leading to an upregulation of transcription factors linked to trophoblast differentiation. These cultures are able to give rise to

28 syncytiotrophoblast-like cells when dissociated and re-plated at low density, forming multinucleated hCG positive patches of cells (Xu et al., 2002). Regardless of their method of differentiation, the apparent mix of villous (hCG-expressing) and extravillous (HLA-G-expressing) cell types derived from hESC colonies suggest that multipotent trophoblast progenitors or individual progenitors of both trophoblast cell types arise during differentiation (Xu et al., 2002; Udayashankar et al., 2011).

A range of trophoblast cell lines is also widely available. These lines are derived from differentiated trophoblast, or from highly invasive choriocarcinomas, and therefore are likely to have key phenotypic differences from TE and the primitive trophoblast.

While terminally differentiated trophoblasts are not able to illustrate the transition from hTSC to terminally differentiated cell type, EVTs provide a look at an intermediate state due to their functional similarity to the implanting TE, denoted by their invasion through and direct interaction with decidual stromal cells. Some studies have used EVT cell lines to model the cell-cell interactions that occur in the initial stages of blastocyst implantation (Genbacev et al., 2003; Hannan and Salamonsen,

2008; Stoikos et al., 2010; Robins et al., 2011). However, it has yet to be established how closely these differentiated trophoblasts represent the implanting polar TE and primitive cytotrophoblast (James et al., 2012). The ability to pick an appropriate trophoblast cell line is confounded by the limited understanding of the adhesion molecule and cytokine repertoires of the implanting TE, as well as the time line of differentiation into various trophoblast lineages from the TE.

Lastly, in an attempt to further improve the accuracy of models used to study early placentation, trophoblast cell lines have been maintained in three-dimensional

29 cultures, successfully generating embryo-free trophoblast vesicles (Robins et al.,

2011). This exciting development may well hold great importance, as TE lineage differentiation is dependent upon the position of each trophoblast cell with respect to the ICM, as seen in both mouse and primate models (Dickson, 1963; Enders, 1976;

Maltepe et al., 2010).

1.1.7. Preeclampsia: A multi-factorial pregnancy complication.

The need to understand human placental development is not without purpose.

Without comprehensive fundamental understanding of the progression of the early trophoblasts of the trophectoderm to ultimately, the placenta, the pathogenesis of many pregnancy complications remain puzzling. Pregnancy complications are commonly associated with placental dysfunction, some even resulting in early termination of the pregnancy. These complications not only affect the viability and health of the fetus, they also endanger the mother’s life.

Preeclampsia is one of the most common and dangerous pregnancy-disorder, and represents the principal cause of maternal morbidity, accounting for 15-20% of pregnancy-related mortalities (Sibai et al., 2005). Preeclampsia is diagnosed when the mother, more than 20 weeks along in the pregnancy, develops both hypertension and proteinuria, which is characterized by significant amounts of protein in the uterine. It is a complex disorder involving multiple organ failure including that of the liver, kidney, lung, and the coagulatory and neural systems. If left untreated, it can develop into eclampsia, the life-threatening occurrence of seizures during pregnancy.

The prognosis for both the mother and fetus in severe preeclampsia is particularly poor in early onset cases (<34 weeks of gestation; Hall et al., 2000). Patients are

30 generally treated for prevention of seizures and control of hypertension, but the lack of understanding of the precise etiology of the condition has hindered the development of preventive and therapeutic measures to effectively treat this disease based on its etiology (Nishizawa et al., 2011). The consensus now is that preeclampsia is a complex polygenetic disorder in which maternal and fetal genes, as well as environmental factors, are involved. While it is recognized that its pathogenesis involves numerous processes, such as oxidative stress, endothelial dysfunction, vasoconstriction, metabolic changes, thrombotic disorders and inflammatory responses, the precise underlying mechanisms have remained elusive

(Roberts and Cooper, 2001; Nishizawa et al., 2011).

Preeclampsia is considered a two-stage disease (Redman and Sargent, 2005; Roberts and Hubel, 2009). The first stage is characterized by inadequate placentation, resulting from shallow invasion of the trophoblasts into the maternal spiral arteries, underscoring the need to understand trophoblast progression in early pregnancy.

Instead of the low-resistance, high-flow system seen in the normal placenta, the preeclamptic placenta is associated with increased resistance with accompanying decreased blood flow, and uneven perfusion, leading to hypoxia, ischemia, and reperfusion injuries in the placenta (Redman and Sargent, 2005; Roberts and Hubel,

2009; Centlow et al., 2011). The second stage of preeclampsia is characterized by general vascular endothelial damage, which eventually affects all maternal organs.

The transition from the first stage to the second is typically seen before 35 weeks of gestation, in early onset preeclampsia. Placentae from early onset preeclampsia more often show histological pathological findings associated with poor placental perfusion, such as infarcts and chronic inflammation (Salafia et al., 1998). In contrast,

31 those from late-onset preeclampsia rarely present these pathological findings.

Unfortunately, placental factors linking the two stages are not well understood.

The placenta is implicated as the primary player in the etiology of this disorder.

Moreover, patients with preeclampsia recover from the associated clinical symptoms after delivery of the placenta. The gene expression profile of the placenta is also consistently altered in preeclampsia, as reported by multiple microarray studies

(Enquobahrie et al., 2008; Founds et al., 2009; Sitras et al., 2009; Centlow et al.,

2011; Meng et al., 2012). While substantial progress has been made in identifying genes involved in the pathogenesis of preeclampsia, there is still considerable debate on the significance of the role of the candidate genes (Chappell and Morgan, 2006).

Additionally, placental gene expression reflects feto–maternal interaction, and could potentially contribute to the maternal phenotype (Oudejans and Dijk, 2008). To that end, several placental factors that may participate in the progression to the second stage have been identified, including soluble fms-like tyrosine kinase 1 (sFlt1) and endoglin (Maynard et al., 2003; Venkatesha et al., 2006).

Preeclampsia has been strikingly associated with a fetal disorder known as intrauterine growth restriction (IUGR), commonly described as poor growth of the fetus while in the uterus. IUGR is a result of heterogeneous causes, including maternal, fetal, and placental factors (Bernstein and Divon, 1997), and has an increased risk of perinatal morbidity and mortality (Bernstein and Gabbe, 1996).

IUGR may be caused by maternal malnutrition or fetal chromosomal abnormalities, which lead to inhibited growth, but it can also strike with an unknown cause. Reviews of current literature indicate that preeclampsia and unexplained IUGR have a

32 common etiological background (Gerretsen et al., 1981; Bernstein and Divon, 1997;

Aardema et al., 2001; Kaufman et al., 2003; Baschat and Hecher, 2004), although this assertion has remained controversial (Villar et al., 2006; Burton et al., 2009). In any case, little is known about their involvement at the molecular level. Nonetheless, a recent microarray study by Nishizawa and colleagues (2011) showed strong correlation between IUGR and preeclampsia; many genes that are known to be upregulated in preeclampsia are also upregulated in IUGR, including anti-angiogenic factors. Moreover, a failure of trophoblast invasion leading to abnormal shallow vascularization and impaired remodeling of the spiral arteries appears to be common ground for these two disorders (Kaufman et al., 2003). It has been posited that such placental defects, as seen in preeclampsia, may thus inhibit the materno-fetal transport of nutrients and oxygen.

33 1.2. EPAS1 and the HIF family.

The EPAS1 gene encodes for the human endothelial PAS-domain containing protein, better known as HIF2α. It is a member of the hypoxia-inducible factor (HIF) family of proteins. This family of proteins is responsive to changes in oxygen levels in the cellular environment.

Oxygen is essential for the survival of all aerobic organisms and is required for aerobic metabolism in order to maintain intracellular energy balance. Aerobic energy metabolism is dependent on oxidative phosphorylation, through which the oxido- reduction energy of the mitochondrial electron transport is converted into the high- energy phosphate bond of ATP. In this process, oxygen serves as the final electron acceptor.

Hypoxia has conventionally been defined as a state of reduced oxygen level, one that is below normal values (Table 1.1A; Koh and Powis, 2012). Since molecular oxygen is essential for the development and growth of multicellular organisms, the ability to sense and respond to low oxygen conditions is critical to survival. When oxygen tensions fall below the normal physiological range, the function of organs, tissues, or cells are restricted, a condition referred to as pathological hypoxia. Pathological hypoxia may occur at high altitude, or due to localized ischemia as a result of disrupted blood flow. On the other hand, the hypoxic environment or hypoxia in first trimester placenta, is normal, and not considered pathological. Nonetheless, physiological oxygen tensions are actually significantly lower than ambient oxygen tensions as a result of the dramatic decrease in blood oxygen content as blood travels throughout the body’s circulation (Table 1.1B; Carreau et al., 2011). Oxygen

34 gradients have many important and beneficial roles in mammalian physiology; hypoxia provides the required extracellular stimulus for proper embryogenesis and wound healing, and helps maintain pluripotency of stem cells (Ezashi et al., 2005;

Dunwoodie et al., 2009; Patterson and Zhang, 2010). Numerous adult stem cell types are in fact maintained in their niche environments at oxygen concentrations between

<1% and 8% (Fig. 1.5; Mohyeldin et al., 2010).

Viewed from a molecular perspective, physiological hypoxia illicits a conserved molecular response. Cells utilize a number of oxygen-sensing pathways, including those that involve the energy and nutrient sensor known as mammalian target of rapamycin (mTOR), the endoplasmic reticulum stress response, the nuclear factor

(NF)-κB transcriptional response, as well as the hypoxia-inducible factor (HIF)- mediated transcriptional response (Lai et al., 2007; Perkins, 2007; Wouters and

Koritzinsky, 2008; Koh and Powis, 2012). These mechanisms promote hypoxia tolerance by activating transcription and inhibiting translation. The HIF-mediated response in particular switches on a large number of genes including those promoting angiogenesis, anaerobic metabolism, and resistance to apoptosis (reviewed by Kaluz et al., 2008; Yoon et al., 2011; Koh and Powis, 2012).

35 Table 1.1 (A) Frequently used definitions describing low oxygen. Ambient oxygen level, termed normoxia, is defined as 20% oxygen concentration. Because physiological tissue oxygen tensions are lower than ambient oxygen tension, this condition is referred to as physiological hypoxia. Extreme hypoxia is known as anoxia, with <0.1% oxygen concentration. (B) Physiological oxygen tensions in a selection of normal human tissue types. Oxygen tension in physiology is lowered as a result of the dramatic decrease in blood oxygen content as it travels throughout the bodyʼs circulation. As a result, multiple tissue types exist and function under physiological hypoxia.

(A) Condition pO2 %O2 (B) Organ Normal pO2 %O2 (mmHg) (mmHg)

Normoxia (ambient) 159 21 Trachea 150 19.7

Physiological 15-68 2-9 Alveoli 110 14.5 hypoxia

Mild hypoxia 8-38 1-5 Arterial blood 100 13.2

Hypoxia <8 <1 Pulmonary arterial blood 40 5.3

Anoxia <0.08 <0.1 Brain 35 4.4

Intestinal tissue 58 7.6

Liver 41 5.4

Kidney 72 9.5

Muscle 29 3.8

Bone marrow 49 6.4

36 Figure 1.5. Oxygen tension measurements in various stem cell niches. Adult stem cell types that are maintained in their niche environments at physiological hypoxia include the mesenchymal stem cell, neural stem cell, as well as the hematopoietic stem cell. Schematic models are depicted based on current available data for the adipose tissue, subventricular zone, and the bone marrow (adapted from Mohyeldin et al., 2010).

37 1.2.1. The canonical HIF pathway: Oxygen-dependent regulation.

The hypoxia-inducible factor (HIF) is a DNA-binding transcription factor that associates with specific nuclear cofactors under hypoxia to transactivate a myriad of genes, triggering various adaptive responses to inadequate oxygen tension. The HIF complex forms through the interaction and heterodimerization of the oxygen-labile

HIF-α subunit and the constitutively expressed HIF-β/ARNT subunit, the aryl hydrocarbon receptor nuclear translocator. Both subunits are part of the basic helix– loop–helix PER-ARNT-SIM (bHLH-PAS) family of transcription factors. Three genes encoding distinct HIF-α isoforms exist in humans: HIF1A, encoding HIF1α;

EPAS1, encoding HIF2α; and HIF3A, which is expressed as multiple HIF3α splice variants (Table 1.2A; Wang et al., 1995; Makino et al., 2002; Maynard et al., 2003).

All HIF-α isoforms possess one oxygen-dependent degradation domain (ODD), as well as one N-terminal transactivation domain (TAD), but only HIF1α and -2α possess an additional transactivation domain at the C-terminal. HIF1α dimerizes with

HIF-β/ARNT forming the HIF-1 transcription factor, and HIF-2 is formed similarly via the association of HIF2α and HIF-β/ARNT. While both HIF-1 and HIF-2 activate transcription, HIF3α has been found to negatively regulate HIF-1 and HIF-2 activity.

Two of the HIF3α isoforms achieve this indirectly by competing for available HIF-β/

ARNT and then only weakly activating transcription, while a third HIF3α isoform inhibits HIF-1 and -2 activity by binding HIF1α and -2α, therefore preventing their heterodimerization with HIF-β/ARNT (reviewed by Greer et al., 2012; Table 1.2B).

Under normal oxygen tension (>5%), prolyl-hydroxylase domain proteins (PHDs) hydroxylate two proline residues within the ODD of HIF-α (Fig. 1.6; Epstein et al.,

38 2001; Ivan et al., 2001; Jaakkola et al., 2001; Masson et al.; 2001). This hydroxylation is recognized by the von Hippel-Lindau tumour suppressor protein

(pVHL) and is subsequently ubiquitylated by the Elongin BC/Cul2/pVHL ubiquitin– ligase complex, thereby marking HIF-α for degradation by the 26S proteasome

(Maxwell et al., 1999; Ohh et al., 2000; Ruas and Poellinger, 2005; Schofield and

Ratcliffe, 2005). PHDs require oxygen for catalytic activity; thus, prolyl- hydroxylation of HIF-α is abrogated under hypoxia (<5% oxygen), allowing HIF-α to escape recognition by the pVHL ubiquitin–ligase complex and accumulate in the nucleus. Similarly, biallelic inactivation of VHL (VHL−/−) has a profound stabilizing effect on HIF-α protein levels, thereby implicating pVHL as a predominant HIF-α antagonist. Additionally, in the presence of oxygen, another hydroxylase-domain protein termed factor inhibiting HIF (FIH) participates in the negative regulation of

HIF-α by hydroxylating an asparagine residue in the C-terminal activation domain, sterically inhibiting interactions between HIF-α and transcriptional co-activators (Fig.

1.6; Arany et al., 1996; Lando et al., 2002; Lisy and Peet, 2008).

39 Table 1.2. (A) Members of the HIF family. There are 3 HIF-α and 1 HIF-β proteins in the HIF family. (B) Associations as well as transcriptional activity of the HIF complexes. The α subunits heterodimerize with the β subunit (ARNT) to form the respective HIF transcription factor complex. HIF-1 and HIF-2 have 2 TADs each, while HIF-3 and HIF-3NEPAS only have 1. IPAS does not activate transcription, but instead prevents transcription by binding HIF1α and -2α thus preventing their dimerization with ARNT (adapted from Dunwoodie, 2009).

A Member Gene Protein

Hif1α HIF1A hypoxia-inducible factor 1, alpha subunit

Hif2α EPAS1 endothelial PAS domain protein 1

Hif3α HIF3A hypoxia-inducible factor 3, alpha subunit

Hif1β ARNT aryl-hydrocarbon receptor nuclear translocator

40 Figure 1.6. HIF regulation under aerobic and hypoxic conditions. Under normoxic conditions, Hif-α is hydroxylated by PHD1-3 and FIH, which targets it for degradation and inhibits its transcriptional activity. In hypoxic conditions, Hif-α heterodimerizes with Hif-β/ ARNT, and binds CBP/p300 to form the HIF transcriptional factor complex; which translocates to the nucleus and binds HREs to activate transcription of target genes (adapted from Dunwoodie, 2009).

41 1.2.2. HIF target genes.

In hypoxic conditions, stable HIF-α dimerizes with HIF-β/ARNT before binding to hypoxia-responsive elements (HREs) comprising a core 5′-[A/G]CGTG-3′ consensus sequence and highly variable flanking sequences in the promoters of HIF-responsive genes (Wenger et al., 2005). FIH-mediated hydroxylation is also reduced when oxygen is depleted, thereby allowing HIFα to interact with the transcriptional co- activators p300/Creb-binding protein (Fig. 1.6; Arany et al., 1996). This transcriptional complex transactivates a specific subset of genes that regulate the cellular adaptive response to hypoxia, including, but not limited to, SLC2A1 (solute carrier family 2, facilitated glucose transporter membrane 1), VEGF (vascular endothelial growth factor), and EPO (erythropoietin; Wang and Semenza, 1993;

Forsythe et al., 1996). SLC2A1 encodes glucose transporter 1 (GLUT1), a transporter protein facilitating the movement of glucose across the plasma membrane in mammalian cells. VEGF is a signaling protein produced by cells that stimulates vasculogenesis and angiogenesis, while EPO is a glycoprotein hormone that controls erythropoiesis, or red blood cell production. These genes, as well as others activated by HIF, are involved in energy metabolism, autophagy, translation inhibition, erythropoiesis and angiogenesis, all of which help promote a tolerance of hypoxia by decreasing the cellular requirement for oxygen and increasing the supply of oxygen to tissues (Kaelin and Ratcliffe, 2008; Anderson et al., 2009).

Historically, HIF target genes have been identified on the basis of one or more of the following strategies: 1) Identification of a functional HRE containing a HIF-binding sequence (HBS); 2) Comparison of patterns of gene expression in HIF-α wild-type

42 and null cells (through genetic knockout or siRNA knockdown); 3) Screening for increased gene expression using VHLnull cells or cells transfected with a HIF-α expression vector (Kaluz et al., 2008). Expression profiling experiments indicated the heterogeneity of gene subsets induced by hypoxia; tumor cell lines as well as normal stromal and epithelial cells display distinct hypoxic profiles (Semenza, 2003). The consensus from various studies is that there is a core set of genes consistently induced by hypoxia, and other variable genes that exhibit cell-type specific induction. This conclusion underscores the importance of studying hypoxic gene expression in a cell- type-dependent context. Hypoxia induces HIF activity in almost all cell types and therefore, HIF alone cannot account for the cell-type specific gene expression.

Instead, the cell-type specific induction by hypoxia appears to be determined by functional interactions of HIF with other type-specific transcription factors (Semenza,

2003; Wenger et al., 2005).

Estimates of the total number of HIF target genes induced by hypoxia, in one or more cell types, vary from more than two hundred (Wenger et al., 2005) to as many as 5% of all human genes (Semenza, 2003). However, not all of them may be regulated directly through HREs, and other transcription factors have been implicated in hypoxia-inducible gene expression as well. HIF-activated transcription factors such as DEC-1 and -2, and ETS-1 also contribute to the pool of hypoxia-induced genes

(Oikawa et al., 2001; Miyazaki et al., 2002). In general, HIF induces expression of proteins that specifically help meet the metabolic and survival needs of hypoxic cells.

Table 1.3 lists representative HIF target genes, grouped according to function

(adapted from Kaluz et al., 2008).

43 Table 1.3. Selected HIF target genes grouped according to their function. HIF induces expression of proteins that specifically help meet metabolic and survival needs of hypoxic cells (adapted from Kaluz et al., 2008).

44 1.2.3. HIF1α vs. HIF2α: Are they redundant?

The existence of three HIF-α isoforms raises questions about the role of each individual isoform in regulation of transcription under hypoxic conditions. While the

α subunits are highly conserved at the protein level, domain structure, and hypoxia- dependent mechanisms of regulation, distinct patterns of cellular expression appear to be responsible for the respective physiological roles of HIF1α and HIF2α (Huang and

Bunn, 2003). For example, both isoforms are abundantly expressed in the kidney but in different cell types; HIF1α is predominantly expressed in epithelial cells whereas

HIF2α is predominantly detected in interstitial fibroblast and endothelial cells

(Rosenberger et al., 2002).

Furthermore, studies have found differences in the phenotypes of HIF1αnull mice and

HIF2αnull mice strains, both of which are embryonic lethal. HIF1αnull mice suffered from cardiac malformations, severe vascular defects, and impaired erythropoiesis, and did not survive beyond embryonic day 11 (Iyer et al., 1998; Kotch et al., 1999;

Yoon et al., 2011). HIF2αnull embryos mostly die by embryonic day 13.5 due to vascular defects (Peng et al., 2000), or bradycardia as a result of deficient catecholamine production (Tian et al., 1998). Some embryos survived gestation and died either as neonates due to impaired lung maturation (Compernolle et al., 2002), or several months after birth due to multi-organ failure mediated by the reactive oxygen species (Scortegagna et al., 2003). This suggests that HIF1α and 2α are functionally non-redundant and are unable to functionally complement each other.

Moreover, several studies have suggested that HIF1α and HIF2α differ in their capability to transactivate hypoxia-inducible genes. While both isoforms regulate the

45 expression of many common genes, HIF1α is found to preferentially activate genes involved in the glycolytic pathway (Hu et al., 2003; Wang et al., 2005), whereas

HIF2α is involved in the regulation of genes important for tumor growth, cell cycle progression, and maintaining stem cell pluripotency, like the proto-oncogene C-MYC

(Gordan et al., 2007) and the stem cell factor OCT4 (Covello et al., 2006). The VEGF receptor 2 gene known as FLK-1 has also been reported to be specifically regulated by HIF2α, although comparison of the FLK-1 HBS with HBSs of HIF1α target genes revealed no obvious difference (Elvert et al., 2003). Intriguingly, while HIF-1 was originally identified by affinity purification with EPO HRE (Wang and Semenza,

1995), recent studies indicate that HIF-2 is the physiological regulator of EPO production (Gruber et al., 2007; Kojima et al., 2007; Kapitsinou et al., 2010). A linkage of familial erythrocytosis further supports this notion with a gain-of-function mutation in HIF2α (Percy et al., 2008).

While a study by Hu and colleagues (2007) has indicated that it is the N-TAD that contributes to target-gene specificity of HIF1α and HIF2α, other factors should be noted in the transactivation of hypoxia-inducible genes. For example, preferential cooperation between the HIF transcription factor complex with certain subsets of transcription factors, co-activators, or co-repressors, as well as tissue-specific expression of HIF-α isoforms may affect the differential gene activation by HIF1α and -2α respectively. Promoter analysis of a subset of genes preferentially regulated by HIF2α revealed a commonality in binding sites for the ETS family of transcription factors. Correspondingly, the knockdown of ELK-1 (the most abundant member of

ETS family) significantly reduced hypoxic induction of the HIF2α-responsive genes

(Aprelikova et al., 2006). Similarly, a separate study found that ELK-1 is required for

46 HIF2α to activate specific target genes such as CITED-2, EPO, and PAI-1 (Hu et al.,

2007). These results demonstrate the need for interaction between the HIF transcription factor complex and appropriate transcriptional cofactors in order for comprehensive expression of hypoxia-inducible genes.

Altogether, a review of the current literature indicates that not only are HIF1α and -2α functionally non-redundant, but are also expressed in a cell-type specific manner, resulting in regulation of differential suites of target genes. Type specific expression, with functional interaction of cooperative factors, determine the consequent activation of their respective target genes, and hence the individual contribution of

HIF1α and HIF2α to overall health and development.

47 1.3. Role of hypoxia and HIF in development.

Much of mammalian embryogenesis occurs at oxygen concentrations of 1-5%, and the role of hypoxia during this time is found to be central to proper embryonic development (reviewed in Dunwoodie, 2009). This hypoxic environment is necessary for the appropriate development of the embryonic vascular system as well as for endochondrial bone formation in the developing embryo (Harvey, 2002; Christoffels et al., 2004; Dai and Rabie, 2007; Hutson and Kirby, 2007; Karsenty, 2008; Patterson and Zhang, 2010). HIF-mediated response to hypoxia is responsible for the regulation of morphogenesis of the developing heart; loss of HIF leads to defective development of the myocardium and adversely affects the vascular endothelium of the heart (Ryan et al., 1998; Compernolle et al., 2003; Krishnan et al., 2008; Ream et al., 2008).

During endochondrial bone formation, HIF is required for chondrogenesis and is a key regulator for mesenchyme condensation and chondrocyte formation (Schipani et al., 2001; Amarilio et al., 2007; Provot et al., 2007).

Apart from embryogenesis, HIF activity also mediates development of the placenta, and is essential for proper trophoblast maturation and activity. One of the main functions of the placenta is to bring oxygenated maternal blood into proximity with deoxygenated fetal blood. Paradoxically, the placenta develops in a low oxygen environment during the first trimester of pregnancy, as a result of trophoblastic plugging of spiral arteries of the endometrial lining (Hustin and Schaaps, 1987).

During this period, oxygen concentrations in the uterine environment range from

1-5% (pO2 0.5-30 mmHg; Okazaki and Maltepe, 2006) allowing for the extensive proliferation of cytotrophoblasts, which upon increased oxygen concentrations, form

48 the invasive cells that give rise to the architecture of the placental vasculature

(Genbacev et al., 1997; Robins et al., 2007). Impaired placental formation and function have been implicated in several pregnancy complications, such as unexplained miscarriage, preeclampsia, IUGR, placental abruption, pre-term labor with intact membranes, premature rupture of the membranes, and stillbirth (Khong et al., 1986 & 1987; Dommisse and Tiltman, 1992; Kim et al., 2002 & 2003; Smith et al., 2004). The changing oxygen environment of the endometrium is thus a key regulator of placental development. Therefore, the ability to sense these changes and to respond appropriately are essential for proper placental function. To this end, there is a large body of work that indicates that HIF is central to this system.

1.3.1. Placentation begins in an environment of physiological hypoxia.

In the early days of placentation and pregnancy, endovascular cytotrophoblasts invade and plug the uterine spiral arterioles, effectively blocking the flow of maternal blood into the intervillous space (IVS). Hustin and Schaaps (1987) first reported that maternal blood flow to the IVS was not fully established until approximately 11 to 12 weeks of gestation, which was later confirmed through Doppler ultrasound (Foidart et al., 1992; Coppens et al., 1996). At about 11 weeks of gestation, the arteriole plugs are displaced and blood flow commences (Hustin and Schaaps, 1987; Burton et al.,

1990). The onset of maternal blood flow is progressive, beginning from the periphery of the placenta and gradually extending towards the centre (Jauniaux et al., 2003a).

The oxygen tension within the IVS thus rises, from 2.5-3% at 8 weeks, to 8.5% at 12 weeks, leading to a burst of oxidative stress in the placenta (Rodesch et al., 1992;

Jauniaux et al., 2000; Jauniaux et al., 2001). Early onset of maternal blood flow into

49 the IVS results in a premature increase in placental oxidative stress, and is associated with early pregnancy loss (Hustin et al., 1990; Jauniaux et al., 2003a & b). This premature onset of blood flow occurs prior to the establishment of robust mechanisms to ameliorate oxidative stress, therefore adequate trophoblast invasion and plugging of the spiral arterioles in early pregnancy are essential for healthy embryonic and placental development.

EPAS1 is expressed most abundantly in the human placenta.

The expression of EPAS1, the gene for HIF2α, across a variety of human tissues has been reported in a study that used whole-genome gene expression arrays to measure the expression of transcripts across a diverse panel of human and mouse tissues. This study therefore presents a gene atlas that describes the normal transcriptome (Su et al., 2002). The tissue-specific pattern of mRNA expression derived from such an atlas could offer important clues about gene function.

Using this tool, it is evident that there is a huge disparity between EPAS1 and HIF1A expression pattern across human tissues (Fig. 1.7). The expansive reach of HIF1A expression is starkly different from the distinct localization of EPAS1 expression to the placenta and lung, suggesting a deviation in role. HIF1A may have a function common to multiple tissues and organs, while EPAS1 has a specialized function unique to placenta and lung. Moreover, while EPAS1 demonstrated its highest expression levels in these two tissues, HIF1A is conversely very lowly expressed in these tissues.

50 EPAS1 displays a substantial expression level in the placenta, and to a lower degree, lung. In comparison, EPAS1 expression in the other human tissues is negligible. For a gene closely tied to oxygen homeostasis, to find it expressed in the lung, a major channel for gas exchange is unsurprising. Similarly, given that the placenta is the major organ through which fetal gas exchange takes places, it is reasonable to find such strong expression of EPAS1. Moreover, EPAS1 has been implicated in regulating trophoblast differentiation as well as proper placental development (see next section).

In a microarray study assessing global gene expression in 8-week placental villi,

EPAS1 was found to be remarkably abundantly expressed, emerging with the 57th strongest signal, out of a total of 48,803 probes (Apps et al., 2011). Similarly, in our lab, previous data from global transcriptome analysis (mRNA-Seq) revealed the strength of EPAS1 expression in hESC-derived trophoblast culture (Fig. 1.8). It emerged as the 70th most abundantly expressed mRNA transcript, and represented the most highly expressed transcription factor, suggesting its involvement in trophoblast and/or placental formation and function. The strength of expression, coupled with the specificity of its tissue localization, strongly indicate a unique function of EPAS1 in the human placenta.

51 Figure 1.7. Expression pattern for EPAS1 and HIF1A across different human tissues. EPAS1 gene expression is by far the highest in the human placenta, followed by the lung, whilst in other human tissues, its expression is negligible in comparison. In contrast, HIF1a is lowly expressed in the lung and placenta, and is instead expressed across a variety of other tissues. The chart was derived using the BioGPS portal conceptualized by Su et al. (2002).

52

EPAS1 Pink histograms/peaks represent individual exons; peak exons; individual represent histograms/peaks Pink expression is low in untreated cells. At 2 days of treatment, treatment, of days 2 cells. At untreated in low is expression EPAS1 by mRNA-Seq in hESC-derived trophoblast culture. trophoblast hESC-derived in mRNA-Seq by EPAS1 iue18 Epeso of Expression 1.8. Figure at Y-axis:value. RPKM. RPKM maximum in measured a expression, to reaching relative is 2275), height RPKM: 6-day 1290; RPKM: (4-day treatment throughout climb to continues and 141), (RPKM: increases expression day 8 (RPKM: 2490).

53 1.3.2. HIF activity impacts placental development.

Previous work in mouse has shown that HIF activity is required for normal development of the placenta, with loss of HIF1α, both HIF1α and -2α, or HIF-β/

ARNT causing catastrophic failure in placenta formation, resulting in embryo lethality early on (Table 1.4; Kozak et al., 1997; Abbott and Buckalew, 2000;

Adelman et al., 2000; Cowden Dahl et al., 2005). Key features of placental defects in mouse models include reduced interaction between the embryo and forming placenta

(chorioallantoic interaction), reduced development of the labyrinthine layer (a feature of the mouse placenta not found in humans), limited vascularization of the placenta, overarching disruption of trophoblast differentiation (including the absence of syncytiotrophoblasts), dramatic reduction in spongiotrophoblasts (a mouse trophoblast cell type), and greater numbers of trophoblast giant cells (another mouse trophoblast cell type). These resulting placental defects offer an insight to the current understanding of how oxygen gradients affect trophoblast proliferation, differentiation, and gene expression.

Cowden Dahl et al. (2005) demonstrated that HIF activity is required for effective chorioallantoic interaction and branching morphogenesis of the mouse placental vasculature. No overt morphological deficiencies were observed in the allantois or chorion, suggesting that HIF might be affecting molecular components of the interaction itself, such as integrins and their ligands (Watson and Cross, 2005). In cases where chorioallantoic interaction does occur, the vascularization of the developing placenta was reduced, thus suggesting that branching morphogenesis of the trophoblasts is impaired and/or that angiogenesis by fetal vascular endothelial

54 cells is defective (Dunwoodie, 2009). Moreover, Abbott and Buckalew (2000) found that HIF-β/ARNTnull mice had a reduced expression of genes involved in the angiogenesis pathway, namely Vegfr2, Ang-1, and Tie-2. However, since the molecular mechanisms by which HIF affects chorioallantoic fusion and fetal vascularization remain unclear, further investigation is necessary to determine whether either or both of these pathways are involved in mediating the role that HIF plays in placental angiogenesis.

In humans, both HIF1α and -2α proteins are constitutively expressed in the placenta.

In the first trimester, HIF1α and -2α have been identified in the syncytiotrophoblast, villous cytotrophoblast, as well as the fetoplacental vascular endothelium (Rajakumar and Conrad, 2000; Genbacev et al., 2001). Their localization is consistent with the low oxygen environment of the placenta during this time, with their abundance decreasing significantly with gestational age (Caniggia et al., 2000; Caniggia and

Winter, 2002; Ietta et al., 2006). Additionally, Ietta et al. (2006) described in detail the dynamic expression of HIF and HIF regulators during placental development; HIF1α and VHL mRNA and protein are present throughout gestation, peaking at 7 to 10 weeks of gestation when the oxygen tension is low, and declining thereafter. The expression of HIF-β/ARNT and HIF2α mRNA and protein are also present and found to be stable throughout gestation, while PHD is dynamically expressed in an oxygen- dependent manner, peaking at 10 to 12 weeks gestation (Ietta et al., 2006). Others report slightly different results, that HIF1α mRNA levels in the placenta does not change across gestation, while HIF2α mRNA abundance increases as gestation progresses (Rajakumar and Conrad, 2000).

55 Meanwhile, comparing placentae from healthy pregnancies at low altitudes (i.e. normoxia) with preeclamptic placentae as well as placentae from pregnancies at high altitudes (i.e. hypoxia) have revealed differences in the abundance of, and presumably activity of HIF proteins. In a study examining pregnancies at high altitude, HIF1α and

VHL mRNA and protein levels were found to be significantly increased in the placentae, whereas HIF2α and HIF-β/ARNT levels were unchanged (Zamudio et al.,

2007a & b). Likewise, HIF1α and HIF2α were over-expressed in the villous placenta of women with preeclampsia compared to normal term pregnancies (Rajakumar et al.,

2001; Rajakumar et al., 2004). Another report by Helske et al. (2001) showed enhanced placental expression of HIF target genes, including VEGF, in pregnancies complicated by preeclampsia and IUGR, which maybe suggestive of a vasculo- adaptive response to chronic hypoxia at the fetoplacental-maternal interface. These results indicate that HIF1α and HIF2α are well placed to regulate placental development throughout gestation in both normal and pathological pregnancies.

56 Table 1.4. Placental defects in mice lacking in components of the HIF transcription complex. Loss of HIF1α, both HIF1α and -2α, or ARNT in the mouse leads to failure in proper placenta formation, resulting in embryo lethality. Key features of placental defects include reduced interaction between the embryo and forming placenta (chorioallantoic interaction), reduced development of the labyrinthine layer, limited vascularization of the placenta, and disrupted trophoblast differentiation (Dunwoodie, 2009).

57 1.3.3. EPAS1 as a candidate gene that confers reproductive success.

The conditions at high altitude present a hypoxic environment, prolonged exposure to which normally results in chronic mountain sickness. When the human body is subjected to low oxygen conditions, it acclimatizes to allow the sustained delivery of normal amounts of oxygen to the organ systems. This involves an increase in hematocrit, through excessive erythrocytosis, to compensate for the insufficient levels of oxygen from the surroundings. However, chronically elevated levels of hematocrit eventually leads to increased blood viscosity, precipitating increased vascular resistance and ultimately compromised tissue oxygenation; a condition known as chronic mountain sickness. Individuals from low-altitude populations who move to live at high altitude thus suffer from a number of potentially lethal diseases specifically related to the low oxygen levels (León-Velarde et al., 2005; Zubieta-

Castillo et al., 2006).

Additionally, residence at high altitude has been shown to have an impact on pregnancy outcomes. Fetal growth is affected; birth weight is reduced by more than

100g per 1,000m elevation gain as a result of slowed fetal growth, or IUGR, rather than shortened gestation (Lichty et al., 1957; Unger et al., 1988; Jensen and Moore,

1997; Lees et al., 2002). Consequently, gestation at high altitude increases the likelihood of infants being born small-for-gestational age (SGA; Julian et al., 2007).

Chronic hypoxia is postulated to be the main cause for such IUGR incidences at high altitude, as birth weight is reduced in animals exposed to hypoxic gas mixtures

(Gilbert and Leturque, 1982). Moreover, other studies also found decreased birth weight in babies born to women with conditions limiting fetal oxygen supply (Novy

58 et al., 1968), as well as greater growth retardation in babies born to women with lower levels of arterial oxygenation and/or uterine blood flow (Moore et al., 1982a & b; Zamudio et al., 1995).

Apart from reduction in birth weight, studies across different populations have found that infant mortality is also affected by altitude. Neonatal mortality rates in Colorado were almost doubled, at an altitude of 2,740m relative to lower altitudes (<2,130m;

McCullough et al., 1977). Furthermore, infants of low birth weight born above

2,730m had higher mortality rates relative to infants of normal birth weight

(McCullough et al., 1977), indicating that low birth weight is not advantageous at high altitude. Similarly in Peru, Mazess (1965) demonstrated that neonatal mortality rate was nearly twofold greater for infants born at high altitude (3,030m) relative to low altitude (338m). More recent studies in Bolivia also show that intrauterine and infant mortality increases with altitude (Giussani, 2002; Keyes et al., 2003).

Likewise, available data indicate that neonatal mortality is higher in highland populations compared with lowland Himalayan regions.

Taken together, these data report that residence at high altitude might be a challenge for reproductive success. It is associated with lower reproductive fitness, evidenced by greater infant mortality rates and/or morbidity in the form of reduced infant birth weight (Julian et al., 2009). It would therefore seem that long-term habitation of high altitude terrain would be challenging, if not fatal, but contrary to that, the high plateaus of Central Asia and the Andes have been inhabited by native populations for generations under hypobaric hypoxia (hypoxia of altitude). Tibetan highlanders, a geographically and genetically distinct population native to the Himalayan range, are

59 in fact particularly resistant to developing chronic mountain sickness (Pei et al., 1989;

León-Velarde et al., 2005). They exhibit little or no increase in hemoglobin concentration with increasing altitude, even as high as 4,000m, and only show moderate increases beyond (Beall et al., 1998). Tibetan women are also less afflicted by pregnancy complications, compared to non-native pregnant women (i.e. Han-

Chinese) living at high altitude - fetal mortality is reduced for the former, and newborns are protected from altitude-associated IUGR (Moore et al., 2001; Tripathy and Gupta, 2005; Yangzom et al., 2008). Moore et al. (2001) found that Tibetan newborns weighed more than Han-Chinese newborns living across the same

2,700-4,800m altitude gradient in the Tibetan Autonomous Region of southwestern

China. Likewise, Yangzom et al. (2008) reported that Tibetan infants weighed more and were less often classified SGA than non-Tibetans (i.e. Han-Chinese and Hui

Muslim) infants born in Lhasa, Tibet.

In trying to understand this ancestrally conferred protection, EPAS1 emerged as the leading candidate gene. Coming from a landmark discovery only two years ago, a variant of the EPAS1 gene in the Tibetan population was postulated to confer fitness to the native highland population (Beall et al., 2010; Simonson et al., 2010; Yi et al.,

2010). In their study, Yi et al. (2010) used exomic sequencing of 50 ethnic Tibetans in order to identify genes showing population-specific allele frequency changes, representing strong candidates for altitude adaptation. EPAS1 emerged with the strongest signal of natural selection, in the form of a single-nucleotide polymorphism

(SNP; C-to-G substitution), which presented at 78% frequency difference between the Tibetan population (87%) and their most recent divergent neighbors, the Han-

Chinese (8.75%). In contrast, the G allele is completely absent in the Danish

60 population, a far removed subset of people. The Tibetan population are thought to have diverged from the Han-Chinese as recently as <10,000 years ago (Aldenderfer and Zhang, 2004), making this SNP the fastest allele frequency change observed in any human gene to date (Yi et al., 2010).

The hypoxic environment at high altitude therefore acts as a selection pressure. An advantageous mutation evolved to confer reproductive success to a sub-population

(i.e. Tibetans) living in a harsh chronic hypoxia environment. In contrast, the Han-

Chinese represents the lowland population most closely related to the Tibetans, but has not undergone selection for altitude-associated hypoxic adaptation.

Impact of SNP on gene expression.

A single-nucleotide polymorphism (SNP) is defined as a variation in the DNA sequence between members of a biological species at a single nucleotide position.

Generally, SNPs occur at a frequency of >1% in a population (Wang et al., 1998). The

International HapMap Project has characterized over 3.1 million human SNPS, indicating a SNP density of approximately one per kilobase (Morizono et al., 2007).

SNPs are not evenly distributed across the genome, but SNPs that affect gene expression occur in all regions of the genome. Generally, SNPs occur much less frequently in coding regions than in non-coding regions of the genome. SNPs found within coding regions of genes have been extensively studied, primarily those resulting in amino acid codon alterations which can lead to variable protein levels and stability (Salavaggione et al., 2005; Beluzic et al., 2006), protein misfolding (Komar et al., 1999), altered protein binding affinities and catalytic properties (Koyano et al.,

61 2004; Kimchi-Sarfaty et al., 2007), improper post-translational modifications (Gentile et al., 2008), and other functional consequences (Gupta and Lee, 2008; reviewed by

Chamary et al., 2006). SNPs in mRNA sequences may also influence mRNA folding and stability (Duan et al., 2003; Capon et al., 2004; Boffa et al., 2008), changing levels of mRNA available for protein translation. Additionally, altered mRNA structure can impact protein synthesis rates (Nackley et al., 2006), and correspondingly affect protein levels.

SNPs located within non-coding regions of the genome are less well understood, and while mostly regarded as non-functional, these variations can alter gene regulatory sequences such as promoters, enhancers, and silencers (Ponomarenko et al., 2002;

Tokuhiro et al., 2003; Wang et al., 2005; reviewed by Knight, 2003 & 2005). SNPs of this kind may alter affinity with the regulatory protein, resulting in differential gene expression (Giacopelli et al., 2004; Egli et al., 2009; Landa et al., 2009), as well as alter efficiency of translation initiation, leading to differential protein expression

(Jacobson et al., 2005). In the most extreme scenario, a gain-of-function SNP conferred promoter-like activity to a non-coding region between the α-globin genes and their upstream regulatory elements (Gobbi et al., 2006), which then interfered with normal activation of downstream genes, consequently leading to a variant form of α thalassemia.

Additionally, SNPs occurring either within splice sites or their auxiliary elements, such as exonic splicing enhancers (ESE), intronic splicing enhancers (ISE), and exonic/intronic splicing silencers (ESS, ISS), can alter or abolish splicing (Krawczak et al., 1992 and 2007), and in doing so, alter the balance of alternative splice forms

62 (Garcia-Blanco et al., 2004). Table 1.5 lists the varying outcomes from modifications of splicing regulatory elements. For example, Pagani et al. (2003) reported that substitutions in a novel regulatory element, described as composite exonic regulatory element of splicing (CERES), of the CFTR gene led to alternative splicing of exon

12, resulting in variable degrees of exon skipping. Skipping of this exon removes part of the first nucleotide-binding domain of CFTR, rendering the protein non-functional.

As another example, a mutation that inactivates an ESE in exon 5 of the MCAD gene causes exon skipping and MCAD deficiency, resulting in a disease that affects mitochondrial function (Nielsen et al., 2007). However, the severity of this disease is alleviated in the presence of a silent A/C polymorphism in the same exon, located 11 bases upstream of the ESE mutation. This is achieved by the C variant which inactivates an adjacent ESS, therefore preventing exon skipping. This particularly revealing example demonstrates the complexity and commonality of the splicing code, and the intricate relationships between disease-causing splicing mutations and disease-modifying natural variants that affect splicing (Wang and Cooper, 2007).

When splicing is altered or abolished, mature mRNA sequence and structure is consequently affected, leading to aberrant protein synthesis. A SNP at a splice donor site of the human MYH genes results in reduced translation efficiency of its transcripts (Yamaguchi et al., 2002). In a splice-site genetic polymorphism of the

KLK12 gene, individuals with the affected genotype have no substantial expression of the KLK12 protein with serine protease activity (Shinmura et al., 2004). Both cases exemplify single nucleotide variations, within splice sites, that cause modifications in translation initiation sites and translation efficiency.

63 Fundamentally, the overall effect of SNPs in splicing regulatory regions basically modulates splice-site and/or splicing efficiency (Pagani and Baralle, 2004). In some cases, this translates to SNPs serving as modifiers of disease severity (Nissim-Rafinia and Kerem, 2005; Wang and Cooper, 2007), much like the earlier described example of MCAD deficiency. SNPs can also impact disease susceptibility; an A/G splice site

SNP in the OAS1 gene, a key antiviral enzyme, is associated with basal enzymatic activity, influencing susceptibility to Type I diabetes (Field et al., 2005). Similarly, an

A/C splice site polymorphism in the first intron of the P2X(7) gene, important for innate immune response, gives rise to a null allele in 1-2% of the Caucasian population (Skarratt et al., 2005), ultimately affecting susceptibility to environmental exposure.

When the natural splice site is altered, the splicing machinery may employ the next available legitimate splice site, leading to a phenomenon known as exon skipping.

Alternatively, the spliceosome may opt for the next best, although illegitimate, splice site in the vicinity, in which case utilizing a cryptic splice site (Krawczak et al.,

2007). For example, intronic mutations in ABCC8 and HADH were identified to create cryptic splice donor sites, leading to the inclusion of an out-of-frame pseudoexon, which eventually resulted in a frameshift and premature termination codon (Flanagan et al., 2013). While the differences between natural and cryptic splice sites have been studied in vitro (Roca et al., 2003 and 2005), it remains largely unclear precisely which properties of the local sequence environment influence the choice of mRNA splicing phenotype.

64 The advent of genomics has led to great leaps in the study of human evolution, unveiling genomic variation through genome-wide scans. In recent years, genome- wide scans for selection have been frequently reported, uncovering several hundred loci which show patterns of variation characteristic of new beneficial mutations that have spread through the population (Bustamante et al., 2005; Sabeti et al., 2007;

Akey et al., 2009; Pickrell et al., 2009; Grossman et al., 2013). Sabeti and colleagues

(2007) showed evidence of selection, within an African population, for a gene linked to the Lassa fever virus. This disease is endemic in Lassa, Nigeria, where 21% of the population show signs of exposure (Richmond and Baglole, 2003). The LARGE gene codes for a glycosylase responsible for post-translational modification of the Lassa fever virus receptor, which has been shown to be critical for virus binding (Kunz et al., 2005). 318 SNPs were found in this gene, 35 of which were common only to the

African population, and a further 3 of that have been identified as functional by current annotation (Sabeti et al., 2007). The authors thus posit that Lassa fever created selective pressure at LARGE.

While most genetic variation has little impact on health, certain variants lead to a myriad of phenotypic effects, and within the right context, can strongly impact disease susceptibility (Sachidanandam et al., 2001). SNPs increase gene diversity, and it is evident that particular SNP alleles confer protection from disease manifestation or progression, whether they are the major allele or variant/minor allele in a population. A SNP that enhances survival will emerge in the population when faced with specific selection pressures, and in this manner, SNPs are contributors to the evolutionary process.

65 Table 1.5. SNPs/Mutations in splicing regulatory elements and resulting outcome. Gene Disease or Variant Effect on splicing/protein Ref. trait affected location and phenotypic outcome

TNNT2 Hypertrophic 5-base deletion Exon 4 skipping. Komamura, cardiomyopathy in intron 3 Py Predisposition to prominent 2004 tract left ventricular hypertrophy.

SHOX Long bone Splice acceptor Activated a cryptic splice Danzig, growth site of first acceptor site in exon 2. 2012 intron

CFTR Cystic fibrosis Splice acceptor Activation of 2 cryptic splice Costantino, site acceptor sites, 2 alternative 2013 splicing products.

CERES Alternative splicing of exon Pagani, (exonic) 12, with varied exon 2003 skipping. Removal of nucleotide-binding domain.

MCAD Medium-chain ESE in exon 5 Exon skipping. Nielsen, acyl-CoA Protein deficiency. 2007 dehydrogenase deficiency Upstream of Inactivates adjacent ESS, ESE in exon 5 prevents exon skipping. Deficiency is alleviated.

KLK12 KLK12 protein Splice donor Splicing abnormality. Shinmura, activity site of intron 4 No substantial expression 2004 of KLK12 with serine protease activity.

OAS1 Type I diabetes Splice site of Splice site moved by 1 Field, 2005 susceptibility intron 6 nucleotide, resulting in longer protein. Basal enzymatic activity.

P2X(7) Innate immunity Splice donor Null allele. Skarratt, site of first 2005 intron

MYH Base excision Splice donor Reduced translation Yamaguchi, repair site of first efficiency. 2002 intron

ABCC8 Hyper- Splice donor Activated cryptic splice Flanagan, insulinaemic site, intronic donor site. Inclusion of out- 2013 hypoglycaemia of-frame pseudoexon, and premature termination.

HADH Hyper- Splice donor Activated cryptic splice Flanagan, insulinaemic site, intronic donor site. Inclusion of out- 2013 hypoglycaemia of-frame pseudoexon, and premature termination.

66 1.4. Experimental hypothesis and approach.

EPAS1 emerged as a subject of interest when it was recently cited as the fastest evolving gene observed in humans to date (Yi et al., 2010). A variant of the gene, containing a SNP, exists in native Tibetans; a population that is protected from chronic mountain sickness despite inhabiting a highland environment low in oxygen, and has been tied to lowered hemoglobin production (Beall et al., 2010; Simonson et al., 2010). While these findings explain why Tibetans are protected from chronic mountain sickness, they also raise the question of how these people compensate for the lack of oxygen if not through increased erythropoiesis.

More strikingly however, the postulated rate of evolution of this gene implied a remarkable population bottleneck, in which a significant percentage of a population that migrated to the highlands must not have survived, or were prevented from reproducing, due to harsh selection pressures. Since the gene in question has been implicated in placental development, it raised the possibility that selection pressure could be acting in utero. I therefore hypothesize that -

A) EPAS1 has a significant role in the development of the human placenta, and that

B) the variant EPAS1 gene contributes to reproductive fitness of the Tibetan population at hypoxic conditions.

We sought to prove this by studying targets of the EPAS1 transcription factor protein in the context of the human trophoblast cell type, the precursor to the placenta. This was achieved through a hESC-derived trophoblast cell culture, maintained under hypoxia (3% oxygen). Targets of EPAS1 were then identified via chromatin

67 immunoprecipitation and sequencing (ChIP-Seq). Bearing in mind that presence of oxygen impacts HIF activity, we validated those targets by studying their in vivo transcriptomic expression in human placental biopsies derived from high and low oxygen uterine environments. These clinical samples were through a collaboration with investigators at the Center for Trophoblast Research, University of Cambridge.

To investigate the role that the variant EPAS1 gene might have in contributing to reproductive fitness at hypoxic conditions, we first characterized the SNP uncovered by Yi et al. (2010), to look for clues as to how it may affect expression of the EPAS1 gene and protein. This was done by inspecting the position of the SNP in relation to splicing regulation as well as assaying for alternatively spliced transcripts. 46 human cell lines, 30 of which were induced pluripotent stem cell lines derived from ethnic

Han-Chinese, were also screened in search of a human cell line with the variant

EPAS1 gene that contains the aforementioned SNP. This would thus enable us to compare expression and activity of the EPAS1 gene with and without the SNP. Hence, we hope to understand how the presence of the SNP might confer reproductive fitness to the native Tibetan population.

68 CHAPTER 2: MATERIALS & METHODS

69 2.1. Cell culture.

Human embryonic stem cell line H1 (WA01, WiCell) was maintained in feeder-free culture using commercially available cell culture media (mTeSRTM1, StemCell

Technologies). Growth media was changed daily, and the volume of media was gradually increased across the days as cells grew more confluent. Upon reaching 80% confluency, cells were passaged at a ratio of 1:12. To dissociate cells, they were incubated with Dispase (1mg/ml; StemCell Technologies) at 37°C for seven minutes.

Dissociation was quenched by flushing with Dulbecco's Modified Eagle Medium:

Nutrient Mixture F-12 Media (DMEM/F-12, GIBCO). Cells were then collected by centrifugation at 500rpm for three minutes. Resulting cell pellet was resuspended in mTeSR to remove cell clumps, and cells were seeded onto tissue culture plates coated with BD MatrigelTM Matrix Basement Membrane (BD Biosciences). For maintenance and propagation, cells were grown at 37°C, at 20% oxygen concentration.

Cells were induced to differentiation into the trophoblast lineage two days post- passage, to allow for cell recovery after the passaging procedure. Differentiation media comprised of 20uM SU5402 (fibroblast growth factor receptor [FGFR] inhibitor; Calbiochem) and 100ng/ml rhBMP4 (recombinant human bone morphogenetic protein 4; R&D Systems), in DMEM/F-12 supplemented with 4ng/ml recombinant human basic fibroblast growth factor (bFGF, GIBCO), 20%

KnockOutTM Serum Replacement (GIBCO), 1mM L-Glutamine (GIBCO) with 35nM

β-mercaptoethanol (GIBCO), and MEM Non-Essential Amino Acids Solution

(GIBCO). Before differentiation media, referred to as BMP4/SU henceforth, was applied to cells, culture was flushed with DMEM/F12 to remove remnants of growth

70 factors present in mTeSRTM1. Cells were incubated in differentiation media for no longer than eight days. For hypoxic treatment, filtered nitrogen gas was piped into cell culture incubators to maintain oxygen at 3% concentration.

For storage of hESCs, confluent cell cultures were dissociated as described above, and centrifuged at 500rpm for three minutes. The cells were resuspended in cryofreezing media mFreSR™ (StemCell Technologies) in the manner recommended by the manufacturer, and aliquoted into cryotubes at 1ml per vial. The cryotubes were placed in “Mr Frosty” containers (Nalgene) buffered with isopropyl alcohol and stored at -80°C overnight, before being transferred into liquid nitrogen for long-term storage. When necessary, cells are taken out of long-term storage and rapidly thawed in a 37°C water bath. Cells were transferred into 10ml of pre-warmed DMEM/F12, and centrifuged at 500rpm for three minutes. The cell pellet was gently resuspended in mTeSR and seeded onto tissue culture plates prepared as previously described.

2.2. Gene expression assay.

2.2.1. RNA extraction.

To extract RNA from cells, 500ul of TRIzol® Reagent (Ambion) was applied to one well of a 6-well culture plate, enough to cover the surface. The reaction was left at room temperature for five minutes, for cells to be lysed and detached from plate surface. Cell lysate was collected in 1.5ml Eppendorf tubes, and 200ul of chloroform was added. Tubes were vortexed until a homogeneous cloudy solution was obtained, and then spun at 13,000rpm for ten minutes. The resulting upper aqueous phase, containing RNA, was transferred to a new tube, and then processed with the RNeasy

Mini Kit (Qiagen) for RNA purification. RNA was resuspended in nuclease-free

71 water (Ambion), and quantified on Nanodrop2000 (Thermo Scientific). For downstream processing, amount of total RNA used from each sample was normalized at 1ug.

2.2.2. Complementary DNA synthesis.

Complementary DNA (cDNA) was synthesized from purified RNA with the High

Capacity cDNA Reverse Transcription Kit (Applied Biosystems). This protocol quantitatively converts total RNA to cDNA, generating single-stranded cDNA suitable for quantitative PCR. The reaction begins with binding of random primers to

RNA, at 25°C for ten minutes. This is followed by reverse transcription by the

TM proprietary MultiScribe Reverse Transcriptase at 37°C for two hours. The final step quenches the reaction by heating to 85°C for five minutes. The resultant single- stranded cDNA can be stored long-term at -20°C. For downstream processing, an aliquot of the cDNA is diluted in nuclease-free water to one-twentieth the original concentration.

2.2.3. Quantitative real-time polymerase chain reaction.

The TaqMan® platform (Applied Biosystems) was used for real-time detection of quantitative polymerase chain reaction (qPCR) product. Using the TaqMan®

Universal PCR Master Mix (Applied Biosystems), reactions were set up with 1ul of cDNA template, along with readily available TaqMan® probes (Applied Biosystems) for the respective human genes. Probes used are included in the Appendix. Reactions were run on the Applied Biosystem ViiA7 thermocycler. The geometric increase in fluorescence signal detected corresponds to the exponential increase of the product, and is used to determine the threshold cycle (Ct) in each reaction.

72 To check for changes in gene expression upon treatment with BMP4/SU, level of expression in treated cells was compared to that of undifferentiated hESCs grown at the same corresponding oxygen concentration. Expression in treated cells grown at

20% oxygen concentration (treated sample) were compared to that of undifferentiated cells (reference sample) also grown at 20% oxygen concentration, and so forth for the

3%-oxygen sample set. GAPDH (glyceraldehyde 3-phosphate dehydrogenase) was used as the reference gene for normalization, since BMP4/SU treatment affects cell morphology and therefore cytoskeletal gene expression, including the common reference gene, β-actin. Although GAPDH is involved in the glycolytic pathway, which is affected by oxygen availability, expression was compared between samples kept at the same oxygen concentration, and not across differing oxygen concentrations.

To derive change in expression level, Ct values from qPCR were analyzed as follows:

ΔCt1 = Ct (target gene in treated sample) - Ct (GAPDH in treated sample)

ΔCt2 = Ct (target gene in reference sample) - Ct (GAPDH in reference sample)

ΔΔCt = ΔCt1 - ΔCt2

2-ΔΔCt = Fold-change in expression

2.3. Assay for alternative transcripts.

To determine transcript sequence generated from EPAS1 in the region of the identified SNP, I screened the EPAS1 mRNA transcripts from H1 hESC-derived trophoblast in the region between exons 2 and 7, within close proximity to the locus of the SNP (located in intron 5).

73 RNA extraction and cDNA synthesis were conducted as previously described.

Primers flanking exons 2 to 7 of the EPAS1 mRNA transcript were designed on CLC

Main Workbench. Primer sequences are included in the Appendix. I amplified this sequence via high fidelity PCR, using the KAPA2G Robust PCR Kit (Kapa

Biosystems). The reaction was initiated at 95°C for three minutes, and was then cycled through 20 seconds at 95°C, 20 seconds at 64°C, and 30 seconds at 72°C, for a total of 35 cycles. PCR products were separated on a 1% agarose gel (UltraPure™

Agarose, Invitrogen) by electrophoresis at 90V for 30 minutes. DNA bands were excised from the gel and purified using QIAquick Gel Extraction Kit.

The resulting DNA was cloned into pCR® 2.1-TOPO® vector (TOPO TA Cloning®

Kit, Invitrogen) and propagated with MAX Efficiency® DH5α™ Competent Cells

(Invitrogen). Colonies were screened and selected for on LB agar using ampicillin- based resistance and IPTG/X-gal-blue/white selection. Positively-transformed colonies were expanded overnight at 37°C in LB broth, and plasmid was purified using Plasmid Mini Kit (Qiagen).

Plasmid DNA was amplified in the same PCR set up as previously described, in search of the sequence between exons 2 and 7 of the EPAS1 transcript. PCR product was again resolved by gel electrophoresis to check for size. Plasmid DNA was sequenced for further analysis. Sequence analysis was performed on CLC Main

Workbench, and sequences were mapped back to the human genome (hg18 assembly,

NCBI Build 36.1).

74 2.4. Screen of human cell lines for presence of SNP.

To aid our search for a human cell line carrying the target EPAS1 SNP, I screened 16 cell lines received from the International Stem Cell Initiative (2011) and another 30 from the laboratory of Lawrence Stanton (Genome Institute of Singapore). DNA, in solution, from the first 16 cell lines were already available in the lab, from an earlier project. The 30 cell lines were received in induced-pluripotent stem cell (iPSCs) form. Coming originally from patients of known Han-Chinese ethnicity, the cells had been reprogrammed to the induced pluripotent stem cell state, in order to generate patient-specific cell lines. For these samples, genomic DNA was harvested from iPSC culture using the DNeasy Kit (Qiagen).

Primers for the regions flanking the SNP were designed on CLC Main Workbench.

Primer sequences are included in the Appendix. High fidelity PCR was performed as previously described, to amplify the sequence within this region. PCR products were separated on 1% agarose by gel electrophoresis at 90V for 30 minutes. Bands of the appropriate size were excised and purified using QIAquick Gel Extraction Kit. DNA was resuspended in nuclease-free water.

All genotyping was conducted on PCR-amplified fragments derived from genomic

DNA, using the BigDye® Terminator v3.1 cycle sequencing kit (Applied

Biosystems). Direct sequencing was performed for all samples on an automated sequencer 3730XL Genetic Analyzer (Applied Biosystems). Sequence analysis was performed on FinchTV (Geospiza Inc.), a DNA sequence chromatogram trace viewer.

75 2.5. Microarray.

RNA extraction was performed as previously described, and assayed for quality

(RNA integrity number, RIN>8.0) using Agilent 2100 Bioanalyzer (Agilent RNA

6000 Pico Kit, Agilent Technologies) in accordance with the manufacturer’s protocol.

300ug of total RNA was processed using Illumina® TotalPrep™ RNA Amplification

Kit (Ambion). This yielded ample synthesized RNA, or cRNA, which was again assayed for quality on the Agilent Bioanalyzer platform. 1ng of cRNA was loaded per lane of the microarray (HumanHT-12 v4 Expression BeadChip Kit, Illumina), and allowed to hybridize for 18 hours at 58°C. Arrays were washed according to the manufacturer’s protocol, and scanned for single-color gene expression signal on

BeadArray Reader (Illumina). Microarray analysis was performed on GeneSpring GX

(Agilent Technologies), using the unpaired student's t-test to determine statistical significance with Benjamini-Hochberg multiple-testing correction (FDR<0.05), and genes demonstrating ≥ 1.5-fold change were reported as statistically significant.

2.6. Immunocytochemistry.

Cells were immuno-stained to confirm protein expression. Since the target proteins are degraded in presence of oxygen, cell culture was harvested on ice to limit enzymatic activity. Cells were fixed in the dish using cold 4% paraformaldehyde

(Sigma-Aldrich) for 30 minutes, and were then washed thrice with phosphate- buffered saline (PBS). Cells were permeabilized using 0.3% Triton X-100 in PBS for five minutes at room temperature, and were then rinsed thrice with blocking buffer comprising 1% bovine serum albumin (Sigma-Aldrich) in PBS.

76 The next step was incubation, with mild agitation, in blocking buffer for 30 minutes at room temperature, to limit non-specific immuno-binding by antibodies. Incubation with primary antibodies was performed for one hour at room temperature, also with mild agitation. Antibodies were used at 1:1000 dilution - EPAS1 (sc-28706, Santa

Cruz Biotech) and HIF1α (sc-13515, Santa Cruz Biotech). Prior to incubation with fluorescent-conjugated secondary antibodies, cells were washed thrice with blocking buffer.

Secondary incubation was then performed for 30 minutes at room temperature, with mild agitation and away from light, due to the light-sensitive nature of the fluorescence conjugates. Secondary antibodies (Molecular Probes®) were used at

1:5000 dilution - Fluorescein Goat Anti-Rabbit IgG (H+L; F-2765) and Alexa Fluor®

488 Donkey Anti-Mouse IgG (H+L; A-21202).

In the final step, cells were stained with 1ug/ml DAPI (4',6-diamidino-2- phenylindole), a nuclear dye, for three minutes at room temperature. Cells were washed thrice with PBS before visualization on the AxioObserver.D1 microscope

(Carl Zeiss International).

2.7. ChIP-Seq.

2.7.1. Chromatin immunoprecipitation.

Cell culture was scaled up to 10-cm dishes to accumulate the necessary number of cells. Cells were harvested on ice and rinsed with cold PBS. To cross-link protein-

DNA complexes, cells were fixed in the dish with 1% formaldehyde (VWR

International), on a shaker at 100rpm for ten minutes at room temperature. Cross-

77 linking reaction was quenched using 375mM glycine. At the end of the reaction, cells were washed twice with cold PBS and scraped off the surface of the dish with a cell scraper. Cells were collected by centrifugation at 3,000rpm for ten minutes.

Lysis was next performed with mild agitation for 30 minutes at 4°C in Cell Lysis

Buffer comprising 10mM Tris-HCl pH8.0, 0.25% Triton X-100, 10mM EDTA pH8.0,

0.1M NaCl, and protease inhibitors. Cellular debris was collected by centrifugation at

2,000rpm for ten minutes. Nuclei were next lysed for 30 minutes at 4°C, also with mild agitation in Nuclear Lysis Buffer comprising 50mM HEPES-KOH pH7.5,

150mM NaCl, 2mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate (NaDOC),

1% SDS, and protease inhibitors. Chromatin pellet was obtained by centrifugation at

20,000rpm for 30 minutes at 4°C.

Chromatin pellet was then resuspended and washed twice by incubating for 30 minutes at 4°C in 0.1% Reduced SDS Buffer (0.1%RSB) comprising 50mM HEPES-

KOH pH7.5, 150mM NaCl, 2mM EDTA, 1% Triton X-100, 0.1% NaDOC, 0.1%

SDS, and protease inhibitors. Chromatin pellet was finally resuspended in 0.1%RSB and sheared to 200-500bp fragments in a water bath sonicator. Debris was removed from solubilized chromatin extract by centrifugation at 13,000rpm for 20 minutes at

4°C.

Chromatin extract was immunoprecipitated with primary antibodies adsorbed on magnetic beads (Dynabeads® Protein G, Invitrogen). 5ug of respective primary antibodies were used per 100ug chromatin extract - EPAS1 (sc-28706x, Santa Cruz

Biotech), HIF1α (PA316521, Pierce), and isotype control rabbit IgG (sc-3888, Santa

Cruz Biotech). Beads were primed with antibodies for two hours at 4°C. Chromatin

78 extract was first pre-cleared by incubation with clean (non antibody-bound) beads for two hours at 4°C. An aliquot of pre-cleared chromatin was set aside as ‘INPUT’ for a quality check procedure, described in detail later.

Pre-cleared chromatin extract was then incubated overnight (<16 hours) with antibody-bound beads at 4°C. At the end of incubation, beads were pelleted on a magnetic stand, and supernatant from each immunoprecipitation reaction was collected, for quality check by Western blot. Immunoprecipitated beads were washed thrice with 0.1%RSB by incubation with mild agitation for five minutes, then rinsed once with Reduced NP40 Wash Buffer (10mM Tris-HCl pH8.0, 0.25M LiCl, 1mM

EDTA, 0.5% NP40, 0.5% NaDOC), and finally with TE Buffer (10mM Tris-HCl pH8.0, 1mM EDTA).

Protein-DNA complexes were eluted from the beads in Elution Buffer (50mM Tris-

HCl pH7.5, 10mM EDTA, 1% SDS) at 68°C for 30 minutes, with agitation at

1,400rpm. Cross-linking was reversed with Proteinase K (Fermentas) in equal parts of Elution Buffer and TE Buffer, at 50°C for four hours, followed by 68°C for six hours. Contaminating RNA was removed using 10ug RNase A (Qiagen) in a 30- minute reaction at 37°C.

DNA precipitation was achieved through phenol/chloroform extraction. Dissolved

DNA was mixed thoroughly with phenol/chloroform/isoamyl alcohol (Calbiochem) by vortexing. The suspension was centrifuged at 13,000rpm for ten minutes. The supernatant was collected, to which chloroform was added. The suspension was again mixed thoroughly by vortexing, followed by a second centrifugation step. To precipitate DNA, the supernatant was mixed with glycogen, 3M sodium acetate, and

79 ethanol, and incubated at -80°C for two hours. DNA precipitate was collected by centrifugation for 30 minutes at 13,000rpm. The pellet was washed once with 75% ethanol, then spun again and finally air dried. The final product, referred to as ChIP-

DNA, was resuspended in nuclease-free water and quantified using the Qubit® dsDNA High Sensitivity Assay Kit (Invitrogen).

2.7.2. Quality check of ChIP-DNA: Western blot and qPCR.

To determine if immunoprecipitation had been successful, the samples were assayed by Western blot. ‘INPUT’ sample from pre-cleared chromatin represents total protein content prior to immunoprecipitation, whereas supernatant from respective immunoprecipitation reactions (IP S/N) would have been cleared of the target protein, as it would have been bound by the respective antibody. If the ChIP had been successful, the target protein would be present in ‘INPUT’, but absent in its corresponding IP S/N.

Protein samples were prepared with NuPAGE® LDS Sample Buffer (Invitrogen) and denatured with 50mM dithiothreitol (DTT; NuPAGE® Sample Reducing Agent,

Invitrogen) at 70°C for ten minutes. Protein gel electrophoresis was carried out on

NuPAGE Novex® 4-12% Bis-Tris gels, at 200V for 50 minutes, using NuPAGE®

MOPS SDS Running Buffer (Invitrogen). Protein standards used were MagicMark™

XP Western Protein Standard (Invitrogen) and SeeBlue® Plus2 Pre-Stained Standard

(Invitrogen). Protein was transferred onto PVDF membranes (Invitrolon™,

Invitrogen) using NuPAGE® Transfer Buffer (Invitrogen), at 25V for 90 minutes.

Protein blots were blocked and washed in 0.1% TWEEN® 20 Tris-buffered saline

(TBS) supplemented with 5% milk (Blotting-Grade Blocker, Bio-Rad). Primary

80 antibodies were used at 1:1000 dilution - EPAS1 (sc-28706x, Santa Cruz Biotech) and HIF1α (PA316521, Pierce). Horseradish peroxidase-linked secondary antibody was used at 1:5000 dilution (anti-rabbit IgG; NA934V, GE Healthcare). Blots were washed once with stringent buffer (0.3% Triton™ X-100 TBS) and rinsed with PBS before visualization using Amersham ECL Prime Western Blotting Detection Reagent

(GE Healthcare).

Another quality check for ChIP is to assay by qPCR for the captured DNA fragments.

Primers for putative HIF binding sites were derived either through a review of the literature or by independent primer design. Four pairs of primers were used in qPCR on the SYBR Green platform. Three of them have been identified to bind to promoter regions of the genes CA9, PHD3, as well as to a putative binding site referred to as

Bind ID 17 (Lau et al., 2007; Schodel et al., 2011). As a negative control, a primer pair binding to the region 1kb upstream of Bind ID 17 was designed. Primer sequences are included in the Appendix.

The template for qPCR was ChIP-DNA from the ‘INPUT’ sample as well as ChIP-

DNA from the immunoprecipitated samples. The binding sites, as specified by the primers, would be enriched in the ChIP-DNA from immunoprecipitated samples when compared to that of the ‘INPUT’. In order to be comparable, amount of template per reaction was kept consistent at 0.1ng DNA. The reaction was set up using SYBR® Green Real-Time PCR Master Mix (Life Technologies), with 0.2uM primer. Real-time detection of the reaction was achieved in the Applied Biosystem

ViiA7 thermocycler. To calculate enrichment, Ct values from qPCR were analyzed as follows:

81 ΔCt1 = Ct (in EPAS1/HIF1α-IP sample) - Ct (Rabbit IgG-IP sample)

ΔCt2 = Ct (INPUT) - Ct (Rabbit IgG-IP sample)

ΔΔCt = ΔCt1 - ΔCt2

2-ΔΔCt = Fold-enrichment

2.7.3. Sequencing.

This protocol describes the preparation of a library of ChIP-DNA in a format compatible with Illumina’s cluster amplification and sequencing platforms. The objective of the protocol is to add a sequencing primer to the 5’-end, as well as adapter sequences to both ends of the DNA fragments. The adapter sequences correspond to the two surface-bound amplification primers on the flow-cells used in the cluster amplification platform, and the sequencing primer corresponds to the primer used in the sequencing-by-synthesis reaction.

15ng of DNA from the EPAS1 ChIP was processed using a modified version of the

ChIP-Seq DNA Sample Prep Kit (Illumina). The first step involves DNA end-repair using components of the kit as recommended by the manufacturer’s protocol. Prior to the following step, DNA was purified using the QIAquick PCR Purification Kit

(Qiagen). The next step was to synthesize an adenine-overhang at the 3’-ends of the

DNA fragments, as described in the kit. The DNA product from this reaction was then purified using the MinElute Reaction Cleanup Kit (Qiagen).

This was followed by ligation of adapters to the ends of the DNA fragments, again as recommended by the manufacturer. As in the previous steps, DNA was again purified, using the QIAquick PCR Purification Kit. Adapter-ligated DNA was next enriched via PCR using the Platinum® Pfx Amplification Kit (Invitrogen). PCR primers were

82 provided in the ChIP-Seq DNA Sample Prep Kit. The reaction was first incubated at

94°C for five minutes, and then cycled through 94°C for 15 seconds, then 55°C for

30 seconds, and 68°C for 60 seconds, for a total of 15 cycles. The DNA product from this reaction was then purified using the MinElute Reaction Cleanup Kit.

The resulting DNA was resolved for size by electrophoresis on a 2% agarose gel

(UltraPure™ Low Melting Point Agarose, Invitrogen) in Tris-acetate buffer (TAE, 1st

Base). The sample was loaded with one lane spacing from DNA ladders (50bp DNA ladder, Invitrogen) to prevent cross-contamination. Electrophoresis was carried out at

120V for 60 minutes.

The gel was next stained with SYBR Green I Nucleic Acid Gel Stain (Invitrogen).

The gel was immersed in a solution of TAE with the gel stain, in a dark box. The set up was incubated at room temperature for 30 minutes, with mild agitation to ensure homogeneous staining of the gel. Visualization was achieved using a UV- transilluminator. The 200bp-300bp section of the DNA smear was excised and weighed. DNA was purified from the gel using the QIAquick Gel Extraction Kit, and eluted in 20ul of elution buffer provided in the kit.

DNA was lastly assayed for quality and quantity using the Agilent 2100 Bioanalyzer

(Agilent DNA 1000 Pico Kit, Agilent Technologies). The sample was diluted to a final format of 15ul of 5nM DNA for sequencing on the Illumina HiSeqTM

Sequencing System.

83 2.7.4. Bioinformatics analysis of ChIP-Seq targets.

Sequences were mapped to reference genome Homo sapiens hg19 (GRCh37) and analyzed using an in-house software package “Control based ChIP-Seq Analysis

Tools” (CCAT; Xu et al., 2010), followed by the Model-based Analysis of ChIP-Seq

(MACS; Zhang et al., 2008).

At the time of writing, peaks were manually identified by viewing sequence tracks on the UCSC genome data browser, and genes in proximity, or with known associations to peak loci, were annotated for binding.

2.8. RNA-Seq.

We interrogated the transcriptomic landscape within the early human placenta to study the in vivo expression of HIF-associated target genes, in the context of differing oxygen levels in the uterine environments.

We received total RNA derived from placental tissue biopsies at the Centre for

Trophoblast Research, University of Cambridge. Prior to the biopsy, gestation age was confirmed by ultrasound. Biopsies are from otherwise normal pregnancies, obtained through chorionic villus sampling (CVS) in the central region of the placenta. Biopsies were immediately preserved cryogenically upon extraction.

In total, 14 samples were grouped into two subsets based on their gestational age. The first subset comprises of 8 samples, collected from 7- to 9-week old placentae. The second subset comprises of 6 samples, collected from 13- to 15-week old placentae.

Since maternal blood flow and the corresponding onset of oxygenation normally

84 begins at 12 weeks, the first subset is referred to as ‘pre-flow’, and the second as

‘post-flow’.

Upon receipt of samples, RNA was assayed for quality using Agilent 2100

Bioanalyzer (Agilent RNA 6000 Pico Kit, Agilent Technologies) in accordance with the manufacturer’s protocol. 2ug of good quality RNA (RIN>7.0) was processed in

96-well format using TruSeq™ RNA Sample Preparation Kit v2 (Illumina). Samples were multiplexed, or barcoded with unique adapters provided by the manufacturer.

Paired-end sequencing on Solexa Genome Analyzer (Illumina) generated 101-base reads, which were then mapped to reference genome Homo sapiens hg19 (GRCh37).

Bioinformatics analysis was performed by Dr. Li Juntao. Sequence reads were aligned using TopHat, and transcript assembly and annotation was performed using

Cufflinks, of the Tuxedo Suite (Trapnell et al., 2012). Differential expression analysis was conducted using edgeR (Robinson et al., 2010). Cluster analysis and visualization of data was achieved through Gene Cluster 3.0 (Eisen et al., 1998).

85 CHAPTER 3: FUNCTIONAL CHARACTERIZATION OF THE TARGET EPAS1 SNP

86 3.1. Introduction.

To investigate the role that the variant EPAS1 gene might have in contributing to reproductive fitness at hypoxic conditions, I first characterized the locus of the target

SNP as defined by Yi et al. (2010). I inspected the position of the SNP in relation to sequence conservation across species, as well as across the genome.

Since the SNP was found in close proximity to a splice site, I speculated that there might be effects on splicing of the mature transcript, and began investigating this by screening the exon-exon junctions of the EPAS1 transcript. This was made possible through the use of RNA-sequencing data established previously (Herath, 2011). In this technique, poly-adenylated mRNA transcripts are captured to analyze expression at the transcriptomic level (mRNA-Seq). These mature transcripts are used to generate smaller fragments known as sequencing reads, which are then mapped to the human genome (hg18 assembly, NCBI Build 36.1). To quantify gene expression, the density or frequency at which reads map to the genome is calculated as ‘reads per kilobase per million mapped reads’ (RPKM). RPKM values of mapped reads can be charted against the genome, and is generally indicative of the expression level for a particular (exon) sequence in the genome. From the mRNA-Seq analysis, I found discrepancies in exon-exon junctions of the EPAS1 mRNA transcript.

This then prompted me to test for alternative splicing, by assaying for alternatively spliced transcripts. Using robust, high fidelity PCR and cloning techniques, I captured three differently spliced transcripts. However, in order to confirm functionality of the

SNP, in perhaps repressing the natural splice site, it was necessary to conduct the

87 assay in a cell line with the SNP. For this purpose, I screened 46 human cell lines for the target G-allele SNP.

88 3.2. Characterizing the locus of the EPAS1 SNP.

The EPAS1 gene comprises 16 exons altogether, with the aforementioned SNP located in the intronic region upstream of the sixth exon. As the precise location of this SNP was not characterized in the original publication (Yi et al., 2010), I investigated this further. Interestingly, I found it to be at chr2:46,441,523; the fifth nucleotide upstream of the 5’ splice site of the sixth exon. Such proximity to the splice site implies it may influence splicing of EPAS1. Moreover, given the C-to-G substitution, there is likely interruption of the pyrimidine-rich tract, usually located about 5 to 40 nucleotides before the 3’ end of the intron (Breathnach and Chambon,

1981; Mount, 1982; Freyer et al., 1989).

At this and immediately adjacent positions, there is sequence conservation as evidenced from cross species comparisons, for the C allele, as well as the T allele

(Fig. 3.1). The genomes of the rabbit (Oryctolagus cuniculus), rat (Rattus norvegicus), mouse (Mus musculus), and the rhesus monkey (Macaca mulatta) have the C allele in this position. The genomes of the elephant (Loxodonta africana), dog

(Canis lupus familiaris), and cow (Bos taurus) have the T allele. This is in contrast to the G allele found in the Tibetan population.

Genome-wide analysis showed that this particular position (five bases 5’ to the start of the exon) has specific base preferences. Base preference indicates the likelihood that any one of the four nucleotide bases adenine, thymine, cytosine, and guanine (A,

T, C, G) occupies a particular position in the genome. Searching across all exons from RefSeq genes in the human genome, we tabulated the occupancy of each nucleotide at positions relative to the 3’ splice acceptor sites. This lead to the

89 observation that at the position of the EPAS1 SNP, the strongest base preference is for

T at 51.1%, followed by C at 31.5%, then A at 8.8%, and finally G at 8.6% (Fig. 3.2).

This thus adds further credence to the idea that the observed C-to-G switch affects splice site function.

90 INTRON EXON 6

Figure 3.1. Locus of EPAS1 SNP shows some sequence conservation. The SNP is at position chr2:46,441,523; 5th nucleotide upstream of the 5ʼ splice site of the 6th exon, indicated by the green arrow. The C allele is found at this position in the genomes of the rabbit (Oryctolagus cuniculus), rat (Rattus norvegicus), mouse (Mus musculus), and the rhesus monkey (Macaca mulatta). The T allele is found at this position in the genomes of the elephant (Loxodonta africana), dog (Canis lupus familiaris), as well as the cow (Bos taurus). In contrast, the Tibetan population has the G allele in this position.

91 Figure 3.2. Locus of EPAS1 SNP has specific base preferences. Analysis of genomic sequence of all internal exons (i.e. those with an intron on either side) represented in RefSeq showed that the 5th position upstream (indicated by position -5) has highest base preference for T at 51.1%, followed by C at 31.5%, then A at 8.8%, and finally G at 8.6%. Position 0: 3ʼ splice acceptor sites.

92 3.3. Alternative splicing between exon 2 and 7.

With the locus of the SNP in such close proximity to the 5’ splice site of the adjacent exon, it seemed reasonable to check for exon-exon junctions, as preliminary evidence for alternative splicing. I next analyzed mRNA-Seq data previously generated in the lab of hESC-derived trophoblast (8 days post-BMP/SU induction). Specifically, I examined reads that map to the 3’ ends of the preceding exon and 5’ end of the following exon. These reads represent those that span exon-exon junctions. The junction between exons 5 and 6 were under-represented compared to that of exons 6 and 7 (Fig. 3.3). More significantly, the junction between exons 4 and 5 was found to be severely under-represented, showing up in only two instances. Lastly but most intriguingly, there was one instance of an exon-exon junction that spanned the 3’ end of exon 4 and the 5’ end of exon 6, altogether skipping exon 5 (indicated by the asterisk; Fig. 3.3). Despite the H1 hESC line used to generate this mRNA-seq data set being homozygous for the C or ‘low altitude’ allele, these read profiles suggested alternative splicing.

To investigate the quality and sequence of EPAS1 transcripts being generated, the region of the EPAS1 transcript between exon 2 and exon 7 was amplified using robust, reverse transcription and high fidelity PCR, and resolved by gel electrophoresis. This resulted in two fragments of distinct sizes; the expected 1,142bp fragment and another fragment approximately 500bp long (Fig. 3.4A). The two fragments, deemed fragment a (1,142bp) and fragment b (~500bp), were extracted from the gel and individually cloned into vectors for propagation. Their DNA was amplified using the same primers as before, in search of the sequence between exons

93 2 and 7 of the EPAS1 transcript. All plasmid clones derived from fragment a presented with a DNA band of approximately 1.1kbp (Fig. 3.4B). Clones derived from fragment b yielded bands of varying sizes, between 400bp to 500bp. Clones were also sequenced for further analysis. Sequences are included in the Appendix.

Analysis of sequences revealed that all clones generated with fragment a yielded canonical transcripts that had been appropriately spliced (i.e. identical to the RefSeq transcript), with no evidence of exon skipping (Fig. 3.4C). In contrast, clones generated with fragment b yielded a variety of alternatively-spliced transcripts. These transcripts differed in the position of their splice junctions, indicating novel splice sites. Two populations of transcripts featured prominently; one in which exon 2 had been spliced to exon 6, and another in which exon 3 had been spliced to exon 6 (Fig.

3.4C). In each of these populations, a novel splice junction had been generated. More pertinently, both scenarios featured a novel splice acceptor site within exon 6, indicating the presence of cryptic splice sites.

Since the SNP was found just five bases upstream of exon 6, I reasoned that, by proximity, it would affect splicing at the exon 5-6 junction more so than it would upstream splicing. Nevertheless, this reasoning would be verified when testing the functionality of the SNP. At this juncture, using synthesized exon 2-7 transcript sequences, splice acceptor sites within exon 6 were mapped back to existing mRNA-

Seq data. In the region of an exon, the detected level of expression is represented as a histogram, and is referred to as the mRNA-Seq footprint of the exon. Within the footprint of exon 6, I mapped the loci of the novel splice acceptor sites (Fig. 3.5).

Both novel splice sites featured the invariant dinucleotide AG, characteristic of the 3’

94 splice acceptor site. The loci of the splice sites were marked with relatively strong peaks downstream, indicating stable and/or strong expression of the following sequence (Fig. 3.5A). Additionally, one of the splicing scenarios preserved the original reading frame of the codons (Fig. 3.5i), while the other did not (Fig. 3.5ii). In the case of (i), 50 codons/amino acids were shaved off the original 68 specified by exon 6. As for (ii), an additional 6 codons were fully omitted, with the 7th only partially removed, leading to a frameshift.

Next, I examined the mRNA-Seq data from chorionic villus samples (see Chapter 5), to assess the expression of the EPAS1 transcript, and any variant isoforms. Bearing in mind that none of the samples were positive for the G allele, it was unsurprising that

EPAS1 expression remained orthodox, with no evidence of variant isoforms, as visualized through de novo transcript assembly, an analysis performed by Dr. Li

(Computational and Systems Biology, GIS).

Lastly, in order to functionally characterize the SNP, the next logical step was to obtain a human cell line with the G-allele SNP, so that we may analyze splicing of

EPAS1 in the presence of the SNP.

95 Figure 3.3. Screening of exon-exon junctions in the EPAS1 mature transcript showed under-representation of exon 4-5 junction and exon 5-6 junction. Pink peaks depict exons 5 and 6 expression, as derived from mRNA-Seq of the EPAS1 transcript. Red bars represent connections between exons as detected through the mapping of sequencing reads to the genome. Red bars between exon 5 peak and exon 6 peak thus represent reads that map to exon 5-6 junction. Exons 4 and 7 are not included, but red bars extending to the right from exon 6 represent reads that map to exon 6-7 junctions, and those that extend to the left from exon 5 represent reads that map to exon 4-5 junction. The number of red bars presented thus indicate the frequency of detection of these exon-exon junction-spanning reads. Exon 5-6 junction was found to be under-represented compared to that of exon 6-7. Exon 4-5 junction was found to be far more drastically under-represented, coming up in only two instances. The red bar marked by the asterisk represents an exon 4-6 junction.

96

Figure 3.4. Alternative splicing between exons 2 and 7 of the EPAS1 gene. (A) The region between exon 2 and exon 7 of the EPAS1 gene was amplified by PCR, and two fragments of distinct sizes were found. The larger fragment was at the expected size of 1,142bp, and the smaller fragment was ~500bp long. (B) When individually cloned and propagated, the larger fragment a yielded clones carrying the full-sized ~1kb fragment, while fragment b yielded varying clones carrying fragments of varying sizes (400bp - 500bp). (C) Upon sequencing, clones that had been transformed with fragment a presented canonical transcripts, with no evidence of exon skipping. On the other hand, clones that had been transformed with fragment b presented a variety of alternatively spliced transcripts. “Other aberrant splicing” represents transcripts that were detected but with no reproducibility. Two populations of alternatively spliced transcripts were found in significance; one at 21%, in which exon 2 is spliced to exon 6, and another at 26%, in which exon 3 is spliced to exon 6.

97 trancript. Loci of novel splice acceptor sites are sites acceptor splice novel of Loci trancript. EPAS1 (A) mRNA-Seq footprint of exon 6 of the the of 6 exon of footprint mRNA-Seq (A) Figure 3.5. Novel splice acceptor sites in exon 6.exon in sites acceptor splice Novel3.5. Figure dinucleotide invariant AG. features intron) of end (3 ʼ site acceptor Splice 6. exon of end 5 ʼ at boundary intron-exonat Sequence (B) (ii). and (i) marked omitted. were 6exon by specified 68original the of codons 50 frame. reading codoninshift no withsequence, to mapped site acceptor splice Novel(i) (ii) The other novel splice acceptor site disrupts the reading frame. Both (i) and (ii) also feature AG at (cyptic) splice acceptor sites.

98 3.4. Screening of human cell lines for the target EPAS1 SNP.

As reported by Yi et al. (2010), the SNP in question exists in the Tibetan population at a frequency of 87%, and in the Han-Chinese population at a frequency of 8.75%.

Based on this statistical probability, I sought to screen 30 locally derived Han-

Chinese cell lines for the presence of the target SNP. I also screened 16 other cell lines, of various ethnic backgrounds, from a collaborative study with the International

Stem Cell Initiative (ISCI).

Having identified the locus of the SNP, primers were designed to flank the intronic region containing the SNP. This region was amplified from genomic DNA of respective cell samples and sequenced to check for the presence of the G allele at the noted SNP position.

All 16 of the ISCI samples were negative for the G allele, and although there was an expected yield of two to three cell lines, out of the 30 Han-Chinese samples (8.75% of 30 samples), none of the samples yielded the targeted SNP. All the cell lines examined had the C allele in the position of the SNP, as opposed to the G allele as found in the Tibetan population. Figure 3.6 illustrates the typical sequencing chromatogram generated from the samples.

99

Figure 3.6. Representative sequencing chromatograms from screened samples. None of the samples yielded the target SNP with the G allele. -5 denotes the position of the SNP, and 0 denotes the 5ʼ splice site of the exon. All the samples had a clear peak for the C allele at the -5 position. Sequencing chromatogram from four samples, P10, P18, P20, and P22, are shown.

100 3.5. Discussion.

As a start to studying the target EPAS1 SNP, we inspected its locus within the gene.

The SNP is found at the fifth nucleotide upstream of the 5’ splice site of the sixth exon. At this locus, there was sequence conservation across some species for the C allele, as well as the T allele, but not the G allele as found in the Tibetan population.

Universally across all RefSeq introns, this position has a ranked preference for bases;

T (51.1%), C (31.5%), A (8.8%), then G (8.6%). Additionally, when I examined occurrence of alternative splicing in proximity to the SNP, we found evidence of cryptic splice acceptor sites within exon 6, just downstream of the SNP. While this finding was promising, I next needed to conduct the same studies in the presence of the G allele, to prove the functionality of the SNP. To that end however, I have been unsuccessful in our search for a human cell line with the target SNP.

3.5.1. Significance of the locus of the SNP.

Based upon the sequence conservation observed at the locus of the SNP, it appears that the C allele is important, whereas the presence of the G allele is novel and unique to the Tibetan population (Fig. 3.1). It is widely accepted that sequence similarities serve as evidence for structural and functional conservation, as well as of evolutionary relationships between the sequences (Kimura, 1979; Cooper and Brown,

2008). Generally, nucleotides with important molecular functions evolve more slowly. Active sites of enzymes, DNA-binding domains of transcription factors, and residues essential for structural maintenance constitute some of the most highly conserved sequences, as changes in amino acid residues are particularly deleterious

(Suckow et al., 1996; Simon et al., 2002). Correspondingly, within cis-regulatory

101 elements of the genome, sites that have a stronger role in regulation accumulate fewer substitutions than those that contribute less (Brown et al., 2007).

Based on our genome-wide analysis, a hierarchy was determined for the base that occupies the position of the SNP. This analysis was done by tabulating the occupancy of a particular nucleotide at positions relative to the 3’ splice acceptor sites. Relative to this position, exonic sequences are found downstream. Exons specify codons for amino acids, and ultimately the nucleic acid sequence for primary protein structure. It stands to reason that if all exonic sequences were compared and tabulated, each position has an equal likelihood of being occupied by any of the four nucleotide bases. This pattern is thus evident in our results (Fig. 3.2). Moreover, it is also evident that just upstream of 3’ splice acceptor sites, there is a greater propensity for T and A nucleotides (>50%) compared to C and G nucleotides, reflecting the region for the pyrimidine-rich tract. A region with a specified recognition or functional motif would indicate explicit base preferences. For example, the consensus sequence of 3’ splice sites has been reported as “(C or A) A G | G T (A or G) A G T” where “|” represents the boundary between intron and exon (Mount, 1982). To that end, the findings of this analysis corroborate fairly well (Fig. 3.2); the consensus sequence at the 3’ splice site is “(C or T) A G | G T (A or G) A G T”. This thus serves to validate our analysis.

Examining the position of the EPAS1 SNP, it was evident that the highest base preference is for T (51.1%), followed by C (31.5%), then A (8.8%), and finally G

(8.6%). The C-to-G switch in this position goes against expectations; a base switch for the least preferred nucleotide could possibly influence splicing, although still speculative at this juncture.

102 Nevertheless, the locus of the SNP is at the 3’ end of what could likely be the pyrimidine-rich tract of that intron (Fig. 3.5). A change to the purine base G as in the

SNP disrupts the pyrimidine sequence, and would likely impact binding by U2 small nuclear ribonucleoprotein (snRNP) auxiliary factor 65 kDa (U2AF65). Meanwhile, cross-species comparison indicates that even with random genetic drift (C/T), the pyrimidine nature of this locus has been preserved (Fig. 3.1).

U2AF65 mediates the first step in splicing, and has been found crucial for vertebrate development (Golling et al., 2002). Deficiencies in this factor are associated with cystic fibrosis (Zuccato et al., 2004), myotonic dystrophy (Tiscornia and Mahadevan,

2000), and some cancers (Davies et al., 1998; Fay et al., 2009; Chen et al., 2011).

Splicing involves a number of trans-acting factors, like the snRNPs, as well as cis- acting determinants, including the splice donor site (5’ end of intron), the splice acceptor site (AG at 3’ end of intron), the branch point sequence (BPS) some 20-35 bases upstream of the splice acceptor site, and also the pyrimidine-rich/ polypyrimidine tract (spanning 4-24 bases) between the BPS and splice acceptor site

(Gao et al., 2008). Pre-mRNA splicing complex assembly is mediated by two specific pre-mRNA-snRNP interactions; the U1 snRNP binds to the splice donor site, and U2 snRNP binds to the BPS. However, prior to these interactions, during the early stages of pre-mRNA splicing, U2AF65 is able to universally recognize and bind degenerate pyrimidine tracts (Ruskin et al., 1988; Zamore et al., 1992), via independent binding ability of two central RNA recognition motifs (Zamore et al., 1992; Banerjee et al.,

2004; Jenkins et al., 2013). U2AF65 then facilitates the ATP-dependent association of

U2 snRNP with pre-mRNA at the branch point (Ruskin et al., 1988). In this way,

U2AF commits the pre-mRNA to the first critical ATP-dependent step of splicing,

103 and its binding is thus often regulated during alternative splicing (Smith and

Valcarcel, 2000).

Although degenerate sequences (splice site signals) mark intron boundaries, U2AF65 is able to recognize these diverse pyrimidine tracts via the action of RRMs. While the exact mechanism remains unclear, it was recently found that RRM1 was more promiscuous for cytosine-containing tracts, and RRM2 was more stringent (Jenkins et al., 2013). This thus implies that while U2AF65 binding may be forgiving (RRM1) to certain alterations in pyrimidine tract sequence, there are some changes that may not be tolerated due to the stringency of RRM2.

It has been shown that interruption of the polypyrimidine tract affects splicing (Freyer et al., 1989; Zuccato et al., 2004), and therefore may also impact phenotypes. A 5- base (CTTCT) deletion polymorphism in intron 3 of TNNT2, one of the genes responsible for hypertrophic cardiomyopathy, has been associated with a predisposition to prominent left ventricular hypertrophy (Komamura et al., 2004).

The deletion allele significantly affected mRNA expression pattern, by skipping exon

4 during splicing, leading the authors to postulate that the five bases were likely part of the intronic polypyrimidine tract, the loss of which therefore impacted splicing.

In the case of the EPAS1 SNP, a G substitution just upstream of the splice acceptor site (region for pyrimidine-rich tract), in a position that was also conserved across species for pyrimidine, could possibly impact U2AF65 binding and consequently splicing. Precisely how this is achieved cannot yet be determined with the current analysis. Although the samples used in this study were derived from the C-allele genotype, the sheer strength of EPAS1 expression in the hESC-derived trophoblasts

104 meant that EPAS1 mRNA was in abundance, and thus brought to light low-level alternative splicing scenarios, including the utilization of cryptic splice sites.

Cryptic splice sites resemble natural splice sites, but are illegitimate, and emerge as the dominant site when the natural splice site has been compromised. According to a hypothesis originally put forth in 1992, the “probability of cryptic splice site utilization is an increasing function of the saturation of the local DNA sequence environment with such motifs” (Krawczak et al., 1992) While one coping mechanism is to opt for the next legitimate splice site and skip exons, in the presence of relevant splicing motifs, cryptic splice site utilization becomes more prevalent (Krawczak et al., 2007). Although the study asserted that this trend was more evident in splice donor sites than in acceptor sites, cryptic splice acceptor sites are not altogether foreign in eukaryotic splicing systems (Roshon et al., 2003). A mutation in the first intron’s splice acceptor site of the SHOX gene, controlling long bone growth, activated a cryptic splice acceptor site in exon 2 (Danzig and Levine, 2012). A study with an approach similar to ours, using reverse transcription and PCR analysis, identified 2 alternative splicing products of the CFTR gene, produced by the activation of two different cryptic acceptor splice sites (Costantino et al., 2013). In contrast to our exon 6 partial exclusion, these cryptic splice acceptor sites resulted in pseudoexon inclusions.

In our analysis, two cryptic splice sites marked by the invariant dinucleotides AG emerged in exon 6, downstream of the natural splice acceptor site (Fig. 3.5). Within the vicinity of the cryptic splice sites (~10 bases upstream), there were clear polypyrimidine tracts (>4 pyrimidines) as well, strengthening the overall splicing

105 motif. Overall, I conclude that should the natural splice acceptor site of intron 5 be compromised, the cryptic sites defined in this study will be utilized in splicing. What remains to be seen however, is if the EPAS1 SNP does indeed compromise the natural splice acceptor site of intron 5, given its proximity. This however requires examination of splicing events in the context of the G-allele SNP.

3.5.2. Impact of cryptic splice sites in exon 6.

Our analysis revealed three differently spliced transcripts; the canonical one that utilized the natural splice acceptor site, and two others which had utilized cryptic splice sites within the adjacent exon. In the non-canonical scenarios, at least 50 codons were omitted from the original 68 specified by exon 6 (Fig. 3.5). The utilization of one of the cryptic splice sites uncovered in this study would lead to a frameshift (Fig. 3.5ii), likely resulting in missense and/or premature termination. In the alternative scenario (Fig. 3.5i), a significant portion of the exonic sequence (50 out of 68 codons) would be removed. When EPAS1 mRNA sequence (as identified on

RefSeq) was mapped using CLC Main Workbench, it was found that exon 6 lay in between two functional domains of the EPAS1 protein; PAS and PAS3 (Fig. 3.7A).

The PAS domain, while first recognized as a signaling module responsible for sensing oxygen, amongst other stimuli (reviewed by Taylor and Zhulin, 1999), is also required for structural integrity of the eventual functional complex (Chapman-Smith et al., 2003). The PAS3 domain of EPAS1 has been found to bind the analogous HIF-

β/ARNT domain, disruption of which interferes with the ability of full-length HIF to respond to hypoxia in living cells (Erbel et al., 2003). In a follow-up study, the authors proposed a new model for the interaction of EPAS1 and HIF-β/ARNT PAS

106 domains in the context of the functional transcription factor (Card et al., 2005). The visualization shows a “potent cooperative interaction” between the PAS domains of both proteins, which may provide additional affinity and/or specificity for the complex (Fig. 3.7B). It also demonstrates the role of linkers between PAS domains in accommodating the associations to form the heterodimer. The authors posit that the

PAS domains form a co-operative, intermolecular tetramer of PAS components that serve to “convert an inherently weak set of protein/protein and protein/DNA interactions into a stronger binding force that permits the formation of a functionally viable complex.”

Should the frame-preserving cryptic splice site defined in this study be utilized during pre-mRNA processing, it would lead to a major loss of the linker region between the

PAS domains of EPAS1. This would presumably lead to inefficient, if at all, association of the PAS domains in forming the heterodimeric HIF transcriptional complex, and could likely inhibit or impede binding to HREs and further gene regulation of downstream targets.

In light of the EPAS1 SNP in the Tibetan population, I now hypothesize that the C-to-

G substitution interferes with U2AF65 binding, altering splicing with the utilization of cryptic splice acceptor sites. At this juncture, two possibilities arise; utilization of the frame-shifting splice site, resulting in a missense protein, or utilization of the frame- preserving splice site, resulting in partial exon-exclusion in the region linking two functional protein domains. Either situation would likely lead to, at best, a compromised ability of the transcription complex to respond to hypoxia. This

107 hypothesis would be verified through assessment of EPAS1 splicing in the context of the G-allele SNP.

108 (A)

(B)

Figure 3.7. Exon 6 as the spacer/linker between two functional domains of EPAS1. (A) EPAS1 mRNA and coding region as identified by RefSeq; 16 exons in EPAS1 mRNA, CDS denotes the coding region. Upstream of the CDS would be the 5ʼ untranslated region (5ʼ UTR), and downstream would be the 3ʼ UTR, followed by the poly-A tail. Within the CDS, four conserved domains have been mapped; PAS, PAS3, HIF-1, and HIF-1a CTAD. Exon 6 mostly makes up the space/linker between the two PAS domains. (B) A model of interaction between HIF-α and HIF-β/ARNT PAS domains, as proposed by Card et al. (2005). Linkers between functional domains accommodate the associations to form the complex. Pre-α refers to α-helical elements that are predicted to form to the N-terminal side of the ARNT and HIF-a PAS-A domains.

109 3.5.3. The SNP screening process.

The search for a human cell type inherently carrying the target SNP fell short. In a study conducted early last year by Buroker et al. (2012), they used a similar genotyping technique, by direct sequencing of PCR-amplified genomic DNA. Their samples were obtained from 162 Han-Chinese and 79 Tibetan patients, of which they successfully identified 2% of the Han-Chinese subgroup and 28% of the Tibetan subgroup to be homozygous for the G-allele SNP. Since the methodology is thus validated, the fruitless search could be attributed to the unfortunately conservative sampling size.

Based on the sequencing chromatograms, all our samples had strong clear peaks for the C allele at the specified position. However, sample P10 represents a typical sample where there was a background peak for another allele, which can be attributed to leaky signaling across the read for the sample (Fig. 3.6.). Overall, the background trace for sample P10 is not as clean as that of the other samples shown. Apart from that, the results clearly indicate that none of the cell lines available were positive for the G-allele.

Alternative methods to overcome this obstacle would be to engineer site-directed mutagenesis, to knock-in the desired gene variant, or to silence the inherent gene and induce artificial expression of the variant gene via a BAC (bacterial artificial chromosome) system. However, these steps are notoriously time-consuming, and not to mention, challenging (Zwaka and Thomson, 2003). For these reasons, I chose not to further pursue this investigation of the SNP. Instead, efforts were focused on elucidating the functional role of EPAS1/HIF2α within the human trophoblast.

110 3.6. Conclusion.

This study had set out to prove the functionality of the C-to-G SNP in the EPAS1 gene. A striking characteristic of this SNP was its prevalence in Tibetan highlanders, a geographically-isolated population that have survived generations of hypobaric hypoxia courtesy of their native habitat. The emergence of the SNP is testimony for the functionality of the sequence element. While studies have associated this SNP with decreased hemoglobin concentration phenotypically (Beall et al., 2010), none have shown the molecular interactions involved in achieving the resultant phenotype.

As a start, this study assessed the likelihood for alternative splicing in the vicinity of the SNP locus. It was evident, through genome analysis across species, as well as genome-wide analysis of base preferences, that this locus was preserved for pyrimidine. Analysis of mRNA-Seq data then unveiled the discrepancies in exon junctions, which led to the search of alternatively-spliced EPAS1 transcripts. This culminated in the discovery of two cryptic splice sites in proximity to the locus of the

SNP. Despite conducting studies in the context of the C-allele SNP, the significant expression of EPAS1 in the hESC-derived trophoblast culture meant that EPAS1 mRNA was in abundance, therefore enabling the identification of the lesser-used cryptic splice sites. Nevertheless, examination of splicing in the presence of the G- allele SNP was necessary to confirm the functionality of the SNP in abolishing the natural splice acceptor site five bases downstream of its position, and the emergence of either or both cryptic splice sites as the default splice acceptor site. Alternatively, in order to investigate if splicing at the cryptic sites does result in phenotypic changes, a functional assay involving the (partial) deletion of exon 6 would be useful.

111 Both possibilities predicted from the cryptic splice sites involve modifications of the affected exon, and either situation would result in attenuation of HIF activity and response to hypoxia. Since the Tibetan population do not display symptoms of altitude sickness, one of which is increased hematocrit, and since the SNP has been associated with decreased hemoglobin (Beall et al., 2010), it could be likely that the

SNP, by compromising HIF activity, in effect serves to mitigate this debilitating condition in the Tibetan population. Hence, these native highlanders are able to escape the dire symptoms. While that may be so, the question still remains: how then do their biological systems operate under prolonged hypoxia?

112 CHAPTER 4: IDENTIFICATION OF EPAS1 TARGETS

113 4.1. Introduction.

This study sought to identify the binding targets of EPAS1/HIF2α in the human trophoblast, in an effort to understand its role in the development of the human placenta and the oxygen-modulation of the uterine environment. As an oxygen- sensitive protein, EPAS1/HIF2α is thought to be a key contributor to the progression of trophoblast cell types, which is modulated by oxygen levels within the uterine environment. The initial climate of physiological hypoxia allows for extensive proliferation of cytotrophoblasts, which upon increased oxygen concentrations, form the invasive cells that give rise to the placental villi (Genbacev et al., 1997; Robins et al., 2007).

Our in vitro system was an hESC-derived trophoblast culture, as previously established in the lab (Luo, 2008). hESCs were induced to differentiate into the trophoblast lineage through the use of growth factor BMP4 and fibroblast growth factor receptor (FGFR)-specific tyrosine kinase inhibitor, SU5402. Luo (2008) presented an extensive list of trophoblast-associated genes, including transcription factors as well as placental genes, that were upregulated within hours of treatment, exemplifying the sensitive nature of the treatment. Additionally, Luo’s work in profiling the transcriptome of the resultant culture showed the lack of expression of genes that mark the ectoderm (SOX1, PAX6, GFAP and NEUROD1), mesoderm (T/

BRACHYURY, PECAM1, MYF5, and MYOD1), endoderm (SOX17, GATA4) and germ cells (SYCP3), this time exemplifying the specificity of the treatment.

Moreover, in a separate thesis, Herath (2011) analyzed the protocol designed by Luo through transcriptome sequencing, and compared the profile of gene expression with

114 published microarray datasets derived from a range of tissue and cell types.

Unsupervised hierarchical clustering indicated that 6-day old cultures from the in vitro differentiation protocol was most similar to the placenta as a tissue type, and most similar to EVTs and cytotrophoblasts, as a cell type. Altogether, much of the earlier work provide evidence in support of our in vitro system as representative of the true trophoblast.

The protocol, like most standard culture protocols, was designed for 20% oxygen conditions. However, since EPAS1/HIF2α is at risk of oxygen-dependent degradation at normoxia, the protocol was modified and subjected to 3% oxygen instead, which also mimics the in vivo uterine environment (Rodesch et al., 1992; Rodesch et al.,

1992; Jauniaux et al., 2000; Jauniaux et al., 2001; Okazaki and Maltepe, 2006).

However, there was a need to first verify and profile the gene expression of the resultant culture at 3% oxygen. This was done by global microarray analysis, with further verification of gene expression by qPCR.

After establishing the trophoblast lineage of the cells, EPAS1/HIF2α protein expression and nuclear localization were first confirmed via immunocytochemistry.

Western blot then verified the ChIP protocol, demonstrating the presence of the target

EPAS1/HIF2α protein in the pre-immunoprecipitation lysate, and the efficient clearing of the target protein from the supernatant of the immunoprecipitation reaction. ChIP-DNA extracts were interrogated by qPCR for putative binding targets of EPAS1/HIF2α, drawn from existing literature (Lau et al., 2007; Schodel et al.,

2011), before being sequenced for physical binding targets. These targets are indicative of the HIF-transcriptional response in the trophoblast cell type, offering insight to the role of EPAS1/HIF2α in placental development.

115 In vitro as well as in vivo expression and upregulation of genes associated with these targets were then assessed. In vitro samples were obtained from 6-day BMP4/SU- treated culture grown in 3% oxygen, and upregulation was assessed by comparison with that cultured in 20% oxygen. In vivo expression was verified by RNA-Seq of 14 chorionic villi samples obtained from collaborators in Cambridge (see Chapter 5).

Samples were categorically split by gestational age; 7- to 9-week old and 13- to 15- week old. Given the understanding that maternal blood perfusion of the placenta occurs between 10 to 12 weeks of gestation (Hustin and Schaaps, 1987; Jauniaux et al., 1992; Jauniaux et al., 1995; Jaffe et al., 1997; Burton et al., 1999), the former was deemed pre-flow, and the latter post-flow. The implication therein was that pre-flow samples were indicative of gene expression in low oxygen (therefore HIF activation), and consequently in post-flow, EPAS1 targets would be downregulated.

116 4.2. In vitro differentiation into trophoblast lineage at 20% and 3% oxygen.

Cell culture, and BMP4- and SU5402-directed differentiation (BMP4/SU) were first observed for variability in growth and maintenance between 20% and 3% oxygen. hESCs were able to proliferate fairly efficiently in 3% oxygen, comparable to that in

20% oxygen. hESCs when grown in culture form circular colonies (Fig. 4.1A). In cultures kept at 20% oxygen, masses of cells were occasionally found growing out from the centre of the colonies after several days, indicative of spontaneous differentiation (Fig. 4.1A[iv]). hESCs cultured in 3% oxygen exhibited far less spontaneous differentiation.

Upon BMP4/SU treatment, cell differentiation began at the periphery of the colony, evidenced by morphology (Fig. 4.1B). Cells take on a flattened form, with increased cell area as well as increased cytoplasmic-nuclear ratio. Over several days, differentiation proceeded in a radial fashion toward the centre of the colony (Fig.

4.1B[ii] & [iv]). Overall, based on morphology, maintenance of cells in culture as well as differentiation ability of cells were observed to be similar at 20% and 3% oxygen.

Since the BMP4/SU treatment had been previously established for 20% oxygen conditions, it was necessary to determine if the treatment would be compatible in 3% oxygen. I determined this by global microarray analysis for gene expression in cells treated for 2 days and for 6 days, in either 3% or 20% oxygen, with biological duplicates, and compared expression with untreated cells grown in 3% and 20% oxygen respectively. Each array had a coverage of 47,323 probes.

117 Upregulated genes (>1.5-fold, p-value<0.05) were examined by Venn analysis to derive overlapping sets of genes that have been upregulated due to hypoxia and/or

BMP4/SU treatment (Fig. 4.2). After subtracting genes upregulated as result of the change in oxygen level, 2,312 genes were found commonly upregulated in 3% and

20% oxygen conditions, after 2-day treatment. Similarly, after 6-day treatment, 3,654 genes were commonly upregulated between the two conditions.

Of the 2,312 genes upregulated at 2 days, many were known trophoblast-associated genes, such as GCM1, CK7, ID2, DLX3, TFAP2C, and TFAP2A. Interestingly,

TFAP2A was found to be also upregulated in untreated hESCs grown in hypoxic culture. By 6 days of treatment, this list grew to include genes for hCG (CGA, CGB,

CGB1, CGB5, and CGB8), syncytin (ERVWE1), the marker for EVTs (HLA-G), and placental growth factor (PGF). In the case of ID2 and TFAP2C, they were no longer found to be upregulated by 6 days, in 20% oxygen, but remained upregulated in 3% oxygen (Fig. 4.2).

To validate the microarray, it was imperative to assay by qPCR for the expression of specific genes. Moreover, this would further establish that BMP4/SU treatment under hypoxia was effectively inducing the trophoblast lineage. Gene expression profiles were also compared to that of normoxic BMP4/SU treatment.

Expression of the pluripotency markers OCT4, SOX2, and NANOG, along with a panel of trophoblast and placental genes were examined (Fig. 4.3). OCT4, SOX2, and

NANOG were similarly downregulated between normoxic and hypoxic BMP4/SU treatment, indicating the loss of the pluripotent stem cell state (Fig. 4.3A). The upregulation of trophoblast transcription factor genes, CDX2, GATA2, GATA3, DLX3,

118 TFAP2A, TFAP2C, HAND1, ID2, and MSX2, in both normoxic and hypoxic BMP4/

SU treatment indicated that cells were differentiating into the trophoblast lineage

(Fig. 4.3B). Expression of transcription factor CDX2 was most strongly upregulated at 2 days, in both treatments, then dropped thereafter. In contrast, most of the other trophoblast transcription factors displayed further upregulation following the initial increase at day 2. Strikingly, ID2 expression showed downregulation at days 6 and 7 when cultured at 20% oxygen, but continued to remain upregulated at 3% oxygen, in line with the microarray analysis.

The placental genes assayed include PGF and CGA, which showed the most remarkable increase in expression upon BMP4/SU treatment, with a significant jump between days 2 and 4 (Fig. 4.3B). Likewise, PGF expression profile demonstrates a similar significant increase between days 2 and 4. Keratin 18 (KRT18), a gene expressed in the placenta, and EPAS1 remains hugely increased throughout the 7 days of treatment, while HIF1A showed a small increase in gene expression in the first 4 days, followed by downregulation thereafter. These trends were similar across both oxygen conditions.

Overall, the BMP4/SU treatment in hypoxic conditions appeared to be as effective as that in normoxic conditions, in its ability to effectively induce the trophoblast lineage in vitro, as evidenced by the upregulation of trophoblast transcription factor genes as well as placental genes. Interestingly, upregulation of CDX2, GATA3, HAND1, and

ID2, were significantly stronger in BMP4/SU treatment at 3% oxygen than it was at

20%.

119 Figure 4.1. Culture of hES and trophoblast-induced cells are comparable between 20% and 3% oxygen conditions. (A) Colony and cell morphology of untreated hESCs, cultured at 20% and 3% oxygen conditions, were found to be similar in early days. After several days, outgrowth of cells from colonies cultured at 20% oxygen was not uncommon (indicated by red arrow), but this was not observed in the 3%-oxygen cultures. (B) Trophoblast-induced cultures (BMP4/SU-treated) displayed cell differentiation beginning at colony periphery, and progressed radially toward the centre. Cells take on a flattened form, with increased cell area as well as increased cytoplasmic-nuclear ratio.

120 Figure 4.2. Venn analysis of upregulated genes. Through microarray analysis, 2,312 genes were found commonly upregulated (>1.5 fold) upon 2-day treatment, while that of 6- day treatment was 3,654 genes. Genes that were potentially upregulated as a result of hypoxia were subtracted; these genes were identified as upregulated in untreated/hESCs cultured at 3% oxygen conditions.

121 Figure 4.3. Expression of pluripotency markers as well as trophoblast markers and placental genes in BMP4/SU treatment at 20% and 3% oxygen conditions. Left column, gene expression profiles for cells treated with BMP4/SU at 20% oxygen concentration; right column, at 3% oxygen concentration. Genes were assayed at 2, 4, 6, and 7 days of treatment. (A) Pluripotency markers OCT4, SOX2, and NANOG are significantly downregulated in both 20% and 3% oxygen conditions. (B) Trophoblast transcription factors and placental genes are similarly upregulated between the two oxygen conditions, with the exception of CDX2, GATA3, HAND1, and ID2 (by studentʼs t-test, **p-value<0.05). The upregulation of these four genes are stronger in hypoxic conditions than in normoxic conditions.

122 4.3. EPAS1 targets by ChIP-Seq.

Prior to ChIP, in vitro protein expression of EPAS1 was confirmed by immuno- staining (Fig. 4.4A). The fluorescence signal was strongest in the nuclei, with expression significantly stronger in cultures kept at 3% oxygen than that at 20% oxygen, suggestive of oxygen-dependent degradation. Western blot of nuclear lysates indicated that EPAS1/HIF2α protein was expressed upon BMP4/SU treatment at 3% oxygen (Fig. 4.4B). In contrast, HIF1α protein was present in untreated hESCs cultured at 3% oxygen, but absent under the treatment. I included the post- immunoprecipitation supernatant (IP S/N) alongside the INPUT fraction (pre-cleared nuclear lysate), to demonstrate antibody efficiency. The results show that the target protein was greatly enriched in the INPUT fraction, and only weakly detected in the

IP S/N, indicating efficient capture by the antibody.

As part of the quality check for ChIP, size resolution of DNA fragments was determined by gel electrophoresis (Fig. 4.4C). The DNA smear indicated an acceptable range of sizes, ranging from 100bp to 500bp, with the strongest signal around 200-300bp. Additionally, prior to sequencing, the ChIP-DNA was validated by qPCR with primers for presently known HIF targets (Fig. 4.4D), with the caveat that none of these targets have previously been described in the trophoblast lineage.

These targets include the promoter region of CA9, HRE of EGLN3, and a novel binding site designated BID17 (Lau et al., 2007; Schodel et al., 2011). We also included a negative control, using primers designed to a region >1kbp upstream of the novel BID17 site. We found 4.9-fold enrichment for BID17, and 2.1-fold enrichment for EGLN3. However, CA9, like the negative control (-)BID17, did not

123 show enrichment, at 0.64-fold and 0.53-fold respectively. As an extra measure of confirmation that these values, since admittedly small, were real, I also conducted

ChIP-qPCR on untreated hESCs kept at 3% oxygen. In all cases, the enrichment values derived from the untreated controls were <0.5-fold.

With all quality checks passed, ChIP-DNA were processed for sequencing. Two libraries were prepared from 6-day BMP4/SU-treated cells grown in 3% oxygen; the

INPUT for control (CHC015), and the EPAS1/HIF2α ChIP-bound fraction

(CHC016). Samples were sequenced on two separate lanes, on the Illumina HiSeqTM

Sequencing System, and mapped to reference genome Homo sapiens hg19

(GRCh37).

While a more thorough analysis of this data set is pending, due to time limits here I only present a preliminary analysis of peaks identified by manual inspection. Visually scanning across the mapped ChIP-seq data on the UCSC Genome Browser I identified the 41 most highly enriched locations in the EPAS1 data that lacked peaks in the control (Table 4.1). As often the most enriched locations are those that are also the most biological relevant I used this analysis to confirm the success of the ChIP- seq. Seven, out of the 41 peaks have been recognized as HIF binding sites in a prior study in a breast cancer cell line, MCF-7 (Schodel et al., 2011); these sites have been associated with SLC2A3, BHLHE40, EGLN1, ALDOA, EGFR, ELF3, and EGLN3.

Peaks were ranked according to strength (peak height), and several peaks are pictorially presented (Figs. 4.5-4.9). These peaks are marked by ENCODE-identified

DNaseI hypersensitivity clusters, which indicate the presence of regulatory elements, as well as histone acetylation (H3K27AC), indicating the presence of active enhancers at least in the cell lines studied which includes undifferentiated human ES

124 cells (Figs. 4.5-4.9). Previously identified transcription factor binding (ChIP-Seq from ENCODE) in close proximity to peaks are also illustrated on figures.

The strongest peak (#1; peak height: 242) is found closest to SLC2A1, also known as

GLUT1, a glucose transporter. The gene is located ~50kb upstream of the binding

(peak) locus (Fig. 4.5). The next strongest peak (#2; peak height: 182) is located next to HIF1A, and overlaps with multiple other transcription factor binding sites (Fig.

4.6). The gene is found immediately upstream of the binding locus. #4 (peak height:

142) is nearest to ARRDC3, and its region has been strongly identified for the

H3K27AC mark (Fig. 4.7). #14 (peak height: 96) is found within 5’ UTR of EGLN1, and has been previously recognized as a HIF2α binding site (Fig. 4.8; Schodel et al.,

2011). #28 (peak height: 77) is found ~40kb downstream of FLT1 (Fig. 4.9).

125 Figure 4.4. Identification of EPAS1/HIF2α in vitro. (A) Immuno-staining of EPAS1/HIF2α in BMP4/SU-treated cells kept at 3% vs.. 20% oxygen. EPAS1 shows nuclear localization in treated cells at 3% oxygen. (B) Western blot of nuclear lysates. EPAS1/HIF2α protein was not present in untreated hESCs cultured at 3% oxygen, but was present upon treatment at 3% oxygen. In contrast, HIF1α was only present in hESCs. Lack of the protein band in IP S/N indicates clearing of the target protein, after capture by the respective antibody. (C) Sonicated DNA fragments, prior to IP, shows a smear of sizes ranging from 100-500bp, with strongest signal at 200-300bp. (D) Fold-enrichment of known HIF2α binding targets in 6-day treated BMP4/SU samples alongside hESCs, as determined by qPCR (**p-value<0.05).

126 Table 4.1. Preliminary analysis of EPAS1 binding sites. 41 peaks were identified and annotated for nearest gene, or known association with genes. Peaks were ranked according to peak height (see Figs. 4.6-4.10). Seven sites have previously been identified as HIF binding sites in MCF7 cells (Schodel et al., 2011).

127 hr ssih oelp ih utpeohr rncito factor transcription other multiple with overlap slight is There . SLC2A1 iue45 SrnetEA1 idn st s~0bdwsra of downstreampartially ~50kb as wellis as site clusters, binding EPAS1hypersensitivity DNaseI Strongest with 4.5. coincides Figure also region peak The ENCODE. on ChIP-Seq by identified as sites binding coincides with histone acetylation (H3K27AC). CHC015: INPUT; CHC016: EPAS1 ChIP.

128 There are multiple overlaps with other transcription factor binding factortranscription otherwith overlaps multiple are There . HIF1A -strongest EPAS1 binding site is just downstream of downstream just is site EPAS1 binding -strongest nd ie (hPSq n NOE. h pa rgo as cicds ih NsI yesniiiy lses a wl a wt hsoe acetylation histone with as well as clusters, hypersensitivity DNaseI with coincides also region peak 2 4.6. Figure The ENCODE). on (ChIP-Seq sites (H3K27AC). CHC015: INPUT; CHC016: EPAS1 ChIP.

129 Figure 4.7. Peak #4 is nearest to ARRDC3. This binding region has been strongly identified for the H3K27AC mark, and has numerous transcription factor binding sites within its proximity. CHC015: INPUT; CHC016: EPAS1 ChIP.

130 Figure 4.8. Peak #14 is found within exon 1 of EGLN1, in the 5ʼ UTR. There are multiple overlaps with other transcription factor binding sites (ChIP-Seq on ENCODE). The peak region also coincides with DNaseI hypersensitivity clusters, as well as with histone acetylation (H3K27AC). CHC015: INPUT; CHC016: EPAS1 ChIP.

131 There are several overlaps with other transcription factor binding on (ChIP-Seq sites binding factor transcription other with overlaps several are There . FLT1 Figure 4.9. Peak #28 is found ~40kb downstream of downstream ~40kb found is #28Peak4.9.Figure INPUT;CHC015: (H3K27AC). acetylation histone withaswell as clusters, hypersensitivity DNaseI with coincides also region peakENCODE). The CHC016: EPAS1 ChIP.

132 4.4. In vitro & in vivo expression of EPAS1 binding targets.

From the global microarray analysis, genes upregulated in 2- and 6-day BMP4/SU- treated cultures grown in 3% oxygen were obtained by comparison with gene expression in that grown in 20% oxygen. Out of the 41 targets from preliminary analysis, 13 genes were found upregulated in hypoxia (Fig. 4.10A); 12 after 6-day

BMP4/SU-treatment at 3% oxygen, and one after 2 days. Since the ChIP samples were derived from 6-day treatment, I focused in vitro upregulation analysis to the 12 genes (Fig. 4.10B). Of the 12 genes, EGFR and ELF3 are previously known HIF2 targets (Schodel et al., 2011).

Next, I assessed in vivo expression of genes associated with EPAS1 binding sites

(Fig. 4.11A). With the exception of PKM2, all genes identified were expressed in the chorionic villus samples, whether pre- and/or post-flow. SLC28A1 was also expressed only minimally (<1 FPKM) in both pre- and post-flow samples. Most of the genes were downregulated post-flow, although only seven were significant (p-value<0.001);

LDHA, COX4I2, CGA, ADM, HIST1H1C, FLT1, and SLC2A1 (Fig. 4.11A).

Individual expression of these genes in each of the in vivo samples is illustrated (Fig.

4.11B). KISS1R is included as well, since expression in post-flow samples is minimal.

Overall, of the genes identified, seven were previously reported as HIF2 binding sites

(Schodel et al., 2011), 12 were upregulated in vitro under hypoxia, and eight were upregulated in vivo pre-flow (Fig. 4.12).

133

Figure 4.10. In vitro expression of EPAS1 targets. (A) Genes upregulated in 3% oxygen vs. 20% oxygen, after 2-day or 6-day BMP4/SU treatment respectively, are overlaid with identified EPAS1 targets. 12 genes associated with EPAS1 binding sites are upregulated in 6d-BMP4/SU treatment at 3% oxygen. (B) Upregulation of the 12 genes associated with EPAS1 binding sites. EGFR and ELF3 have previously been identified as HIF2 targets. Horizontal axis: gene; vertical axis: fold change.

134 Figure 4.11. In vivo expression of EPAS1 targets by RNA-Seq. (A) Pre- vs. post-flow expression of genes associated with or nearest to EPAS1 binding regions (*** indicates p- value<0.001). AVE PRE-FLOW: average expression in 7- to 9-week placentae; AVE POST- FLOW: average expression in 13- to 15-week placentae; horizontal axis: gene; vertical axis: expression level in FPKM. (B) Individual expression of genes differentially expressed between pre- and post-flow (p-value<0.001). KISS1R is also included since post-flow expression is minimal. #: potentially novel isoform, with at least one splice junction shared with a reference transcript.

135 Figure 4.12. 35 genes associated with EPAS1 binding sites. Seven were identified prior by Schodel et al. (2011), 12 were upregulated in vitro under hypoxia, and eight were downregulated in vivo post-flow.

136 4.5. Discussion.

4.5.1. Establishing the trophoblast lineage in 3% oxygen.

Here I have established entry into the trophoblast lineage in our hypoxic in vitro system, as evidenced by the upregulation of trophoblast transcription factors and accompanying downregulation of pluripotency markers OCT4, SOX2, and NANOG.

The upregulation of trophoblast transcription factor genes appeared in a temporal fashion, where CDX2, a transcription factor marking the early trophoblast in the TE, was most strongly upregulated at 2 days of treatment, while that of the others,

GATA2, GATA3, DLX3, TFAP2A, TFAP2C, HAND1, ID2, and MSX2, were further upregulated after 2 days. The other placental genes, CGA, KRT18, PGF, and EPAS1, were also further upregulated after 2 days.

The BMP4/SU protocol developed in our lab has successfully demonstrated rapid upregulation of transcription factors associated with trophoblast emergence, including

CDX2, GATA2, GATA3, HAND1, ID2, MSX2, and TFAP2A and TFAP2C (Luo, 2008).

However, given the oxygen-sensitive nature of EPAS1/HIF2α, there was a need to keep the in vitro system under hypoxia. There was thus a concern for BMP4/SU treatment efficacy since multiple studies have demonstrated that low oxygen tensions hinder differentiation in hESC cultures, and is preferential for the maintenance of a highly proliferative, pluripotent population of hESCs (Ezashi et al., 2005; Forristal et al., 2010). Indeed, I observed higher tendency for spontaneous differentiation, by morphology, in hESCs cultured at 20% oxygen than at 3% oxygen, as evidenced by the outgrowth of cell masses from the centre of the cell colony (Fig. 4.1).

137 Nevertheless, it appeared that BMP4/SU-induction of the trophoblast lineage was effective even in hypoxic conditions (3% oxygen). Both microarray and qPCR analysis demonstrated the upregulation of trophoblast/placenta-associated genes, including transcription factors (Fig. 4.2 and 4.3). Gene expression profiles of these genes remained largely similar between the two oxygen conditions, with the exception of CDX2, GATA3, HAND1, and ID2, showing surprisingly greater upregulation in hypoxic BMP4/SU treatment. PGF and hCG genes were not represented in the microarray as upregulated at 2 days, and only emerged after 6 days, comparable to their qPCR expression profiles, which presented a significant increase after 2 days. Syncytin, or ERVWE1, as well as HLA-G, also only emerged as upregulated after 6 days, although this was not verified by qPCR. By microarray analysis, TFAP2C had dropped out of the upregulated genes list in 6-day treatment at

20% oxygen, but this was not mirrored in qPCR analysis; TFAP2C expression profile was comparable in 3% and 20% oxygen.

CDX2 and GATA3 were more strongly upregulated under hypoxic conditions than under normoxic conditions across all time-points. Since CDX2 is a well-known marker of the early trophectoderm (Kimber et al., 2008), which coincidentally develops in a low oxygen environment, the BMP4/SU treatment under hypoxia was conceivably more conducive for the emergence and maintenance of this early cell type. Similarly, GATA3, a CDX2-regulator also found abundantly in trophectoderm

(Home et al., 2009), has its expression beginning at the 4-cell embryo stage, within an environment of low oxygen. It is thus reasonable to say that while low oxygen preserves pluripotency, under a condition of induced or controlled differentiation, we

138 are mimicking the in vivo environment, which overall gives enhanced maintenance of the trophoblast progenitors.

Interestingly, HAND1 and ID2 also showed stronger upregulation in hypoxic conditions. ID2 expression was in fact downregulated at 6 and 7 days in normoxic

BMP4/SU treatment. Both of these genes have been postulated to be involved in early trophoblast differentiation and cytotrophoblast development (Knofler et al., 1998;

Janatpour et al., 2000). While the exact mechanism of their function in the trophoblast cell type is yet to be fully understood, their expression patterns have provided clues to their possible roles. For example, ID2 expression was relatively increased in cases of compromised cytotrophoblast differentiation (Janatpour et al.,

2000). The authors also reported that cytotrophoblasts transduced to constitutively express ID2 retained characteristics of undifferentiated cells, displaying inhibited invasion capabilities. HAND1 expression, on the other hand, was undetectable in the villous trophoblast, but was abundantly expressed in cytotrophoblastic cell lines

(Knofler et al., 1998). My observation of increased ID2 and HAND1 expression under hypoxic conditions, especially when tied to the downregulation of ID2 under normoxic BMP4/SU treatment, indicates that the treatment in low oxygen appears to support the presence of the earlier cytotrophoblast. This is in line with the findings by

Genbacev and colleagues (1997), in which they found that hypoxic culture (2% oxygen) supported the proliferation of cytotrophoblasts while normoxic culture (20% oxygen) promoted differentiation. Regardless, in this study, I do not attempt to make any definitive claims about the specific trophoblast cell sub-type achieved in our in vitro system. To that end, there is a separate project underway in the lab, which aims

139 to design an in vitro system capable of maintaining the syncytiotrophoblast and/or the cytotrophoblast.

Altogether, the gene expression assays indicates that BMP4/SU treatment under hypoxia was effective in repressing pluripotency markers and switching on trophoblast transcription factor genes, as well as in the production of placenta- specific genes, such as CGA and PGF.

4.5.2. Verification of ChIP.

Having ascertained mRNA and protein expression of EPAS1/HIF2α and HIF1α in our in vitro system, I confirmed that EPAS1/HIF2α was the dominant HIF protein in the trophoblast-induced culture, being upregulated to a greater extent than HIF1α, which additionally, was only transiently upregulated during treatment (Fig. 4.3B). This trend of EPAS1/HIF2α mRNA abundance over that of HIF1α, in the context of the human placenta, was not unlike that which had previously been reported by Su and colleagues (2002). This observation was supported by protein expression, in which

EPAS1/HIF2α was markedly abundant upon treatment, but not in the case of HIF1α

(Fig 4.4B).

Contrary to recent reports linking HIF proteins and the maintenance of pluripotency in hESCs (Covello et al., 2006; Forristal et al., 2010; Das et al., 2012), I found no significant nucleic presence of EPAS1/HIF2α in our hESC culture at 3% oxygen, which is likely due to the length of time the culture was kept in hypoxia. I conducted short-term cultures, where the cells were incubated in 3% oxygen for no longer than 7 days, which is in stark contrast to the study by Forristal et al. (2010), where they conducted long-term cultures of more than three passages. Even so, the authors

140 reported transient (<3 passages) nuclear localization of HIF1α in hESC culture at 5% oxygen, while EPAS1/HIF2α displayed an initial cytoplasmic localization, before translocating to the nucleus following long-term (>3 passages) culture at 5% oxygen.

My observations of HIF protein expression in hESCs cultured at hypoxia were similar to theirs; I also report nuclear localization of HIF1α in short-term hypoxic culture, but am unable to confirm neither the cytoplasmic expression of EPAS1/HIF2α at short- term nor the nuclear expression at long-term, since I only tested the nuclear lysates of short-term culture.

Nevertheless, I confirmed EPAS1/HIF2α protein expression in vitro, with nuclear localization (Fig 4.4A). I first validated the ChIP, by qPCR using primers retrieved from previous studies (Lau et al., 2007; Schodel et al., 2011). These results, in fold- enrichment, indicate the interaction of the DNA binding site with the antibody- protein/ChIP complex; the higher the fold-enrichment, the stronger the affinity of the protein for the binding site.

The results of the ChIP-qPCR indicated that there was binding at the PHD3 and

BID17 sites, but not at the CA9 and (-)BID17 sites. The primers used were designed to flank the hypoxia-responsive elements (HRE) of PHD3 and CA9 (Lau et al., 2007), while that of BID17 was designed against a novel binding site, whose corresponding gene is yet unclear (Schodel et al., 2011). For this reason, I designed yet another pair of primers, (-)BID17, for the region 1kbp upstream of the identified BID17, as a negative control.

Since the fold-enrichment values were not particularly high, I included a ChIP set conducted on untreated hESCs for comparison. The untreated hESCs represented a

141 null scenario, with no expression of target protein, and hence no binding, resulting in a <1 fold-enrichment. As evident in Fig. 4.4D, the fold-enrichment in treated samples for PHD3 and BID17 were significantly higher than that in untreated samples, indicating with conviction that there was protein binding to these sites. In contrast, a

HIF1β ChIP tested in the same manner resulted in non-significant fold-enrichment between treated and untreated samples (data not shown).

The PHD3 gene has been demonstrated to be regulated by HIF proteins, through its

HRE located in its first intron (Pescador et al., 2005). Preliminary results (#38; Table

4.1) now indicate that likewise, in our in vitro system, EPAS1/HIF2α also binds to the same HRE of PHD3, also known as EGLN3, the gene for one of the HIF prolyl hydroxylases (PHD). While PHDs are responsible for hydroxylation of HIFα proteins, targeting them for recognition and protein degradation under normal oxygen conditions, PHD3/EGLN3 is itself a HIF target and is markedly induced by hypoxia

(Appelhoff et al., 2004; Aprelikova et al., 2004), suggesting that it may participate in a feedback loop to mitigate the HIF response in hypoxia (Minamishima et al., 2009).

4.5.3. Assessment of EPAS1/HIF2α targets in the human trophoblast.

Apart from EGLN3, through a preliminary manual analysis at the time of writing,

ChIP-Seq of EPAS1/HIF2α ultimately derived 41 binding sites (Table 4.1). 35 known genes associated with, or nearest to, the identified sites were deemed potential

EPAS1/HIF2α regulatory targets (Fig. 4.12). Of these genes, seven had already been identified as HIF2 targets (Schodel et al., 2011), which I find a positive indicator of the results presented.

142 Twelve genes were found by global microarray analysis to be upregulated in vitro, and coupled with ChIP identification of binding sites, this indicates possible active regulation by the bound EPAS1/HIF2α transcription factor. In contrast, the other 12 genes identified by ChIP-Seq but not upregulated in vitro are likely not under active regulation in the human trophoblast. Meanwhile, only eight of the 35 potential targets were verified in vivo, indicated by significant change in expression under oxygenation (downregulated post-flow), and out of that eight, only CGA and

HIST1H1C remained consistent between in vitro (upregulated in hypoxic culture) and in vivo pattern of regulation. Regardless, this does not mean that the other targets were inaccurately identified. Rather, this reflects the disparity between cell types achieved in culture and that extracted from tissue samples. While the in vitro system manages to approximate the cell types found in the tissue, it might also boil down to activation of different targets during different phases of placental development, wherein the in vitro cells are less mature than that of the in vivo.

This list of targets included proponents of energy metabolism, such as glucose transporters (SLC2A1 and SLC2A3), which serve to increase glucose availability for glycolysis. Glycolysis compensates for energy production in low oxygen, and correspondingly, glycolytic enzymes (ALDOA, GAPDH, LDHA, and PKM2) were also identified as prospective targets. Additionally, genes involved in angiogenesis, such as the vascular endothelial growth factor receptor (VEGFR/Flt1) and adrenomedullin (ADM), a vasodilator, were found.

As verified by hypoxic upregulation in vitro, potential EPAS1/HIF2α targets in the trophoblast include EGFR, HIST1H1C, ELF3, MAPK4, KRT19, MID1, CGA, NANS,

143 ARRDC3, CLK3, AMOTL2, and IDH2 (Fig. 4.10B). EGFR and ELF3 were reported by Schodel et al. (2011) as HIF targets in MCF7 cells. In our hESC-derived trophoblast culture, EGFR was upregulated ~31-fold while ELF3 was upregulated

~7-fold (Fig. 4.10B), confirming EPAS1/HIF2α-mediated activation in vitro.

Meanwhile, examining pre-flow upregulation, where EPAS1 is expected to be active within the low oxygen placental environment of the first trimester, CGA, HIST1H1C,

LDHA, ADM, FLT1, SLC2A1 and KISS1R also emerge as potential EPAS1/HIF2α targets in the trophoblast (Fig. 4.11B). These genes represent strong contenders as they were downregulated/unexpressed post-flow, in a presumably oxygen-rich environment, one in which EPAS1/HIF2α would be subject to oxygen-dependent degradation and have its activity compromised. For example, peak #35 was found right next to ADM, a gene significantly downregulated post-flow (4.11B). ADM expression is relatively specific to the placenta (by BioGPS; Su et al., 2002), and although it did not emerge in our microarray analysis as upregulated in vitro, it has been found to be induced by hypoxia in placental explant culture (Marinoni et al.,

2011). Altogether, this hypoxic induction, in vivo downregulation post-flow, and identification of EPAS1/HIF2α binding in proximity to the gene, are evidence of

ADM as a HIF2 target in the human trophoblast.

On the other hand, COX4I2 (cytochrome c oxidase subunit IV isoform 2), was instead downregulated pre-flow. This gene codes for a subunit of cytochrome c oxidase

(COX), the terminal enzyme of the mitochondrial respiratory chain (Huttemann et al.,

2001). Increased expression and function of COX in highly oxygenated tissue, like lung and trachea, has been tied to its protective function, in preventing oxidative

144 damage by accelerating the final electron transport step, which decrease the available electrons for free radical formation (Huttemann et al., 2007). This thus might explain the increased expression of COX4I2 post-flow, upon perfusion and oxygenation of the placenta.

Regulation of COX4I2 in hypoxic conditions is thought to be mediated either through a novel conserved oxygen-responsive element found in the promoter (Huttemann et al., 2007), or via HIF1 transcription activity (Fukuda et al., 2007). One of the HIF1- targeted HREs (5’-CTGTGTGCACGTA-3’) on COX4I2 is found in intron 1 (Fukuda et al., 2007), and sits well within the ChIP-identified EPAS1/HIF2α binding locus.

Interestingly, although verification at the protein level is necessary to prove corresponding changes in protein presence, the decreased expression of COX4I2 pre- flow (low oxygen climate) might suggest a transcriptional repressor role for EPAS1/

HIF2α.

4.5.4. EPAS1 regulation and preeclampsia.

Preeclampsia is a condition plagued by placental insufficiencies, which ultimately results in inefficient perfusion of the tissue, likely leading to dysregulation of oxygen- sensitive gene expression. Given the capacity of EPAS1/HIF2α in mediating gene expression in the trophoblast, coupled with its oxygen-sensitive nature, it was logical to wonder if EPAS1/HIF2α is implicated in the dysregulation of gene expression during preeclampsia.

FLT1, a gene notorious for its implication in placental development and preeclampsia, emerged as an EPAS1/HIF2α target in the human trophoblast. The binding region was found ~40kb away from the gene, and it was also significantly

145 downregulated in post-flow placenta; average expression was 215.0 FPKM pre-flow and 13.6 FPKM post-flow. Additionally, LEP, a gene frequently associated with preeclampsia, is found ~221kb away from peak #5 (Table 4.1). LEP is also significantly downregulated post-flow, indicating oxygen-sensitive regulation (see

Chapter 5). It is thus not beyond reason to speculate if LEP expression in the placenta is also EPAS1/HIF2α-mediated.

Since the analysis presented is preliminary, there are likely numerous other EPAS1/

HIF2α targets in the trophoblast yet to be identified, with the possibility that more genes in that list are implicated in preeclampsia. More extensive and in-depth bioinformatics analysis is in progress to uncover those genes. Ultimately, it remains to be seen if EPAS1/HIF2α-mediated gene expression has an overreaching influence on genes associated with preeclampsia. At this juncture, this notion of EPAS1/HIF2α as a key player in gene dysregulation during preeclampsia is novel and altogether exciting.

Nevertheless, in the next chapter, I explore global gene expression in the chorionic villus samples to uncover genes differentially expressed between pre- and post-flow, conditions which act as a measure of oxygen-mediated regulation, and their association with placental development and pregnancy health.

146 4.5. Concluding remarks.

In this chapter, I presented the approach and results of identifying EPAS1/HIF2α transcription factor targets in the human trophoblast. Given the ethical concerns surrounding the use of, as well as, the limited accessibility to healthy human placental tissue, we believe the best foreseeable method to elucidating the HIF-mediated response in the early placenta is to maintain a simulated in vitro system, derived from pluripotent hESCs, that is capable of supporting the trophoblast cell type. To that end, many regard the BMP4-driven/FGF2-inhibited hESC culture as a reliable model of trophoblast differentiation (Schulz et al., 2008; reviewed by Ezashi et al., 2012).

Using a hESC-derived trophoblast culture kept at hypoxia, EPAS1/HIF2α binding sites were established by ChIP-Seq. Associated genes were identified, and both in vitro and in vivo expression as well as regulation were assessed.

At the time of writing, binding sites that had been sequenced were being computationally assessed by CCAT (Xu et al., 2010) and MACS (Zhang et al., 2008) to derive a more comprehensive list with 95% confidence (p-value<0.05). From this list, biologically-relevant targets will be identified using the strategy described in this chapter; by examining gene expression in the relevant conditions. Since we can simulate the conditions in which EPAS1/HIF2α is present and active (low oxygen culture; pre-flow placenta), as well as when EPAS1/HIF2α is not active (high oxygen culture; post-flow placenta), we are able to sidestep the need for EPAS1/HIF2α knockout and over-expression. Moreover, such molecular manipulation drives the cell further from its biological context, possibly leading to inaccurate identification or misrepresentation of targets. In addition, using the CCAT-/MACS-identified hits, de

147 novo motif analysis will be conducted, which could amount to a convincing assessment of the experiment, since the HIF-binding sequence has been identified and reviewed in the literature.

As with all ChIP-based approaches to identifying transcription factor gene targets, further molecular validation is obligatory. One approach is by mutagenesis of the

HIF-bound sequence, followed by assessment of subsequent gene expression to ascertain the gene target tied to respective EPAS1/HIF2α binding sites. At present analysis, association of gene with transcription factor binding site is by observation of proximity as well as previous gene annotations.

It is hoped that from these attempts, we can glean the first insight to the HIF transcriptional network of genes responding to early development in the placenta.

Many previous studies have identified HIF gene targets in a variety of tissue and cell types (Wang and Semenza, 1993; Forsythe et al., 1996; Kaluz et al., 2008; Schodel et al., 2011). The study by Schodel et al. (2011) identified 500 HIF-binding sites in a breast cancer cell line, and by overlapping with high-resolution mapping of DNase1 hypersensitivity sites, their results provide a comprehensive pan-genomic analysis of the HIF-mediated transcriptional response to hypoxia, at least in MCF7 cells.

However, it is essential to specify these targets within a cell-specific or biological context, as evidence has shown that the HIF transcriptional network may or may not activate all of its targets, depending on the active HIF isoform present, its regulators, as well as as its co-activators (Semenza, 2003; Wenger et al., 2005). In this regard, it would also be prudent to identify co-activators of EPAS1/HIF2α in the trophoblast cell type. Moreover, studying the HIF transcription network in the trophoblast cell

148 type would enhance our understanding of early placental development, which begins in a low oxygen environment.

149 CHAPTER 5: GLOBAL TRANSCRIPTOMIC ANALYSIS OF THE PRE- AND POST-FLOW PLACENTA

150 5.1. Introduction.

Examination of the global transcriptomic landscape of the early placenta offers insight into the biological processes occurring within the tissue. Genes being expressed are strong indicators of active pathways as well as cellular interactions ongoing in the tissue. Comparison across changing environments represents an excellent strategy to understand developmental and coping mechanisms which are activated during this window of change. During the transition from first to second trimester, a significant change takes place in the uterine environment. The placental vasculature, although fully formed by the end of the first trimester, is virtually devoid of maternal blood due to the presence of trophoblastic plugs in the spiral arteries

(Hustin and Schaaps, 1987; Burton et al., 1999). During this time, which spans the entire embryonic period (up to 8 weeks) and the first few weeks of the fetal period, the embryo/fetus is undergoing extensive growth and virtually all organogenesis.

Without blood flow, nutrition is derived from secretions of the uterine glands, in what is known as histiotrophic nutrition (Burton et al., 2002). At approximately 12 weeks, the trophoblastic plugs are broken down, thus unleashing the flow of blood (Hustin and Schaaps, 1987; Jauniaux et al., 1992; Jauniaux et al., 1995; Jaffe et al., 1997;

Burton et al., 1999). The sudden presence of blood drastically modifies the uterine environment; the blood brings with it oxygenation, as well as hemotrophic nutrition, which is the exchange of blood-borne materials between the maternal and fetal circulations (Wooding and Flint, 1994).

An analysis of pre-flow versus post-flow chorionic villi would provide an opportunity to explore transcriptional dynamics in response to oxygen within trophoblast under a

151 physiological setting. With the abundance of expression of EPAS1 in chorionic villi

(Apps et al., 2011), presumably this physiological response would be mediated, in part, through HIF2α. Through collaboration (Cheryl Lee, Andrew Sharkey, and

Ashley Moffett), I was able to access two sets of RNA samples, derived from chorionic villus of normal pregnancies, one from near the end of the first trimester (7- week to 9-week old; pre-flow), and the other from the start of the second trimester

(13-week to 15-week old; post-flow). We achieved a comprehensive assessment of gene expression through massively parallel sequencing of poly-adenylated mRNA

(i.e. mRNA-Seq). The changes evident in the expression of these genes, between pre- flow and post-flow samples, therefore indicate the biological processes triggered by the onset of maternal blood flow, as well as processes actively taking place during this time of changing oxygen concentration in the placenta. mRNA-Seq permits the accurate quantification of gene expression from read counts, as the number of RNA-Seq reads generated from a transcript is directly proportional to its relative abundance in the sample. However, the number of reads generated from a transcript is influenced by the length of the transcript. Therefore, to calculate the correct expression level of each transcript, reads that map to a particular transcript are tallied, and this count is then normalized by the length of that respective transcript. To compare the expression level of a transcript across runs, the counts must be normalized for the total yield of the machine. The commonly used quantifier

‘fragments per kilobase of transcript per million mapped fragments’ (FPKM) incorporates these two normalization steps to ensure that expression levels for different genes and transcripts can be compared across runs (Trapnell et al., 2012).

FPKM was used as opposed to ReadsPKM since we generated paired-end mRNA-Seq,

152 thus the fragment equates to counting of two reads, one from either end of the fragment.

We thus obtained the levels of gene expression, quantified as FPKM, in the placental samples. Comparing between pre-flow and post-flow, we determined a list of the genes which were differentially expressed across the two conditions (p-value<0.001).

These genes were then screened, via gene ontology analysis, for their associations with, as well as function in, biological pathways and cellular interactions. Genes with the highest expression levels in the placenta were also identified. Additionally, amongst the differentially expressed genes were numerous genes known to be implicated in preeclamptic pregnancies, a complication characterized by high blood pressure and proteinuria in the expectant mother.

153 5.2 mRNA-seq data generation.

The samples were derived from placental biopsies in Cambridge, UK under local

Institutional Review Board approval. Prior to the biopsy, gestational age was confirmed by ultrasound. The samples were from otherwise normal pregnancies that were being electively terminated. Biopsies were taken from the central region of the placenta through a chorionic villus sampling (CVS)-like procedure and immediately cryogenically preserved. Total RNA was isolated in Cambridge by our collaborators and subsequently shipped on dry ice to Singapore.

In total, 15 samples were grouped into two subsets based on gestational age (Table

5.1). The first subset comprises of 8 samples, collected from 7- to 9-week old placentae. The second subset comprises of 7 samples, collected from 13- to 15-week old placentae. Since maternal blood flow and the corresponding onset of oxygenation normally begins at 12 weeks, the first subset is referred to as ‘pre-flow’, and the second as ‘post-flow’. All RNA Integrity Number (RIN) values as determined by the

Agilent Bioanalyzer were 6.0 or greater. After paired-end library construction and bar coding, the 15 samples were run across 5 lanes on the Illumina HiSeq 2000 with each lane generating 154 to 200 million 101-base paired-end reads per lane. For each sample, approximately 40-60 million paired reads were generated (Table 5.1). With each cDNA fragment sequenced 101 bases into either end, there would be 202 bases of sequence per fragment. Thus each RNA sample was sequenced to a depth of approximately 10 billion bases. Mapping to the genome was in the ~70-80% range

(Table 5.1) which is consistent with other human mRNA-Seq experiments.

154 Left and right mapped reads were tallied, normalized, and presented as FPKM values for each unique transcript, indicating strength of expression. The top 10% of all genes expressed in the tissue had an average expression (across all samples) of 14.7 -

91,698.3 FPKM. To exemplify the use of this data set in interrogating gene expression, I will first outline the expression levels and profiles of some of the most highly expressed genes; the hCG genes (CGA and CGB).

5.2.1. Human chorionic gonadotropin genes.

CGA, the gene coding for the alpha subunit of hCG (Fiddes and Goodman, 1979), also gives rise to the alpha subunit of luteinizing hormone (LH), follicle-stimulating hormone (FSH), and thyroid-stimulating hormone (TSH). These glycoprotein hormones are heterodimeric, with their unique beta subunit conveying their individual biological functions. The beta subunit of hCG is encoded by six highly similar genes, CGB, CGB1, CGB2, CGB5, CGB7, and CGB8, arranged in tandem and inverted pairs on chromosome 19 (Policastro et al., 1986; Talmadge et al., 1984; Bo and Boime, 1992).

When gene expression was ranked based upon average FPKM across samples, CGA, along with three CGB genes, turned up as the top four (i.e. most abundantly expressed) genes in pre-flow placentae, while in post-flow placentae, they placed within top 10 genes (Fig. 5.1). CGA was the gene most highly expressed in both pre- and post-flow conditions, and across all samples. In pre-flow samples, CGA yielded an average of 149,334.26 FPKM, that is, ~14.9% of all polyadenylated transcripts.

Expression of CGB genes in pre-flow placentae stood as follows; CGB8 came in

155 second with 36,856.39 FPKM, followed by CGB5 (36,272.68 FPKM), and CGB

(28,500.73 FPKM).

156 15 samples were obtained through a CVS-like procedure, with gestational age determined age gestationalwith procedure, CVS-like a through obtained were samples 15

multiplex with prepped were Libraries sequencing. for information. library and Tablesamples mRNA-Seq5.1. prep library before (ng/ul), quantity and RNA RNA first inmeasured were collaborators accompanying Samples from with(ng/ul). received format, concentrationwere Samples ultrasound. by number) and integrity RNA reads), mapped (RIN; Right quality # forreads; mapped checked Left (# Left tallied # (tallied were also reads right and left were sequencing, paired-end During ID. mapped unique a assigned was that libraryeach and (Illumina), TruSeq using Reads adapters sequencing. of depth indicating reads), Right # reads; presented as a percentage of total left/right reads (Left %; Right %), with >70% indicating well-prepped library.

157 Figure 5.1. Expression of hCG genes. CGA and all six CGB genes were highly expressed in pre-flow placentae, with CGA, CGB8, CGB5, and CGB making up the top four genes. Average FPKM in pre-flow placentae: CGA, 149,334.26 FPKM; CGB8, 36,856.39 FPKM; CGB5, 36,272.68 FPKM; CGB, 28.500.73 FPKM. Horizontal axis: gene; vertical axis: FPKM.

158 5.3. Differential analysis and gene ontology.

The transcriptomes of all samples were compared to identify differentially-expressed genes. The initial unsupervised clustering analysis of all 15 samples clearly clustered the majority of samples according to their pre-flow or post-flow nominal status. The one exception was RHM077 which clustered with the pre-flow samples. Flow was not measured in these placentae prior to CVS and the exact timing of trophoblast rupture is known to vary across a couple of weeks. RHM077 represents one of the earliest ‘post-flow’ samples (13 weeks + 0 days). Based on its clustering with pre- flow, I assumed this sample represented the not totally unexpected occurrence of a later onset trophoblast rupture. With its later staging but apparent pre-flow status, I chose to exclude this sample from all subsequent analysis.

The unsupervised clustering of all remaining samples, produced two subsets, aligning according to gestational age (Fig. 5.2). The first subset encompassed all samples of gestational age between 7 to 9 weeks (i.e. pre-flow), while the second subset included all samples that were at 13 to 15 weeks gestation (i.e. post-flow). Between these two subsets, 681 genes were differentially expressed (FDR<0.001; Fig. 5.2; complete list included in Appendix, Supplementary Table 1).

I then focused our gene ontology analysis based on this exclusive set of 681 genes, in order to examine gene sets representing major biological processes (Fig. 5.3). Blood coagulation pathways featured prominently across various databases (Complement and coagulation cascades, KEGG; Hemostasis, REACTOME; Blood coagulation,

PANTHER; p-value<0.05). Cell surface receptors and interactions, as well as adhesion molecules, were also represented in the gene ontology analysis (Integrins

159 and other cell-surface receptors, BBID; CAMs, KEGG; ECM-receptor interaction,

KEGG; p-value<0.05). Pathways of the immune system also emerged in the analysis

(Signaling in immune system, REACTOME; Antigen processing and presentation,

KEGG; p-value<0.05). The exhaustive list of genes is included in the Appendix

(Supplementary Table 2). The variation in expression of these genes therefore indicate the biological processes triggered by the onset of maternal blood flow, as well as processes actively taking place during this time of changing oxygen concentration in the placenta.

160 nuevsdcutrn lge ape codn ogsainl g (.. w2 7 7w+2: (e.g. age gestational to according samples aligned clustering Unsupervised Figure 5.2. Heatmap of differentially expressed genes. expressed differentially of Heatmap 5.2. Figure weeks and 2 days). 681 genes were differentially expressed between ʻ pre-flow ʼ and post-flow subsets (FDR<0.001).

161 ), followed by followed ), -6 -value = 6.03 x 10 x 6.03 = -value p ). -5 -value = 6.33 x 10 p ) and “Hemostasis” ( -5 -value = 2.56 x 10 p “Complement and coagulation cascades” was most highly represented ( represented highly most was cascades” coagulation and “Complement iue53 Gn nooyanalysis. ontology Gene 5.3. Figure “Integrins and other cell-surface receptors” (

162 5.3.1. Hemoglobin genes.

The hemoglobin genes HBZ (hemoglobin zeta subunit) and HBE1 (hemoglobin epsilon subunit 1) emerged as the two most differentially expressed genes (p-values =

4.74 x 10-32 and 1.42 x 10-29 respectively; Fig. 5.4A; Supplementary Table 1). Both genes were, on average, expressed approximately 230-times higher in pre-flow than in post-flow placental samples. This then prompted us to assess the expression of other hemoglobin genes. With the exception of the hemoglobin gene HBD (delta subunit), all five of the remaining known hemoglobin genes were found to be expressed in our samples (Fig. 5.5A); HBG1 (gamma subunit 1), HBG2 (gamma subunit 2), HBA1 (alpha subunit 1), HBA2 (alpha subunit 2), and HBB (beta subunit).

Hemoglobin genes encode the various polypeptide chains that make up the heterotetramer protein, with the specific combination of subunits defining the biological relevance and function of the hemoglobin protein (Collins and Weissman,

1984; Karlsson and Nienhuis, 1985; Higgs et al., 1989). Adult hemoglobin (HbA,

α2β2) comprises two alpha chains with two beta chains, fetal hemoglobin (HbF, α2γ2) constitutes two alpha chains and two gamma chains, while embryonic hemoglobin is predominantly made up of two epsilon chains and two zeta chains (ζ2ε2; Hecht et al.,

1966).

A secondary analysis determined contribution, by expression, of each individual hemoglobin gene (Fig. 5.4B). Samples were analyzed individually, with expression of each hemoglobin gene charted as a percentage of total hemoglobin gene expression.

A distinct trend of hemoglobin switching was observed; where in pre-flow placentae,

HBZ and HBE1 make up the predominant hemoglobin gene expression, while in post-

163 flow placentae, HBG2 alone makes up 40-60% of total hemoglobin gene expression, with no significant contribution by HBZ and HBE1.

164

Figure 5.4. Expression of hemoglobin genes indicates hemoglobin switching. (A) Scatter plots of hemoglobin genes expression. HBZ and HBE1 were differentially expressed between pre-flow and post-flow samples (*** indicates p-value<0.001). Horizontal axis: gestational age; vertical axis: FPKM. (B) Proportional analysis of hemoglobin gene expression, indicates that in pre-flow placentae, HBZ and HBE1 make up the predominant hemoglobin gene expression, while in post-flow placentae, HBG2 alone makes up 40-60% of total hemoglobin gene expression, with no significant contribution by HBZ and HBE1. Contribution of HBB gene expression across all samples fall under 10%.

165 5.3.2. Blood coagulation.

Several genes involved in the blood coagulation cascade were found differentially expressed between pre-flow and post-flow placentae. Three genes in particular,

PROCR, TFPI, and TFPI2, were significantly downregulated in the post-flow subset, while another three, A2M, F3, and F13A1, were significantly upregulated (Fig. 5.5A).

The gene F3 (coagulation factor III) encodes tissue factor (TF; Morrissey et al., 1987;

Spicer et al., 1987; Mackman et al., 1989), the cell surface glycoprotein which functions as the high-affinity receptor for factor VII, ultimately forming the factor

VIIa-TF complex (Konigsberg et al., 2001). Briefly, the coagulation process begins with the formation of a factor VIIa-TF complex, which then activates additional proteases, factors IX and X, ultimately leading to the formation of a fibrin clot (Rao and Rapaport, 1987; Davie et al., 1991).

The F13A1 gene encodes the A1 polypeptide of the A subunit of coagulation factor

XIII, the last protease precursor to become activated in the blood coagulation cascade

(Ichinose and Davie, 1988; Board et al., 1988). Factor XIII, a heterotetramer of 2 A and 2 B subunits, is activated upon cleavage by thrombin, yielding the active enzyme factor XIIIa (reviewed in Muszbek et al., 2011). This enzyme catalyzes the crosslinking between fibrin molecules, forming the stable fibrin clot (Adany and

Bardos, 2003; Bereczky et al., 2005).

TFPI (tissue factor pathway inhibitor) is implicated in the process through its role in inhibiting the activated factor X, forming the Xa-TFPI complex, which subsequently also inhibits the VIIa-TF complex (Broze and Miletich, 1987; Broze et al., 1988 and

1990; Girard et al., 1989; Novotny et al., 1989).

166 Meanwhile, the TFPI2 gene was designated to a previously known placental protein that was found to be homologous to TFPI (Sprecher et al., 1994; Petersen et al.,

1996). This protein, originally isolated from the placenta, exhibits serine protease inhibitory activity (Butzow et al., 1988; Miyagi et al., 1994). TFPI2 mRNA, along with evidence of protein synthesis (Seppala et al., 1983), has been identified in the syncytiotrophoblast from first trimester to term placenta (Udagawa et al., 1998), specifically on the surface of microvilli and the membrane of the endoplasmic reticulum (Udagawa et al., 2002). In contrast, TFPI is widely distributed in fetal and placental tissues such as syncytiotrophoblasts, cytotrophoblasts, extravillous trophoblasts and the vascular endothelium, while TFPI2 is abundant only in syncytiotrophoblasts (Edstrom et al., 2000; Udagawa et al., 2002; Aharon et al.,

2004).

Within the blood coagulation pathways, there are multiple other inhibitors that regulate the processes, including protein C (PROC), a serine protease that inactivates factors Va and VIIIa. PROC activity is enhanced when it binds its membrane-bound receptor (PROCR; Walker, 1981; Walker et al., 1987). However, PROCR also circulates in soluble form, and when it binds activated PROC in this form, it inhibits

PROC activity instead (Van de Wouwer et al., 2004).

A2M (alpha-2-macroglobulin) also functions as an inhibitor of coagulation by blocking thrombin and factor Xa (Downing et al., 1978; Ellis et al., 1982). It can conversely impede fibrinolysis by inhibiting plasmin (PLG) and kallikrein (KLK1B;

Harpel, 1970 and 1977; Schapira et al., 1982; van der Graaf et al., 1983; Kaufman et al., 1991). A2M may additionally act as a carrier protein through its binding with numerous growth factors and cytokines (reviewed in Borth, 1992).

167 (A)

(B)

Figure 5.5. Expression of genes involved in the blood coagulation cascade. (A) Expression levels determined by mRNA-Seq. PROCR, TFPI, and TFPI2, were significantly downregulated in the post-flow subset. A2M, F3, and F13A1, were significantly upregulated (*** indicates p-value<0.001). Horizontal axis: gene; vertical axis: FPKM. (B) KEGG pathway for complement and coagulation cascades (complement cascade not shown). F3/F7: factor VIIa-TF complex; TFPI: tissue factor pathway inhibitor; F13: factor XIII; PROC/PROCR: endothelial protein C (receptor); F5: factor Vl; F8: factor VIII; A2M: alpha-2-macroglobulin; PLG: plasmin; KLK1B: kallikrein.

168 5.3.3. Complement system.

Components of the complement cascade, C1QA, C1QB, C1QC, C1R, CFD, and C7, along with the complement receptor C3AR1, were upregulated, while inhibitors of the cascade, CD55 and CD59, were downregulated (Fig. 5.6).

The complement cascade can be triggered by three pathways; classical, alternative, or lectin pathway (reviewed in Janeway et al., 2001; Fig. 5.6B). The classical pathway is initiated by activation of the C1 complex, which comprises three subcomponents;

C1q, C1r, and C1s. The genes C1QA, C1QB, and C1QC, make up the C1q subcomponent (Sellar et al., 1991), while C1R specifies the C1r subcomponent

(Nguyen et al., 1988).

All three pathways converge at the C3-convertase stage. In the alternative pathway,

CFD (complement factor D) facilitates the formation of C3-convertase (Fig. 5.6B).

C3-convertase can be inhibited by DAF (delay accelerating factor), specified by the

CD55 gene (Lublin et al., 1987). Past this stage, the complement cascade forms a membrane attack complex (MAC), comprising C6, C7, C8, and multiple C9 components, ultimately leading to lysis of the pathogenic cell. MAC formation can be inhibited by CD59, a complement regulatory protein (Huang et al., 2006).

Another inflammatory outcome is also possible, mediated by C3a, C4a, and C5a, which trigger chemotaxis as well as increase vascular permeability and smooth muscle contraction (Janeway et al., 2001). This is facilitated by membrane receptors for these components; C3ar1 and C5r1.

169 (A)

(B) Complement cascade

Figure 5.6. Expression of genes involved in the complement cascade. (A) Expression levels determined by mRNA-Seq. Components of the complement system, C1q (C1QA, C1QB, C1QC), C1R, CFD, and C7, are upregulated post-flow, while inhibitors CD55/DAF and CD59 are downregulated. C3AR1, a complement receptor, is also upregulated. (*** indicates p-value<0.001). Horizontal axis: gene; vertical axis: FPKM. (B) KEGG pathway for complement and coagulation cascades (only complement cascade is shown). C1Q: q subcomponent of C1 complex; C1R: r subcomponent; C7: complement component 7; DF: complement factor D (CFD); DAF: delay accelerating factor/CD55; CD59: complement regulatory protein; C3AR1: complement component 3a receptor 1.

170 5.4. Genes most abundantly expressed in the placenta.

As a means to understand biological processes important in placental development, we also assessed genes that were most highly expressed in our samples. The top 30 genes expressed in pre-flow samples ranged from 429.04 FPKM to 149,334.26

FPKM (Table 5.2), and in post-flow ranged from 136.74 FPKM to 14,850.41 FPKM

(Table 5.3).

5.4.1. Chorionic somatomammotropin hormone 1.

CSH1 (chorionic somatomammotropin hormone 1) was strongly expressed in pre- flow placentae (4,757.77 average FPKM), and it was significantly upregulated post- flow (12,282.67 average FPKM; Fig. 5.7). Expression levels in post-flow samples were comparable to that of CGA and CGB, coming in second only to CGA (Table

5.3).

Also known as human placental lactogen, the protein encoded by CSH1 is a member of the somatotropin/prolactin family of hormones (Josimovich et al., 1963; Seeburg et al., 1977; Shine et al., 1977; Martial et al., 1979). CSH1 is expressed mainly in the placenta, playing an important role in growth control. It functions similarly to the human growth hormone, modifying the metabolism of the mother during pregnancy, in order to facilitate supply of energy to the fetus (reviewed in Newbern and

Freemark, 2011).

5.4.2. Growth differentiation factor 15.

GDF15 (growth differentiation factor 15) is abundantly expressed in both pre-

(6,080.27 average FPKM) and post-flow (894.56 average FPKM) placentae, with

171 significant downregulation in post-flow samples (Fig. 5.7). In pre-flow placentae, expression of GDF15 was the highest after the family of hCG genes (Table 5.2).

This gene is found abundantly expressed in human placenta (Fairlie et al., 1999), with its expression dysregulated both in preeclampsia and in diabetic pregnancies (Sugulle et al., 2009). The product of the gene belongs to the transforming growth factor beta

(TGFβ) superfamily, which has a role in regulating inflammatory and apoptotic pathways in injured tissues and during disease processes (Hsiao et al., 2000; Xu et al.,

2006).

5.4.3. Peptidylprolyl isomerase B.

The average expression of PPIB (peptidylprolyl isomerase B) in pre-flow samples was 1,502.67 FPKM. This was significantly downregulated (501.89 average FPKM) in the post-flow subset (Fig. 5.7). The protein encoded by the PPIB gene, also known as cyclophilin B (CypB), is mainly localized to the endoplasmic reticulum, extracellular space, and nucleus, with a role in secretory pathways (Peddada et al.,

1992; Price et al., 1994).

5.4.4. Epstein-Barr virus induced 3.

In our samples, the expression of EBI3 (Epstein-Barr virus induced 3) ranked 15th most abundant in the pre-flow subset (1,233.42 average FPKM; Table 5.2), but was downregulated in the post-flow subset (361.28 average FPKM; Fig. 5.7). EBI3 is a secreted glycoprotein, a member of the hematopoietin receptor family related to the p40-subunit of interleukin-12 (IL12). It is found as a subunit in two distinct heterodimeric cytokines, interleukin-27 (IL27) and -35 (IL35; Devergne et al., 1997;

Pflanz et al., 2002), and has a role in regulating cell-mediated immune responses. It is

172 expressed in the first trimester placenta with localization to the syncytiotrophoblast layer (Coulomb-L'Herminé et al., 2007). Interestingly, the other subunit of IL27, specified by the gene IL27, is also downregulated in post-flow samples, although not significantly (data not shown).

5.4.5. Ferritin light polypeptide.

FTL (ferritin light polypeptide) expression was upregulated in post-flow samples

(1,412.57 average FPKM), from 658.00 average FPKM in pre-flow (Fig. 5.7). The

FTL gene encodes the light subunit of the ferritin protein, the major intracellular iron storage protein, keeping iron in a soluble and non-toxic state (Munro et al., 1988;

Gasparini et al., 1997). Abnormal expression of the FTL gene has also been observed in fetuses conceived by assisted reproductive techniques, namely in vitro fertilization and intracytoplasmic sperm injection (Zhang et al., 2008).

173 Table 5.2. Pre-flow expression of genes. Average FPKM values of top 30 genes expressed in pre-flow samples.

174 Table 5.3. Post-flow expression of genes. Average FPKM values of top 30 genes expressed in post-flow samples.

175 Figure 5.7. Genes abundantly expressed in the placenta. These genes were expressed abundantly in the placental samples, ranking within the top 30 in both pre-flow and post-flow subsets. CSH1 and FTL were upregulated post-flow, while the other three genes, GDF15, PPIB, and EBI3, were downregulated post-flow (*** indicates p-value<0.001). Horizontal axis: gene; vertical axis: FPKM.

176 5.5. Genes associated with preeclampsia and other pregnancy complications.

We found at least seven genes known to be associated with preeclampsia, that were differentially expressed between pre- and post-flow placental samples. LEP, INHA,

INHBA, and ENG were downregulated in post-flow placental samples, while SPP1, and HGF were upregulated (Fig. 5.8). A novel isoform of FLT1, not mapped to

Ensembl transcripts, was also found downregulated, through a separate analysis of novel genes and/or isoforms expression (analysis by Dr. Li Juntao; refer to Fig. 4.11).

These genes are early-onset placental genes, whose dysregulation have been implicated in unsatisfactory pregnancy outcomes.

5.5.1. Leptin.

The product of the LEP (leptin) gene is an adipose-derived hormone that plays a key role in the regulation of energy intake and expenditure (Ingalls et al., 1950; Zhang et al., 1994; Isse et al., 1995; ). The major source of leptin is adipose tissue, but it can also be produced by syncytiotrophoblasts in the placenta and ovaries (Masuzaki et al.,

1995; Cioffi et al., 1997; Hassink et al., 1997; Masuzaki et al., 1997; Senaris et al.,

1997; Henson et al., 1998). Placental LEP expression is markedly augmented in preeclamptic pregnancies (Mise et al., 1998), as well as pregnancies complicated with diabetes mellitus (Lepercq et al., 1998; Lea et al., 2000).

5.5.2. Inhibin genes.

The INHA (inhibin, alpha) gene encodes the alpha subunit of inhibins A and B protein complexes, while the INHBA (inhibin, beta A) gene encodes the beta subunit of inhibin A (Stewart et al., 1986). The inhibin alpha subunit combines with either the beta A or beta B subunit to form a complex that negatively regulates follicle-

177 stimulating hormone (FSH) secretion from the pituitary gland (Welt et al., 1997;

Lahlou et al., 1999; Welt et al., 2003). Additionally, the beta A subunit forms a homodimer, known as activin A, and also heterodimerizes with the beta B subunit, forming activin AB, both of which stimulate FSH secretion (Ling et al., 1986; Vale et al., 1986). Inhibin A, the complex comprising INHA and INHBA, and activin A are glycoproteins secreted during pregnancy (Schneider-Kolsky et al., 2002), circulating in higher amounts in women with gestational hypertension and/or preeclampsia, as a result of increased placental production (Manuelpillai et al., 2001; Florio et al., 2002;

Casagrandi et al., 2003).

5.5.3. Endoglin and fms-related tyrosine kinase 1.

ENG (endoglin), a cell surface receptor for TGF-β1 (Lastres et al., 1996), exhibits increased expression in preeclamptic pregnancies (Yang et al., 2008; Tannetta et al.,

2013). ENG was also found to induce (in pregnant rats) vascular permeability and hypertension, mimicking maternal symptoms of preeclampsia (Venkatesha et al.,

2006). It is postulated to act in a concerted manner with sFlt1 to cause severe preeclampsia, with restriction of fetal growth (Maynard et al., 2003; Venkatesha et al., 2006).

FLT1 (fms-related tyrosine kinase 1) is a receptor tyrosine kinase belonging to the family of vascular endothelial growth factor receptors (VEGFR). It is involved in angiogenesis and vasculogenesis, and found expressed in vascular endothelial cells, placental trophoblasts (Helske et al., 2001) and peripheral blood monocytes

(reviewed in Shibuya, 2006). Known isoforms include a full-length transmembrane receptor isoform, as well as a shortened, soluble isoform (i.e. sFlt1; Shibuya et al.,

178 1990; Kendall and Thomas, 1993). Excess levels of the latter has been associated with the onset of preeclampsia (Koga et al., 2003; Maynard et al., 2003; Levine et al.,

2004).

5.5.4. Secreted phosphoprotein 1.

The SPP1 gene specifies secreted phosphoprotein 1 (SPP1), also known as osteopontin (OPN), a protein involved in bone mineralization (Kiefer et al., 1989;

Reinholt et al., 1990; Choi et al., 2008). SPP1/OPN also upregulates expression of the pro-inflammatory cytokines, interleukin-12 (IL12) and interferon-gamma (IFN-γ), and reduces production of the anti-inflammatory cytokine, interleukin-10 (IL10;

Ashkar et al., 2000; Wang et al., 2008). SPP1 expression was found upregulated in preeclamptic placenta (Gabinskaya et al., 1998; Tsoi et al., 2003).

5.5.5. Hepatocyte growth factor.

HGF (hepatocyte growth factor) regulates cell growth, cell migration, and morphogenesis by activating a tyrosine kinase signaling cascade (Bottaro et al., 1991;

Naldini et al., 1991; Rubin et al., 1993). HGF is known to have a major role in embryonic organ development, adult organ regeneration, and wound healing

(reviewed in Skibinski, 2003). HGF gene expression was diminished in preeclampsia

(Wolf et al., 1991; Kauma et al., 1997).

179 Figure 5.8. Expression of genes implicated in preeclamptic pregnancies. LEP, INHA, INHBA, and ENG were downregulated in post-flow placental samples, while SPP1, and HGF were upregulated (*** indicates p-value<0.001). Horizontal axis: gene; vertical axis: FPKM. A novel isoform of FLT1, not mapped to Ensembl transcripts, was also found downregulated.

180 5.6. Discussion.

Through quantifiable transcriptomic analysis, we were afforded unique insight into human placental gene expression across a time frame of early pregnancy development. This window traverses the earliest major environmental change in the site of gestation, triggered by the onset of maternal blood flow into the placenta, entailing a switch in oxygen conditions as well as mode of nutrition.

We approached the study by means of differential analysis, choosing to focus on genes whose expression levels were changing significantly between two selected time points. These time points are reflective of the changing uterine environment; the earlier time point was during the first trimester of pregnancy, prior to the flow of maternal blood into the placenta, while the later time point was at the beginning of the second trimester, just after the onset of maternal blood flow. Our rationale behind this approach was that given the altered conditions, fluctuations in gene expression would be indicative of biological processes and pathways activated or repressed, in wake of the introduction of maternal blood, responsible either for adaptation to the new environment, or perhaps even the commencement of the next phase of development.

5.6.1. Hemoglobin switching by the placenta.

Our data has the potential to identify genes that are changing significantly during this transition. The top two genes that emerged with the most significant difference in expression levels were the hemoglobin genes HBZ and HBE1. Upon assessment of the expression of other genes in the hemoglobin family, it was evident that the expression levels reflected a switch from embryonic hemoglobin (ζ2ε2) to fetal

181 th th hemoglobin (HbF, α2γ2), at approximately 9 to 12 weeks of pregnancy. The zeta and epsilon chains, specified by HBZ and HBE1 respectively, were no longer being expressed with any significance, falling to <1%, from the 13th week onwards.

It was first postulated by Brinkman and colleagues (1934) that “human hemoglobin is changing from one form to the other”, and to date it is recognized that the first major hemoglobin switch occurs as ζ- and ϵ-globin expression ceases, and α- and γ-globin synthesis begins, leading to production of HbF (Peschle et al., 1985; Ley et al., 1989).

These events are coincident with the transition of the site of erythropoiesis from the yolk sac, in the early embryo, to the fetal liver (Migliaccio et al., 1986). The second switch involves the perinatal decline of HbF synthesis coupled with the increased synthesis of the adult form of hemoglobin (HbA, α2β2). Unlike the well-studied fetal- to-adult globin transition, little is understood about the control of the embryonic-to- fetal globin transition. Nevertheless, multiple studies have indicated, by means of comparative globin gene expression, that the embryonic-to-fetal globin switch occurs after 10 weeks of gestation. One study found that the relative abundance of ϵ-globin mRNA transcripts in mid-gestational human fetal liver was less than 1% that of γ- globin, and ζ-globin transcripts less than 5% that of α-globin (Ley et al., 1989). More recently, Al-Mufti and colleagues (2000) reported that 25% of erythroblasts express the ζ-chain at 10 weeks, but decreased exponentially with gestation to less than 1% by 17 weeks. Similarly, expression of the ϵ-chain decreased from 97% of erythroblasts at 10 weeks to less than 1% by 25 weeks. In contrast, expression of the

γ-chain increased from about 30% at 10 weeks to 90% by 16 weeks. From our study meanwhile, we established that this switch happened as early as the 13th week of gestation.

182 The embryonic to fetal hemoglobin switch has never been associated with the pre- and post- flow placenta. Here I have shown that this switch tightly correlates with this trophoblast plug rupture-induced physiological change. As the major physiological change that the chorionic villi see is a switch from low oxygen to oxygen-rich maternal blood, it is intriguing to note that embryonic hemoglobins have a higher affinity for oxygen than do fetal or adult hemoglobins (Brittain, 2002) with 4, 12 and

6 mm Hg for the three human embryonic proteins compared to 20 and 26 mm Hg for the human fetal and adult proteins. I would thus speculate that this embryonic to fetal hemoglobin switch is functionally linked to trophoblast plug rupture. With low oxygen prior to rupture, the placenta requires hemoglobin with higher affinity to capture oxygen compared to when oxygen-enriched maternal blood is in immediate contact with the syncytiotrophoblast.

Additionally, EPAS1 has been implicated in hemoglobin regulation (Beall et al.,

2010; Yi et al., 2010), although the exact molecular mechanism has not been elucidated. The presence of oxygen post-flow, and its subsequent effect on EPAS1 activity, could also likely mediate the changes seen in hemoglobin gene expression.

However, because our hESC-derived trophoblast culture (see Chapter 4) does not express hemoglobin, we are unable to confirm if EPAS1 directly targets hemoglobin genes, but this does not exclude the possibility of EPAS1 regulating hemoglobin gene expression in chorionic villi.

Overall, our results fully corroborate the current understanding of hemoglobin switching. Our data reveals gene expression in chorionic villi of the placenta, admittedly comprising pooled populations of various cell types. Although the role of the placenta in erythropoiesis has yet to be clearly established, recent studies have

183 suggested that it has a broader role as a hematopoietic organ than previously appreciated, functioning as an integral component of both primitive (embryonic) and definitive (fetal/adult) hematopoiesis (Rhodes et al., 2008; Barcena et al., 2009a &

2009b; Robin et al., 2009; van Handel et al., 2010). Barcena and colleagues (2009) successfully isolated, from human placental chorionic villi, multipotent primitive hematopoietic progenitors as well as intermediate progenitors, which is indicative of active hematopoiesis. The population of these progenitor cells increased throughout early gestation, peaking at 5 to 8 weeks, then decreasing more than 7-fold from the 9th week onwards. Additionally, van Handel et al. (2010) identified ζ-globin+ primitive erythroid cells in placental villi between 5 to 7 weeks of development. Our findings are thus consistent with earlier studies, in which we found predominant expression of embryonic globin genes in human chorionic villi prior to 9 weeks of gestation.

Moreover, the loss of ζ- and ϵ-globin expression from 13 weeks onwards was met with predominant expression of α- and γ-globin genes, signifying the onset of fetal hematopoiesis. Our results therefore provide evidence in support of the placenta as a hematopoietic organ.

5.6.2. Changing state of the placenta: Anti-coagulatory to pro-coagulatory.

We also report the emergence of genes involved in the blood coagulation cascade as candidates important during the transition from first to second trimester. In the case of

PROCR, transcriptional data bears little insight to its activity as it is the protein form that dictates its contribution to the coagulation process. As a protein, it could act either as an enhancer in its membrane-bound form (Walker, 1981; Walker et al.,

1987), or as a hindrance, in its soluble form (Van de Wouwer et al., 2004).

184 Meanwhile, the upregulation of F3 and F13A1, components of the blood coagulation cascade (Morrissey et al., 1987; Spicer et al., 1987; Mackman et al., 1989;

Konigsberg et al., 2001), as well as the downregulation TFPI and TFPI2, both inhibitors of the TF-dependent pathway of the cascade (Broze and Miletich, 1987;

Girard et al., 1989; Novotny et al., 1989; Petersen et al., 1996; Udagawa et al., 2002), testifies to the increased activity of the blood coagulation cascade in the second trimester. F3, coding for tissue factor (TF), the potent initiator of the TF-dependent pathway, has also been found strongly expressed in syncytiotrophoblasts (Reverdiau et al., 1995; Aharon et al., 2004). Some suggest that this procoagulant tendency may reflect a physiological need for immediate inhibition of hemorrhage in the placental intervillous space (Lanir et al., 2003; Dusse et al., 2006). Moreover, the presence of coagulant activity in the amniotic fluid during the second and third trimester is a well- known phenomenon (Lockwood et al., 2000; Uszynski et al., 2001). Our data accordingly indicates that second-trimester placental villi tend toward a procoagulant state, in response to the onslaught of maternal blood, likely in a bid to minimize risk of hemorrhage.

Conversely, in the first trimester, the upregulation of TFPI and TFPI2 reflects an anti- coagulant environment during this time. Having previously established the expression of TFPI2 in syncytiotrophoblasts (Udagawa et al., 1998), Udagawa and colleagues

(2002) later examined the association between the presence of the syncytiotrophoblast layer and coagulation using clinical specimens, revealing that fibrin deposits were specifically observed in regions devoid of the syncytiotrophoblast layer in placental villi. Such local inhibitory mechanisms limit coagulation and fibrin deposition, in favor of patency for intervillous blood flow

185 (Udagawa et al., 2002; Lanir et al., 2003). This anti-coagulant environment is composed in preparation for the onset of maternal blood flow, thus enhancing the flow of blood into the placenta. This phenomenon is well in agreement with endovascular trophoblast invasion and remodeling of the spiral artery into dilated inelastic tubes, reducing flow resistance and increasing uteroplacental perfusion to the fetus (Brosens et al., 1967; Robertson and Manning, 1974; Robertson, 1976;

Hirano et al., 2002).

The increased expression of the anti-protease A2M suggests a propensity toward the inhibition of the blood coagulation cascade in the second trimester, conflicting with the earlier justification of a procoagulant state. On the contrary however, A2M, an anti-protease, inactivates a large variety of proteases, not exclusive to the components of the blood coagulation cascade alone. It also inhibits plasmin and kallikrein

(Harpel, 1970 and 1977; Schapira et al., 1982; van der Graaf et al., 1983; Kaufman et al., 1999), hampering fibrinolysis, a process in which the fibrin clot, the product of coagulation, is broken down. In this regard, A2M is also a contributor to the procoagulant state of the second trimester placenta.

5.6.3. Complement system and pregnancy.

Complement system activation has been associated with pregnancy (Baines et al.,

1974; Hopkinson and Powell, 1992; Richani et al., 2005). Complement products are deposited on placental tissue, in both physiological as well as pathological pregnancy

(Faulk et al., 1980; Weir, 1981; Tedesco et al., 1990). It has been postulated that the complement system plays a key role in protecting the fetus, at the feto-maternal

186 interface, from potential pathogens invading the uteroplacental unit (Girardi et al.,

2006; Girardi et al., 2011).

Our data, beginning at ~8 weeks and leading to ~14 weeks, indicates upregulation of the complement pathway during second trimester, indicated by the increased expression of complement components C1q, C1r and CFD, which culminate in the activation of C3-convertase through both alternative and classical pathways. At this juncture, we find downregulation of DAF/CD55, an inhibitor of C3-convertase, as well as upregulation of the C3a complement receptor. Further down the cascade, C7 was also found upregulated, while the inhibitory CD59 was downregulated (Fig. 5.6).

Overall, our data suggests that activity of the complement system increases between first and second trimester, coinciding with the onset of maternal blood. The activation of the complement system during this time could likely provide immuno-protection from potential pathogens found in maternal circulation.

However, this increased activity has to be properly controlled and/or limited, as excessive complement activation in the placenta has been associated with unsatisfactory pregnancy outcomes (Sinha et al., 1984; Shamonki et al., 2007; Lynch et al., 2008; Buurma et al., 2012). Trophoblasts express complement regulatory proteins, such as DAF/CD55 and CD59, from at least 6 weeks of gestation until term, likely serving to regulate the complement pathway (Holmes et al., 1992; Xu et al.,

2000). Dysregulation of the complement pathway has been associated with preeclampsia, IUGR, and fetal rejection (Sinha et al., 1984; Tedesco et al., 1990;

Girardi et al., 2006; Lynch et al., 2008; Buurma et al., 2012), although the origins of dysregulation remain unknown. Sinha et al. (1984) found increased deposition of

C1q, C3d, and C9 in preeclamptic tissue. Similarly, Wang et al. (2012) demonstrated

187 elevated C3 deposition in preeclamptic placentae, and also reported that complement

3a receptor activation is a key mechanism to sFlt1 secretion and subsequent decreased angiogenesis. An earlier study had also linked complement activation with increased sFlt1, a potent anti-angiogenic protein associated with placental dysfunction (Girardi et al., 2006). Buurma et al. (2012) found upregulated mRNA expression of complement regulatory proteins CD55 and CD59 in preeclampsia. It was postulated that complement regulatory proteins were upregulated in response to excessive complement activation, but their regulatory capacity may nevertheless be overwhelmed, leading to development of pathological pregnancy (Denny et al.,

2012).

In conclusion, there is evidence for increased classical pathway activation and altered complement regulation in preeclampsia, leading to angiogenic factor imbalance.

These data indicate that immunological mechanisms are operating within the chorionic villi of the placenta during preeclampsia, suggesting that among other mechanisms, immune processes may be operative in the pathophysiology of this disease (Sinha et al., 1984).

5.6.4. Recognizing important placental genes.

In our study, we next chose to focus on genes that were highly expressed in the placenta, with the understanding that the magnitude of expression is reflective of their role in the development and maintenance of pregnancy. These include the hCG genes, a family well known and equally well studied for their involvement in pregnancy, as well as five other genes, CSH1, GDF15, PPIB, EBI3, and FTL.

188 It was no surprise that genes specifying the human chorionic gonadotropin hormone

(hCG) emerged as most abundant. hCG is essential to pregnancy success, by playing a part in the production of steroid hormones and other growth factors in the corpus luteum (Jaffe et al., 1969; Hay and Lopata, 1988; Rao, 2001), in the modulation of blastocyst implantation as well as human trophoblast differentiation and fusion

(Srisuparp et al., 2001; Yang et al., 2003; Handschuh et al., 2007), in uterine vascularization and angiogenesis (Zygmunt et al., 2002; Berndt et al., 2006; Bourdiec et al., 2012), and uterine quiescence and immunological adaptation during pregnancy

(Toth et al., 2001). hCG is synthesized by the trophoblast shortly after fertilization, and later, by placental tissue (Ruddon et al., 1981; Hay, 1988; Hay and Lopata, 1988). Levels of hCG in maternal serum rise rapidly during early gestation, peaking at 10 to 12 weeks, declining steadily thereafter (Diczfalusy, 1953; Midgley and Jaffe, 1968). In the third and final trimester, hCG level rises in a gradual, yet significant, manner until term

(Boothby et al., 1983; Hay, 1988). Multiple studies have reported that placental levels of hCG-β peak at 9 to 10 weeks of pregnancy, and fall thereafter up to the 16th week

(Bo and Boime, 1992; Jameson and Hollenberg, 1993; Miller-Lindholm et al., 1997).

Our results show a general trend of increasing levels of hCG expression as gestational age approaches 9 weeks, and at 13 weeks onward, indicate a decrease in expression levels, across CGA and all CGB genes, well in accordance with previously observed patterns.

Similarly, we established expression profiles for five other genes whose respective expressions were highest in the placental samples. These genes include those that have long been known to be expressed during pregnancy, such as CSH1 and FTL, as

189 well as more recent discoveries of placental genes, like GDF15 and EBI3. The effects of CSH1 work in parallel, increasing maternal food intake, and at the same time promoting maternal beta-cell expansion and insulin production to defend against development of gestational diabetes mellitus (Mannik et al., 2012). Based on our results, CSH1 expression in the placenta increased as pregnancy entered the second trimester, in keeping with the metabolic demands of the growing fetus, which is thus in line with its known functions (reviewed in Newbern and Freemark, 2011; Mannik et al., 2012). Likewise, our results show that the expression of FTL increased with gestational age, thereby increasing cellular iron stores to meet the demands of the growing fetus (Scholl, 2005). The FTL expression identified in our samples represents placental isoferritin, the intracellular iron storage protein found in the cytoplasm of placental syncytiotrophoblasts (Li et al., 2008). Levels of placental isoferritin in circulation have been correlated with pregnancy outcome. Multiple studies reported that elevated levels of placental isoferritin during the second trimester were predictive of early spontaneous preterm delivery (Goepel et al., 1988;

Maymon et al., 1989; Goldenberg et al., 1996; Tamura et al., 1996), although this was likely a reflection of an acute-phase reaction to subclinical infections, which are closely associated with premature delivery (Tamura et al., 1996).

Meanwhile, for GDF15, although its expression has been identified in the placenta

(Fairlie et al., 1999; Herse et al., 2007; Sugulle et al., 2009), its biological role during pregnancy has yet to be understood. Nevertheless, it is known to promote angiogenesis, under hypoxic conditions, in human umbilical vein endothelial cells, by modulating HIF1α signaling (Song et al., 2012). It thus stands to reason that GDF15 could likely be a contributor to angiogenesis in placental development during first

190 trimester, when oxygen is low. Moreover, its dysregulation in preeclamptic, as well as diabetic, pregnancies signifies its importance in a successful and healthy pregnancy

(Herse et al., 2007; Sugulle et al., 2009).

EBI3 too is highly expressed in placental cells during pregnancy (Devergne et al.,

2001), and also differentially expressed in preeclamptic placentae (Enquobahrie et al.,

2008; Kang et al., 2011; Mayor-Lynn et al., 2011; Nishizawa et al., 2011).

Interestingly, EBI3 and IL27, which together make up the heterodimeric cytokine

IL27, were both more lowly expressed post-flow, although the change in the latter was not significant. IL27 has a regulatory role in the adaptive immune response, which is responsible for recognizing foreign antigens (Janeway et al., 2001). In pregnancy, the maternal immune system faces a major challenge in the tolerance of fetal allo-antigens, encoded by paternal genes. The downregulation of EBI3, could therefore likely help maintain maternal immune tolerance toward the fetus (Coulomb-

L’Hermine et al., 2007; Louwen et al., 2012).

Additionally, our study also demonstrated potential in identifying novel genes important in pregnancy. Our findings revealed the expression of genes previously not recognized during pregnancy, for example PPIB. The PPIB gene thus far is only postulated to be implicated in pregnancy through its function as a chaperone for the nuclear transport and action of the somatotropin/prolactin hormone family (Rycyzyn et al., 2000; Fang et al., 2010). Given the integral position of this hormone family in pregnancy, it stands to reason that its regulator would emerge as an essential gene transcribed during pregnancy.

191 5.6.5. Influence of the uterine environment on preeclampsia pathogenesis.

Of the genes already discussed, three are known to be implicated in pregnancy complications; GDF15 expression is dysregulated in preeclampsia and diabetic pregnancies (Sugulle et al., 2009), PROCR mutations have been associated with late fetal loss (Franchi et al., 2001), and factor XIII (F13A1) deficiencies can result in recurrent spontaneous miscarriage (Ichinose et al., 2005; Asahina et al., 2007). Our differential analysis unveiled at least six other genes associated with preeclampsia, revealing the influence of the uterine environment on their regulation.

Firstly, LEP, INHA, INHBA, ENG, and FLT1 (albeit a novel isoform) were downregulated in the second trimester. While the placental expression of LEP has been widely demonstrated (Hassink et al., 1997; Masuzaki et al., 1997; Senaris et al.,

1997; Henson et al., 1998), no study thus far has tracked its expression levels throughout the course of pregnancy. Regardless, LEP expression is markedly augmented in preeclamptic placenta in comparison to gestational age-matched healthy controls (Mise et al., 1998). Multiple studies also suggest a close correlation between fetal growth restriction and elevated LEP levels found in maternal serum

(Schubring et al., 1996; Mise et al., 1998; Lepercq et al., 1999; Ong et al., 1999).

Overall, LEP is postulated to be involved in the regulation of fetal growth, placental angiogenesis, and mobilization of maternal fat (Wauters et al., 2000; Hoggard et al.,

2001), implying that dysregulation of leptin metabolism and/or leptin function in the placenta would result in the pathologies of pregnancy disorders, such as gestational diabetes, IUGR, and preeclampsia (Bajoria et al., 2002; Sagawa et al., 2002).

192 Interestingly, in a study done to elucidate the mechanisms for augmented placental production of leptin in preeclampsia, the authors found that leptin secretion was enhanced in extended hypoxic conditions (Mise et al., 1998). The authors thus proposed that placental insufficiencies, which plague preeclampsia, lead to chronic hypoxia of the trophoblast cells, therefore resulting in augmented LEP production. In this scenario, the same placental insufficiencies also result in chronic disturbance of nutrient and oxygen supply to the fetus, thereby leading to IUGR. Our study offers support to their hypothesis in the form of oxygen-regulated expression of placental

LEP; we observed downregulation of LEP in the second trimester, following perfusion of the placenta by oxygenated maternal blood. Whether or not increased oxygen levels directly triggered this change however remains to be seen, since LEP expression is also induced by hCG (Maymo et al., 2012), which we found downregulated as well.

INHA and INHBA, together make up inhibin A, while a homodimer of INHBA forms activin A (Stewart et al., 1986). Several studies have showed that levels of inhibin A peaked at 8 to 10 weeks gestation, declining thereafter, but rose gradually again throughout the third trimester (Muttukrishna et al., 1995; Illingworth et al., 1996;

Birdsall et al., 1997). Our data demonstrates a similar trend for the expression of

INHA and INHBA in the first and second trimester. While inhibins and activins have been shown to regulate gonadotropin-releasing hormone, hCG, and progesterone secretion from cultured human placental cells (Petraglia et al., 1989), the biological role of activin A and inhibin A in early pregnancy however remain unclear.

Nevertheless, further in vivo studies have revealed that activin A is capable of inducing cell migration and invasion, thus regulating human cytotrophoblast

193 differentiation (Caniggia et al., 1997; Bearfield et al., 2005). However, evidence has shown elevated placental expression of both inhibin A and activin A in preeclampsia

(Jackson et al., 2000; Manuelpillai et al., 2001; Florio et al., 2002; Casagrandi et al.,

2003; Yu et al., 2012), which is inconsistent with the invasion-promoting effect of activin A, given the restricted invasive properties of trophoblast cells in preeclamptic placenta (Yu et al., 2012). On the other hand, Yu and colleagues (2012) recently demonstrated that high levels of activin A, acting via Nodal signaling, induce apoptosis in trophoblast cells (Munir et al., 2004; Nadeem et al., 2011), which could conceivably lead to the preeclamptic outcome. Yet, it might be prudent to note that it is not fully understood if activin A and inhibin A play a role in the pathology of preeclampsia, or if their expression is merely a manifestation of the disease. To that end, an understanding of their regulation would be necessary to elucidate the underlying mechanism of their function. Our findings mark the first step, in which we can begin to understand how the uterine environment regulates their expression, and how this might go awry in a condition like preeclampsia.

Endoglin (ENG) and sFlt1 are both endogenous anti-angiogenic factors (Park et al.,

1994; Toporsian et al., 2005). Endoglin, expressed in vascular endothelial cells and placental trophoblasts, is a surface receptor for TGF-β1 (Lastres et al., 1996), and in this manner it impairs binding of TGF-β1 to its other receptors, thus leading to dysregulation of TGF-β1 signaling and subsequent effect on vasodilation (Yang et al.,

2008). sFlt1 meanwhile binds to VEGF and placental growth factor (PGF), blocking their angiogenic effect on vascular tissue (Park et al., 1994; Shibuya, 2006; Yang et al., 2008). PGF plays an important role in enhancing VEGF-driven angiogenesis; the

194 loss of which impairs angiogenesis, wound healing, and inflammatory responses

(Autiero et al., 2003; Nagy et al., 2003).

Our data demonstrates the downregulation of endoglin and sFlt1 in the second trimester, which in preeclamptic pregnancies, were upregulated in both placenta and serum (Koga et al., 2003; Maynard et al., 2003; Venkatesha et al., 2006; Yang et al.,

2008). Excess levels of endoglin and sFlt1 may offset VEGF, PGF, and TGF-β1 activity, impairing maternal vascular endothelial functions and inducing glomerular endotheliosis (Maynard et al., 2003; Venkatesha et al., 2006), both of which have been implicated as major components in the pathogenesis of preeclampsia (Levine et al., 2004; Levine et al., 2006). Moreover, Yang et al., (2008) also demonstrated that trophoblast cultures from preeclamptic placentae produced more endoglin and sFlt1 than those from healthy placentae, and that additionally, hypoxic conditions promoted endoglin and sFlt1 production. In the context of preeclampsia, with its concomitant placental insufficiencies, the hypoxic conditions may only exacerbate the pathogenesis.

Interestingly, thrombin, a mediator of blood coagulation, has been associated with sFlt1 regulation (Lockwood et al., 2007; Zhao et al., 2012). We report that the placental environment undergoes a switch from anti- to pro-coagulatory state, implying the increase in activity of thrombin between first and second trimesters.

Although this is not reflected in gene expression, the precursors to thrombin and its inhibitors are accordingly regulated, facilitating the switch to a pro-coagulatory environment. Thrombin has been found to promote sFlt1 expression in decidual cells

(Lockwood et al., 2007) as well as trophoblasts (Zhao et al., 2012), and with the decreased activity/presence of thrombin in first trimester, this indicates that sFlt1 is

195 adherently regulated and not found in excess. This sits well with the development of the placenta, as during this time, the activity of the anti-angiogenic sFlt1 would be detrimental. It would otherwise promote preeclampsia by interfering with local vascular transformation. This exemplifies the interconnectedness between components of the biology of pregnancy, and thus the need for holistic regulation, as dysregulation of one component (i.e. blood coagulation) has far-reaching repercussions (i.e. angiogenesis, which is outside of blood coagulation) that may affect the health of the entire system (i.e. pregnancy). Moreover, the increased activity of the coagulation cascade in the second trimester would promote sFlt1 expression, and it is thus also necessary for fine regulation to ensure that sFlt1 is not produced in excess, as the placenta continues to grow and remodel throughout pregnancy. This regulation, which we postulate to be oxygen-mediated EPAS1 regulation of FLT1, is implied in our findings as even with increased activity of blood coagulation, sFlt1 expression is downregulated.

Secondly, our study demonstrates the upregulation of SPP1 and HGF in the second trimester. The SPP1/OPN protein is an extracellular matrix protein implicated in cell adhesion, spreading, and invasion (Franzen and Heinegard, 1985; Fisher et al., 1987;

Johnson et al., 2003). It is expressed in the placenta, with evidence for its role in regulation of placental invasiveness (Young et al., 1990; Gabinskaya et al., 1998;

Briese et al., 2005). In first trimester placenta, SPP1/OPN was present in the villous trophoblast, with strongest expression in the cytotrophoblast. Briese and colleagues

(2005) also reported that SPP1/OPN presented strong expression in EVTs at the invasion front of normal placenta, colocalizing with CEACAM1 (carcinoembryonic antigen-related cell adhesion molecule 1). Both SPP1/OPN and CEACAM1 interact

196 with integrin β3, playing an important role in the implantation and invasion process

(Brummer et al., 2001; Damsky et al., 1994). SPP1/OPN however is not expressed in the syncytial trophoblast (Daiter et al., 1996).

Our results show an upregulation of SPP1 from the 13th week onwards, upon perfusion of the placenta, which indicate an affinity for invasiveness in the presence of oxygenated maternal blood, therein implying the ability of the trophoblast in oxygen sensing and in response, placental remodeling. While the mechanism employed by trophoblast cells for oxygen sensing is yet to be fully understood

(Hunkapiller and Fisher, 2010), an investigation of the effect of oxygen on trophoblast invasion, both endovascular and interstitial trophoblast invasion were reduced in arteries at 3% oxygen, compared to arteries at 17% oxygen (Crocker et al.,

2005). Logically then, the presence of oxygen accompanying maternal blood flow into the placenta would elicit an increase in invasion-mediating proteins in response.

HGF expression has been found in the syncytial and extravillous trophoblasts, as well as the mesenchyme cells of the villous core, while the expression of its receptor Met was localized mainly to cytotrophoblasts (Wolf et al., 1991; Kauma et al., 1997).

Thus it was postulated that placental HGF acts in a paracrine manner on trophoblasts expressing Met. HGF, by nature of its potential in stimulating the invasion of multiple epithelial cell types (Tsarfaty et al., 1994; Kalluri and Neilson, 2003), is also thought to regulate placental trophoblast invasion. In fact, HGF-knockout mice were found to be embryonic-lethal due to abnormal placental development (Uehara et al., 1995). A study done by Kauma et al. (1999) showed that HGF expression was diminished in placentae from preeclamptic pregnancies, in line with the inadequate trophoblast invasion that is characteristic of this condition. The authors thus proposed that

197 decreased placental production of HGF in preeclampsia provides a potential mechanism for shallow trophoblast invasion that is seen in this pregnancy disorder

(Kauma et al., 1999). Our findings demonstrate the upregulation of HGF at the beginning of the second trimester, and as with the case of SPP1, increased HGF during this time indicates invasiveness on the part of the trophoblast for placental remodeling, in response to the presence of maternal blood.

198 5.7. Concluding remarks.

Like our study, two other studies have sought to understand the dynamic regulation of genes throughout pregnancy (Mikheev et al., 2008; Sitras et al., 2012). Sitras and colleagues (2012) juxtaposed placental gene expression between late first trimester

(9-12 weeks) and that at term, while Mikheev and colleagues (2008) compared expression at mid-first trimester (6-8 weeks) and second trimester (approximately 16 weeks). Our study is notable in that we characterized gene expression across the shortest time period (7-9 weeks vs.. 13-15 weeks), capturing the changing biology occurring during the most drastic transformation of the uterine environment - placental perfusion and oxygenation.

In summary, our study derived a set of genes showing significant change in expression level between late first trimester and early second trimester of pregnancy, two time points which were chosen to contrast the situation prior to and after the perfusion of the placenta by maternal blood. Of the genes identified, we found six genes implicated in the blood coagulation cascade, and nine associated with the complement cascade. Additionally, we were able to determine the first-to-second trimester expression profile of multiple genes known to be important in pregnancy.

Many of these genes also have prior implications in preeclampsia and other pregnancy complications, including gestational diabetes and IUGR.

The data set from this study stands strongly and independently as a tool to further understand the intrauterine environment and its concomitant changes throughout pregnancy. From our assessment of differential gene expression, we postulate that the blood-devoid first trimester placenta is placed in an anti-coagulant state, later

199 switching to a pro-coagulant state during the second trimester in the presence of maternal blood. The former setting facilitates entry of blood into the placenta, allowing a smooth undisturbed flow, while in the latter, the placenta readies itself to limit blood loss, due to trauma or otherwise. In addition, our data unveiled the production of hemoglobin genes, by cells of the chorionic villi, switching from embryonic to fetal globin between first and second trimester, supporting the role of the placenta as a hematopoietic organ. We were also able to analyze the expression of a well known set of pregnancy-associated genes, the hCG family, and in doing so, identified the expression profiles of respective genes throughout the transition between first and second trimester.

Our findings highlighted the top most highly expressed genes in the placenta that were differentially expressed across this transition. Four of these genes, CSH1,

GDF15, EBI3, and FTL, have already been recognized as important in pregnancy, while the last, PPIB, is a novel finding on our part. We postulate that the increased expression of CSH1 and FTL in the second trimester was in keeping with the metabolic demands of the growing fetus. GDF15 and EBI3, both recently identified in the placenta, remain unresolved with regard to their biological role in pregnancy, but with our data outlining their downregulation in second trimester, we gain an insight to their response to placental perfusion, and clues to their roles in the placenta.

Likewise for PPIB, the question of its role and regulation persists. Admittedly, our study uncovers as many questions as it resolved to answer, as with many other instrumental studies.

Lastly, our data also shed light on the placental expression of genes known to be dysregulated in preeclampsia. Based on what is known of respective genes, our study

200 substantiates the standing hypotheses for the function and roles they play in the pathogenesis of preeclampsia. A preliminary look indicates that many genes which were downregulated post-flow (i.e. higher in first trimester: GDF15, PROCR, LEP,

INH, INHA, ENG, FLT1, CD55, and CD59) have been implicated in preeclampsia, with augmented expression observed in the preeclamptic as compared to healthy placenta. This suggests an incomplete switch on the part of the placental cells from the pre-flow state to post-flow, with insufficient downregulation of those genes, which could be attributed to inadequate trophoblast invasion during placental development, and thus ineffective perfusion.

Multiple studies have proposed the use of select molecules as predictors of preeclampsia (Lynch et al., 2008; Lopez-Novoa, 2007; Leanos-Miranda et al., 2013), but we propose that a more extensive and comprehensive list of contributors to the pathogenesis of preeclampsia be considered. More pertinently, the genes that have so far been associated with the pathogenesis of preeclampsia also emerge in our analysis, as being differentially expressed between pre- and post-flow. Apart from signifying the potential of our list in identifying indicators of preeclampsia, this also suggests that regulation during this stage of placental development (trophoblast plug rupture, and the transition from pre- to post-flow and its accompanying change in oxygen), is crucial to determining the health of the pregnancy, and if preeclampsia develops.

The pinnacle of this study would be to derive a comprehensive set of gene markers, measured from maternal serum, which exhibit dynamic expression between two time points (i.e. pre- and post-flow), that are reflective of the physiological changes

201 occurring in the uteroplacental unit. This set of markers would thus be used as the gold standard for preeclampsia prediction in pregnancy diagnosis.

202 CHAPTER 6: CLOSING REMARKS

203 Due to ethical restrictions, the only means of studying the human placenta is by surveying and interrogating discarded tissue. Even then, access to such material is limited. Animal models, although useful in the right context, bear stark differences in reproductive physiology. As such, there is a scarcity of appropriate in vivo models for the trophoblast lineage. Inevitably, in vitro systems become the popular choice, and by manipulating growth conditions, including oxygen, efforts remain to capture and maintain various trophoblast cell types, and subsequently study their progression of differentiation down the lineage.

Exploiting the in vitro system, we investigated the role of a key transcription factor in the trophoblast by identifying its gene targets within a biological context. With this work, we are starting to see the EPAS1-regulated pathways in the trophoblast, and this finding is set to expand with the more comprehensive list currently being analyzed. In conjunction with the data obtained from the human placental transcriptome, it would be most intriguing to study how EPAS1-regulated genes are changing across first and second trimester placenta, tying it in with hypoxia and the reproductive success the Tibetans are observing despite their low oxygen habitat.

While we anticipate further verification, we were fortunate enough to gain access to human tissue samples, which overall led to a two-pronged approach to studying oxygen-mediated gene regulation in the placenta. The information laid out in this thesis is only the tip of the iceberg. With access to the placental transcriptome across a crucial developmental milestone, we can not only begin to understand the changes in the early placenta, we can then apply this understanding to address pregnancy complications, such as preeclampsia. The work achieved in this dissertation presents

204 a novel and rich resource useful not only for human reproduction and developmental biologists, but also for more niche fields like hematopoiesis.

The vast expanse of information extracted from the global transcriptomic analysis will lead to countless options for further data mining; one of which, presently underway in the lab, is the search for novel genes and/or isoforms expressed in the early placenta. Given the sample design, we are also able to track how the expression levels of these novel sequences change with progression into the second trimester, thereby presenting with the first means of understanding their role and association with the placenta. Ideally, the secondary layer of analysis would then be to compare against the diseased or compromised tissue.

Bearing in mind the complexity, coupled with the transience of the placenta as an organ, the pace at which it forms and then performs its multitude of tasks is remarkable. The effort to fully understand the placenta will be no less than a dizzying maze of information and postulations. We believe, and we hope, the evidence and ideas presented in this dissertation will go a long way to this cause.

205 BIBLIOGRAPHY

Aardema, M., Oosterhof, H., Timmer, A., van Rooy, I., & Aarnoudse, J. (2001). Uterine artery Doppler flow and uteroplacental vascular pathology in normal pregnancies and pregnancies complicated by pre-eclampsia and small for gestational age fetuses. Placenta , 22 (5), 405-11. Abbott, B., & Buckalew, A. (2000). Placental defects in ARNT-knockout conceptus correlate with localized decreases in VEGF-R2, Ang-1, and Tie-2. Developmental dynamics : an official publication of the American Association of Anatomists , 219 (4), 526-38. Aboagye-Mathiesen, G., Laugesen, J., Zdravkovic, M., & Ebbesen, P. (1996). Isolation and characterization of human placental trophoblast subpopulations from first-trimester chorionic villi. Clinical and diagnostic laboratory immunology , 3 (1), 14-22. Adány, R., & Bárdos, H. (2003). Factor XIII subunit A as an intracellular transglutaminase. Cellular and molecular life sciences : CMLS , 60 (6), 1049-60. Adelman, D., Gertsenstein, M., Nagy, A., Simon, M., & Maltepe, E. (2000). Placental cell fates are regulated in vivo by HIF-mediated hypoxia responses. Genes & Development , 14 (24), 3191-203. Ago, T., & Sadoshima, J. (2006). GDF15, a cardioprotective TGF-beta superfamily protein. Circulation research , 98 (3), 294-7. Aharon, A., Brenner, B., Katz, T., Miyagi, Y., & Lanir, N. (2004). Tissue factor and tissue factor pathway inhibitor levels in trophoblast cells: implications for placental hemostasis. Thrombosis and haemostasis , 92 (4), 776-86. Akey, J., Zhang, G., Zhang, K., Jin, L., & Shriver, M. (2002). Interrogating a high-density SNP map for signatures of natural selection. Genome Research , 12 (12), 1805-14. Aldenderfer, M., & Yinong, Z. (2004, Jan 1). The prehistory of the Tibetan Plateau to the seventh century AD: perspectives and research from China and the West since 1950. Journal of World Prehistory . Amarilio, R., Viukov, S., Sharir, A., Eshkar-Oren, I., Johnson, R., & Zelzer, E. (2007). HIF1alpha regulation of Sox9 is necessary to maintain differentiation of hypoxic prechondrogenic cells during early skeletogenesis. Development (Cambridge, England) , 134 (21), 3917-28. Amit, M., Shariki, C., Margulets, V., & Itskovitz-Eldor, J. (2004). Feeder layer- and serum-free culture of human embryonic stem cells. Biology of reproduction , 70 (3), 837-45. Appelhoff, R., Tian, Y.-M., Raval, R., Turley, H., Harris, A., Pugh, C., et al. (2004). Differential function of the prolyl hydroxylases PHD1, PHD2, and PHD3 in the regulation of hypoxia-inducible factor. The Journal of biological chemistry , 279 (37), 38458-65. Apps, R., Sharkey, A., Gardner, L., Male, V., Trotter, M., Miller, N., et al. (2011). Genome-wide expression profile of first trimester villous and extravillous human trophoblast cells. Placenta , 32 (1), 33-43. Aprelikova, O., Chandramouli, G., Wood, M., Vasselli, J., Riss, J., Maranchie, J., et al. (2004). Regulation of HIF prolyl hydroxylases by hypoxia-inducible factors. Journal of cellular biochemistry , 92 (3), 491-501. Aprelikova, O., Wood, M., Tackett, S., Chandramouli, G., & Barrett, J. (2006). Role of ETS transcription factors in the hypoxia-inducible factor-2 target gene selection. Cancer Research , 66 (11), 5641-7. Arany, Z., Huang, L., Eckner, R., Bhattacharya, S., Jiang, C., Goldberg, M., et al. (1996). An essential role for p300/CBP in the cellular response to hypoxia. Proceedings of the National Academy of Sciences of the United States of America , 93 (23), 12969-73. Arnold, S., & Robertson, E. (2009). Making a commitment: cell lineage allocation and axis patterning in the early mouse embryo. Nature reviews Molecular cell biology , 10 (2), 91-103. Asahina, T., Kobayashi, T., Takeuchi, K., & Kanayama, N. (2007). Congenital blood coagulation factor XIII deficiency and successful deliveries: a review of the literature. Obstetrical & gynecological survey , 62 (4), 255-60.

206 Ashkar, S., Weber, G., Panoutsakopoulou, V., Sanchirico, M., Jansson, M., Zawaideh, S., et al. (2000). Eta-1 (osteopontin): an early component of type-1 (cell-mediated) immunity. Science (New York, NY) , 287 (5454), 860-4. Autiero, M., Waltenberger, J., Communi, D., Kranz, A., Moons, L., Lambrechts, D., et al. (2003). Role of PlGF in the intra- and intermolecular cross talk between the VEGF receptors Flt1 and Flk1. Nature Medicine , 9 (7), 936-43. Bützow, R., Huhtala, M., Bohn, H., Virtanen, I., & Seppälä, M. (1988). Purification and characterization of placental protein 5. Biochemical and Biophysical Research Communications , 150 (1), 483-90. Bárcena, A., Kapidzic, M., Muench, M., Gormley, M., Scott, M., Weier, J., et al. (2009). The human placenta is a hematopoietic organ during the embryonic and fetal periods of development. Developmental biology , 327 (1), 24-33. Bárcena, A., Muench, M., Kapidzic, M., & Fisher, S. (2009). A new role for the human placenta as a hematopoietic site throughout gestation. Reproductive sciences (Thousand Oaks, Calif) , 16 (2), 178-87. Babaie, Y., Herwig, R., Greber, B., Brink, T., Wruck, W., Groth, D., et al. (2007). Analysis of Oct4- dependent transcriptional networks regulating self-renewal and pluripotency in human embryonic stem cells. Stem cells (Dayton, Ohio) , 25 (2), 500-10. Bai, Q., Assou, S., Haouzi, D., Ramirez, J.-M., Monzo, C., Becker, F., et al. (2012). Dissecting the first transcriptional divergence during human embryonic development. Stem cell reviews , 8 (1), 150-62. Baines, M., Millar, K., & Mills, P. (1974). Studies of complement levels in normal human pregnancy. Obstetrics and gynecology , 43 (6), 806-10. Bajoria, R., Sooranna, S., Ward, B., & Chatterjee, R. (2002). Prospective function of placental leptin at maternal-fetal interface. Placenta , 23 (2-3), 103-15. Bakker, G., Hoogerbrugge, J., Rommerts, F., & van der Molen, H. (1981). Lutropin-dependent protein phosphorylation and steroidogenesis in rat tumour Leydig cells. The Biochemical journal , 198 (2), 339-46. Banerjee, H., Rahn, A., Gawande, B., Guth, S., Valcarcel, J., & Singh, R. (2004). The conserved RNA recognition motif 3 of U2 snRNA auxiliary factor (U2AF 65) is essential in vivo but dispensable for activity in vitro. Rna , 10 (2), 240-53. Bania, J., Gatti, E., Lelouard, H., David, A., Cappello, F., Weber, E., et al. (2003). Human cathepsin S, but not cathepsin L, degrades efficiently MHC class II-associated invariant chain in nonprofessional APCs. Proceedings of the National Academy of Sciences of the United States of America , 100 (11), 6664-9. Baraibar, M., Barbeito, A., Muhoberac, B., & Vidal, R. (2012). A mutant light-chain ferritin that causes neurodegeneration has enhanced propensity toward oxidative damage. Free radical biology & medicine , 52 (9), 1692-7. Barak, Y., Nelson, M., Ong, E., Jones, Y., Ruiz-Lozano, P., Chien, K., et al. (1999). PPAR gamma is required for placental, cardiac, and adipose tissue development. Molecular Cell , 4 (4), 585-95. Barnikol-Watanabe, S., Gross, N., Götz, H., Henkel, T., Karabinos, A., Kratzin, H., et al. (1994). Human protein NEFA, a novel DNA binding/EF-hand/leucine zipper protein. Molecular cloning and sequence analysis of the cDNA, isolation and characterization of the protein. Biological chemistry Hoppe-Seyler , 375 (8), 497-512. Baschat, A., & Hecher, K. (2004). Fetal growth restriction due to placental disease. Seminars in perinatology , 28 (1), 67-80. Baumgrass, R., Williamson, M., & Price, P. (1997). Identification of peptide fragments generated by digestion of bovine and human osteocalcin with the lysosomal proteinases cathepsin B, D, L, H, and S. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research , 12 (3), 447-55. Beall, C. (2007). Two routes to functional adaptation: Tibetan and Andean high-altitude natives. Proceedings of the National Academy of Sciences of the United States of America , 104 Suppl 1, 8655-60.

207 Beall, C., Brittenham, G., Strohl, K., Blangero, J., Williams-Blangero, S., Goldstein, M., et al. (1998). Hemoglobin concentration of high-altitude Tibetans and Bolivian Aymara. American journal of physical anthropology , 106 (3), 385-400. Beall, C., Cavalleri, G., Deng, L., Elston, R., Gao, Y., Knight, J., et al. (2010). Natural selection on EPAS1 (HIF2alpha) associated with low hemoglobin concentration in Tibetan highlanders. Proceedings of the National Academy of Sciences of the United States of America , 107 (25), 11459-64. Bearfield, C., Jauniaux, E., Groome, N., Sargent, I., & Muttukrishna, S. (2005). The secretion and effect of inhibin A, activin A and follistatin on first-trimester trophoblasts in vitro. European journal of endocrinology / European Federation of Endocrine Societies , 152 (6), 909-16. Beattie, G., Lopez, A., Bucay, N., Hinton, A., Firpo, M., King, C., et al. (2005). Activin A maintains pluripotency of human embryonic stem cells in the absence of feeder layers. Stem cells (Dayton, Ohio) , 23 (4), 489-95. Beers, C., Burich, A., Kleijmeer, M., Griffith, J., Wong, P., & Rudensky, A. (2005). Cathepsin S controls MHC class II-mediated antigen presentation by epithelial cells in vivo. Journal of immunology (Baltimore, Md : 1950) , 174 (3), 1205-12. Beluzić, R., Cuk, M., Pavkov, T., Fumić, K., Barić, I., Mudd, S., et al. (2006). A single mutation at Tyr143 of human S-adenosylhomocysteine hydrolase renders the enzyme thermosensitive and affects the oxidation state of bound cofactor nicotinamide-adenine dinucleotide. The Biochemical journal , 400 (2), 245-53. Bereczky, Z., Katona, E., & Muszbek, L. (2003). Fibrin stabilization (factor XIII), fibrin structure and thrombosis. Pathophysiology of haemostasis and thrombosis , 33 (5-6), 430-7. Bernardo, A., Faial, T., Gardner, L., Niakan, K., Ortmann, D., Senner, C., et al. (2011). BRACHYURY and CDX2 mediate BMP-induced differentiation of human and mouse pluripotent stem cells into embryonic and extraembryonic lineages. Cell stem cell , 9 (2), 144-55. Berndt, S., Perrier d'Hauterive, S., Blacher, S., Péqueux, C., Lorquet, S., Munaut, C., et al. (2006). Angiogenic activity of human chorionic gonadotropin through LH receptor activation on endothelial and epithelial cells of the endometrium. FASEB journal : official publication of the Federation of American Societies for Experimental Biology , 20 (14), 2630-2. Bernstein, P., & Divon, M. (1997). Etiologies of fetal growth restriction. Clinical obstetrics and gynecology , 40 (4), 723-9. Bilban, M., Haslinger, P., Prast, J., Klinglmüller, F., Woelfel, T., Haider, S., et al. (2009). Identification of novel trophoblast invasion-related genes: heme oxygenase-1 controls motility via peroxisome proliferator-activated receptor gamma. Endocrinology , 150 (2), 1000-13. Birdsall, M., Ledger, W., Groome, N., Abdalla, H., & Muttukrishna, S. (1997). Inhibin A and activin A in the first trimester of human pregnancy. The Journal of clinical endocrinology and metabolism , 82 (5), 1557-60. Bischof, P., & Campana, A. (2000). Molecular mediators of implantation. Baillière's best practice & research Clinical obstetrics & gynaecology , 14 (5), 801-14. Bischof, P., Meisser, A., & Campana, A. (2000). Paracrine and autocrine regulators of trophoblast invasion--a review. Placenta , 21 Suppl A, S55-60. Black, D. (2003, Jan 1). Mechanisms of alternative pre-messenger RNA splicing. Annual review of biochemistry . Black, S., Arnaud, F., Palmarini, M., & Spencer, T. (2010). Endogenous retroviruses in trophoblast differentiation and placental development. American journal of reproductive immunology (New York, NY : 1989) , 64 (4), 255-64. Bo, M., & Boime, I. (1992). Identification of the transcriptionally active genes of the chorionic gonadotropin beta gene cluster in vivo. The Journal of biological chemistry , 267 (5), 3179-84. Board, P., Webb, G., McKee, J., & Ichinose, A. (1988). Localization of the coagulation factor XIII A subunit gene (F13A) to chromosome bands 6p24----p25. Cytogenetics and cell genetics , 48 (1), 25-7. Boffa, M., Maret, D., Hamill, J., Bastajian, N., Crainich, P., Jenny, N., et al. (2008). Effect of single nucleotide polymorphisms on expression of the gene encoding thrombin-activatable fibrinolysis inhibitor: a functional analysis. Blood , 111 (1), 183-9.

208 Bogovich, K., Richards, J., & Reichert, L. (1981). Obligatory role of luteinizing hormone (LH) in the initiation of preovulatory follicular growth in the pregnant rat: specific effects of human chorionic gonadotropin and follicle-stimulating of hormone on LH receptors and steroidogenesis in theca, granulosa, and luteal cells. Endocrinology , 109 (3), 860-7. Bole, D., Hendershot, L., & Kearney, J. (1986). Posttranslational association of immunoglobulin heavy chain binding protein with nascent heavy chains in nonsecreting and secreting hybridomas. The Journal of cell biology , 102 (5), 1558-66. Boothby, M., Kukowska, J., & Boime, I. (1983). Imbalanced synthesis of human choriogonadotropin alpha and beta subunits reflects the steady state levels of the corresponding mRNAs. The Journal of biological chemistry , 258 (15), 9250-3. Borth, W. (1992). Alpha 2-macroglobulin, a multifunctional binding protein with targeting characteristics. FASEB journal : official publication of the Federation of American Societies for Experimental Biology , 6 (15), 3345-53. Bottaro, D., Rubin, J., Faletto, D., Chan, A., Kmiecik, T., Vande Woude, G., et al. (1991). Identification of the hepatocyte growth factor receptor as the c-met proto-oncogene product. Science (New York, NY) , 251 (4995), 802-4. Bourdiec, A., Shao, R., Rao, C., & Akoum, A. (2012). Human chorionic gonadotropin triggers angiogenesis via the modulation of endometrial stromal cell responsiveness to interleukin 1: a new possible mechanism underlying embryo implantation. Biology of reproduction , 87 (3), 66. Boyd, J. and Hamilton, W. (1970). The human placenta. Cambridge: W. Heffer & Sons Ltd. Brümmer, J., Ebrahimnejad, A., Flayeh, R., Schumacher, U., Löning, T., Bamberger, A., et al. (2001). cis Interaction of the cell adhesion molecule CEACAM1 with integrin beta(3). The American journal of pathology , 159 (2), 537-46. Breathnach, R., & Chambon, P. (1981). Organization and expression of eucaryotic split genes coding for proteins. Annual review of biochemistry , 50, 349-83. Briese, J., Oberndörfer, M., Pätschenik, C., Schulte, H., Makrigiannakis, A., Löning, T., et al. (2005). Osteopontin is colocalized with the adhesion molecule CEACAM1 in the extravillous trophoblast of the human placenta and enhances invasion of CEACAM1-expressing placental cells. The Journal of clinical endocrinology and metabolism , 90 (9), 5407-13. Briese, J., Oberndörfer, M., Schulte, H., Löning, T., & Bamberger, A. (2005). Osteopontin expression in gestational trophoblastic diseases: correlation with expression of the adhesion molecule, CEACAM1. International journal of gynecological pathology : official journal of the International Society of Gynecological Pathologists , 24 (3), 271-6. Brinkman, R., Wildschut, A., & Wittermans, A. (1934). On the occurrence of two kinds of haemoglobin in normal human blood. The Journal of physiology , 80 (4), 377-87. Brittain, T. (2002). Molecular aspects of embryonic hemoglobin function. Molecular aspects of medicine , 23 (4), 293-342. Brosens, I., Robertson, W., & Dixon, H. (1967). The physiological response of the vessels of the placental bed to normal pregnancy. The Journal of pathology and bacteriology , 93 (2), 569-79. Brown, C., Johnson, D., & Sidow, A. (2007). Functional architecture and evolution of transcriptional elements that drive gene coexpression. Science (New York, NY) , 317 (5844), 1557-60. Brown, J. (2000). Exploiting the hypoxic cancer cell: mechanisms and therapeutic strategies. Molecular medicine today , 6 (4), 157-62. Broze, G., & Miletich, J. (1987). Isolation of the tissue factor inhibitor produced by HepG2 hepatoma cells. Proceedings of the National Academy of Sciences of the United States of America , 84 (7), 1886-90. Broze, G., Girard, T., & Novotny, W. (1990). Regulation of coagulation by a multivalent Kunitz-type inhibitor. Biochemistry , 29 (33), 7539-46. Broze, G., Warren, L., Novotny, W., Higuchi, D., Girard, J., & Miletich, J. (1988). The lipoprotein- associated coagulation inhibitor that inhibits the factor VII-tissue factor complex also inhibits factor Xa: insight into its possible mechanism of action. Blood , 71 (2), 335-43.

209 Buroker, N., Ning, X., Zhou, Z., Li, K., & Cen…, W. (2012, Jan 1). EPAS1 and EGLN1 associations with high altitude sickness in Han and Tibetan Chinese at the Qinghai–Tibetan Plateau. Blood Cells . Burton, G., Jauniaux, E., & Watson, A. (1999). Maternal arterial connections to the placental intervillous space during the first trimester of human pregnancy: the Boyd collection revisited. YMOB , 181 (3), 718-24. Burton, G., Watson, A., Hempstock, J., Skepper, J., & Jauniaux, E. (2002). Uterine glands provide histiotrophic nutrition for the human fetus during the first trimester of pregnancy. The Journal of clinical endocrinology and metabolism , 87 (6), 2954-9. Burton, G., Yung, H.-W., Cindrova-Davies, T., & Charnock-Jones, D. (2009). Placental endoplasmic reticulum stress and oxidative stress in the pathophysiology of unexplained intrauterine growth restriction and early onset preeclampsia. Placenta , 30 Suppl A, S43-8. Bustamante, C., Fledel-Alon, A., Williamson, S., Nielsen, R., Hubisz, M., Glanowski, S., et al. (2005). Natural selection on protein-coding genes in the human genome. Nature , 437 (7062), 1153-7. Buurma, A., Cohen, D., Veraar, K., Schonkeren, D., Claas, F., Bruijn, J., et al. (2012). Preeclampsia is characterized by placental complement dysregulation. Hypertension , 60 (5), 1332-7. Caniggia, I., & Winter, J. (2002). Adriana and Luisa Castellucci Award lecture 2001. Hypoxia inducible factor-1: oxygen regulation of trophoblast differentiation in normal and pre-eclamptic pregnancies--a review. Placenta , 23 Suppl A, S47-57. Caniggia, I., Lye, S., & Cross, J. (1997). Activin is a local regulator of human cytotrophoblast cell differentiation. Endocrinology , 138 (9), 3976-86. Caniggia, I., Winter, J., Lye, S., & Post, M. (2000). Oxygen and placental development during the first trimester: implications for the pathophysiology of pre-eclampsia. Placenta , 21 Suppl A, S25-30. Capon, F., Allen, M., Ameen, M., Burden, A., Tillman, D., Barker, J., et al. (2004). A synonymous SNP of the corneodesmosin gene leads to increased mRNA stability and demonstrates association with psoriasis across diverse ethnic groups. Human molecular genetics , 13 (20), 2361-8. Card, P., Erbel, P., & Gardner, K. (2005). Structural basis of ARNT PAS-B dimerization: use of a common beta-sheet interface for hetero- and homodimerization. Journal of molecular biology , 353 (3), 664-77. Carlson, B.M. (1994). Human embryology and developmental biology. Michigan: Mosby. Carreau, A., El Hafny-Rahbi, B., Matejuk, A., Grillon, C., & Kieda, C. (2011). Why is the partial oxygen pressure of human tissues a crucial parameter? Small molecules and hypoxia. Journal of cellular and molecular medicine , 15 (6), 1239-53. Carter, A. (2007). Animal models of human placentation--a review. Placenta , 28 Suppl A, S41-7. Cartwright, J., Kenny, L., Dash, P., Crocker, I., Aplin, J., Baker, P., et al. (2002). Trophoblast invasion of spiral arteries: a novel in vitro model. Placenta , 23 (2-3), 232-5. Casagrandi, D., Bearfield, C., Geary, J., Redman, C., & Muttukrishna, S. (2003). Inhibin, activin, follistatin, activin receptors and beta-glycan gene expression in the placental tissue of patients with pre-eclampsia. Molecular human reproduction , 9 (4), 199-203. Centlow, M., Wingren, C., Borrebaeck, C., Brownstein, M., & Hansson, S. (2011). Differential gene expression analysis of placentas with increased vascular resistance and pre-eclampsia using whole- genome microarrays. Journal of pregnancy , 2011, 472354. Chamary, J., Parmley, J., & Hurst, L. (2006, Jan 1). Hearing silence: non-neutral evolution at synonymous sites in mammals. Nature reviews Genetics . Chapman-Smith, A., Lutwyche, J., & Whitelaw, M. (2004). Contribution of the Per/Arnt/Sim (PAS) domains to DNA binding by the basic helix-loop-helix PAS transcriptional regulators. The Journal of biological chemistry , 279 (7), 5353-62. Chappell, S., & Morgan, L. (2006). Searching for genetic clues to the causes of pre-eclampsia. Clinical science (London, England : 1979) , 110 (4), 443-58. Chen, Y., Zhang, L., & Jones, K. (2011). SKIP counteracts p53-mediated apoptosis via selective regulation of p21Cip1 mRNA splicing. Genes & Development , 25 (7), 701-16.

210 Cho, W., Choi, J., Park, C.-H., Yoon, S.-O., Jeoung, D.-I., Kim, Y.-M., et al. (2008). Expression of CD320 in human B cells in addition to follicular dendritic cells. BMB reports , 41 (12), 863-7. Choi, S., Kim, J., Kang, E.-J., Lee, S.-W., Park, M.-C., Park, Y.-B., et al. (2008). Osteopontin might be involved in bone remodelling rather than in inflammation in ankylosing spondylitis. Rheumatology (Oxford, England) , 47 (12), 1775-9. Christoffels, V., Burch, J., & Moorman, A. (2004). Architectural plan for the heart: early patterning and delineation of the chambers and the nodes. Trends in cardiovascular medicine , 14 (8), 301-7. Chung, D., Ohashi, K., Watanabe, M., Miyasaka, N., & Hirosawa, S. (2000). Mannose trimming targets mutant alpha(2)-plasmin inhibitor for degradation by the proteasome. The Journal of biological chemistry , 275 (7), 4981-7. Cioffi, J., Van Blerkom, J., Antczak, M., Shafer, A., Wittmer, S., & Snodgrass, H. (1997). The expression of leptin and its receptors in pre-ovulatory human follicles. Molecular human reproduction , 3 (6), 467-72. Clancy, S. (2008, Jan 1). RNA splicing: introns, exons and spliceosome. Nature Education . Clegg, J. (1980). Embryonic hemoglobin: sequence of the epsilon and zeta chains. Texas reports on biology and medicine , 40, 23-8. Collins, F., & Weissman, S. (1984). The molecular genetics of human hemoglobin. Progress in nucleic acid research and molecular biology , 31, 315-462. Collins, F., Brooks, L., & Chakravarti, A. (1998, Jan 1). A DNA polymorphism discovery resource for research on human genetic variation. Genome Research . Compernolle, V., Brusselmans, K., Acker, T., Hoet, P., Tjwa, M., Beck, H., et al. (2002). Loss of HIF-2alpha and inhibition of VEGF impair fetal lung maturation, whereas treatment with VEGF prevents fatal respiratory distress in premature mice. Nature Medicine , 8 (7), 702-10. Compernolle, V., Brusselmans, K., Franco, D., Moorman, A., Dewerchin, M., Collen, D., et al. (2003). Cardia bifida, defective heart development and abnormal neural crest migration in embryos lacking hypoxia-inducible factor-1alpha. Cardiovascular research , 60 (3), 569-79. Cooke, B., Lindh, M., & Janszen, F. (1977). Effect of lutropin on phosphorylation of endogenous proteins in testis leydig cells. Correlation with testosterone production. The Biochemical journal , 168 (1), 43-8. Cooper, G., & Brown, C. (2008). Qualifying the relationship between sequence conservation and molecular function. Genome Research , 18 (2), 201-5. Coppens, M., Loquet, P., Kollen, M., De Neubourg, F., & Buytaert, P. (1996). Longitudinal evaluation of uteroplacental and umbilical blood flow changes in normal early pregnancy. Ultrasound in obstetrics & gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology , 7 (2), 114-21. Costantino, L., Rusconi, D., Soldà, G., Seia, M., Paracchini, V., Porcaro, L., et al. (2013, Jan 24). Fine Characterization of the Recurrent c.1584+18672A>G Deep-Intronic Mutation in the CFTR Gene. American journal of respiratory cell and molecular biology . Coulomb-L'Herminé, A., Larousserie, F., Pflanz, S., Bardel, E., Kastelein, R., & Devergne, O. (2007). Expression of interleukin-27 by human trophoblast cells. Placenta , 28 (11-12), 1133-40. Covello, K., Kehler, J., Yu, H., Gordan, J., Arsham, A., Hu, C.-J., et al. (2006). HIF-2alpha regulates Oct-4: effects of hypoxia on stem cell function, embryonic development, and tumor growth. Genes & Development , 20 (5), 557-70. Cowden Dahl, K., Fryer, B., Mack, F., Compernolle, V., Maltepe, E., Adelman, D., et al. (2005). Hypoxia-inducible factors 1alpha and 2alpha regulate trophoblast differentiation. Molecular and cellular biology , 25 (23), 10479-91. Crocker, I., Wareing, M., Ferris, G., Jones, C., Cartwright, J., Baker, P., et al. (2005). The effect of vascular origin, oxygen, and tumour necrosis factor alpha on trophoblast invasion of maternal arteries in vitro. The Journal of pathology , 206 (4), 476-85. Cross, J., Anson-Cartwright, L., & Scott, I. (2002). Transcription factors underlying the development and endocrine functions of the placenta. Recent progress in hormone research , 57, 221-34.

211 Dai, J., & Rabie, A. (2007). VEGF: an essential mediator of both angiogenesis and endochondral ossification. Journal of dental research , 86 (10), 937-50. Daiter, E., Omigbodun, A., Wang, S., Walinsky, D., Strauss, J., Hoyer, J., et al. (1996). Cell differentiation and endogenous cyclic adenosine 3',5'-monophosphate regulate osteopontin expression in human trophoblasts. Endocrinology , 137 (5), 1785-90. Damsky, C., Librach, C., Lim, K., Fitzgerald, M., McMaster, M., Janatpour, M., et al. (1994). Integrin switching regulates normal trophoblast invasion. Development (Cambridge, England) , 120 (12), 3657-66. Danzig, J., & Levine, M. (2012). A novel intronic mutation in SHOX causes short stature by disrupting a splice acceptor site: direct demonstration of aberrant splicing by expression of a minigene in HEK-293T cells. Journal of pediatric endocrinology & metabolism : JPEM , 25 (9-10), 889-95. Das, C., Kumar, V., Gupta, S., & Kumar, S. (2002). Network of cytokines, integrins and hormones in human trophoblast cells. Journal of reproductive immunology , 53 (1-2), 257-68. Das, P., Ezashi, T., Schulz, L., Westfall, S., Livingston, K., & Roberts, R. (2007). Effects of fgf2 and oxygen in the bmp4-driven differentiation of trophoblast from human embryonic stem cells. Stem cell research , 1 (1), 61-74. David, V., Hochstenbach, F., Rajagopalan, S., & Brenner, M. (1993). Interaction with newly synthesized and retained proteins in the endoplasmic reticulum suggests a chaperone function for human integral membrane protein IP90 (calnexin). The Journal of biological chemistry , 268 (13), 9585-92. Davie, E., Fujikawa, K., & Kisiel, W. (1991). The coagulation cascade: initiation, maintenance, and regulation. Biochemistry , 30 (43), 10363-70. Davies, R., Calvio, C., Bratt, E., Larsson, S., Lamond, A., & Hastie, N. (1998). WT1 interacts with the splicing factor U2AF65 in an isoform-dependent manner and can be incorporated into spliceosomes. Genes & Development , 12 (20), 3217-25. De Gobbi, M., Viprakasit, V., Hughes, J., Fisher, C., Buckle, V., Ayyub, H., et al. (2006). A regulatory SNP causes a human genetic disease by creating a new transcriptional promoter. Science (New York, NY) , 312 (5777), 1215-7. Denko, N., Fontana, L., Hudson, K., Sutphin, P., Raychaudhuri, S., Altman, R., et al. (2003). Investigating hypoxic tumor physiology through gene expression patterns. Oncogene , 22 (37), 5907-14. Denny, K., Woodruff, T., Taylor, S., & Callaway, L. (2013). Complement in pregnancy: a delicate balance. American journal of reproductive immunology (New York, NY : 1989) , 69 (1), 3-11. Devergne, O., Birkenbach, M., & Kieff, E. (1997). Epstein-Barr virus-induced gene 3 and the p35 subunit of interleukin 12 form a novel heterodimeric hematopoietin. Proceedings of the National Academy of Sciences of the United States of America , 94 (22), 12041-6. Devergne, O., Coulomb-L'Herminé, A., Capel, F., Moussa, M., & Capron, F. (2001). Expression of Epstein-Barr virus-induced gene 3, an interleukin-12 p40-related molecule, throughout human pregnancy: involvement of syncytiotrophoblasts and extravillous trophoblasts. The American journal of pathology , 159 (5), 1763-76. DICKSON, A. (1963). TROPHOBLASTIC GIANT CELL TRANSFORMATION OF MOUSE BLASTOCYTS. Journal of reproduction and fertility , 6, 465-6. DICZFALUSY, E. (1953). Chorionic gonadotrophin and oestrogens in the human placenta. Acta endocrinologica Supplementum , 11 (12), 1-175. Dimitriadis, E., Nie, G., Hannan, N., Paiva, P., & Salamonsen, L. (2010). Local regulation of implantation at the human fetal-maternal interface. The International journal of developmental biology , 54 (2-3), 313-22. Dommisse, J., & Tiltman, A. (1992). Placental bed biopsies in placental abruption. British journal of obstetrics and gynaecology , 99 (8), 651-4. Douglas, G., VandeVoort, C., Kumar, P., Chang, T.-C., & Golos, T. (2009). Trophoblast stem cells: models for investigating trophectoderm differentiation and placental development. Endocrine Reviews , 30 (3), 228-40.

212 Downing, M., Bloom, J., & Mann, K. (1978). Comparison of the inhibition of thrombin by three plasma protease inhibitors. Biochemistry , 17 (13), 2649-53. Drummond, D., Raval, A., & Wilke, C. (2006). A single determinant dominates the rate of yeast protein evolution. Molecular biology and evolution , 23 (2), 327-37. Duan, J., Wainwright, M., Comeron, J., Saitou, N., Sanders, A., Gelernter, J., et al. (2003). Synonymous mutations in the human dopamine receptor D2 (DRD2) affect mRNA stability and synthesis of the receptor. Human molecular genetics , 12 (3), 205-16. Dunwoodie, S. (2009). The role of hypoxia in development of the Mammalian embryo. Developmental Cell , 17 (6), 755-73. Dusse, L., Carvalho, M., Cooper, A., & Lwaleed, B. (2006). Tissue factor and tissue factor pathway inhibitor: a potential role in pregnancy and obstetric vascular complications? Clinica chimica acta; international journal of clinical chemistry , 372 (1-2), 43-6. Earl, P., Moss, B., & Doms, R. (1991). Folding, interaction with GRP78-BiP, assembly, and transport of the human immunodeficiency virus type 1 envelope protein. Journal of virology , 65 (4), 2047-55. Edstrom, C., Calhoun, D., & Christensen, R. (2000). Expression of tissue factor pathway inhibitor in human fetal and placental tissues. Early human development , 59 (2), 77-84. Egli, R., Southam, L., Wilkins, J., Lorenzen, I., Pombo-Suarez, M., Gonzalez, A., et al. (2009). Functional analysis of the osteoarthritis susceptibility-associated GDF5 regulatory polymorphism. Arthritis and rheumatism , 60 (7), 2055-64. Eisen, M., Spellman, P., Brown, P., & Botstein, D. (1998). Cluster analysis and display of genome- wide expression patterns. Proceedings of the National Academy of Sciences of the United States of America , 95 (25), 14863-8. Ellis, V., Scully, M., MacGregor, I., & Kakkar, V. (1982). Inhibition of human factor Xa by various plasma protease inhibitors. Biochimica et biophysica acta , 701 (1), 24-31. Elvert, G., Kappel, A., Heidenreich, R., Englmeier, U., Lanz, S., Acker, T., et al. (2003). Cooperative interaction of hypoxia-inducible factor-2alpha (HIF-2alpha ) and Ets-1 in the transcriptional activation of vascular endothelial growth factor receptor-2 (Flk-1). The Journal of biological chemistry , 278 (9), 7520-30. Enders, A. (1976). Cytology of human early implantation. Research in reproduction , 8 (5), 1-2. Enders, A. (1989). Trophoblast differentiation during the transition from trophoblastic plate to lacunar stage of implantation in the rhesus monkey and human. The American journal of anatomy , 186 (1), 85-98. Enquobahrie, D., Meller, M., Rice, K., Psaty, B., Siscovick, D., & Williams, M. (2008). Differential placental gene expression in preeclampsia. American journal of obstetrics and gynecology , 199 (5), 566.e1-11. Epstein, A., Gleadle, J., McNeill, L., Hewitson, K., O'Rourke, J., Mole, D., et al. (2001). C. elegans EGL-9 and mammalian homologs define a family of dioxygenases that regulate HIF by prolyl hydroxylation. Cell , 107 (1), 43-54. Erbel, P., Card, P., Karakuzu, O., Bruick, R., & Gardner, K. (2003). Structural basis for PAS domain heterodimerization in the basic helix--loop--helix-PAS transcription factor hypoxia-inducible factor. Proceedings of the National Academy of Sciences of the United States of America , 100 (26), 15504-9. Evans, M., & Kaufman, M. (1981). Establishment in culture of pluripotential cells from mouse embryos. Nature , 292 (5819), 154-6. Ezashi, T., Das, P., & Roberts, R. (2005). Low O2 tensions and the prevention of differentiation of hES cells. Proceedings of the National Academy of Sciences of the United States of America , 102 (13), 4783-8. Ezashi, T., Telugu, B., & Roberts, R. (2012). Model systems for studying trophoblast differentiation from human pluripotent stem cells. Cell and tissue research , 349 (3), 809-24. Fairlie, W., Moore, A., Bauskin, A., Russell, P., Zhang, H., & Breit, S. (1999). MIC-1 is a novel TGF- beta superfamily cytokine associated with macrophage activation. Journal of leukocyte biology , 65 (1), 2-5.

213 Fang, F., Zheng, J., Galbaugh, T., Fiorillo, A., Hjort, E., Zeng, X., et al. (2010). Cyclophilin B as a co- regulator of prolactin-induced gene expression and function in breast cancer cells. Journal of molecular endocrinology , 44 (6), 319-29. Faulk, W., Jarret, R., Keane, M., Johnson, P., & Boackle, R. (1980). Immunological studies of human placentae: complement components in immature and mature chorionic villi. Clinical and experimental immunology , 40 (2), 299-305. Fay, J., Kelehan, P., Lambkin, H., & Schwartz, S. (2009). Increased expression of cellular RNA- binding proteins in HPV-induced neoplasia and cervical cancer. Journal of medical virology , 81 (5), 897-907. Fiddes, J., & Goodman, H. (1979). Isolation, cloning and sequence analysis of the cDNA for the alpha- subunit of human chorionic gonadotropin. Nature , 281 (5730), 351-6. Field, L., Bonnevie-Nielsen, V., Pociot, F., Lu, S., Nielsen, T., & Beck-Nielsen, H. (2005). OAS1 splice site polymorphism controlling antiviral enzyme activity influences susceptibility to type 1 diabetes. Diabetes , 54 (5), 1588-91. Filipowicz, W., Bhattacharyya, S., & Sonenberg, N. (2008). Mechanisms of post-transcriptional regulation by microRNAs: are the answers in sight? Nature reviews Genetics , 9 (2), 102-14. Fisher, L., Hawkins, G., Tuross, N., & Termine, J. (1987). Purification and partial characterization of small proteoglycans I and II, bone sialoproteins I and II, and osteonectin from the mineral compartment of developing human bone. The Journal of biological chemistry , 262 (20), 9702-8. Fisher, S., Cui, T., Zhang, L., Hartman, L., Grahl, K., Zhang, G., et al. (1989). Adhesive and degradative properties of human placental cytotrophoblast cells in vitro. The Journal of cell biology , 109 (2), 891-902. Flanagan, S., Xie, W., Caswell, R., Damhuis, A., Vianey-Saban, C., Akcay, T., et al. (2013). Next- generation sequencing reveals deep intronic cryptic ABCC8 and HADH splicing founder mutations causing hyperinsulinism by pseudoexon activation. American journal of human genetics , 92 (1), 131-6. Florio, P., Ciarmela, P., Luisi, S., Palumbo, M., Lambert-Messerlian, G., Severi, F., et al. (2002). Pre- eclampsia with fetal growth restriction: placental and serum activin A and inhibin A levels. Gynecological endocrinology : the official journal of the International Society of Gynecological Endocrinology , 16 (5), 365-72. Foidart, J., Hustin, J., Dubois, M., & Schaaps, J. (1992). The human placenta becomes haemochorial at the 13th week of pregnancy. The International journal of developmental biology , 36 (3), 451-3. Forristal, C., Wright, K., Hanley, N., Oreffo, R., & Houghton, F. (2010). Hypoxia inducible factors regulate pluripotency and proliferation in human embryonic stem cells cultured at reduced oxygen tensions. Reproduction (Cambridge, England) , 139 (1), 85-97. Forsythe, J., Jiang, B., Iyer, N., Agani, F., Leung, S., Koos, R., et al. (1996). Activation of vascular endothelial growth factor gene transcription by hypoxia-inducible factor 1. Molecular and cellular biology , 16 (9), 4604-13. Founds, S., Conley, Y., Lyons-Weiler, J., Jeyabalan, A., Hogge, W., & Conrad, K. (2009). Altered global gene expression in first trimester placentas of women destined to develop preeclampsia. Placenta , 30 (1), 15-24. Franchi, F., Biguzzi, E., Cetin, I., Facchetti, F., Radaelli, T., Bozzo, M., et al. (2001). Mutations in the thrombomodulin and endothelial protein C receptor genes in women with late fetal loss. British journal of haematology , 114 (3), 641-6. Franzén, A., & Heinegård, D. (1985). Isolation and characterization of two sialoproteins present only in bone calcified matrix. The Biochemical journal , 232 (3), 715-24. Freyer, G., O'Brien, J., & Hurwitz, J. (1989). Alterations in the polypyrimidine sequence affect the in vitro splicing reactions catalyzed by HeLa cell-free preparations. The Journal of biological chemistry , 264 (25), 14631-7. Fukuda, R., Zhang, H., Kim, J.-w., Shimoda, L., Dang, C., & Semenza, G. (2007). HIF-1 regulates cytochrome oxidase subunits to optimize efficiency of respiration in hypoxic cells. Cell , 129 (1), 111-22.

214 Gabinskaya, T., Salafia, C., Gulle, V., Holzman, I., & Weintraub, A. (1998). Gestational age-dependent extravillous cytotrophoblast osteopontin immunolocalization differentiates between normal and preeclamptic pregnancies. American journal of reproductive immunology (New York, NY : 1989) , 40 (5), 339-46. Gambino, Y., Maymó, J., Pérez Pérez, A., Calvo, J., Sánchez-Margalet, V., & Varone, C. (2012). Elsevier Trophoblast Research Award lecture: Molecular mechanisms underlying estrogen functions in trophoblastic cells--focus on leptin expression. Placenta , 33 Suppl, S63-70. Gao, K., Masuda, A., Matsuura, T., & Ohno, K. (2008). Human branch point consensus sequence is yUnAy. Nucleic Acids Research , 36 (7), 2257-67. Garcia-Blanco, M., Baraniak, A., & Lasda, E. (2004). Alternative splicing in disease and therapy. Nature biotechnology , 22 (5), 535-46. Gasparini, P., Calvano, S., Memeo, E., Bisceglia, L., & Zelante, L. (1997). Assignment of ferritin L gene (FTL) to human chromosome band 19q13.3 by in situ hybridization. Annales de génétique , 40 (4), 227-8. Genbacev, O., Donne, M., Kapidzic, M., Gormley, M., Lamb, J., Gilmore, J., et al. (2011). Establishment of human trophoblast progenitor cell lines from the chorion. Stem cells (Dayton, Ohio) , 29 (9), 1427-36. Genbacev, O., Krtolica, A., Kaelin, W., & Fisher, S. (2001). Human cytotrophoblast expression of the von Hippel-Lindau protein is downregulated during uterine invasion in situ and upregulated by hypoxia in vitro. Developmental biology , 233 (2), 526-36. Genbacev, O., Prakobphol, A., Foulk, R., Krtolica, A., Ilic, D., Singer, M., et al. (2003). Trophoblast L- selectin-mediated adhesion at the maternal-fetal interface. Science (New York, NY) , 299 (5605), 405-8. Genbacev, O., Zhou, Y., Ludlow, J., & Fisher, S. (1997). Regulation of human placental development by oxygen tension. Science (New York, NY) , 277 (5332), 1669-72. Gentile, S., Martin, N., Scappini, E., Williams, J., Erxleben, C., & Armstrong, D. (2008). The human ERG1 channel polymorphism, K897T, creates a phosphorylation site that inhibits channel activity. Proceedings of the National Academy of Sciences of the United States of America , 105 (38), 14704-8. Gerhard, D., Wagner, L., Feingold, E., Shenmen, C., Grouse, L., Schuler, G., et al. (2004). The status, quality, and expansion of the NIH full-length cDNA project: the Mammalian Gene Collection (MGC). Genome Research , 14 (10B), 2121-7. Gerretsen, G., Huisjes, H., & Elema, J. (1981). Morphological changes of the spiral arteries in the placental bed in relation to pre-eclampsia and fetal growth retardation. British journal of obstetrics and gynaecology , 88 (9), 876-81. Giaccia, A., Simon, M., & Johnson, R. (2004). The biology of hypoxia: the role of oxygen sensing in development, normal function, and disease. Genes & Development , 18 (18), 2183-94. Giacopelli, F., Marciano, R., Pistorio, A., Catarsi, P., Canini, S., Karsenty, G., et al. (2004). Polymorphisms in the osteopontin promoter affect its transcriptional activity. Physiological Genomics , 20 (1), 87-96. Gibbs, R., Belmont, J., Hardenbol, P., Willis, T., & Yu…, F. (2003, Jan 1). The international HapMap project. Nature. Gilbert, S. F. (2000). Developmental biology. Sunderland, MA: Sinauer Associates. Gilbert, M., & Leturque, A. (1982). Fetal weight and its relationship to placental blood flow and placental weight in experimental intrauterine growth retardation in the rat. Journal of developmental physiology , 4 (4), 237-46. Girard, T., Warren, L., Novotny, W., Bejcek, B., Miletich, J., & Broze, G. (1989). Identification of the 1.4 kb and 4.0 kb messages for the lipoprotein associated coagulation inhibitor and expression of the encoded protein. Thrombosis research , 55 (1), 37-50. Girardi, G., Bulla, R., Salmon, J., & Tedesco, F. (2006). The complement system in the pathophysiology of pregnancy. Molecular immunology , 43 (1-2), 68-77. Girardi, G., Prohászka, Z., Bulla, R., Tedesco, F., & Scherjon, S. (2011). Complement activation in animal and human pregnancies as a model for immunological recognition. Molecular immunology , 48 (14), 1621-30.

215 Girardi, G., Yarilin, D., Thurman, J., Holers, V., & Salmon, J. (2006). Complement activation induces dysregulation of angiogenic factors and causes fetal rejection and growth restriction. The Journal of experimental medicine , 203 (9), 2165-75. Glisovic, T., Bachorik, J., Yong, J., & Dreyfuss, G. (2008, Jan 1). RNA-binding proteins and post- transcriptional gene regulation. Febs Letters . Goepel, E., Ulmer, H., & Neth, R. (1988). Premature labor contractions and the value of serum ferritin during pregnancy. Gynecologic and obstetric investigation , 26 (4), 265-73. Goldenberg, R., Tamura, T., DuBard, M., Johnston, K., Copper, R., & Neggers, Y. (1996). Plasma ferritin and pregnancy outcome. American journal of obstetrics and gynecology , 175 (5), 1356-9. Golling, G., Amsterdam, A., Sun, Z., Antonelli, M., Maldonado, E., Chen, W., et al. (2002). Insertional mutagenesis in zebrafish rapidly identifies genes essential for early vertebrate development. Nature Genetics , 31 (2), 135-40. Gordan, J., Bertout, J., Hu, C.-J., Diehl, J., & Simon, M. (2007). HIF-2alpha promotes hypoxic cell proliferation by enhancing c-myc transcriptional activity. Cancer cell , 11 (4), 335-47. Gossler, A. (1992). Early embryonic development of animals (Hennig, W., Nover, L., and Scheer, U. eds). Berlin, New York: Springer-Verlag. Gottwein, E., Cai, X., & Cullen, B. (2006, Jan 1). A novel assay for viral microRNA function identifies a single nucleotide polymorphism that affects Drosha processing. Journal of virology . Greber, B., Lehrach, H., & Adjaye, J. (2008). Control of early fate decisions in human ES cells by distinct states of TGFbeta pathway activity. Stem cells and development , 17 (6), 1065-77. Greber, B., Lehrach, H., & Adjaye, J. (2007). Fibroblast growth factor 2 modulates transforming growth factor beta signaling in mouse embryonic fibroblasts and human ESCs (hESCs) to support hESC self-renewal. Stem cells (Dayton, Ohio) , 25 (2), 455-64. Greer, S., Metcalf, J., Wang, Y., & Ohh, M. (2012). The updated biology of hypoxia-inducible factor. The EMBO Journal , 31 (11), 2448-60. Grossman, S., Andersen, K., Shlyakhter, I., Tabrizi, S., Winnicki, S., Yen, A., et al. (2013). Identifying recent adaptations in large-scale genomic data. Cell , 152 (4), 703-13. Gruber, M., Hu, C.-J., Johnson, R., Brown, E., Keith, B., & Simon, M. (2007). Acute postnatal ablation of Hif-2alpha results in anemia. Proceedings of the National Academy of Sciences of the United States of America , 104 (7), 2301-6. Gu, Y., Lewis, D., & Wang, Y. (2008). Placental productions and expressions of soluble endoglin, soluble fms-like tyrosine kinase receptor-1, and placental growth factor in normal and preeclamptic pregnancies. The Journal of clinical endocrinology and metabolism , 93 (1), 260-6. Guillemot, F., Nagy, A., Auerbach, A., Rossant, J., & Joyner, A. (1994). Essential role of Mash-2 in extraembryonic development. Nature , 371 (6495), 333-6. Gupta, P., & Lee, K. (2008, Jan 1). Silent mutations result in HlyA hypersecretion by reducing intracellular HlyA protein aggregates. Biotechnology and bioengineering . Hüttemann, M., Kadenbach, B., & Grossman, L. (2001). Mammalian subunit IV isoforms of cytochrome c oxidase. Gene , 267 (1), 111-23. Hüttemann, M., Lee, I., Liu, J., & Grossman, L. (2007). Transcription of mammalian cytochrome c oxidase subunit IV-2 is controlled by a novel conserved oxygen responsive element. The FEBS journal , 274 (21), 5737-48. Haas, I., & Wabl, M. (1983). Immunoglobulin heavy chain binding protein. Nature , 306 (5941), 387-9. Hall, D., Odendaal, H., Steyn, D., & Grové, D. (2000). Expectant management of early onset, severe pre-eclampsia: maternal outcome. BJOG : an international journal of obstetrics and gynaecology , 107 (10), 1252-7. Hamatani, T., Carter, M., Sharov, A., & Ko, M. (2004). Dynamics of global gene expression changes during mouse preimplantation development. Developmental Cell , 6 (1), 117-31. Handschuh, K., Guibourdenche, J., Tsatsaris, V., Guesnon, M., Laurendeau, I., Evain-Brion, D., et al. (2007). Human chorionic gonadotropin produced by the invasive trophoblast but not the villous

216 trophoblast promotes cell invasion and is down-regulated by peroxisome proliferator-activated receptor-gamma. Endocrinology , 148 (10), 5011-9. Hannan, N., & Salamonsen, L. (2008). CX3CL1 and CCL14 regulate extracellular matrix and adhesion molecules in the trophoblast: potential roles in human embryo implantation. Biology of reproduction , 79 (1), 58-65. Harpel, P. (1970, Jan 1). HUMAN PLASMA ALPHA 2-MACROGLOBULIN AN INHIBITOR OF PLASMA KALLIKREIN. The Journal of experimental medicine . Harpel, P. (1977, Jan 1). Plasmin inhibitor interactions. The effectiveness of alpha2-plasmin inhibitor in the presence of alpha2-macroglobulin. The Journal of experimental medicine . Harun, R., Ruban, L., Matin, M., Draper, J., Jenkins, N., Liew, G., et al. (2006). Cytotrophoblast stem cell lines derived from human embryonic stem cells and their capacity to mimic invasive implantation events. Human reproduction (Oxford, England) , 21 (6), 1349-58. Harvey, R. (2002). Patterning the vertebrate heart. Nature reviews Genetics , 3 (7), 544-56. Hassink, S., de Lancey, E., Sheslow, D., Smith-Kirwin, S., O'Connor, D., Considine, R., et al. (1997). Placental leptin: an important new growth factor in intrauterine and neonatal development? Pediatrics , 100 (1), E1. Hay, D. (1988). Placental histology and the production of human choriogonadotrophin and its subunits in pregnancy. British journal of obstetrics and gynaecology , 95 (12), 1268-75. Hay, D., & Lopata, A. (1988). Chorionic gonadotropin secretion by human embryos in vitro. The Journal of clinical endocrinology and metabolism , 67 (6), 1322-4. Hecht, F., Motulsky, A., Lemire, R., & Shepard, T. (1966). Predominance of hemoglobin Gower 1 in early human embryonic development. Science (New York, NY) , 152 (3718), 91-2. Hellman, R., Vanhove, M., Lejeune, A., Stevens, F., & Hendershot, L. (1999). The in vivo association of BiP with newly synthesized proteins is dependent on the rate and stability of folding and not simply on the presence of sequences that can bind to BiP. The Journal of cell biology , 144 (1), 21-30. Helske, S., Vuorela, P., Carpén, O., Hornig, C., Weich, H., & Halmesmäki, E. (2001). Expression of vascular endothelial growth factor receptors 1, 2 and 3 in placentas from normal and complicated pregnancies. Molecular human reproduction , 7 (2), 205-10. Henson, M., Swan, K., & O'Neil, J. (1998). Expression of placental leptin and leptin receptor transcripts in early pregnancy and at term. Obstetrics and gynecology , 92 (6), 1020-8. Herath, Wishva Bandara. (2011). Trancriptomics Of Early Human Development. (Doctoral dissertation). Retrieved from ScholarBank@NUS. Herrera, E. (2002). Implications of dietary fatty acids during pregnancy on placental, fetal and postnatal development--a review. Placenta , 23 Suppl A, S9-19. Herse, F., Dechend, R., Harsem, N., Wallukat, G., Janke, J., Qadri, F., et al. (2007). Dysregulation of the circulating and tissue-based renin-angiotensin system in preeclampsia. Hypertension , 49 (3), 604-11. Hertig, A., Rock, J., & Adams, E. (1956). A description of 34 human ova within the first 17 days of development. The American journal of anatomy , 98 (3), 435-93. Hertig, A., Rock, J., & Adams, E., & Menkin, M. (1959). Thirty-four fertilized human ova, good, bad and indifferent, recovered from 210 women of known fertility; a study of biologic wastage in early human pregnancy. Pediatrics , 23 (1 Part 2), 202-11. Herzog, M. (1909). A contribution to our knowledge of the earliest known stages of placentation and embryonic development in man. Am. J. Anat., 9 (1), 361–400. doi: 10.1002/aja.1000090112 Higgs, D., Vickers, M., Wilkie, A., Pretorius, I., Jarman, A., & Weatherall, D. (1989). A review of the molecular genetics of the human alpha-globin gene cluster. Blood , 73 (5), 1081-104. Hirano, H., Imai, Y., & Ito, H. (2002). Spiral artery of placenta: development and pathology- immunohistochemical, microscopical, and electron-microscopic study. The Kobe journal of medical sciences, 48 (1-2), 13-23. Hogan, B., Beddington, R., Constantini, F., and Lacy, E. (1994). Manipulating the mouse embryo a laboratory manual. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press.

217 Hoggard, N., Haggarty, P., Thomas, L., & Lea, R. (2001). Leptin expression in placental and fetal tissues: does leptin have a functional role? Biochemical Society transactions , 29 (Pt 2), 57-63. Holmes, C., Simpson, K., Okada, H., Okada, N., Wainwright, S., Purcell, D., et al. (1992). Complement regulatory proteins at the feto-maternal interface during human placental development: distribution of CD59 by comparison with membrane cofactor protein (CD46) and decay accelerating factor (CD55). European journal of immunology , 22 (6), 1579-85. Home, P., Ray, S., Dutta, D., Bronshteyn, I., Larson, M., & Paul, S. (2009). GATA3 is selectively expressed in the trophectoderm of peri-implantation embryo and directly regulates Cdx2 gene expression. The Journal of biological chemistry , 284 (42), 28729-37. Hopkinson, N., & Powell, R. (1992). Classical complement activation induced by pregnancy: implications for management of connective tissue diseases. Journal of clinical pathology , 45 (1), 66-7. Hoya-Arias, R., Tomishima, M., Perna, F., Voza, F., & Nimer, S. (2011). L3MBTL1 deficiency directs the differentiation of human embryonic stem cells toward trophectoderm. Stem cells and development , 20 (11), 1889-900. Hsiao, E., Koniaris, L., Zimmers-Koniaris, T., Sebald, S., Huynh, T., & Lee, S. (2000). Characterization of growth-differentiation factor 15, a transforming growth factor beta superfamily member induced following liver injury. Molecular and cellular biology , 20 (10), 3742-51. Hu, C.-J., Sataur, A., Wang, L., Chen, H., & Simon, M. (2007). The N-terminal transactivation domain confers target gene specificity of hypoxia-inducible factors HIF-1alpha and HIF-2alpha. Molecular biology of the cell , 18 (11), 4528-42. Hu, C.-J., Wang, L.-Y., Chodosh, L., Keith, B., & Simon, M. (2003). Differential roles of hypoxia- inducible factor 1alpha (HIF-1alpha) and HIF-2alpha in hypoxic gene regulation. Molecular and cellular biology , 23 (24), 9361-74. Huang, L., & Bunn, H. (2003). Hypoxia-inducible factor and its biomedical relevance. The Journal of biological chemistry , 278 (22), 19575-8. Huang, Y., Qiao, F., Abagyan, R., Hazard, S., & Tomlinson, S. (2006). Defining the CD59-C9 binding interaction. The Journal of biological chemistry , 281 (37), 27398-404. Hubé, F., Reverdiau, P., Iochmann, S., Trassard, S., Thibault, G., & Gruel, Y. (2003). Demonstration of a tissue factor pathway inhibitor 2 messenger RNA synthesis by pure villous cytotrophoblast cells isolated from term human placentas. Biology of reproduction , 68 (5), 1888-94. Hunkapiller, N., & Fisher, S. (2008). Chapter 12. Placental remodeling of the uterine vasculature. Methods in enzymology , 445, 281-302. Hustin, J., & Schaaps, J. (1987). Echographic [corrected] and anatomic studies of the maternotrophoblastic border during the first trimester of pregnancy. YMOB , 157 (1), 162-8. Hustin, J., Jauniaux, E., & Schaaps, J. (1990). Histological study of the materno-embryonic interface in spontaneous abortion. Placenta , 11 (6), 477-86. Hutson, M., & Kirby, M. (2007). Model systems for the study of heart development and disease. Cardiac neural crest and conotruncal malformations. Seminars in cell & developmental biology , 18 (1), 101-10. Hyman, A., & Simons, K. (2011). The new cell biology: Beyond HeLa cells. Nature , 480 (7375), 34. Hyslop, L., Stojkovic, M., Armstrong, L., Walter, T., Stojkovic, P., Przyborski, S., et al. (2005). Downregulation of NANOG induces differentiation of human embryonic stem cells to extraembryonic lineages. Stem cells (Dayton, Ohio) , 23 (8), 1035-43. Ichinose, A., & Davie, E. (1988). Characterization of the gene for the a subunit of human factor XIII (plasma transglutaminase), a blood coagulation factor. Proceedings of the National Academy of Sciences of the United States of America , 85 (16), 5829-33. Ichinose, A., Asahina, T., & Kobayashi, T. (2005). Congenital blood coagulation factor XIII deficiency and perinatal management. Current drug targets , 6 (5), 541-9. Ietta, F., Wu, Y., Winter, J., Xu, J., Wang, J., Post, M., et al. (2006). Dynamic HIF1A regulation during human placental development. Biology of reproduction , 75 (1), 112-21.

218 Illingworth, P., Groome, N., Duncan, W., Grant, V., Tovanabutra, S., Baird, D., et al. (1996). Measurement of circulating inhibin forms during the establishment of pregnancy. The Journal of clinical endocrinology and metabolism , 81 (4), 1471-5. INGALLS, A., DICKIE, M., & SNELL, G. (1950). Obese, a new mutation in the house mouse. The Journal of heredity , 41 (12), 317-8. International Stem Cell Initiative, Amps, K., Andrews, P., Anyfantis, G., Armstrong, L., Avery, S., et al. (2011). Screening ethnically diverse human embryonic stem cells identifies a chromosome 20 minimal amplicon conferring growth advantage. Nature biotechnology , 29 (12), 1132-44. Isse, N., Ogawa, Y., Tamura, N., Masuzaki, H., Mori, K., Okazaki, T., et al. (1995). Structural organization and chromosomal assignment of the human obese gene. The Journal of biological chemistry , 270 (46), 27728-33. Itskovitz-Eldor, J., Schuldiner, M., Karsenti, D., Eden, A., Yanuka, O., Amit, M., et al. (2000). Differentiation of human embryonic stem cells into embryoid bodies compromising the three embryonic germ layers. Molecular medicine (Cambridge, Mass) , 6 (2), 88-95. Ivan, M., Kondo, K., Yang, H., Kim, W., Valiando, J., Ohh, M., et al. (2001). HIFalpha targeted for VHL-mediated destruction by proline hydroxylation: implications for O2 sensing. Science (New York, NY) , 292 (5516), 464-8. Iyer, N., Kotch, L., Agani, F., Leung, S., Laughner, E., Wenger, R., et al. (1998). Cellular and developmental control of O2 homeostasis by hypoxia-inducible factor 1 alpha. Genes & Development , 12 (2), 149-62. Jaakkola, P., Mole, D., Tian, Y., Wilson, M., Gielbert, J., Gaskell, S., et al. (2001). Targeting of HIF- alpha to the von Hippel-Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation. Science (New York, NY) , 292 (5516), 468-72. Jackson, M., Cohen-Doyle, M., Peterson, P., & Williams, D. (1994). Regulation of MHC class I transport by the molecular chaperone, calnexin (p88, IP90). Science (New York, NY) , 263 (5145), 384-7. Jackson, N., Biddolph, S., Ledger, W., Groome, N., & Manek, S. (2000). Inhibin expression in normal and pre-eclamptic placental tissue. International journal of gynecological pathology : official journal of the International Society of Gynecological Pathologists , 19 (3), 219-24. Jacobson, E., Concepcion, E., Oashi, T., & Tomer, Y. (2005, Jan 1). … ' disease-associated Kozak sequence single-nucleotide polymorphism enhances the efficiency of CD40 gene translation: a case for translational pathophysiology. Endocrinology . Jaffe, R., Jauniaux, E., & Hustin, J. (1997). Maternal circulation in the first-trimester human placenta-- myth or reality? YMOB , 176 (3), 695-705. Jaffe, R., Lee, P., & Midgley, A. (1969). Serum gonadotropins before, at the inception of, and following human pregnancy. The Journal of clinical endocrinology and metabolism , 29 (9), 1281-3. James, D., Levine, A., Besser, D., & Hemmati-Brivanlou, A. (2005). TGFbeta/activin/nodal signaling is necessary for the maintenance of pluripotency in human embryonic stem cells. Development (Cambridge, England) , 132 (6), 1273-82. James, J., Carter, A., & Chamley, L. (2012). Human placentation from nidation to 5 weeks of gestation. Part I: What do we know about formative placental development following implantation? Placenta , 33 (5), 327-34. James, J., Carter, A., & Chamley, L. (2012). Human placentation from nidation to 5 weeks of gestation. Part II: Tools to model the crucial first days. Placenta , 33 (5), 335-42. James, J., Stone, P., & Chamley, L. (2005). Cytotrophoblast differentiation in the first trimester of pregnancy: evidence for separate progenitors of extravillous trophoblasts and syncytiotrophoblast. Reproduction (Cambridge, England) , 130 (1), 95-103. James, J., Stone, P., & Chamley, L. (2007). The isolation and characterization of a population of extravillous trophoblast progenitors from first trimester human placenta. Human reproduction (Oxford, England) , 22 (8), 2111-9. Jameson, J., & Hollenberg, A. (1993). Regulation of chorionic gonadotropin gene expression. Endocrine Reviews , 14 (2), 203-21.

219 Janatpour, M., McMaster, M., Genbacev, O., Zhou, Y., Dong, J., Cross, J., et al. (2000). Id-2 regulates critical aspects of human cytotrophoblast differentiation, invasion and migration. Development (Cambridge, England) , 127 (3), 549-58. Janatpour, M., Utset, M., Cross, J., Rossant, J., Dong, J., Israel, M., et al. (1999). A repertoire of differentially expressed transcription factors that offers insight into mechanisms of human cytotrophoblast differentiation. Developmental genetics , 25 (2), 146-57. Janeway, C.A. Jr., Travers, P., Walport, M., et al. (2001). The major histocompatibility complex and its functions. In 5th edition, Immunobiology: The Immune System in Health and Disease (Available from: http://www.ncbi.nlm.nih.gov/books/NBK27156/). New York: Garland Science. Janeway, C.A. Jr., Travers, P., Walport, M., et al. (2001). The humoral immune response. In 5th edition, Immunobiology: The Immune System in Health and Disease (Available from: http:// www.ncbi.nlm.nih.gov/books/NBK27156/). New York: Garland Science. Janeway, C.A. Jr., Travers, P., Walport, M., et al. (2001). The complement system and innate immunity. In 5th edition, Immunobiology: The Immune System in Health and Disease (Available from: http:// www.ncbi.nlm.nih.gov/books/NBK27100/). New York: Garland Science. Janszen, F., Cooke, B., van Driel, J., & van der Molen, H. (1978). Regulation of the synthesis of lutropin-induced protein in rat testis Leydig cells. The Biochemical journal , 170 (1), 9-15. Jauniaux, E., Greenwold, N., Hempstock, J., & Burton, G. (2003). Comparison of ultrasonographic and Doppler mapping of the intervillous circulation in normal and abnormal early pregnancies. Fertility and sterility , 79 (1), 100-6. Jauniaux, E., Hempstock, J., Greenwold, N., & Burton, G. (2003). Trophoblastic oxidative stress in relation to temporal and regional differences in maternal placental blood flow in normal and abnormal early pregnancies. The American journal of pathology , 162 (1), 115-25. Jauniaux, E., Jurkovic, D., & Campbell, S. (1995). Current topic: in vivo investigation of the placental circulations by Doppler echography. Placenta , 16 (4), 323-31. Jauniaux, E., Jurkovic, D., Campbell, S., & Hustin, J. (1992). Doppler ultrasonographic features of the developing placental circulation: Correlation with anatomic findings. YMOB , 166 (2), 585-7. Jauniaux, E., Watson, A., & Burton, G. (2001). Evaluation of respiratory gases and acid-base gradients in human fetal fluids and uteroplacental tissue between 7 and 16 weeks' gestation. YMOB , 184 (5), 998-1003. Jauniaux, E., Watson, A., Hempstock, J., Bao, Y., Skepper, J., & Burton, G. (2000). Onset of maternal arterial blood flow and placental oxidative stress. A possible factor in human early pregnancy failure. The American journal of pathology , 157 (6), 2111-22. Jenkins, J., Agrawal, A., Gupta, A., Green, M., & Kielkopf, C. (2013, Feb 1). U2AF65 adapts to diverse pre-mRNA splice sites through conformational selection of specific and promiscuous RNA recognition motifs. Nucleic Acids Research . Jensen, G., & Moore, L. (1997). The effect of high altitude and other risk factors on birthweight: independent or interactive effects? American journal of public health , 87 (6), 1003-7. Johnson, G., Burghardt, R., Bazer, F., & Spencer, T. (2003). Osteopontin: roles in implantation and placentation. Biology of reproduction , 69 (5), 1458-71. Johnson, M., & Ziomek, C. (1981). Induction of polarity in mouse 8-cell blastomeres: specificity, geometry, and stability. The Journal of cell biology , 91 (1), 303-8. Jokimaa, V., Oksjoki, S., Kujari, H., Vuorio, E., & Anttila, L. (2001). Expression patterns of cathepsins B, H, K, L and S in the human endometrium. Molecular human reproduction , 7 (1), 73-8. Josimovich, J., Atwood, B., & Goss, D. (1963). Luteotrophic, immunologic and electrophoretic properties of human placental lactogen. Endocrinology , 73, 410-20. Julian, C., Vargas, E., Armaza, J., Wilson, M., Niermeyer, S., & Moore, L. (2007). High-altitude ancestry protects against hypoxia-associated reductions in fetal growth. Archives of disease in childhood Fetal and neonatal edition , 92 (5), F372-7. Julian, C., Wilson, M., & Moore, L. (2009). Evolutionary adaptation to high altitude: A view from in utero. American Journal of Human Biology , 21 (5), 614-622.

220 Jurisicova, A., Antenos, M., Kapasi, K., Meriano, J., & Casper, R. (1999). Variability in the expression of trophectodermal markers beta-human chorionic gonadotrophin, human leukocyte antigen-G and pregnancy specific beta-1 glycoprotein by the human blastocyst. Human reproduction (Oxford, England) , 14 (7), 1852-8. Jurisicova, A., Detmar, J., & Caniggia, I. (2005). Molecular mechanisms of trophoblast survival: from implantation to birth. Birth defects research Part C, Embryo today : reviews , 75 (4), 262-80. Kaelin, W., & Ratcliffe, P. (2008). Oxygen sensing by metazoans: the central role of the HIF hydroxylase pathway. Molecular Cell , 30 (4), 393-402. Kalluri, R., & Neilson, E. (2003). Epithelial-mesenchymal transition and its implications for fibrosis. The Journal of clinical investigation , 112 (12), 1776-84. Kaluz, S., Kaluzová, M., & Stanbridge, E. (2008). Regulation of gene expression by hypoxia: integration of the HIF-transduced hypoxic signal at the hypoxia-responsive element. Clinica chimica acta; international journal of clinical chemistry , 395 (1-2), 6-13. Kang, J., Song, H., Yoon, J., Park, D., Kim, S., Lee, K., et al. (2011). Preeclampsia leads to dysregulation of various signaling pathways in placenta. Journal of hypertension , 29 (5), 928-36. Kapitsinou, P., Liu, Q., Unger, T., Rha, J., Davidoff, O., Keith, B., et al. (2010). Hepatic HIF-2 regulates erythropoietic responses to hypoxia in renal anemia. Blood , 116 (16), 3039-48. Karlsson, S., & Nienhuis, A. (1985). Developmental regulation of human globin genes. Annual review of biochemistry , 54, 1071-108. Karsenty, G. (2008). Transcriptional control of skeletogenesis. Annual review of genomics and human genetics , 9, 183-96. Kaufman, N., Page, J., Pixley, R., Schein, R., Schmaier, A., & Colman, R. (1991). Alpha 2- macroglobulin-kallikrein complexes detect contact system activation in hereditary angioedema and human sepsis. Blood , 77 (12), 2660-7. Kauma, S., Bae-Jump, V., & Walsh, S. (1999). Hepatocyte growth factor stimulates trophoblast invasion: a potential mechanism for abnormal placentation in preeclampsia. The Journal of clinical endocrinology and metabolism , 84 (11), 4092-6. Kauma, S., Hayes, N., & Weatherford, S. (1997). The differential expression of hepatocyte growth factor and met in human placenta. The Journal of clinical endocrinology and metabolism , 82 (3), 949-54. Keene, J. (2007, Jan 1). RNA regulons: coordination of post-transcriptional events. Nature reviews Genetics . Kendall, R., & Thomas, K. (1993). Inhibition of vascular endothelial cell growth factor activity by an endogenously encoded soluble receptor. Proceedings of the National Academy of Sciences of the United States of America , 90 (22), 10705-9. Keyes, L., Armaza, J., Niermeyer, S., Vargas, E., Young, D., & Moore, L. (2003). Intrauterine growth restriction, preeclampsia, and intrauterine mortality at high altitude in Bolivia. Pediatric research , 54 (1), 20-5. Khong, T., De Wolf, F., Robertson, W., & Brosens, I. (1986). Inadequate maternal vascular response to placentation in pregnancies complicated by pre-eclampsia and by small-for-gestational age infants. British journal of obstetrics and gynaecology , 93 (10), 1049-59. Khong, T., Liddell, H., & Robertson, W. (1987). Defective haemochorial placentation as a cause of miscarriage: a preliminary study. British journal of obstetrics and gynaecology , 94 (7), 649-55. Kiefer, M., Bauer, D., & Barr, P. (1989). The cDNA and derived amino acid sequence for human osteopontin. Nucleic Acids Research , 17 (8), 3306. Kielkopf, C., Lücke, S., & Green, M. (2004). U2AF homology motifs: protein recognition in the RRM world. Genes & Development , 18 (13), 1513-26. Kim, P., Lu, L., Xia, Y., & Gerstein, M. (2006). Relating three-dimensional structures to protein networks provides evolutionary insights. Science (New York, NY) , 314 (5807), 1938-41. Kim, Y., Bujold, E., Chaiworapongsa, T., Gomez, R., Yoon, B., Thaler, H., et al. (2003). Failure of physiologic transformation of the spiral arteries in patients with preterm labor and intact membranes. YMOB , 189 (4), 1063-9.

221 Kim, Y., Chaiworapongsa, T., Gomez, R., Bujold, E., Yoon, B., Rotmensch, S., et al. (2002). Failure of physiologic transformation of the spiral arteries in the placental bed in preterm premature rupture of membranes. YMOB , 187 (5), 1137-42. Kimber, S., Sneddon, S., Bloor, D., El-Bareg, A., Hawkhead, J., Metcalfe, A., et al. (2008). Expression of genes involved in early cell fate decisions in human embryos and their regulation by growth factors. Reproduction (Cambridge, England) , 135 (5), 635-47. Kimchi-Sarfaty, C., Oh, J., Kim, I., & Sauna…, Z. (2007, Jan 1). A" silent" polymorphism in the MDR1 gene changes substrate specificity. Science (New York, NY) . Kimura, M. (1979). The neutral theory of molecular evolution. Scientific American , 241 (5), 98-100, 102, 108 passim. Knöfler, M., Meinhardt, G., Vasicek, R., Husslein, P., & Egarter, C. (1998). Molecular cloning of the human Hand1 gene/cDNA and its tissue-restricted expression in cytotrophoblastic cells and heart. Gene , 224 (1-2), 77-86. Knittler, M., Dirks, S., & Haas, I. (1995). Molecular chaperones involved in protein degradation in the endoplasmic reticulum: quantitative interaction of the heat shock cognate protein BiP with partially folded immunoglobulin light chains that are degraded in the endoplasmic reticulum. Proceedings of the National Academy of Sciences of the United States of America , 92 (5), 1764-8. Koga, K., Osuga, Y., Yoshino, O., Hirota, Y., Ruimeng, X., Hirata, T., et al. (2003). Elevated serum soluble vascular endothelial growth factor receptor 1 (sVEGFR-1) levels in women with preeclampsia. The Journal of clinical endocrinology and metabolism , 88 (5), 2348-51. Koh, M., & Powis, G. (2012). Passing the baton: the HIF switch. Trends in biochemical sciences , 37 (9), 364-72. Kojima, I., Tanaka, T., Inagi, R., Kato, H., Yamashita, T., Sakiyama, A., et al. (2007). Protective role of hypoxia-inducible factor-2alpha against ischemic damage and oxidative stress in the kidney. Journal of the American Society of Nephrology : JASN , 18 (4), 1218-26. Komamura, K., Iwai, N., Kokame, K., Yasumura, Y., Kim, J., Yamagishi, M., et al. (2004). The role of a common TNNT2 polymorphism in cardiac hypertrophy. Journal of human genetics , 49 (3), 129-33. Komar, A., Lesnik, T., & Reiss, C. (1999, Jan 1). Synonymous codon substitutions affect ribosome traffic and protein folding during in vitro translation. Febs Letters . Konigsberg, W., Kirchhofer, D., Riederer, M., & Nemerson, Y. (2001). The TF:VIIa complex: clinical significance, structure-function relationships and its role in signaling and metastasis. Thrombosis and haemostasis , 86 (3), 757-71. Kotch, L., Iyer, N., Laughner, E., & Semenza, G. (1999). Defective vascularization of HIF-1alpha-null embryos is not associated with VEGF deficiency but with mesenchymal cell death. Developmental biology , 209 (2), 254-67. Koyano, S., Kurose, K., Saito, Y., Ozawa, S., Hasegawa, R., Komamura, K., et al. (2004). Functional characterization of four naturally occurring variants of human pregnane X receptor (PXR): one variant causes dramatic loss of both DNA binding activity and the transactivation of the CYP3A4 promoter/ enhancer region. Drug metabolism and disposition: the biological fate of chemicals , 32 (1), 149-54. Kozak, K., Abbott, B., & Hankinson, O. (1997). ARNT-deficient mice and placental differentiation. Developmental biology , 191 (2), 297-305. KraÈutler, B., Konrat, R., & Bister, K. (2001, Jan 1). Structure, function, and dynamics of the dimerization and DNA-binding domain of oncogenic transcription factor v-Myc. J. Mol. Biol . Krampl, E., Lees, C., Bland, J., Espinoza Dorado, J., Moscoso, G., & Campbell, S. (2000). Fetal biometry at 4300 m compared to sea level in Peru. Ultrasound in obstetrics & gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology , 16 (1), 9-18. Krawczak, M., Reiss, J., & Cooper, D. (1992). The mutational spectrum of single base-pair substitutions in mRNA splice junctions of human genes: causes and consequences. Human genetics , 90 (1-2), 41-54. Krawczak, M., Thomas, N., Hundrieser, B., Mort, M., Wittig, M., Hampe, J., et al. (2007). Single base- pair substitutions in exon-intron junctions of human genes: nature, distribution, and consequences for mRNA splicing. Human mutation , 28 (2), 150-8.

222 Krishnan, J., Ahuja, P., Bodenmann, S., Knapik, D., Perriard, E., Krek, W., et al. (2008). Essential role of developmentally activated hypoxia-inducible factor 1alpha for cardiac morphogenesis and function. Circulation research , 103 (10), 1139-46. Kroll, K., Otte, S., Hirschfeld, G., Barnikol-Watanabe, S., Götz, H., Sternbach, H., et al. (1999). Heterologous overexpression of human NEFA and studies on the two EF-hand calcium-binding sites. Biochemical and Biophysical Research Communications , 260 (1), 1-8. Krupinski, P., Chickarmane, V., & Peterson, C. (2011). Simulating the mammalian blastocyst-- molecular and mechanical interactions pattern the embryo. PLoS computational biology , 7 (5), e1001128. Krylov, D., Wolf, Y., Rogozin, I., & Koonin, E. (2003). Gene loss, protein sequence divergence, gene dispensability, expression level, and interactivity are correlated in eukaryotic evolution. Genome Research , 13 (10), 2229-35. Kunz, S., Rojek, J., Kanagawa, M., Spiropoulou, C., Barresi, R., Campbell, K., et al. (2005). Posttranslational modification of alpha-dystroglycan, the cellular receptor for arenaviruses, by the glycosyltransferase LARGE is critical for virus binding. Journal of virology , 79 (22), 14282-96. Lacroix, M., Guibourdenche, J., Frendo, J., Muller, F., & Evain-Brion, D. (2002). Human placental growth hormone--a review. Placenta , 23 Suppl A, S87-94. Lahlou, N., Chabbert-Buffet, N., Christin-Maitre, S., Le Nestour, E., Roger, M., & Bouchard, P. (1999). Main inhibitor of follicle stimulating hormone in the luteal-follicular transition: inhibin A, oestradiol, or inhibin B? Human reproduction (Oxford, England) , 14 (5), 1190-3. Lai, E., Teodoro, T., & Volchuk, A. (2007). Endoplasmic reticulum stress: signaling the unfolded protein response. Physiology (Bethesda, Md) , 22, 193-201. Landa, I., Ruiz-Llorente, S., & Montero-Conde…, C. (2009, Jan 1). The variant rs1867277 in FOXE1 gene confers thyroid cancer susceptibility through the recruitment of USF1/USF2 transcription factors. PLoS genetics . Lando, D., Peet, D., Whelan, D., Gorman, J., & Whitelaw, M. (2002). Asparagine hydroxylation of the HIF transactivation domain a hypoxic switch. Science (New York, NY) , 295 (5556), 858-61. Lanir, N., Aharon, A., & Brenner, B. (2003). Haemostatic mechanisms in human placenta. Best practice & research Clinical haematology , 16 (2), 183-95. Lastres, P., Letamendía, A., Zhang, H., Rius, C., Almendro, N., Raab, U., et al. (1996). Endoglin modulates cellular responses to TGF-beta 1. The Journal of cell biology , 133 (5), 1109-21. Lau, K., Tian, Y.-M., Raval, R., Ratcliffe, P., & Pugh, C. (2007). Target gene selectivity of hypoxia- inducible factor-α in renal cancer cells is conveyed by post-DNA-binding mechanisms. British Journal of Cancer , 96 (8), 1284-1292. Lea, R., Howe, D., Hannah, L., Bonneau, O., Hunter, L., & Hoggard, N. (2000). Placental leptin in normal, diabetic and fetal growth-retarded pregnancies. Molecular human reproduction , 6 (8), 763-9. Leaños-Miranda, A., Campos-Galicia, I., Ramírez-Valenzuela, K., Chinolla-Arellano, Z., & Isordia- Salas, I. (2013, Mar 4). Circulating Angiogenic Factors and Urinary Prolactin as Predictors of Adverse Outcomes in Women With Preeclampsia. Hypertension . Lee, Y., Brewer, J., Hellman, R., & Hendershot, L. (1999). BiP and immunoglobulin light chain cooperate to control the folding of heavy chain and ensure the fidelity of immunoglobulin assembly. Molecular biology of the cell , 10 (7), 2209-19. Lenter, M., & Vestweber, D. (1994). The integrin chains beta 1 and alpha 6 associate with the chaperone calnexin prior to integrin assembly. The Journal of biological chemistry , 269 (16), 12263-8. Lepercq, J., Cauzac, M., Lahlou, N., Timsit, J., Girard, J., Auwerx, J., et al. (1998). Overexpression of placental leptin in diabetic pregnancy: a critical role for insulin. Diabetes , 47 (5), 847-50. Lepercq, J., Lahlou, N., Timsit, J., Girard, J., & Mouzon, S. (1999). Macrosomia revisited: ponderal index and leptin delineate subtypes of fetal overgrowth. American journal of obstetrics and gynecology , 181 (3), 621-5. Leung, T., Hall, C., Rajendran, M., Spurr, N., & Lim, L. (1992). The human heat-shock genes HSPA6 and HSPA7 are both expressed and localize to chromosome 1. Genomics , 12 (1), 74-9.

223 León-Velarde, F., Maggiorini, M., Reeves, J., Aldashev, A., Asmus, I., Bernardi, L., et al. (2005). Consensus statement on chronic and subacute high altitude diseases. High Altitude Medicine & Biology , 6 (2), 147-57. Levine, R., Lam, C., Qian, C., Yu, K., Maynard, S., Sachs, B., et al. (2006). Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. The New England journal of medicine , 355 (10), 992-1005. Levine, R., Maynard, S., Qian, C., Lim, K.-H., England, L., Yu, K., et al. (2004). Circulating angiogenic factors and the risk of preeclampsia. The New England journal of medicine , 350 (7), 672-83. Lewis, S., Janus, T., Lorand, L., & Shafer, J. (1985). Regulation of formation of factor XIIIa by its fibrin substrates. Biochemistry , 24 (24), 6772-7. Ley, T., Maloney, K., Gordon, J., & Schwartz, A. (1989). Globin gene expression in erythroid human fetal liver cells. The Journal of clinical investigation , 83 (3), 1032-8. Li, L., Zhang, X., Kovacic, S., Long, A., Bourque, K., Wood, C., et al. (2000). Identification of a human follicular dendritic cell molecule that stimulates germinal center B cell growth. The Journal of experimental medicine , 191 (6), 1077-84. Li, Y.-Q., Yan, H., & Bai, B. (2008). Change in iron transporter expression in human term placenta with different maternal iron status. European journal of obstetrics, gynecology, and reproductive biology , 140 (1), 48-54. LICHTY, J., TING, R., BRUNS, P., & DYAR, E. (1957). Studies of babies born at high altitudes. I. Relation of altitude to birth weight. AMA journal of diseases of children , 93 (6), 666-9. Ling, N., Ying, S., Ueno, N., Shimasaki, S., Esch, F., Hotta, M., et al. (1986). Pituitary FSH is released by a heterodimer of the beta-subunits from the two forms of inhibin. Nature , 321 (6072), 779-82. Lisy, K., & Peet, D. (2008). Turn me on: regulating HIF transcriptional activity. Cell Death and Differentiation , 15 (4), 642-9. Lockwood, C., Krikun, G., Runic, R., Schwartz, L., Mesia, A., & Schatz, F. (2000). Progestin- epidermal growth factor regulation of tissue factor expression during decidualization of human endometrial stromal cells. The Journal of clinical endocrinology and metabolism , 85 (1), 297-301. Lockwood, C., Toti, P., Arcuri, F., Norwitz, E., Funai, E., Huang, S.-T., et al. (2007). Thrombin regulates soluble fms-like tyrosine kinase-1 (sFlt-1) expression in first trimester decidua: implications for preeclampsia. The American journal of pathology , 170 (4), 1398-405. Lopata, A., Bentin-Ley, U., & Enders, A. (2002). "Pinopodes" and implantation. Reviews in endocrine & metabolic disorders , 3 (2), 77-86. Louwen, F., Muschol-Steinmetz, C., Reinhard, J., Reitter, A., & Yuan, J. (2012). A lesson for cancer research: placental microarray gene analysis in preeclampsia. Oncotarget , 3 (8), 759-73. Lublin, D., Lemons, R., Le Beau, M., Holers, V., Tykocinski, M., Medof, M., et al. (1987). The gene encoding decay-accelerating factor (DAF) is located in the complement-regulatory locus on the long arm of chromosome 1. The Journal of experimental medicine , 165 (6), 1731-6. Ludwig, T., Bergendahl, V., Levenstein, M., Yu, J., Probasco, M., & Thomson, J. (2006). Feeder- independent culture of human embryonic stem cells. Nature methods , 3 (8), 637-46. Luo, Wenlong. (2008). A Human Embryonic Stem Cell–based Model of Trophoblast Formation. (Doctoral dissertation). Retrieved from ScholarBank@NUS. López-Novoa, J. (2007). Soluble endoglin is an accurate predictor and a pathogenic molecule in pre- eclampsia. Nephrology, dialysis, transplantation : official publication of the European Dialysis and Transplant Association - European Renal Association , 22 (3), 712-4. Lynch, A., Murphy, J., Byers, T., Gibbs, R., Neville, M., Giclas, P., et al. (2008). Alternative complement pathway activation fragment Bb in early pregnancy as a predictor of preeclampsia. American journal of obstetrics and gynecology , 198 (4), 385.e1-9. Männik, J., Vaas, P., Rull, K., Teesalu, P., & Laan, M. (2012). Differential placental expression profile of human Growth Hormone/Chorionic Somatomammotropin genes in pregnancies with pre-eclampsia and gestational diabetes mellitus. Molecular and cellular endocrinology , 355 (1), 180-7.

224 Ma, G., Roth, M., Groskopf, J., Tsai, F., Orkin, S., Grosveld, F., et al. (1997). GATA-2 and GATA-3 regulate trophoblast-specific gene expression in vivo. Development (Cambridge, England) , 124 (4), 907-14. Mackman, N., Morrissey, J., Fowler, B., & Edgington, T. (1989). Complete sequence of the human tissue factor gene, a highly regulated cellular receptor that initiates the coagulation protease cascade. Biochemistry , 28 (4), 1755-62. Makino, Y., Kanopka, A., Wilson, W., Tanaka, H., & Poellinger, L. (2002). Inhibitory PAS domain protein (IPAS) is a hypoxia-inducible splicing variant of the hypoxia-inducible factor-3alpha locus. The Journal of biological chemistry , 277 (36), 32405-8. Maltepe, E., Bakardjiev, A., & Fisher, S. (2010). The placenta: transcriptional, epigenetic, and physiological integration during development. The Journal of clinical investigation , 120 (4), 1016-25. Maltepe, E., Krampitz, G., Okazaki, K., Red-Horse, K., Mak, W., Simon, M., et al. (2005). Hypoxia- inducible factor-dependent histone deacetylase activity determines stem cell fate in the placenta. Development (Cambridge, England) , 132 (15), 3393-403. Mamo, S., Sargent, C., Affara, N., Tesfaye, D., El-Halawany, N., Wimmers, K., et al. (2006). Transcript profiles of some developmentally important genes detected in bovine oocytes and in vitro- produced blastocysts using RNA amplification and cDNA microarrays. Reproduction in domestic animals = Zuchthygiene , 41 (6), 527-34. Manalo, D., Rowan, A., Lavoie, T., Natarajan, L., Kelly, B., Ye, S., et al. (2005). Transcriptional regulation of vascular endothelial cell responses to hypoxia by HIF-1. Blood , 105 (2), 659-69. Manuelpillai, U., Schneider-Kolsky, M., Dole, A., & Wallace, E. (2001). Activin A and activin receptors in gestational tissue from preeclamptic pregnancies. The Journal of endocrinology , 171 (1), 57-64. Marchand, M., Horcajadas, J., Esteban, F., McElroy, S., Fisher, S., & Giudice, L. (2011). Transcriptomic signature of trophoblast differentiation in a human embryonic stem cell model. Biology of reproduction , 84 (6), 1258-71. Margetic, S., Gazzola, C., Pegg, G., & Hill, R. (2002). Leptin: a review of its peripheral actions and interactions. International journal of obesity and related metabolic disorders : journal of the International Association for the Study of Obesity , 26 (11), 1407-33. Marinoni, E., Pacioni, K., Sambuchini, A., Moscarini, M., Letizia, C., & DI Iorio, R. (2011). Regulation by hypoxia of adrenomedullin output and expression in human trophoblast cells. European journal of obstetrics, gynecology, and reproductive biology , 154 (2), 146-50. Martial, J., Hallewell, R., Baxter, J., & Goodman, H. (1979). Human growth hormone: complementary DNA cloning and expression in bacteria. Science (New York, NY) , 205 (4406), 602-7. Martin, G. (1981). Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proceedings of the National Academy of Sciences of the United States of America , 78 (12), 7634-8. Martin, R. (2008, Jan 1). Evolution of placentation in primates: implications of mammalian phylogeny. Evolutionary Biology . Masson, N., Willam, C., Maxwell, P., Pugh, C., & Ratcliffe, P. (2001). Independent function of two destruction domains in hypoxia-inducible factor-alpha chains activated by prolyl hydroxylation. The EMBO Journal , 20 (18), 5197-206. Masuzaki, H., Ogawa, Y., Isse, N., Satoh, N., Okazaki, T., Shigemoto, M., et al. (1995). Human obese gene expression. Adipocyte-specific expression and regional differences in the adipose tissue. Diabetes , 44 (7), 855-8. Masuzaki, H., Ogawa, Y., Sagawa, N., Hosoda, K., Matsumoto, T., Mise, H., et al. (1997). Nonadipose tissue production of leptin: leptin as a novel placenta-derived hormone in humans. Nature Medicine , 3 (9), 1029-33. Matin, M., Walsh, J., Gokhale, P., Draper, J., Bahrami, A., Morton, I., et al. (2004). Specific knockdown of Oct4 and beta2-microglobulin expression by RNA interference in human embryonic stem cells and embryonic carcinoma cells. Stem cells (Dayton, Ohio) , 22 (5), 659-68.

225 Maxwell, P., Wiesener, M., Chang, G., Clifford, S., Vaux, E., Cockman, M., et al. (1999). The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature , 399 (6733), 271-5. Maymon, R., Bahari, C., & Moroz, C. (1989). Placental isoferritin measured by a specific monoclonal antibody as a predictive marker for preterm contraction outcome. Obstetrics and gynecology , 74 (4), 597-9. Maymó, J., Pérez Pérez, A., Maskin, B., Dueñas, J., Calvo, J., Sánchez Margalet, V., et al. (2012). The Alternative Epac/cAMP pathway and the MAPK pathway mediate hCG induction of leptin in placental cells. PLoS ONE , 7 (10), e46216. Maynard, M., Qi, H., Chung, J., Lee, E., Kondo, Y., Hara, S., et al. (2003). Multiple splice variants of the human HIF-3 alpha locus are targets of the von Hippel-Lindau E3 ubiquitin ligase complex. The Journal of biological chemistry , 278 (13), 11032-40. Maynard, S., Min, J.-Y., Merchan, J., Lim, K.-H., Li, J., Mondal, S., et al. (2003). Excess placental soluble fms-like tyrosine kinase 1 (sFlt1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia. The Journal of clinical investigation , 111 (5), 649-58. Mayor-Lynn, K., Toloubeydokhti, T., Cruz, A., & Chegini, N. (2011). Expression profile of microRNAs and mRNAs in human placentas from pregnancies complicated by preeclampsia and preterm labor. Reproductive sciences (Thousand Oaks, Calif) , 18 (1), 46-56. Mazess, R. (1965). Neonatal mortality and altitude in Peru. American journal of physical anthropology , 23 (3), 209-13. McCullough, R., Reeves, J., & Liljegren, R. (1977). Fetal growth retardation and increased infant mortility at high altitude. Obstetrical & gynecological survey , 32 (7), 596-8. Meister, G., & Tuschl, T. (2004, Jan 1). Mechanisms of gene silencing by double-stranded RNA. Nature . Meng, T., Chen, H., Sun, M., Wang, H., Zhao, G., & Wang, X. (2012). Identification of differential gene expression profiles in placentas from preeclamptic pregnancies versus normal pregnancies by DNA microarrays. Omics : a journal of integrative biology , 16 (6), 301-11. Midgley, A., & Jaffe, R. (1968). Regulation of human gonadotropins. II. Disappearance of human chorionic gonadotropin following delivery. The Journal of clinical endocrinology and metabolism , 28 (12), 1712-8. Migliaccio, G., Migliaccio, A., Petti, S., Mavilio, F., Russo, G., Lazzaro, D., et al. (1986). Human embryonic hemopoiesis. Kinetics of progenitors and precursors underlying the yolk sac----liver transition. The Journal of clinical investigation , 78 (1), 51-60. Mikheev, A., Nabekura, T., Kaddoumi, A., Bammler, T., Govindarajan, R., Hebert, M., et al. (2008). Profiling gene expression in human placentae of different gestational ages: an OPRU Network and UW SCOR Study. Reproductive sciences (Thousand Oaks, Calif) , 15 (9), 866-77. Miller-Lindholm, A., LaBenz, C., Ramey, J., Bedows, E., & Ruddon, R. (1997). Human chorionic gonadotropin-beta gene expression in first trimester placenta. Endocrinology , 138 (12), 5459-65. Minamishima, Y., Moslehi, J., Padera, R., Bronson, R., Liao, R., & Kaelin, W. (2009). A feedback loop involving the Phd3 prolyl hydroxylase tunes the mammalian hypoxic response in vivo. Molecular and cellular biology , 29 (21), 5729-41. Mise, H., Sagawa, N., Matsumoto, T., Yura, S., Nanno, H., Itoh, H., et al. (1998). Augmented placental production of leptin in preeclampsia: possible involvement of placental hypoxia. The Journal of clinical endocrinology and metabolism , 83 (9), 3225-9. Miyagi, Y., Koshikawa, N., Yasumitsu, H., Miyagi, E., Hirahara, F., Aoki, I., et al. (1994). cDNA cloning and mRNA expression of a serine proteinase inhibitor secreted by cancer cells: identification as placental protein 5 and tissue factor pathway inhibitor-2. Journal of biochemistry , 116 (5), 939-42. Miyazaki, K., Kawamoto, T., Tanimoto, K., Nishiyama, M., Honda, H., & Kato, Y. (2002). Identification of functional hypoxia response elements in the promoter region of the DEC1 and DEC2 genes. The Journal of biological chemistry , 277 (49), 47014-21. Mohamed, M., & Sloane, B. (2006). Cysteine cathepsins: multifunctional enzymes in cancer. Nature reviews Cancer , 6 (10), 764-75.

226 Mohyeldin, A., Garzón-Muvdi, T., & Quiñones-Hinojosa, A. (2010). Oxygen in stem cell biology: a critical component of the stem cell niche. Cell stem cell , 7 (2), 150-61. Moore, L., Jahnigen, D., Rounds, S., Reeves, J., & Grover, R. (1982). Maternal hyperventilation helps preserve arterial oxygenation during high-altitude pregnancy. Journal of applied physiology: respiratory, environmental and exercise physiology , 52 (3), 690-4. Moore, L., Rounds, S., Jahnigen, D., Grover, R., & Reeves, J. (1982). Infant birth weight is related to maternal arterial oxygenation at high altitude. Journal of applied physiology: respiratory, environmental and exercise physiology , 52 (3), 695-9. Morizono, T., Nagashima, A., Ohnishi, Y., & Sekine…, A. (2007, Jan 1). A second generation human haplotype map of over 3.1 million SNPs. Nature . Morrissey, J., Fakhrai, H., & Edgington, T. (1987). Molecular cloning of the cDNA for tissue factor, the cellular receptor for the initiation of the coagulation protease cascade. Cell , 50 (1), 129-35. Mossman, H. (1991). Classics revisited: Comparative morphogenesis of the fetal membranes and accessory uterine structures. Placenta , 12 (1), 1-5. Mount, S. (1982). A catalogue of splice junction sequences. Nucleic Acids Research , 10 (2), 459-72. Mount, S. (1982, Jan 1). A catalogue of splice junction sequences. Nucleic Acids Research . Munir, S., Xu, G., Wu, Y., Yang, B., Lala, P., & Peng, C. (2004). Nodal and ALK7 inhibit proliferation and induce apoptosis in human trophoblast cells. The Journal of biological chemistry , 279 (30), 31277-86. Munro, H., Aziz, N., Leibold, E., Murray, M., Rogers, J., Vass, J., et al. (1988). The ferritin genes: structure, expression, and regulation. Annals of the New York Academy of Sciences , 526, 113-23. Munro, S., & Pelham, H. (1986). An Hsp70-like protein in the ER: identity with the 78 kd glucose- regulated protein and immunoglobulin heavy chain binding protein. Cell , 46 (2), 291-300. Muszbek, L., Bereczky, Z., Bagoly, Z., Komáromi, I., & Katona, É. (2011). Factor XIII: a coagulation factor with multiple plasmatic and cellular functions. Physiological reviews , 91 (3), 931-72. Muttukrishna, S., Fowler, P., George, L., Groome, N., & Knight, P. (1996). Changes in peripheral serum levels of total activin A during the human menstrual cycle and pregnancy. The Journal of clinical endocrinology and metabolism , 81 (9), 3328-34. Muttukrishna, S., George, L., Fowler, P., Groome, N., & Knight, P. (1995). Measurement of serum concentrations of inhibin-A (alpha-beta A dimer) during human pregnancy. Clinical endocrinology , 42 (4), 391-7. Nachman, R., & Harpel, P. (1976). Platelet alpha2-macroglobulin and alpha1-antitrypsin. The Journal of biological chemistry , 251 (15), 4512-21. Nackley, A., Shabalina, S., Tchivileva, I., & Satterfield…, K. (2006, Jan 1). Human catechol-O- methyltransferase haplotypes modulate protein expression by altering mRNA secondary structure. Science (New York, NY) . Nadeem, L., Munir, S., Fu, G., Dunk, C., Baczyk, D., Caniggia, I., et al. (2011). Nodal signals through activin receptor-like kinase 7 to inhibit trophoblast migration and invasion: implication in the pathogenesis of preeclampsia. The American journal of pathology , 178 (3), 1177-89. Nagy, J., Dvorak, A., & Dvorak, H. (2003). VEGF-A(164/165) and PlGF: roles in angiogenesis and arteriogenesis. Trends in cardiovascular medicine , 13 (5), 169-75. Naldini, L., Weidner, K., Vigna, E., Gaudino, G., Bardelli, A., Ponzetto, C., et al. (1991). Scatter factor and hepatocyte growth factor are indistinguishable ligands for the MET receptor. The EMBO Journal , 10 (10), 2867-78. Newbern, D., & Freemark, M. (2011). Placental hormones and the control of maternal metabolism and fetal growth. Current opinion in endocrinology, diabetes, and obesity , 18 (6), 409-16. Nguyen, V., Tosi, M., Gross, M., Cohen-Haguenauer, O., Jegou-Foubert, C., de Tand, M., et al. (1988). Assignment of the complement serine protease genes C1r and C1s to chromosome 12 region 12p13. Human genetics , 78 (4), 363-8.

227 Nickerson, D., Taylor, S., Weiss, K., Clark, A., Hutchinson, R., Stengård, J., et al. (1998). DNA sequence diversity in a 9.7-kb region of the human lipoprotein lipase gene. Nature Genetics , 19 (3), 233-40. Nielsen, K., Sørensen, S., Cartegni, L., Corydon, T., Doktor, T., Schroeder, L., et al. (2007). Seemingly neutral polymorphic variants may confer immunity to splicing-inactivating mutations: a synonymous SNP in exon 5 of MCAD protects from deleterious mutations in a flanking exonic splicing enhancer. American journal of human genetics , 80 (3), 416-32. Nishizawa, H., Ota, S., Suzuki, M., Kato, T., Sekiya, T., Kurahashi, H., et al. (2011). Comparative gene expression profiling of placentas from patients with severe pre-eclampsia and unexplained fetal growth restriction. Reproductive Biology and Endocrinology , 9, 107. Nissim-Rafinia, M., & Kerem, B. (2005). The splicing machinery is a genetic modifier of disease severity. Trends in genetics : TIG , 21 (9), 480-3. Noonan, E., Place, R., Rasoulpour, R., Giardina, C., & Hightower, L. (2007). Cell number-dependent regulation of Hsp70B' expression: evidence of an extracellular regulator. Journal of cellular physiology , 210 (1), 201-11. Norwitz, E., Schust, D., & Fisher, S. (2001). Implantation and the survival of early pregnancy. The New England journal of medicine , 345 (19), 1400-8. Novinec, M., Grass, R., Stark, W., Turk, V., Baici, A., & Lenarcic, B. (2007). Interaction between human cathepsins K, L, and S and elastins: mechanism of elastinolysis and inhibition by macromolecular inhibitors. The Journal of biological chemistry , 282 (11), 7893-902. Novotny, W., Girard, T., Miletich, J., & Broze, G. (1989). Purification and characterization of the lipoprotein-associated coagulation inhibitor from human plasma. The Journal of biological chemistry , 264 (31), 18832-7. Ohh, M., Park, C., Ivan, M., Hoffman, M., Kim, T., Huang, L., et al. (2000). Ubiquitination of hypoxia-inducible factor requires direct binding to the beta-domain of the von Hippel-Lindau protein. Nature cell biology , 2 (7), 423-7. Oh-I, S., Shimizu, H., Satoh, T., Okada, S., Adachi, S., Inoue, K., et al. (2006). Identification of nesfatin-1 as a satiety molecule in the hypothalamus. Nature , 443 (7112), 709-12. Oikawa, M., Abe, M., Kurosawa, H., Hida, W., Shirato, K., & Sato, Y. (2001). Hypoxia induces transcription factor ETS-1 via the activity of hypoxia-inducible factor-1. Biochemical and Biophysical Research Communications , 289 (1), 39-43. Olden, K., Pratt, R., Jaworski, C., & Yamada, K. (1979). Evidence for role of glycoprotein carbohydrates in membrane transport: specific inhibition by tunicamycin. Proceedings of the National Academy of Sciences of the United States of America , 76 (2), 791-5. Ong, K., Ahmed, M., Sherriff, A., Woods, K., Watts, A., Golding, J., et al. (1999). Cord blood leptin is associated with size at birth and predicts infancy weight gain in humans. ALSPAC Study Team. Avon Longitudinal Study of Pregnancy and Childhood. The Journal of clinical endocrinology and metabolism , 84 (3), 1145-8. Oudejans, C., & van Dijk, M. (2008). Placental gene expression and pre-eclampsia. Placenta , 29 Suppl A, S78-82. Pál, C., Papp, B., & Hurst, L. (2001). Highly expressed genes in yeast evolve slowly. Genetics , 158 (2), 927-31. Pagani, F., & Baralle, F. (2004). Genomic variants in exons and introns: identifying the splicing spoilers. Nature reviews Genetics , 5 (5), 389-96. Pagani, F., Stuani, C., Tzetis, M., Kanavakis, E., Efthymiadou, A., Doudounakis, S., et al. (2003). New type of disease causing mutations: the example of the composite exonic regulatory elements of splicing in CFTR exon 12. Human molecular genetics , 12 (10), 1111-20. Pakula, R., Melchior, A., Denys, A., Vanpouille, C., Mazurier, J., & Allain, F. (2007). Syndecan-1/ CD147 association is essential for cyclophilin B-induced activation of p44/42 mitogen-activated protein kinases and promotion of cell adhesion and chemotaxis. Glycobiology , 17 (5), 492-503. Panayi, G., Corrigall, V., & Henderson, B. (2004). Stress cytokines: pivotal proteins in immune regulatory networks; Opinion. Current opinion in immunology , 16 (4), 531-4.

228 Park, J., Chen, H., Winer, J., Houck, K., & Ferrara, N. (1994). Placenta growth factor. Potentiation of vascular endothelial growth factor bioactivity, in vitro and in vivo, and high affinity binding to Flt-1 but not to Flk-1/KDR. The Journal of biological chemistry , 269 (41), 25646-54. Parsian, A., Sheren, J., Tao, T., Goswami, P., Malyapa, R., Van Rheeden, R., et al. (2000). The human Hsp70B gene at the HSPA7 locus of chromosome 1 is transcribed but non-functional. Biochimica et biophysica acta , 1494 (1-2), 201-5. Patterson, A., & Zhang, L. (2010). Hypoxia and fetal heart development. Current molecular medicine , 10 (7), 653-66. Peddada, L., McPherson, J., Law, R., Wasmuth, J., Youderian, P., & Deans, R. (1992). Somatic cell mapping of the human cyclophilin B gene (PPIB) to chromosome 15. Cytogenetics and cell genetics , 60 (3-4), 219-21. Pei, S., Chen, X., Si Ren, B., Liu, Y., Cheng, X., Harris, E., et al. (1989). Chronic mountain sickness in Tibet. The Quarterly journal of medicine , 71 (266), 555-74. Peiffer, I., Belhomme, D., Barbet, R., Haydont, V., Zhou, Y.-P., Fortunel, N., et al. (2007). Simultaneous differentiation of endothelial and trophoblastic cells derived from human embryonic stem cells. Stem cells and development , 16 (3), 393-402. Peng, J., Zhang, L., Drysdale, L., & Fong, G. (2000). The transcription factor EPAS-1/hypoxia- inducible factor 2alpha plays an important role in vascular remodeling. Proceedings of the National Academy of Sciences of the United States of America , 97 (15), 8386-91. Perkins, N. (2007). Integrating cell-signalling pathways with NF-kappaB and IKK function. Nature reviews Molecular cell biology , 8 (1), 49-62. Pescador, N., Cuevas, Y., Naranjo, S., Alcaide, M., Villar, D., Landázuri, M., et al. (2005). Identification of a functional hypoxia-responsive element that regulates the expression of the egl nine homologue 3 (egln3/phd3) gene. The Biochemical journal , 390 (Pt 1), 189-97. Peschle, C., Mavilio, F., Carè, A., Migliaccio, G., Migliaccio, A., Salvo, G., et al. (1985). Haemoglobin switching in human embryos: asynchrony of zeta----alpha and epsilon----gamma-globin switches in primitive and definite erythropoietic lineage. Nature , 313 (5999), 235-8. Petanceska, S., Canoll, P., & Devi, L. (1996). Expression of rat cathepsin S in phagocytic cells. The Journal of biological chemistry , 271 (8), 4403-9. Petersen, L., Bjørn, S., Olsen, O., Nordfang, O., Norris, F., & Norris, K. (1996). Inhibitory properties of separate recombinant Kunitz-type-protease-inhibitor domains from tissue-factor-pathway inhibitor. European journal of biochemistry / FEBS , 235 (1-2), 310-6. Petersen, L., Sprecher, C., Foster, D., Blumberg, H., Hamamoto, T., & Kisiel, W. (1996). Inhibitory properties of a novel human Kunitz-type protease inhibitor homologous to tissue factor pathway inhibitor. Biochemistry , 35 (1), 266-72. Petraglia, F., Aguzzoli, L., Gallinelli, A., Florio, P., Zonca, M., Benedetto, C., et al. (1995). Hypertension in pregnancy: changes in activin A maternal serum concentration. Placenta , 16 (5), 447-54. Petraglia, F., Sawchenko, P., Lim, A., Rivier, J., & Vale, W. (1987). Localization, secretion, and action of inhibin in human placenta. Science (New York, NY) , 237 (4811), 187-9. Petraglia, F., Vaughan, J., & Vale, W. (1989). Inhibin and activin modulate the release of gonadotropin- releasing hormone, human chorionic gonadotropin, and progesterone from cultured human placental cells. Proceedings of the National Academy of Sciences of the United States of America , 86 (13), 5114-7. Pflanz, S., Timans, J., Cheung, J., Rosales, R., Kanzler, H., Gilbert, J., et al. (2002). IL-27, a heterodimeric cytokine composed of EBI3 and p28 protein, induces proliferation of naive CD4(+) T cells. Immunity , 16 (6), 779-90. Pickrell, J., Coop, G., Novembre, J., Kudaravalli, S., Li, J., Absher, D., et al. (2009). Signals of recent positive selection in a worldwide sample of human populations. Genome Research , 19 (5), 826-37. Pijnenborg, R., Bland, J., Robertson, W., Dixon, G., & Brosens, I. (1981). The pattern of interstitial trophoblastic invasion of the myometrium in early human pregnancy. Placenta , 2 (4), 303-16.

229 Pijnenborg, R., Robertson, W., Brosens, I., & Dixon, G. (1981). Review article: trophoblast invasion and the establishment of haemochorial placentation in man and laboratory animals. Placenta , 2 (1), 71-91. Policastro, P., Daniels-McQueen, S., Carle, G., & Boime, I. (1986). A map of the hCG beta-LH beta gene cluster. The Journal of biological chemistry , 261 (13), 5907-16. Policastro, P., Ovitt, C., Hoshina, M., Fukuoka, H., Boothby, M., & Boime, I. (1983). The beta subunit of human chorionic gonadotropin is encoded by multiple genes. The Journal of biological chemistry , 258 (19), 11492-9. Ponomarenko, J., Orlova, G., Merkulova, T., Gorshkova, E., Fokin, O., Vasiliev, G., et al. (2002). rSNP_Guide: an integrated database-tools system for studying SNPs and site-directed mutations in transcription factor binding sites. Human mutation , 20 (4), 239-48. Pouysségur, J., Dayan, F., & Mazure, N. (2006). Hypoxia signalling in cancer and approaches to enforce tumour regression. Nature , 441 (7092), 437-43. Price, E., Jin, M., Lim, D., Pati, S., Walsh, C., & McKeon, F. (1994). Cyclophilin B trafficking through the secretory pathway is altered by binding of cyclosporin A. Proceedings of the National Academy of Sciences of the United States of America , 91 (9), 3931-5. Provot, S., Zinyk, D., Gunes, Y., Kathri, R., Le, Q., Kronenberg, H., et al. (2007). Hif-1alpha regulates differentiation of limb bud mesenchyme and joint development. The Journal of cell biology , 177 (3), 451-64. Quadros, E., Nakayama, Y., & Sequeira, J. (2009). The protein and the gene encoding the receptor for the cellular uptake of transcobalamin-bound cobalamin. Blood , 113 (1), 186-92. Quinn, L., Latham, S., & Kalionis, B. (2000). The homeobox genes MSX2 and MOX2 are candidates for regulating epithelial-mesenchymal cell interactions in the human placenta. Placenta , 21 Suppl A, S50-4. Rajagopalan, S., & Brenner, M. (1994). Calnexin retains unassembled major histocompatibility complex class I free heavy chains in the endoplasmic reticulum. The Journal of experimental medicine , 180 (1), 407-12. Rajakumar, A., & Conrad, K. (2000). Expression, ontogeny, and regulation of hypoxia-inducible transcription factors in the human placenta. Biology of reproduction , 63 (2), 559-69. Rajakumar, A., Brandon, H., Daftary, A., Ness, R., & Conrad, K. (2004). Evidence for the functional activity of hypoxia-inducible transcription factors overexpressed in preeclamptic placentae. Placenta , 25 (10), 763-9. Rajakumar, A., Whitelock, K., Weissfeld, L., Daftary, A., Markovic, N., & Conrad, K. (2001). Selective overexpression of the hypoxia-inducible transcription factor, HIF-2alpha, in placentas from women with preeclampsia. Biology of reproduction , 64 (2), 499-506. Rao, C. (2001). Tropic effects of LH and hCG on early pregnancy events in women's reproductive tract. Early pregnancy (Online) , 5 (1), 18-9. Rao, L., & Rapaport, S. (1987). Studies of a mechanism inhibiting the initiation of the extrinsic pathway of coagulation. Blood , 69 (2), 645-51. Rawn, S., & Cross, J. (2008). The evolution, regulation, and function of placenta-specific genes. Annual review of cell and developmental biology , 24, 159-81. Ream, M., Ray, A., Chandra, R., & Chikaraishi, D. (2008). Early fetal hypoxia leads to growth restriction and myocardial thinning. American journal of physiology Regulatory, integrative and comparative physiology , 295 (2), R583-95. Redman, C., & Sargent, I. (2005). Latest advances in understanding preeclampsia. Science (New York, NY) , 308 (5728), 1592-4. Reinholt, F., Hultenby, K., Oldberg, A., & Heinegård, D. (1990). Osteopontin--a possible anchor of osteoclasts to bone. Proceedings of the National Academy of Sciences of the United States of America , 87 (12), 4473-5. Reverdiau, P., Jarousseau, A., Thibault, G., Khalfoun, B., Watier, H., Lebranchu, Y., et al. (1995). Tissue factor activity of syncytiotrophoblast plasma membranes and tumoral trophoblast cells in culture. Thrombosis and haemostasis , 73 (1), 49-54.

230 Rhodes, K., Gekas, C., Wang, Y., Lux, C., Francis, C., Chan, D., et al. (2008). The emergence of hematopoietic stem cells is initiated in the placental vasculature in the absence of circulation. Cell stem cell , 2 (3), 252-63. Richani, K., Soto, E., Romero, R., Espinoza, J., Chaiworapongsa, T., Nien, J., et al. (2005). Normal pregnancy is characterized by systemic activation of the complement system. The journal of maternal- fetal & neonatal medicine : the official journal of the European Association of Perinatal Medicine, the Federation of Asia and Oceania Perinatal Societies, the International Society of Perinatal Obstetricians , 17 (4), 239-45. Richardson, B., Langland, R., Bachurski, C., Richards, R., Kessler, C., Cheng, Y., et al. (2000). Activator protein-2 regulates human placental lactogen gene expression. Molecular and cellular endocrinology , 160 (1-2), 183-92. Richmond, J., & Baglole, D. (2003). Lassa fever: epidemiology, clinical features, and social consequences. BMJ (Clinical research ed) , 327 (7426), 1271-5. Roberts, J., & Cooper, D. (2001). Pathogenesis and genetics of pre-eclampsia. Lancet , 357 (9249), 53-6. Roberts, R., & Fisher, S. (2011). Trophoblast stem cells. Biology of reproduction , 84 (3), 412-21. Robertson, W. (1976). Uteroplacental vasculature. Journal of clinical pathology Supplement (Royal College of Pathologists) , 10, 9-17. Robertson, W., & Manning, P. (1974). Elastic tissue in uterine blood vessels. The Journal of pathology , 112 (4), 237-43. Robin, C., Bollerot, K., Mendes, S., Haak, E., Crisan, M., Cerisoli, F., et al. (2009). Human placenta is a potent hematopoietic niche containing hematopoietic stem and progenitor cells throughout development. Cell stem cell , 5 (4), 385-95. Robins, J., Heizer, A., Hardiman, A., Hubert, M., & Handwerger, S. (2007). Oxygen tension directs the differentiation pathway of human cytotrophoblast cells. Placenta , 28 (11-12), 1141-6. Robins, J., Morgan, J., Krueger, P., & Carson, S. (2011). Bioengineering anembryonic human trophoblast vesicles. Reproductive sciences (Thousand Oaks, Calif) , 18 (2), 128-35. Robinson, M., McCarthy, D., & Smyth, G. (2010). edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics (Oxford, England) , 26 (1), 139-40. Roca, X., Sachidanandam, R., & Krainer, A. (2005). Determinants of the inherent strength of human 5' splice sites. Rna , 11 (5), 683-98. Roca, X., Sachidanandam, R., & Krainer, A. (2003). Intrinsic differences between authentic and cryptic 5' splice sites. Nucleic Acids Research , 31 (21), 6321-33. Rodesch, F., Simon, P., Donner, C., & Jauniaux, E. (1992). Oxygen measurements in endometrial and trophoblastic tissues during early pregnancy. Obstetrics and gynecology , 80 (2), 283-5. Rosenberger, C., Mandriota, S., Jürgensen, J., Wiesener, M., Hörstrup, J., Frei, U., et al. (2002). Expression of hypoxia-inducible factor-1alpha and -2alpha in hypoxic and ischemic rat kidneys. Journal of the American Society of Nephrology : JASN , 13 (7), 1721-32. Roshon, M., DeGregori, J., & Ruley, H. (2003). Gene trap mutagenesis of hnRNP A2/B1: a cryptic 3' splice site in the neomycin resistance gene allows continued expression of the disrupted cellular gene. BMC genomics , 4 (1), 2. Rossant, J., & Cross, J. (2001). Placental development: lessons from mouse mutants. Nature reviews Genetics , 2 (7), 538-48. Rossant, J., & Tam, P. (2009). Blastocyst lineage formation, early embryonic asymmetries and axis patterning in the mouse. Development (Cambridge, England) , 136 (5), 701-13. Ruas, J., & Poellinger, L. (2005). Hypoxia-dependent activation of HIF into a transcriptional regulator. Seminars in cell & developmental biology , 16 (4-5), 514-22. Rubin, J., Bottaro, D., & Aaronson, S. (1993). Hepatocyte growth factor/scatter factor and its receptor, the c-met proto-oncogene product. Biochimica et biophysica acta , 1155 (3), 357-71.

231 Ruddon, R., Hartle, R., Peters, B., Anderson, C., Huot, R., & Stromberg, K. (1981). Biosynthesis and secretion of chorionic gonadotropin subunits by organ cultures of first trimester human placenta. The Journal of biological chemistry , 256 (22), 11389-92. Ruskin, B., Zamore, P., & Green, M. (1988). A factor, U2AF, is required for U2 snRNP binding and splicing complex assembly. Cell , 52 (2), 207-19. Ryan, H., Lo, J., & Johnson, R. (1998). HIF-1 alpha is required for solid tumor formation and embryonic vascularization. The EMBO Journal , 17 (11), 3005-15. Rycyzyn, M., Reilly, S., O'Malley, K., & Clevenger, C. (2000). Role of cyclophilin B in prolactin signal transduction and nuclear retrotranslocation. Molecular endocrinology (Baltimore, Md) , 14 (8), 1175-86. Sabeti, P., Varilly, P., Fry, B., Lohmueller, J., Hostetter, E., Cotsapas, C., et al. (2007). Genome-wide detection and characterization of positive selection in human populations. Nature , 449 (7164), 913-8. Sachidanandam, R., Weissman, D., & Schmidt…, S. (2001, Jan 1). A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms. Nature . Sagawa, N., Yura, S., Itoh, H., Mise, H., Kakui, K., Korita, D., et al. (2002). Role of leptin in pregnancy--a review. Placenta , 23 Suppl A, S80-6. Sairam, M., & Li, C. (1975). Human pituitary lutropin. Isolation, properties, and the complete amino acid sequence of the beta-subunit. Biochimica et biophysica acta , 412 (1), 70-81. Salafia, C., Ghidini, A., Lopèz-Zeno, J., & Pezzullo, J. (1998). Uteroplacental pathology and maternal arterial mean blood pressure in spontaneous prematurity. Journal of the Society for Gynecologic Investigation , 5 (2), 68-71. Salavaggione, O., Wang, L., Wiepert, M., Yee, V., & Weinshilboum, R. (2005). Thiopurine S- methyltransferase pharmacogenetics: variant allele functional and comparative genomics. Pharmacogenetics and genomics , 15 (11), 801-15. Schapira, M., Scott, C., & Colman, R. (1982). Contribution of plasma protease inhibitors to the inactivation of kallikrein in plasma. The Journal of clinical investigation , 69 (2), 462-8. Schenke-Layland, K., Angelis, E., Rhodes, K., Heydarkhan-Hagvall, S., Mikkola, H., & Maclellan, W. (2007). Collagen IV induces trophoectoderm differentiation of mouse embryonic stem cells. Stem cells (Dayton, Ohio) , 25 (6), 1529-38. Schipani, E., Ryan, H., Didrickson, S., Kobayashi, T., Knight, M., & Johnson, R. (2001). Hypoxia in cartilage: HIF-1alpha is essential for chondrocyte growth arrest and survival. Genes & Development , 15 (21), 2865-76. Schmid, H., Sauerbrei, R., Schwarz, G., Weber, E., Kalbacher, H., & Driessen, C. (2002). Modulation of the endosomal and lysosomal distribution of cathepsins B, L and S in human monocytes/ macrophages. Biological chemistry , 383 (7-8), 1277-83. Schneider-Kolsky, M., Manuelpillai, U., Waldron, K., Dole, A., & Wallace, E. (2002). The distribution of activin and activin receptors in gestational tissues across human pregnancy and during labour. Placenta , 23 (4), 294-302. Schodel, J., Oikonomopoulos, S., Ragoussis, J., Pugh, C., Ratcliffe, P., & Mole, D. (2011). High- resolution genome-wide mapping of HIF-binding sites by ChIP-seq. Blood , 117 (23), e207-e217. Schofield, C., & Ratcliffe, P. (2005). Signalling hypoxia by HIF hydroxylases. Biochemical and Biophysical Research Communications , 338 (1), 617-26. Scholl, T. (2005). Iron status during pregnancy: setting the stage for mother and infant. The American journal of clinical nutrition , 81 (5), 1218S-1222S. Schrag, J., Bergeron, J., Li, Y., Borisova, S., Hahn, M., Thomas, D., et al. (2001). The Structure of calnexin, an ER chaperone involved in quality control of protein folding. Molecular Cell , 8 (3), 633-44. Schubring, C., Kiess, W., Englaro, P., Rascher, W., & Blum, W. (1996). Leptin concentrations in amniotic fluid, venous and arterial cord blood and maternal serum: high leptin synthesis in the fetus and inverse correlation with placental weight. European journal of pediatrics , 155 (9), 830. Schulz, L., Ezashi, T., Das, P., Westfall, S., Livingston, K., & Roberts, R. (2008). Human embryonic stem cells as models for trophoblast differentiation. Placenta , 29 Suppl A, S10-6.

232 Schwarz, G., Boehncke, W.-H., Braun, M., Schröter, C., Burster, T., Flad, T., et al. (2002). Cathepsin S activity is detectable in human keratinocytes and is selectively upregulated upon stimulation with interferon-gamma. The Journal of investigative dermatology , 119 (1), 44-9. Scortegagna, M., Ding, K., Oktay, Y., Gaur, A., Thurmond, F., Yan, L.-J., et al. (2003). Multiple organ pathology, metabolic abnormalities and impaired homeostasis of reactive oxygen species in Epas1-/- mice. Nature Genetics , 35 (4), 331-40. Seeburg, P., Shine, J., Martial, J., Ullrich, A., Baxter, J., & Goodman, H. (1977). Nucleotide sequence of part of the gene for human chorionic somatomammotropin: purification of DNA complementary to predominant mRNA species. Cell , 12 (1), 157-65. Sellar, G., Blake, D., & Reid, K. (1991). Characterization and organization of the genes encoding the A-, B- and C-chains of human complement subcomponent C1q. The complete derived amino acid sequence of human C1q. The Biochemical journal , 274 ( Pt 2), 481-90. Semenza, G. (2002). Signal transduction to hypoxia-inducible factor 1. Biochemical pharmacology , 64 (5-6), 993-8. Semenza, G. (2003). Targeting HIF-1 for cancer therapy. Nature reviews Cancer , 3 (10), 721-32. Senner, C., & Hemberger, M. (2010). Regulation of early trophoblast differentiation - lessons from the mouse. Placenta , 31 (11), 944-50. Señarís, R., Garcia-Caballero, T., Casabiell, X., Gallego, R., Castro, R., Considine, R., et al. (1997). Synthesis of leptin in human placenta. Endocrinology , 138 (10), 4501-4. Seppälä, M., Rutanen, E., Siiteri, J., Wahlström, T., Koistinen, R., Pietilä, R., et al. (1983). Immunologic and biological properties and clinical significance of placental proteins PP5 and PP12. Annals of the New York Academy of Sciences , 417, 368-82. Shamonki, J., Salmon, J., Hyjek, E., & Baergen, R. (2007). Excessive complement activation is associated with placental injury in patients with antiphospholipid antibodies. American journal of obstetrics and gynecology , 196 (2), 167.e1-5. Shi, G., Munger, J., Meara, J., Rich, D., & Chapman, H. (1992). Molecular cloning and expression of human alveolar macrophage cathepsin S, an elastinolytic cysteine protease. The Journal of biological chemistry , 267 (11), 7258-62. Shi, G., Webb, A., Foster, K., Knoll, J., Lemere, C., Munger, J., et al. (1994). Human cathepsin S: chromosomal localization, gene structure, and tissue distribution. The Journal of biological chemistry , 269 (15), 11530-6. Shibuya, M. (2006). Vascular endothelial growth factor receptor-1 (VEGFR-1/Flt-1): a dual regulator for angiogenesis. Angiogenesis , 9 (4), 225-30; discussion 231. Shibuya, M., Yamaguchi, S., Yamane, A., Ikeda, T., Tojo, A., Matsushime, H., et al. (1990). Nucleotide sequence and expression of a novel human receptor-type tyrosine kinase gene (flt) closely related to the fms family. Oncogene , 5 (4), 519-24. Shine, J., Seeburg, P., Martial, J., Baxter, J., & Goodman, H. (1977). Construction and analysis of recombinant DNA for human chorionic somatomammotropin. Nature , 270 (5637), 494-9. Shinmura, K., Tao, H., Yamada, H., Kataoka, H., Sanjar, R., Wang, J., et al. (2004). Splice-site genetic polymorphism of the human kallikrein 12 (KLK12) gene correlates with no substantial expression of KLK12 protein having serine protease activity. Human mutation , 24 (3), 273-4. Shiu, R., Pouyssegur, J., & Pastan, I. (1977). Glucose depletion accounts for the induction of two transformation-sensitive membrane proteinsin Rous sarcoma virus-transformed chick embryo fibroblasts. Proceedings of the National Academy of Sciences of the United States of America , 74 (9), 3840-4. Shome, B., & Parlow, A. (1973). The primary structure of the hormone-specific, beta subunit of human pituitary luteinizing hormone (hLH). The Journal of clinical endocrinology and metabolism , 36 (3), 618-21. Sibai, B., Dekker, G., & Kupferminc, M. (2005). Pre-eclampsia. Lancet , 365 (9461), 785-99. Simon, A., Stone, E., & Sidow, A. (2002). Inference of functional regions in proteins by quantification of evolutionary constraints. Proceedings of the National Academy of Sciences of the United States of America , 99 (5), 2912-7.

233 Simonson, T., Yang, Y., Huff, C., Yun, H., Qin, G., Witherspoon, D., et al. (2010). Genetic evidence for high-altitude adaptation in Tibet. Science (New York, NY) , 329 (5987), 72-5. Sinha, D., Wells, M., & Faulk, W. (1984). Immunological studies of human placentae: complement components in pre-eclamptic chorionic villi. Clinical and experimental immunology , 56 (1), 175-84. Sitras, V., Fenton, C., Paulssen, R., Vårtun, Å., & Acharya, G. (2012). Differences in gene expression between first and third trimester human placenta: a microarray study. PLoS ONE , 7 (3), e33294. Sitras, V., Paulssen, R., Grønaas, H., Leirvik, J., Hanssen, T., Vårtun, A., et al. (2009). Differential placental gene expression in severe preeclampsia. Placenta , 30 (5), 424-33. Skarratt, K., Fuller, S., Sluyter, R., Dao-Ung, L.-P., Gu, B., & Wiley, J. (2005). A 5' intronic splice site polymorphism leads to a null allele of the P2X7 gene in 1-2% of the Caucasian population. Febs Letters , 579 (12), 2675-8. Skibinski, G. (2003). The role of hepatocyte growth factor/c-met interactions in the immune system. Archivum immunologiae et therapiae experimentalis , 51 (5), 277-82. Smith, C., & Valcárcel, J. (2000). Alternative pre-mRNA splicing: the logic of combinatorial control. Trends in biochemical sciences , 25 (8), 381-8. Smith, G., Crossley, J., Aitken, D., Pell, J., Cameron, A., Connor, J., et al. (2004). First-trimester placentation and the risk of antepartum stillbirth. JAMA : the journal of the American Medical Association , 292 (18), 2249-54. Song, H., Yin, D., & Liu, Z. (2012). GDF-15 promotes angiogenesis through modulating p53/HIF-1α signaling pathway in hypoxic human umbilical vein endothelial cells. Molecular biology reports , 39 (4), 4017-22. Sowter, H., Raval, R., Moore, J., Ratcliffe, P., & Harris, A. (2003). Predominant role of hypoxia- inducible transcription factor (Hif)-1alpha versus Hif-2alpha in regulation of the transcriptional response to hypoxia. Cancer Research , 63 (19), 6130-4. Spicer, E., Horton, R., Bloem, L., Bach, R., Williams, K., Guha, A., et al. (1987). Isolation of cDNA clones coding for human tissue factor: primary structure of the protein and cDNA. Proceedings of the National Academy of Sciences of the United States of America , 84 (15), 5148-52. Spitalieri, P., Cortese, G., Pietropolli, A., Filareto, A., Dolci, S., Klinger, F., et al. (2009). Identification of multipotent cytotrophoblast cells from human first trimester chorionic villi. Cloning and stem cells , 11 (4), 535-56. Sprecher, C., Kisiel, W., Mathewes, S., & Foster, D. (1994). Molecular cloning, expression, and partial characterization of a second human tissue-factor-pathway inhibitor. Proceedings of the National Academy of Sciences of the United States of America , 91 (8), 3353-7. Srisuparp, S., Strakova, Z., & Fazleabas, A. (2001). The role of chorionic gonadotropin (CG) in blastocyst implantation. Archives of medical research , 32 (6), 627-34. Staun-Ram, E., & Shalev, E. (2005). Human trophoblast function during the implantation process. Reproductive Biology and Endocrinology , 3, 56. Stewart, A., Milborrow, H., Ring, J., Crowther, C., & Forage, R. (1986). Human inhibin genes. Genomic characterisation and sequencing. Febs Letters , 206 (2), 329-34. Stoikos, C., Salamonsen, L., Hannan, N., O'Connor, A., Rombauts, L., & Dimitriadis, E. (2010). Activin A regulates trophoblast cell adhesive properties: implications for implantation failure in women with endometriosis-associated infertility. Human reproduction (Oxford, England) , 25 (7), 1767-74. Stojkovic, M., Lako, M., Stojkovic, P., Stewart, R., Przyborski, S., Armstrong, L., et al. (2004). Derivation of human embryonic stem cells from day-8 blastocysts recovered after three-step in vitro culture. Stem cells (Dayton, Ohio) , 22 (5), 790-7. Strausberg, R., Feingold, E., Grouse, L., Derge, J., Klausner, R., Collins, F., et al. (2002). Generation and initial analysis of more than 15,000 full-length human and mouse cDNA sequences. Proceedings of the National Academy of Sciences of the United States of America , 99 (26), 16899-903. Su, A., Cooke, M., Ching, K., Hakak, Y., Walker, J., Wiltshire, T., et al. (2002). Large-scale analysis of the human and mouse transcriptomes. Proceedings of the National Academy of Sciences of the United States of America , 99 (7), 4465-70.

234 Suckow, J., Markiewicz, P., Kleina, L., Miller, J., Kisters-Woike, B., & Müller-Hill, B. (1996). Genetic studies of the Lac repressor. XV: 4000 single amino acid substitutions and analysis of the resulting phenotypes on the basis of the protein structure. Journal of molecular biology , 261 (4), 509-23. Sugulle, M., Dechend, R., Herse, F., Weedon-Fekjaer, M., Johnsen, G., Brosnihan, K., et al. (2009). Circulating and placental growth-differentiation factor 15 in preeclampsia and in pregnancy complicated by diabetes mellitus. Hypertension , 54 (1), 106-12. Sun, G., Yan, J., Noltner, K., Feng, J., Li, H., & Sarkis…, D. (2009, Jan 1). SNPs in human miRNA genes affect biogenesis and function. Rna . Sutherland, A., Speed, T., & Calarco, P. (1990). Inner cell allocation in the mouse morula: the role of oriented division during fourth cleavage. Developmental biology , 137 (1), 13-25. Suzuki, N., Yamada, T., Matsuoka, K., Hiraoka, N., Iwamaru, Y., & Hata, J. (1999). Functional expressions of fms and M-CSF during trophoectodermal differentiation of human embryonal carcinoma cells. Placenta , 20 (2-3), 203-11. Taleb, S., Cancello, R., Clément, K., & Lacasa, D. (2006). Cathepsin s promotes human preadipocyte differentiation: possible involvement of fibronectin degradation. Endocrinology , 147 (10), 4950-9. Talmadge, K., Boorstein, W., & Fiddes, J. (1983). The human genome contains seven genes for the beta-subunit of chorionic gonadotropin but only one gene for the beta-subunit of luteinizing hormone. DNA (Mary Ann Liebert, Inc) , 2 (4), 281-9. Talmadge, K., Boorstein, W., Vamvakopoulos, N., Gething, M., & Fiddes, J. (1984). Only three of the seven human chorionic gonadotropin beta subunit genes can be expressed in the placenta. Nucleic Acids Research , 12 (22), 8415-36. Tamura, T., Goldenberg, R., Johnston, K., Cliver, S., & Hickey, C. (1996). Serum ferritin: a predictor of early spontaneous preterm delivery. Obstetrics and gynecology , 87 (3), 360-5. Tanaka, S., Kunath, T., Hadjantonakis, A., Nagy, A., & Rossant, J. (1998). Promotion of trophoblast stem cell proliferation by FGF4. Science (New York, NY) , 282 (5396), 2072-5. Tannetta, D., Dragovic, R., Gardiner, C., Redman, C., & Sargent, I. (2013). Characterisation of syncytiotrophoblast vesicles in normal pregnancy and pre-eclampsia: expression of flt-1 and endoglin. PLoS ONE , 8 (2), e56754. Tarkowski, A., & Wróblewska, J. (1967). Development of blastomeres of mouse eggs isolated at the 4- and 8-cell stage. Journal of embryology and experimental morphology , 18 (1), 155-80. Taylor, B., & Zhulin, I. (1999, Jan 1). PAS domains: internal sensors of oxygen, redox potential, and light. Microbiology and Molecular Biology Reviews . Tedesco, F., Radillo, O., Candussi, G., Nazzaro, A., Mollnes, T., & Pecorari, D. (1990). Immunohistochemical detection of terminal complement complex and S protein in normal and pre- eclamptic placentae. Clinical and experimental immunology , 80 (2), 236-40. Teng, M., Wang, Y., Wang, G., Jung, J., Edenberg, H., Sanford, J., et al. (2011). Prioritizing single- nucleotide variations that potentially regulate alternative splicing. BMC proceedings , 5 Suppl 9, S40. Thomson, J., Itskovitz-Eldor, J., Shapiro, S., Waknitz, M., Swiergiel, J., Marshall, V., et al. (1998). Embryonic stem cell lines derived from human blastocysts. Science (New York, NY) , 282 (5391), 1145-7. Tian, H., Hammer, R., Matsumoto, A., Russell, D., & McKnight, S. (1998). The hypoxia-responsive transcription factor EPAS1 is essential for catecholamine homeostasis and protection against heart failure during embryonic development. Genes & Development , 12 (21), 3320-4. Tiscornia, G., & Mahadevan, M. (2000). Myotonic dystrophy: the role of the CUG triplet repeats in splicing of a novel DMPK exon and altered cytoplasmic DMPK mRNA isoform ratios. Molecular Cell , 5 (6), 959-67. Tjoelker, L., Seyfried, C., Eddy, R., Byers, M., Shows, T., Calderon, J., et al. (1994). Human, mouse, and rat calnexin cDNA cloning: identification of potential calcium binding motifs and gene localization to human chromosome 5. Biochemistry , 33 (11), 3229-36. Tokuhiro, S., Yamada, R., Chang, X., Suzuki, A., Kochi, Y., Sawada, T., et al. (2003). An intronic SNP in a RUNX1 binding site of SLC22A4, encoding an organic cation transporter, is associated with rheumatoid arthritis. Nature Genetics , 35 (4), 341-8.

235 Toporsian, M., Gros, R., Kabir, M., Vera, S., Govindaraju, K., Eidelman, D., et al. (2005). A role for endoglin in coupling eNOS activity and regulating vascular tone revealed in hereditary hemorrhagic telangiectasia. Circulation research , 96 (6), 684-92. Toth, P., Lukacs, H., Gimes, G., Sebestyen, A., Pasztor, N., Paulin, F., et al. (2001). Clinical importance of vascular LH/hCG receptors--a review. Reproductive biology , 1 (2), 5-11. Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D., et al. (2012). Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nature protocols , 7 (3), 562-78. Tripathy, V., & Gupta, R. (2005). Birth weight among Tibetans at different altitudes in India: are Tibetans better protected from IUGR? American journal of human biology : the official journal of the Human Biology Council , 17 (4), 442-50. Trounson, A.O., Gardner, D.K., Baker, G., Barnes, F.L., Bongso, A., Bourne, H., et al. (2000). Handbook of in vitro fertilization. Boca Raton, London, New York, Washington, D.C.: CRC Press. Tsarfaty, I., Rong, S., Resau, J., Rulong, S., da Silva, P., & Vande Woude, G. (1994). The Met proto- oncogene mesenchymal to epithelial cell conversion. Science (New York, NY) , 263 (5143), 98-101. Tsoi, S., Cale, J., Bird, I., & Kay, H. (2003). cDNA microarray analysis of gene expression profiles in human placenta: up-regulation of the transcript encoding muscle subunit of glycogen phosphorylase in preeclampsia. Journal of the Society for Gynecologic Investigation , 10 (8), 496-502. TSUI, K., & WAYE, M. (2005, Jan 1). A haplotype map of the human genome. Nature . Turpeenniemi-Hujanen, T., Rönnberg, L., Kauppila, A., & Puistola, U. (1992). Laminin in the human embryo implantation: analogy to the invasion by malignant cells. Fertility and sterility , 58 (1), 105-13. Udagawa, K., Miyagi, Y., Hirahara, F., Miyagi, E., Nagashima, Y., Minaguchi, H., et al. (1998). Specific expression of PP5/TFPI2 mRNA by syncytiotrophoblasts in human placenta as revealed by in situ hybridization. Placenta , 19 (2-3), 217-23. Udagawa, K., Yasumitsu, H., Esaki, M., Sawada, H., Nagashima, Y., Aoki, I., et al. (2002). Subcellular localization of PP5/TFPI-2 in human placenta: a possible role of PP5/TFPI-2 as an anti-coagulant on the surface of syncytiotrophoblasts. Placenta , 23 (2-3), 145-53. Udayashankar, R., Baker, D., Tuckerman, E., Laird, S., Li, T., & Moore, H. (2011). Characterization of invasive trophoblasts generated from human embryonic stem cells. Human reproduction (Oxford, England) , 26 (2), 398-406. Uehara, Y., Minowa, O., Mori, C., Shiota, K., Kuno, J., Noda, T., et al. (1995). Placental defect and embryonic lethality in mice lacking hepatocyte growth factor/scatter factor. Nature , 373 (6516), 702-5. Unger, C., Weiser, J., McCullough, R., Keefer, S., & Moore, L. (1988). Altitude, low birth weight, and infant mortality in Colorado. JAMA : the journal of the American Medical Association , 259 (23), 3427-32. Ushizawa, K., Herath, C., Kaneyama, K., Shiojima, S., Hirasawa, A., Takahashi, T., et al. (2004). cDNA microarray analysis of bovine embryo gene expression profiles during the pre-implantation period. Reproductive Biology and Endocrinology , 2, 77. Uszyński, M., Zekanowska, E., Uszyński, W., & Kuczyński, J. (2001). Tissue factor (TF) and tissue factor pathway inhibitor (TFPI) in amniotic fluid and blood plasma: implications for the mechanism of amniotic fluid embolism. European journal of obstetrics, gynecology, and reproductive biology , 95 (2), 163-6. Vale, W., Rivier, J., Vaughan, J., McClintock, R., Corrigan, A., Woo, W., et al. (1986). Purification and characterization of an FSH releasing protein from porcine ovarian follicular fluid. Nature , 321 (6072), 776-9. Vallier, L., Alexander, M., & Pedersen, R. (2005). Activin/Nodal and FGF pathways cooperate to maintain pluripotency of human embryonic stem cells. Journal of cell science , 118 (Pt 19), 4495-509. Vallier, L., Reynolds, D., & Pedersen, R. (2004). Nodal inhibits differentiation of human embryonic stem cells along the neuroectodermal default pathway. Developmental biology , 275 (2), 403-21.

236 Van de Wouwer, M., Collen, D., & Conway, E. (2004). Thrombomodulin-protein C-EPCR system: integrated to regulate coagulation and inflammation. Arteriosclerosis, thrombosis, and vascular biology , 24 (8), 1374-83. van der Graaf, F., Koedam, J., & Bouma, B. (1983). Inactivation of kallikrein in human plasma. The Journal of clinical investigation , 71 (1), 149-58. Van Handel, B., Prashad, S., Hassanzadeh-Kiabi, N., Huang, A., Magnusson, M., Atanassova, B., et al. (2010). The first trimester human placenta is a site for terminal maturation of primitive erythroid cells. Blood , 116 (17), 3321-30. Vargas, M., Vargas, E., Julian, C., Armaza, J., Rodriguez, A., Tellez, W., et al. (2007). Determinants of blood oxygenation during pregnancy in Andean and European residents of high altitude. American journal of physiology Regulatory, integrative and comparative physiology , 293 (3), R1303-12. Vassena, R., Boué, S., González-Roca, E., Aran, B., Auer, H., Veiga, A., et al. (2011). Waves of early transcriptional activation and pluripotency program initiation during human preimplantation development. Development (Cambridge, England) , 138 (17), 3699-709. Venkatesha, S., Toporsian, M., Lam, C., Hanai, J.-i., Mammoto, T., Kim, Y., et al. (2006). Soluble endoglin contributes to the pathogenesis of preeclampsia. Nature Medicine , 12 (6), 642-9. Villar, J., Carroli, G., Wojdyla, D., Abalos, E., Giordano, D., Ba'aqeel, H., et al. (2006). Preeclampsia, gestational hypertension and intrauterine growth restriction, related or independent conditions? American journal of obstetrics and gynecology , 194 (4), 921-31. Voellmy, R., Ahmed, A., Schiller, P., Bromley, P., & Rungger, D. (1985). Isolation and functional analysis of a human 70,000-dalton heat shock protein gene segment. Proceedings of the National Academy of Sciences of the United States of America , 82 (15), 4949-53. Walker, F. (1981). Regulation of activated protein C by protein S. The role of phospholipid in factor Va inactivation. The Journal of biological chemistry , 256 (21), 11128-31. Walker, F., Chavin, S., & Fay, P. (1987). Inactivation of factor VIII by activated protein C and protein S. Archives of biochemistry and biophysics , 252 (1), 322-8. Wall, D., Hirsh, A., Fraser, H., Kumm, J., Giaever, G., Eisen, M., et al. (2005). Functional genomic analysis of the rates of protein evolution. Proceedings of the National Academy of Sciences of the United States of America , 102 (15), 5483-8. Wang, C., Hsueh, A., & Erickson, G. (1981). LH stimulation of estrogen secretion by cultured rat granulosa cells. Molecular and cellular endocrinology , 24 (1), 17-28. Wang, D., Fan, J., Siao, C., Berno, A., & Young…, P. (1998, Jan 1). Large-scale identification, mapping, and genotyping of single-nucleotide polymorphisms in the human genome. Science (New York, NY) . Wang, G., & Semenza, G. (1993). Characterization of hypoxia-inducible factor 1 and regulation of DNA binding activity by hypoxia. The Journal of biological chemistry , 268 (29), 21513-8. Wang, G., & Semenza, G. (1995). Purification and characterization of hypoxia-inducible factor 1. The Journal of biological chemistry , 270 (3), 1230-7. Wang, G., Jiang, B., Rue, E., & Semenza, G. (1995). Hypoxia-inducible factor 1 is a basic-helix-loop- helix-PAS heterodimer regulated by cellular O2 tension. Proceedings of the National Academy of Sciences of the United States of America , 92 (12), 5510-4. Wang, G., Zhang, H., Zhao, Y., Li, J., Cai, J., Wang, P., et al. (2005). Noggin and bFGF cooperate to maintain the pluripotency of human embryonic stem cells in the absence of feeder layers. Biochemical and Biophysical Research Communications , 330 (3), 934-42. Wang, G.-S., & Cooper, T. (2007). Splicing in disease: disruption of the splicing code and the decoding machinery. Nature reviews Genetics , 8 (10), 749-61. Wang, K., & Denhardt, D. (2008). Osteopontin: role in immune regulation and stress responses. Cytokine & growth factor reviews , 19 (5-6), 333-45. Wang, Q., Piotrowska, K., Ciemerych, M., Milenkovic, L., Scott, M., Davis, R., et al. (2004). A genome-wide study of gene activity reveals developmental signaling pathways in the preimplantation mouse embryo. Developmental Cell , 6 (1), 133-44.

237 Wang, V., Davis, D., Haque, M., Huang, L., & Yarchoan, R. (2005). Differential gene up-regulation by hypoxia-inducible factor-1alpha and hypoxia-inducible factor-2alpha in HEK293T cells. Cancer Research , 65 (8), 3299-306. Wang, X., Tomso, D., Chorley, B., Cho, H.-Y., Cheung, V., Kleeberger, S., et al. (2007). Identification of polymorphic antioxidant response elements in the human genome. Human molecular genetics , 16 (10), 1188-200. Wang, X., Tomso, D., Liu, X., & Bell, D. (2005). Single nucleotide polymorphism in transcriptional regulatory regions and expression of environmentally responsive genes. Toxicology and applied pharmacology , 207 (2 Suppl), 84-90. Ward, A., & Cooper, T. (2010). The pathobiology of splicing. The Journal of pathology , 220 (2), 152-63. Watson, E., & Cross, J. (2005). Development of structures and transport functions in the mouse placenta. Physiology (Bethesda, Md) , 20, 180-93. Wauters, M., Considine, R., & Van Gaal, L. (2000). Human leptin: from an adipocyte hormone to an endocrine mediator. European journal of endocrinology / European Federation of Endocrine Societies , 143 (3), 293-311. Wegmann, F., Petri, B., Khandoga, A., Moser, C., Khandoga, A., Volkery, S., et al. (2006). ESAM supports neutrophil extravasation, activation of Rho, and VEGF-induced vascular permeability. The Journal of experimental medicine , 203 (7), 1671-7. Weir, P. (1981). Immunofluorescent studies of the uteroplacental arteries in normal pregnancy. British journal of obstetrics and gynaecology , 88 (3), 301-7. Welt, C., Martin, K., Taylor, A., Lambert-Messerlian, G., Crowley, W., Smith, J., et al. (1997). Frequency modulation of follicle-stimulating hormone (FSH) during the luteal-follicular transition: evidence for FSH control of inhibin B in normal women. The Journal of clinical endocrinology and metabolism , 82 (8), 2645-52. Welt, C., Pagan, Y., Smith, P., Rado, K., & Hall, J. (2003). Control of follicle-stimulating hormone by estradiol and the inhibins: critical role of estradiol at the hypothalamus during the luteal-follicular transition. The Journal of clinical endocrinology and metabolism , 88 (4), 1766-71. Wenger, R., Stiehl, D., & Camenisch, G. (2005). Integration of oxygen signaling at the consensus HRE. Science's STKE : signal transduction knowledge environment , 2005 (306), re12. Wolf, H., Zarnegar, R., Oliver, L., & Michalopoulos, G. (1991). Hepatocyte growth factor in human placenta and trophoblastic disease. The American journal of pathology , 138 (4), 1035-43. Wooding, F.B.P., and Flint, A.P.F. (1994). Placentation. In Lamming GE, ed., Marshall’s physiology of reproduction (233-460). London: Chapman & Hall. Wouters, B., & Koritzinsky, M. (2008). Hypoxia signalling through mTOR and the unfolded protein response in cancer. Nature reviews Cancer , 8 (11), 851-64. Xu, C., Mao, D., Holers, V., Palanca, B., Cheng, A., & Molina, H. (2000). A critical role for murine complement regulator crry in fetomaternal tolerance. Science (New York, NY) , 287 (5452), 498-501. Xu, H., Handoko, L., Wei, X., Ye, C., Sheng, J., Wei, C.-L., et al. (2010). A signal-noise model for significance analysis of ChIP-seq with negative control. Bioinformatics (Oxford, England) , 26 (9), 1199-204. Xu, J., Kimball, T., Lorenz, J., Brown, D., Bauskin, A., Klevitsky, R., et al. (2006). GDF15/MIC-1 functions as a protective and antihypertrophic factor released from the myocardium in association with SMAD protein activation. Circulation research , 98 (3), 342-50. Xu, R.-H., Chen, X., Li, D., Li, R., Addicks, G., Glennon, C., et al. (2002). BMP4 initiates human embryonic stem cell differentiation to trophoblast. Nature biotechnology , 20 (12), 1261-4. Xu, R.-H., Peck, R., Li, D., Feng, X., Ludwig, T., & Thomson, J. (2005). Basic FGF and suppression of BMP signaling sustain undifferentiated proliferation of human ES cells. Nature methods , 2 (3), 185-90. Xu, R.-H., Sampsell-Barron, T., Gu, F., Root, S., Peck, R., Pan, G., et al. (2008). NANOG is a direct target of TGFbeta/activin-mediated SMAD signaling in human ESCs. Cell stem cell , 3 (2), 196-206.

238 Yamaguchi, S., Shinmura, K., Saitoh, T., Takenoshita, S., Kuwano, H., & Yokota, J. (2002). A single nucleotide polymorphism at the splice donor site of the human MYH base excision repair genes results in reduced translation efficiency of its transcripts. Genes to cells : devoted to molecular & cellular mechanisms , 7 (5), 461-74. Yang, M., Lei, Z., & Rao, C. (2003). The central role of human chorionic gonadotropin in the formation of human placental syncytium. Endocrinology , 144 (3), 1108-20. Yi, X., Liang, Y., Huerta-Sanchez, E., Jin, X., Cuo, Z., Pool, J., et al. (2010). Sequencing of 50 human exomes reveals adaptation to high altitude. Science (New York, NY) , 329 (5987), 75-8. Yoon, D., Ponka, P., & Prchal, J. (2011). Hypoxia. 5. Hypoxia and hematopoiesis. American journal of physiology Cell physiology , 300 (6), C1215-22. Young, M., Kerr, J., Termine, J., Wewer, U., Wang, M., McBride, O., et al. (1990). cDNA cloning, mRNA distribution and heterogeneity, chromosomal location, and RFLP analysis of human osteopontin (OPN). Genomics , 7 (4), 491-502. Yu, L., Li, D., Liao, Q.-P., Yang, H.-X., Cao, B., Fu, G., et al. (2012). High levels of activin A detected in preeclamptic placenta induce trophoblast cell apoptosis by promoting nodal signaling. The Journal of clinical endocrinology and metabolism , 97 (8), E1370-9. Yu, P., Pan, G., Yu, J., & Thomson, J. (2011). FGF2 sustains NANOG and switches the outcome of BMP4-induced human embryonic stem cell differentiation. Cell stem cell , 8 (3), 326-34. Zamore, P., Patton, J., & Green, M. (1992). Cloning and domain structure of the mammalian splicing factor U2AF. Nature , 355 (6361), 609-14. Zamudio, S., Kovalenko, O., Vanderlelie, J., Illsley, N., Heller, D., Belliappa, S., et al. (2007). Chronic hypoxia in vivo reduces placental oxidative stress. Placenta , 28 (8-9), 846-53. Zamudio, S., Palmer, S., Droma, T., Stamm, E., Coffin, C., & Moore, L. (1995). Effect of altitude on uterine artery blood flow during normal pregnancy. Journal of applied physiology: respiratory, environmental and exercise physiology , 79 (1), 7-14. Zamudio, S., Postigo, L., Illsley, N., Rodriguez, C., Heredia, G., Brimacombe, M., et al. (2007). Maternal oxygen delivery is not related to altitude- and ancestry-associated differences in human fetal growth. The Journal of physiology , 582 (Pt 2), 883-95. Zamudio, S., Wu, Y., Ietta, F., Rolfo, A., Cross, A., Wheeler, T., et al. (2007). Human placental hypoxia-inducible factor-1alpha expression correlates with clinical outcomes in chronic hypoxia in vivo. The American journal of pathology , 170 (6), 2171-9. Zegers, D., Beckers, S., Mertens, I., Van Gaal, L., & Van Hul, W. (2011). Association between polymorphisms of the Nesfatin gene, NUCB2, and obesity in men. Molecular Genetics and Metabolism , 103 (3), 282-6. Zhang, P., Zucchelli, M., Bruce, S., Hambiliki, F., Stavreus-Evers, A., Levkov, L., et al. (2009). Transcriptome profiling of human pre-implantation development. PLoS ONE , 4 (11), e7844. Zhang, X., Li, L., Jung, J., Xiang, S., Hollmann, C., & Choi, Y. (2001). The distinct roles of T cell- derived cytokines and a novel follicular dendritic cell-signaling molecule 8D6 in germinal center-B cell differentiation. Journal of immunology (Baltimore, Md : 1950) , 167 (1), 49-56. Zhang, Y., Liu, T., Meyer, C., Eeckhoute, J., Johnson, D., Bernstein, B., et al. (2008). Model-based analysis of ChIP-Seq (MACS). Genome Biology , 9 (9), R137. Zhang, Y., Proenca, R., Maffei, M., Barone, M., Leopold, L., & Friedman, J. (1994). Positional cloning of the mouse obese gene and its human homologue. Nature , 372 (6505), 425-32. Zhang, Y., Zhang, Y.-L., Feng, C., Wu, Y.-T., Liu, A.-X., Sheng, J.-Z., et al. (2008). Comparative proteomic analysis of human placenta derived from assisted reproductive technology. Proteomics , 8 (20), 4344-56. Zhao, Y., Koga, K., Osuga, Y., Nagai, M., Izumi, G., Takamura, M., et al. (2012). Thrombin enhances soluble Fms-like tyrosine kinase 1 expression in trophoblasts; possible involvement in the pathogenesis of preeclampsia. Fertility and sterility , 98 (4), 917-21. Zubieta-Castillo, G., Zubieta-Calleja, G., & Zubieta-Calleja, L. (2006). Chronic mountain sickness: the reaction of physical disorders to chronic hypoxia. Journal of physiology and pharmacology : an official journal of the Polish Physiological Society , 57 Suppl 4, 431-42.

239 Zuccato, E., Buratti, E., Stuani, C., Baralle, F., & Pagani, F. (2004). An intronic polypyrimidine-rich element downstream of the donor site modulates cystic fibrosis transmembrane conductance regulator exon 9 alternative splicing. The Journal of biological chemistry , 279 (17), 16980-8. Zwaka, T., & Thomson, J. (2003). Homologous recombination in human embryonic stem cells. Nature biotechnology , 21 (3), 319-21. Zygmunt, M., Herr, F., Keller-Schoenwetter, S., Kunzi-Rapp, K., Münstedt, K., Rao, C., et al. (2002). Characterization of human chorionic gonadotropin as a novel angiogenic factor. The Journal of clinical endocrinology and metabolism , 87 (11), 5290-6.

240 APPENDIX

Primers used:

Forward Reverse Amplification of:

GGAGAGGAGGAAGGAGAA GCTTGGTGAATAGGAAGT EPAS1 exons 2 to 7

TTCCCATCCTTCCACATC GTTCAGTTCCACACCTTT SNP site

ACCTGCCCCTCACTCCACCCCC GCGGCTGACTGTGGGGTGTCC CA9 binding site

AGTGTCCGTTCCCAGCTCAG TAGGCACAGTAAACAGGCCC PHD3 binding site

GGGTGAAGGTGACAGCAAGT ACACAGACTCAACGTGCCA BID17 binding site

GTCACCACGCCTAATTTT TTCCCTTTCCCATCCTCT (-)BID17 binding site

TaqMan® probes used:

Gene Probe ID

CDX2 Hs01078080_m1

CGA Hs00985275_g1

DLX3 Hs00270938_m1

EPAS1 Hs01026149_m1

GATA2 Hs00231119_m1

GATA3 Hs00231122_m1

HAND1 Hs02330376_s1

HIF1A Hs00153153_m1

ID2 Hs04187239_m1

KRT18 Hs02827483_g1

MSX2 Hs00751239_s1

NANOG Hs04260366_g1

OCT4/POU5F1 Hs00999632_g1

PGF Hs00182176_m1

SOX2 Hs01053049_s1

TFAP2A Hs01029413_m1

TFAP2C Hs00231476_m1

241 Sequence for transcript skipping exon 3:

CCGGTGGGTGCGGGGGACAGGAGACGGAGGTGTTCTATGAGCTGGCCCAT GAGCTGCCTCTGCCCCACAGTGTGAGCTCCCATCTGGACAAGGCCTCCATC ATGCGACTGGCAATCAGCTTCCTGCGAACACACAAGCTCCTCTCCTCAGTT TGCTCTGAAAACGAGTCCGAAGCCGAAGCTGACCAG:CATGGACATGAAG TTCACCTACTGTGATGACAGAATCACAGAACTGATTGGTTACCACCCTGAG GAGCTGCTTGGCCGCTCAGCCTATGAATTCTACCATGCGCTAGACTCCGAG AACATGACCAAGAGTCACCAGAACTTGTGCACCAAGGGTCAGGTAGTAAG TGGCCAGTACCGGATGCTCGCAAAGCATGGGGGCTACGTGTGGCTGGAGA CCCAGGGGACGGTCATCTACAACCCTCGCAACAAGGGCGAATTCCAGCAC ACTGGCGGCCGTTACTAGTGGATCCGAGCTCGGTACCAAGCTTGGCGTAAT CATGGTCATAGCTGTTTCCTGTGTGAAATTGTTATCCGCTCACAATTCCACA CAACATACGAGCCGGAAGCATAAAGTGTAAAGCCTGGGGTGCCTAATGAG TGAGCTAACTCACATTAATTGCGTTGCGCTCACTGCCCGCTTTCCAGTCGG GAAACCTGTCGTGCCAGCTGCATTAATGAATCGGCCAACGCGCGGGGAGA GGCGGTTTGCGTATTGGGCGCTCTTCCGCTTCCTCGCTCACTGACTCGCTG CGCTCGGTCGTTCGGCTGCGGCGAGCGGTATCAGCTCACTCAAAGGCGGT AATACGGTTATCCACAGAATCAGGGGATAACGCAGGAAAGAACATGTGAG CAAAAGGCCAGCAAAAGGCCAGGAACCGTAAAAAGGCCGCGTTGCTGGC GTTTTTCCATAGGCTCCGCCCCCCTGACGAGCATCACAAAAATCGACGCTC AAGTCAGAGGTGGCGAAACCCGACAGGACTATAAAGATACCAGGCGTTTC CCCCTGGAAGCTCCCTCGTGCGCTCTCCTGTTCCGACCCTGCCGCTTACCG GATACCTGTCCGCTTTTCTCCCTTCGGGAAGCGTGGCGCTTTCCCAAAGCT CACGCTGTAGGGATCTCAGTTCG

Sequence for transcript skipping exon 2 (partial) and exon 3:

CGGTGCCGTGCGGCGGACAAGGAGACGGAGGTGTTCTATGAGCTGGCCCA TGAGCTGCCTCTGCCCCACAGTGTGAGCTCCCATCTGGACAAGGCCTCCAT CATGCGACTGGCAATCAGCTTCCTGCGAACACACAAG:ACCTTCCTGAGCC GCCACAGCATGGACATGAAGTTCACCTACTGTGATGACAGAATCACAGAA CTGATTGGTTACCACCCTGAGGAGCTGCTTGGCCGCTCAGCCTATGAATTC TACCATGCGCTAGACTCCGAGAACATGACCAAGAGTCACCAGAACTTGTG CACCAAGGGTCAGGTAGTAAGTGGCCAGTACCGGATGCTCGCAAAGCATG GGGGCTACGTGTGGCTGGAGACCCAGGGGACGGTCATCTACAACCCTCGC AACAAGGGCGAATTCCAGCACACTGGCGGCCGTTACTAGTGGATCCGAGC TCGGTACCAAGCTTGGCGTAATCATGGTCATAGCTGTTTCCTGTGTGAAATT GTTATCCGCTCACAATTCCACACAACATACGAGCCGGAAGCATAAAGTGTA AAGCCTGGGGTGCCTAATGAGTGAGCTAACTCACATTAATTGCGTTGCGCT CACTGCCCGCTTTCCAGTCGGGAAACCTGTCGTGCCAGCTGCATTAATGAA TCGGCCAACGCGCGGGGAGAGGCGGTTTGCGTATTGGGCGCTCTTCCGCT TCCTCGCTCACTGACTCGCTGCGCTCGGTCGTTCGGCTGCGGCGAGCGGTA TCAGCTCACTCAAAGGCGGTAATACGGTTATCCACAGAATCAGGGGATAAC GCAGGAAAGAACATGTGAGCAAAAGGCCAGCAAAAGGCCAGGAACCGTA AAAAGGCCGCGTTGCTGGCGTTTTTCCATAGGCTCCGCCCCCCTGACGAGC

242 ATCACAAAAATCGACGCTCAAGTCAGAGGTGGCGAAACCCGACAGGACTA TAAAGATACCAGGCGTTTCCCCCTGGAAGCTCCCTCGTGCGCTCTCCTGTT CCGACCCTGCCGCTTACCGGATACCTGTCCGCCTTTCTCCCTTCGGGAAGC GTGGCGCTTTCTCATAGCTCACGCTGTAGGTATCTCAGTTCGGGGTAAGGT CGTTCGCTCCAAGCTGGGCTTGGTGGCCCGAACCCCCCGTTCAGCCCGAA CGGTGGGCCTTTTT

243 Supplementary Table 1. 681 genes differentially expressed between pre- and post-flow placentae.

Gene p-value Gene p-value

HBZ 4.74E-32 FCER1G 2.54E-13

HBE1 1.42E-29 BMP5 2.50E-13

SPCS3 3.58E-24 RNU1-6 2.90E-13

TUSC3 6.03E-17 PSAP 3.01E-13

MS4A4A 5.36E-17 LEP 2.81E-13

NCF1C 3.80E-17 RNU1-9 2.90E-13

CLEC3B 4.82E-17 CGB7 3.17E-13

CYP19A1 8.57E-16 RNASE1 3.30E-13

RNASE1 1.25E-15 FOLR2 4.48E-13

CGB2 1.46E-15 SEPP1 5.50E-13

LNPEP 4.98E-15 GDF15 6.41E-13

FCGR2B 5.52E-15 C1QA 7.46E-13

CD164 9.55E-15 SPCS3 1.01E-12

TMEM176B 1.53E-14 NUCB2 1.09E-12

CANX 1.66E-14 NUCB2 1.46E-12

ATF3 1.53E-14 C1QB 1.52E-12

PDPN 1.89E-14 RNU1-4 2.14E-12

TFPI 2.00E-14 RNU1-5 2.14E-12

ADAM10 1.52E-14 FAM162B 1.90E-12

PDIA6 2.21E-14 RNU1-1 2.14E-12

MAOA 2.68E-14 RNU1-7 2.14E-12

CADM3 4.54E-14 RNU1-3 2.14E-12

SDC1 5.07E-14 RNU1-2 2.05E-12

INHBA 7.22E-14 RNU1-8 2.13E-12

SPARCL1 8.40E-14 GGTA1P 2.55E-12

RNU4-2 9.23E-14 C1QA 3.04E-12

PORCN 9.00E-14 NCF1 3.35E-12

ABP1 1.73E-13 SNORD97 4.03E-12

PDPN 2.06E-13 RNASE1 4.31E-12

HSPA7 2.26E-13 NEURL 4.62E-12

244 Gene p-value Gene p-value

CGA 4.88E-12 CYP19A1 7.58E-11

CD320 5.15E-12 SERP1 7.71E-11

CLDN10 7.16E-12 NCF1B 7.82E-11

KDELR2 7.94E-12 VSIG4 8.56E-11

YIPF6 9.32E-12 RPL8 9.21E-11

HLA-A 9.22E-12 CD59 9.16E-11

TMBIM6 1.02E-11 HSD17B2 9.49E-11

C7orf71 1.43E-11 TREM1 9.77E-11

HIGD1B 1.70E-11 TMBIM6 1.04E-10

CGB7 1.74E-11 HLA-C 1.15E-10

MPEG1 1.80E-11 UCA1 1.25E-10

FCGR2C 1.98E-11 DARC 1.40E-10

RARRES2 2.03E-11 TMEM167A 1.55E-10

RAB7A 2.08E-11 LYVE1 1.71E-10

AIG1 2.20E-11 RGCC 1.74E-10

CYP19A1 2.19E-11 C1R 1.71E-10

INHA 2.17E-11 MS4A6A 1.86E-10

FCGR1A 2.42E-11 ETF1 2.35E-10

PSG5 2.52E-11 SVEP1 2.40E-10

ST3GAL4 2.91E-11 MS4A6A 2.57E-10

A2M 3.26E-11 GAB2 2.78E-10

C1QC 3.34E-11 Metazoa_SRP 2.98E-10

SRSF5 3.77E-11 ABI3BP 3.10E-10

TMEM100 3.83E-11 CGB5 3.26E-10

CGB2 4.54E-11 TCL6 3.52E-10

APOD 4.85E-11 NUCB2 3.59E-10

FCGR2A 6.00E-11 CGB1 3.79E-10

F13A1 6.22E-11 AC016722.2 4.00E-10

PORCN 6.51E-11 SPCS2 4.22E-10

ENTPD1 7.04E-11 SAT1 4.19E-10

245 Gene p-value Gene p-value

GNAS 4.70E-10 LVRN 2.24E-09

CGB1 4.99E-10 UQCRC1 2.31E-09

C8orf4 4.96E-10 LAPTM5 2.40E-09

WNT3A 5.19E-10 NEURL 2.70E-09

CDC42SE1 6.26E-10 C3AR1 2.73E-09

EMP1 6.24E-10 PTGDS 2.81E-09

CD37 6.23E-10 WASH7P 2.95E-09

RCN1 6.46E-10 CGB1 3.07E-09

RPS27A 6.52E-10 LAIR1 3.10E-09

ERO1L 6.72E-10 IGFBP4 3.34E-09

AREG 6.81E-10 SVEP1 3.33E-09

HPGDS 8.16E-10 RNU2-2 4.20E-09

AREGB 8.80E-10 RASSF4 4.25E-09

TMED2 8.77E-10 AC083884.8 4.35E-09

CD68 8.74E-10 PORCN 4.34E-09

WDR1 9.07E-10 MT1G 4.46E-09

GLRX 9.54E-10 TYROBP 4.63E-09

C4orf26 9.49E-10 REEP5 4.72E-09

CYTL1 1.01E-09 CD36 4.81E-09

CADM3 1.05E-09 C8orf42 5.02E-09

AL773572.7 1.20E-09 SEMA3B 5.00E-09

ALDH1A1 1.19E-09 SNORA26 4.99E-09

CGB 1.18E-09 GPR34 5.55E-09

CCR7 1.25E-09 PTPN21 5.76E-09

FKBP11 1.56E-09 TCL6 7.41E-09

PSG6 1.60E-09 CD36 7.86E-09

CFD 1.77E-09 TCF21 7.85E-09

HERPUD1 1.88E-09 LHB 7.99E-09

IL6ST 1.88E-09 CD14 8.20E-09

WNT2 2.03E-09 IFITM2 9.21E-09

246 Gene p-value Gene p-value

ANXA5 9.45E-09 TFPI2 3.61E-08

GPNMB 9.69E-09 HLA-B 3.65E-08

PSG5 9.89E-09 CD53 3.71E-08

PROCR 9.80E-09 KYNU 3.78E-08

FPR1 9.88E-09 CD82 4.00E-08

IGSF6 1.02E-08 CD68 4.04E-08

PLAGL1 1.06E-08 ALPL 4.08E-08

HLA-B 1.10E-08 LYVE1 4.19E-08

METTL7B 1.14E-08 PRELID1 4.34E-08

ALPL 1.24E-08 SRSF5 4.53E-08

TMSB4X 1.27E-08 LMAN2 4.56E-08

CARD17 1.31E-08 KCNA7 4.61E-08

ADAM10 1.46E-08 SEC11C 4.87E-08

ZNHIT1 1.45E-08 CD163 4.94E-08

MBD6 1.49E-08 TREML2 4.93E-08

ILVBL 1.52E-08 FCGRT 5.04E-08

TMED10 1.65E-08 SLC5A6 5.03E-08

CALU 1.77E-08 LY6D 5.34E-08

ARL1 2.10E-08 CSDE1 5.71E-08

SRGN 2.23E-08 MAN1A2 5.66E-08

ABP1 2.49E-08 HN1L 5.71E-08

SDF2L1 2.50E-08 SLCO2B1 5.69E-08

PRG2 2.53E-08 HCST 5.76E-08

POLDIP3 2.88E-08 TSPAN9 5.89E-08

C15orf29 2.91E-08 RPL41 5.93E-08

HLA-B 3.00E-08 KYNU 6.10E-08

CSHL1 3.26E-08 PSG8 6.29E-08

PDPN 3.66E-08 RPL21 6.28E-08

IFITM3 3.62E-08 ITGB2 6.37E-08

NUCB2 3.66E-08 RP11-495P10.9 6.42E-08

247 Gene p-value Gene p-value

PSG5 6.47E-08 RNASEH2C 1.26E-07

PSG6 6.61E-08 ABCG2 1.27E-07

CAPRIN1 6.82E-08 TUSC3 1.30E-07

C7 6.81E-08 CANX 1.30E-07

HES2 6.89E-08 SMPDL3B 1.31E-07

FBN1 7.04E-08 C11orf58 1.33E-07

GPT2 7.27E-08 TMED7 1.42E-07

RER1 7.54E-08 AL773572.7 1.49E-07

ADAM12 7.57E-08 BACE2 1.50E-07

SEMA3B 7.97E-08 IL33 1.49E-07

TFPI 7.95E-08 PSG7 1.49E-07

APOE 8.36E-08 CD163 1.51E-07

PMP22 8.35E-08 ALPP 1.52E-07

MT1G 8.73E-08 C2orf40 1.53E-07

CD63 8.80E-08 SRSF5 1.56E-07

GAS6 9.03E-08 TFPI 1.58E-07

SIGLEC1 9.01E-08 PLP2 1.60E-07

CYP19A1 9.34E-08 HBG2 1.58E-07

CD93 9.44E-08 HPGDS 1.60E-07

SLC16A3 9.42E-08 SPP1 1.59E-07

ELL2 9.62E-08 ILVBL 1.57E-07

RPLP0 9.62E-08 ANO1 1.60E-07

TXNIP 9.79E-08 EZR 1.73E-07

HBB 1.00E-07 MFAP5 1.76E-07

MCFD2 1.03E-07 NPIPL3 1.80E-07

FBLN5 1.09E-07 IGFBP2 1.92E-07

HSD3B1 1.17E-07 FILIP1L 1.94E-07

CALM1 1.18E-07 RASA1 1.94E-07

TCL6 1.19E-07 ENG 1.96E-07

TMSB4XP6 1.22E-07 CD34 1.97E-07

248 Gene p-value Gene p-value

PLAGL1 1.98E-07 CD63 3.14E-07

FKBP11 2.02E-07 PLEKHA6 3.25E-07

SDHAP2 2.01E-07 GPAA1 3.49E-07

AL773572.7 2.03E-07 PLTP 3.53E-07

SLC30A2 2.04E-07 MMP2 3.54E-07

LVRN 2.10E-07 RP11-76P2.3 3.56E-07

CYBB 2.18E-07 OLAH 3.61E-07

MAF 2.26E-07 PSG5 3.64E-07

HGF 2.28E-07 PIK3IP1 4.03E-07

PAM 2.30E-07 TUSC3 4.13E-07

CD63 2.33E-07 CXCL1 4.12E-07

PLA2G7 2.34E-07 CCRL2 4.32E-07

IFI16 2.37E-07 RNASE1 4.44E-07

NUCB2 2.36E-07 CLDN7 4.52E-07

FCGR1A 2.38E-07 PDLIM3 4.57E-07

ARHGAP17 2.36E-07 PSG9 4.56E-07

NDN 2.45E-07 CD4 4.56E-07

RHOA 2.46E-07 REEP5 4.62E-07

PSG2 2.48E-07 COX4I2 4.73E-07

RBP7 2.56E-07 TMED9 4.72E-07

CD63 2.61E-07 AP006285.6 4.96E-07

DDX3X 2.68E-07 HPGD 5.03E-07

PSG9 2.73E-07 CDH11 5.01E-07

PSG3 2.72E-07 RP11-495P10.8 5.24E-07

DAB2 2.78E-07 NR2F1 5.43E-07

PLAC1 2.88E-07 ITGBL1 5.43E-07

AC004878.6 2.87E-07 CCL2 5.48E-07

SAR1A 3.08E-07 CD36 5.50E-07

LGALS9 3.11E-07 MMRN2 5.62E-07

SOD3 3.14E-07 GJA4 5.60E-07

249 Gene p-value Gene p-value

G0S2 5.61E-07 MARK2 8.77E-07

HBE1 5.59E-07 SNRPG 8.79E-07

TRAM1 5.72E-07 TCL6 8.75E-07

C4orf36 5.82E-07 SEMA7A 9.14E-07

AZIN1 5.94E-07 GPRC5C 9.61E-07

KLF2 5.97E-07 OLFML3 9.73E-07

SPI1 5.99E-07 AGPAT5 9.67E-07

MLEC 5.96E-07 FHL1 9.72E-07

CTDSP1 6.10E-07 MANF 9.71E-07

PRG2 6.08E-07 C1GALT1 9.91E-07

SPP1 6.24E-07 NUCB2 1.01E-06

YIPF6 6.73E-07 CSF1 1.01E-06

SLC16A3 6.78E-07 PROCR 1.02E-06

FBLN5 6.76E-07 GPR32 1.03E-06

ENPP2 6.89E-07 PISD 1.03E-06

SLC16A3 6.93E-07 SLC44A1 1.06E-06

SEL1L3 7.15E-07 NBL1 1.06E-06

FBXL16 7.37E-07 APOLD1 1.06E-06

RHOBTB1 7.39E-07 PPIB 1.09E-06

LYPD3 7.52E-07 TMEM59 1.09E-06

LAMA2 7.82E-07 GIMAP4 1.08E-06

SULT1A1 8.15E-07 UBE2D3 1.10E-06

FTH1 8.13E-07 GMFG 1.13E-06

CSH1 8.25E-07 TGFBI 1.14E-06

FCGR3B 8.32E-07 TUSC3 1.14E-06

C1QTNF6 8.32E-07 TYROBP 1.17E-06

AK2 8.38E-07 DNAJC3 1.17E-06

CALM1 8.56E-07 BST2 1.19E-06

PSCA 8.68E-07 LGALS14 1.19E-06

AREG 8.67E-07 TMED10 1.20E-06

250 Gene p-value Gene p-value

EGFL7 1.19E-06 ANG 1.64E-06

RPS13 1.20E-06 HYOU1 1.64E-06

HIST1H1C 1.21E-06 TFPI 1.68E-06

PSG8 1.25E-06 HGF 1.68E-06

HDLBP 1.25E-06 CAST 1.69E-06

EIF4A2 1.26E-06 RPS20 1.70E-06

RSPO3 1.26E-06 LAMP2 1.73E-06

CD55 1.27E-06 EBI3 1.80E-06

LPAR6 1.29E-06 PLEC 1.80E-06

CTSS 1.29E-06 LVRN 1.82E-06

MORF4L1 1.29E-06 NNMT 1.85E-06

TNNT1 1.30E-06 SEMA3B 1.86E-06

KYNU 1.32E-06 KIAA0182 1.91E-06

GPT2 1.33E-06 PTPN21 1.92E-06

RAB7A 1.37E-06 P4HB 1.96E-06

SEL1L3 1.40E-06 PLAGL1 1.96E-06

ADCY3 1.40E-06 PSG5 2.05E-06

RFTN1 1.42E-06 PTEN 2.04E-06

PSG1 1.42E-06 CSH1 2.11E-06

WASH3P 1.43E-06 FAM83D 2.11E-06

ST3GAL4 1.48E-06 CYP19A1 2.17E-06

FKBP2 1.49E-06 PLEK 2.17E-06

ANG 1.52E-06 IL6ST 2.17E-06

PORCN 1.53E-06 GRAMD1B 2.26E-06

PHYHIP 1.56E-06 RPLP2 2.31E-06

C6orf89 1.57E-06 CDC42 2.37E-06

RP11-676F20.1 1.58E-06 MLEC 2.40E-06

PORCN 1.61E-06 LARP4B 2.46E-06

SSR3 1.63E-06 CYTH4 2.52E-06

ITM2B 1.63E-06 GAS6 2.53E-06

251 Gene p-value Gene p-value

GOLT1A 2.55E-06 RPL8 3.42E-06

SLC25A6 2.57E-06 MATN2 3.47E-06

ATP5B 2.62E-06 ARHGAP4 3.56E-06

RMRP 2.64E-06 HIST1H2BK 3.59E-06

SEMA6A 2.65E-06 VTN 3.61E-06

RPN2 2.68E-06 TAC3 3.71E-06

VPS28 2.71E-06 NCK2 3.78E-06

INSL4 2.75E-06 MMP2 3.86E-06

HLA-DRA 2.78E-06 ZEB2 3.90E-06

AC107016.1 2.77E-06 NDUFA4L2 3.94E-06

ENPP2 2.79E-06 SLC39A8 4.00E-06

HLA-G 2.85E-06 TREM1 3.99E-06

PLXNA3 2.89E-06 CD74 4.02E-06

SLC7A5 2.90E-06 MRC1L1 4.08E-06

RNASE6 2.91E-06 FKBP2 4.20E-06

TPT1-AS1 2.93E-06 CTSS 4.25E-06

C1QTNF6 2.97E-06 SRA1 4.25E-06

AL773572.7 2.99E-06 CTSZ 4.28E-06

MFSD12 3.00E-06 PVRL3 4.29E-06

AIF1 3.10E-06 VSIG4 4.38E-06

SPTLC3 3.14E-06 GYPB 4.41E-06

ITGB2 3.15E-06 ARSD 4.50E-06

PLA2G2A 3.17E-06 TMEM123 4.52E-06

FBN1 3.18E-06 SNRPG 4.58E-06

TUBB 3.19E-06 BMP1 4.60E-06

PSG1 3.25E-06 GUCY1B3 4.62E-06

CD44 3.35E-06 EIF4G2 4.63E-06

CLN3 3.35E-06 TSPAN4 4.71E-06

DNAJB11 3.35E-06 CD74 4.70E-06

CPQ 3.38E-06 PDIA3 4.70E-06

252 Gene p-value Gene p-value

C4orf26 4.74E-06 SH2D5 6.10E-06

MEG3 4.74E-06 HEYL 6.30E-06

CD36 5.04E-06 CCBP2 6.42E-06

WNT2 5.06E-06 RPL32 6.56E-06

PLAGL1 5.09E-06 MMP2 6.60E-06

HMGN4 5.08E-06 BASP1 6.62E-06

SAR1B 5.08E-06 NPL 6.65E-06

PSG4 5.11E-06 INHBA 6.77E-06

MTUS1 5.14E-06 KRT7 6.81E-06

ADM 5.15E-06 PSG1 6.84E-06

PITPNA 5.18E-06 Metazoa_SRP 6.85E-06

SPDYC 5.29E-06 EIF4E2 6.89E-06

LHFP 5.31E-06 AIM1 7.08E-06

TMEM59 5.45E-06 ERBB2 7.17E-06

FAM156B 5.45E-06 PRSS29P 7.17E-06

ADORA3 5.55E-06 FAM65B 7.16E-06

CSH2 5.56E-06 AP1S2 7.25E-06

CD63 5.61E-06 CAPG 7.60E-06

SH3BGRL 5.65E-06 HLA-B 7.62E-06

HK2 5.66E-06 EEF1A1 7.65E-06

SPP1 5.77E-06 INPPL1 7.65E-06

PRSS8 5.76E-06 RP11-104D21.3 7.78E-06

CD36 5.81E-06 CHD4 7.78E-06

SIGLEC6 5.84E-06 ST3GAL4 7.94E-06

RPLP0 5.84E-06 STT3B 8.01E-06

IGFBP5 5.90E-06 PDE4DIP 8.03E-06

BACE2 5.97E-06 HSPA5 8.20E-06

LDHA 6.05E-06 RP11-10A14.5 8.29E-06

ENPP2 6.10E-06 TIMP2 8.35E-06

RAB10 6.11E-06 FDFT1 8.46E-06

253 Gene p-value Gene p-value

MESDC2 8.69E-06 MTNR1B 1.16E-05

RP11-24F11.2 8.69E-06 FTL 1.17E-05

PSG11 8.83E-06 NCF1B 1.19E-05

STAB1 8.92E-06 INMT 1.21E-05

AC073624.1 9.06E-06 MEG3 1.22E-05

SH3YL1 9.10E-06 BZW2 1.21E-05

MT1H 9.27E-06 CTSD 1.22E-05

TCF25 9.32E-06 TSPAN4 1.21E-05

PCNP 9.36E-06 MAFK 1.23E-05

MXRA7 9.35E-06 DHRS9 1.23E-05

RPL7 9.43E-06 MAP3K13 1.24E-05

P4HB 9.48E-06 SPTLC3 1.24E-05

RAB3IL1 9.55E-06 NPEPPS 1.24E-05

HDGF 9.68E-06 SLC5A6 1.26E-05

RPL21 9.73E-06 CD52 1.29E-05

MAFB 9.79E-06 BZW2 1.29E-05

ACER3 1.00E-05 PTPN6 1.32E-05

ZDHHC3 1.01E-05 SLC9A9 1.32E-05

GALNT2 1.01E-05 TGM2 1.34E-05

VSIG4 1.01E-05 ERVW-1 1.35E-05

NUDT21 1.04E-05 TGFBI 1.38E-05

LAMC3 1.05E-05 RAB2A 1.41E-05

ANXA4 1.07E-05 F3 1.41E-05

HLA-C 1.08E-05 ARRB2 1.42E-05

EIF3C 1.10E-05 ARPC3 1.44E-05

CCDC47 1.13E-05 UBE2K 1.45E-05

CSHL1 1.13E-05 CNPY2 1.45E-05

MMP14 1.14E-05 RANGRF 1.46E-05

AEBP1 1.16E-05 AP2M1 1.46E-05

MBD2 1.16E-05 IGFBP2 1.46E-05

254 Gene p-value

SCARB2 1.47E-05

RP11-462G2.1 1.50E-05

AMFR 1.49E-05

KLHL5 1.51E-05

PDGFRB 1.57E-05

PAMR1 1.56E-05

SRSF4 1.57E-05

COL5A3 1.58E-05

AAMP 1.60E-05

TPPP3 1.60E-05

TPD52 1.61E-05

WDR1 1.62E-05

MAN1C1 1.63E-05

GDPD5 1.64E-05

BLVRA 1.68E-05

FSTL1 1.69E-05

MAOA 1.69E-05

DLK1 1.69E-05

ATP6V0B 1.73E-05

CCT6A 1.72E-05

RP11-76P2.2 1.73E-05

255 Supplementary Table 2. List of pathways and genes from gene ontology analysis.

256