The Roles of Sox4 and Med15 in the Development and Maintenance of Pancreatic Β-Cells

THE ROLES OF SOX4 AND MED15 IN THE DEVELOPMENT AND MAINTENANCE OF PANCREATIC β-CELLS

ERIC ERYUE XU

B.Sc., The University of British Columbia, 2007

A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF

THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES

(Cell and Developmental Biology)

THE UNIVERSITY OF BRITISH COLUMBIA

(Vancouver)

December 2016

Abstract

Pancreatic beta-cells (β-cells) are essential for the maintenance of blood glucose homeostasis, as

the primary insulin-secreting cells of the body. During embryogenesis, β-cells differentiate from

pancreatic progenitor cells, and following birth, these cells re-enter the cell-cycle and proliferate

to maintain a sufficient adult population of β-cells. Transcription factors (TFs) such as neurogenin3 (Neurog3) are essential for endocrine cell specification within the pancreas, while other TFs are required in adult β-cells to maintain their function. Despite the identification of many TFs throughout β-cell development, how TFs regulate the transition between cell states,

and how these TFs engage the RNA Polymerase II holoenzyme to regulate transcription is

unknown. To address these questions, this thesis examines the role of Sry-related HMG-box 4

(SOX4) and Mediator 15 (MED15) in β-cell development and the adult β-cell state.

Work in this thesis has established that in mice, SOX4 is expressed in pancreatic progenitor cells and cooperates with NEUROG3 to activate Neurog3 expression. This demonstrated a

requirement for SOX4 in endocrine progenitor cell specification. SOX4 continued to be

expressed in endocrine specified cells, and was essential for Neurod1 and Pax4 induction, TFs required for β-cell specification. High-fat diet (HFD)-fed mice with inducible SOX4 deletion in

β-cells also exhibited glucose intolerance, due to decreased β-cell mass and replication rate. Loss of Sox4 led to the upregulation of the cell-cycle inhibitor Cdkn1a, a gene that prevents G1-S cell- cycle transition. Additionally, Med15 deletion in pancreatic progenitors demonstrated reduced

NEUROG3 expression, and reduced endocrine cell numbers. MED15 deletion following endocrine specification also led to reduced β-cell numbers. Finally, Ins1-Cre facilitated deletion

of MED15 in β-cells revealed that its function varied depending on cell-state, with compromised

β-cell function if expression is lost during β-cell maturation. These data are the first to determine

when SOX4 is required for pancreatic endocrine specification in mice and which targets are directly regulated by SOX4. In addition, the first known in-vivo role for MED15 in mammals is identified, demonstrating that it is an indispensable factor for β-cell differentiation and function.

Collectively, these findings contribute to the understanding of how TFs regulate β-cell states.

iii

Preface

Animal studies were reviewed and approved by the University of British Columbia Committee on Animal Care under protocols A13-0184 and A14-0163.

All studies in this thesis were conceived and designed by E.E. Xu and F.C. Lynn. All experiments, animal monitoring, and data analysis were conducted by E.E. Xu with technical assistance as follows:

Chapter 3: N.A.J. Krentz performed human embryonic stem cell culture and expression profiling. S. Tan performed luciferase assays. S.Z. Chow performed luciferase assay and pancreatic quantification experiments. M. Tang performed embryonic pancreas dissections and expression profiling. C. Nian performed western blots. F.C. Lynn performed TALEN-mediated gene editing of mPAC L20 cells, in-situ hybridization, and chromatin immunoprecipitation experiments.

Chapter 4: T. Speckmann and P.V. Sabatini provided technical assistance with mouse monitoring and experiments. C. Nian assisted with islet isolations.

Chapter 5: A.Z. Kadhim performed immunofluorescence analysis of Ins1-Cre; Med15 mouse islets. T. Speckmann performed western blots. C. Nian assisted with islet isolations and western blots. M. Tang assisted with animal monitoring and immunostaining of Pdx1-Cre; Med15 pancreatic sections.

Experiments presented in Chapter 3 have been published as Xu E.E., Krentz N.A.J., Tan S., Chow S.Z., Tang M., Nian C., and Lynn F.C. (2015). SOX4 cooperates with neurogenin 3 to regulate endocrine pancreas formation in mouse models. Diabetologia.58: 1013-1023. E.E. Xu performed experiments, analyzed data, and wrote the manuscript.

Studies described in chapter 4 are peer-reviewed and in revision as Xu E.E., Speckmann T., Nian C., and Lynn F.C. (2016) SOX4 allows facultative expansion of β-cell mass through repression of Cdkn1a. Diabetes. E.E. Xu performed experiments, analyzed data, and wrote the manuscript. T. Speckmann performed western blot experiments. C. Nian performed islet isolations.

Table of Contents

Abstract ...... ii Preface ...... iv Table of Contents ...... v List of Tables ...... ix List of Figures ...... x List of Abbreviations ...... xii Acknowledgements ...... xiii 1. Introduction ...... 1 1.1 Diabetes Mellitus...... 1 1.2 A Brief History of Diabetes Research ...... 1 1.3 Systemic Insulin Signaling and Insulin Secretion ...... 2 1.4 Diagnosis of Diabetes...... 3 1.5 Aetiology of Diabetes Mellitus ...... 3 1.6 Current Treatments for Diabetes Mellitus ...... 5 1.6 Early Pancreas Budding and Development ...... 7 1.7 Tip and Trunk Formation ...... 9 1.8 Endocrine Cell Specification ...... 10 1.9 Endocrine Cell Differentiation ...... 11 1.10 Human pancreas development ...... 15 1.10 β-cell Maturation and Identity ...... 16 1.11 Regulation of β-cell Identity and Function ...... 17 1.12 β-cell Differentiation Strategies ...... 20 1.13 Maintenance of Adult β-cell Mass ...... 21 1.14 Modulators of β-cell Replication ...... 22 1.15 β-cell Cycle Gene Regulation ...... 23 1.16 SOX Factors in Pancreas Development ...... 24 1.17 Mediator and Transcriptional Regulation ...... 27 1.18 Thesis Objectives ...... 30 2 Materials and Methods ...... 32

2.1 Chemicals ...... 32 2.2 Animals ...... 32 2.3 Diet ...... 32 2.4 Mouse Strains ...... 32 2.5 Timed Matings ...... 33 2.6 Tamoxifen Administration ...... 33 2.7 Physiological Measurements ...... 33 2.8 Embryonic Pancreas Dissection ...... 34 2.9 Adult Pancreas Collection ...... 34 2.10 Cell Culture ...... 35 2.11 Fluorescence Activated Cell Sorting ...... 35 2.12 Islet Isolation ...... 35 2.13 Human Islet Culture ...... 36 2.14 804G ECM Matrix Preparation ...... 36 2.15 Glucose Stimulated Insulin Secretion Assay ...... 37 2.16 Insulin ELISA ...... 37 2.17 Seahorse Metabolism Assay...... 38 2.18 Cell Lysate Collection and Quantitation ...... 39 2.19 Western Blot ...... 40 2.20 Chromatin Immunoprecipitation ...... 41 2.21 Gene Expression Studies ...... 42 2.22 RNA-Sequencing ...... 44 2.23 Bioinformatics Analysis ...... 44 2.24 Tissue Processing for Histology ...... 45 2.25 Immunostaining ...... 46 2.26 TSA Staining ...... 47 2.27 EdU Staining ...... 47 2.28 TUNEL Staining ...... 47 2.29 Morphometric Analysis ...... 48 2.30 Adenovirus Generation and Adenoviral Infection ...... 48 2.31 Luciferase Assay ...... 49

2.32 Gene Targeting in mPAC-L20 Cells ...... 49 2.33 Human Embryonic Stem Cell Culture ...... 50 2.34 Statistics ...... 51 3. SOX4 Is Required for Pancreatic Endocrine Cell Development ...... 56 3.1 Introduction ...... 56 3.2 Sox4 is Dynamically Expressed during Mouse Pancreas Formation ...... 57 3.3 Validation of SOX4 Mouse Models ...... 62 3.4 Sox4 Activates the Neurog3 Promoter and Induces its Expression ...... 67 3.5 Sox4 Regulates Endocrine Cell Differentiation Downstream of Neurog3 ...... 68 3.6 Loss of Sox4 Does Not Change the Number of NEUROG3 Lineage Cells ...... 69 3.7 Sox4 Regulates Factors Downstream of Neurogenin3 ...... 73 3.8 Discussion ...... 78 4. SOX4 Is Required for the Maintenance of Adult β-cell Population via Proliferation ..... 82 4.1 Introduction ...... 82 4.2 SOX4 Expression is the Highest of All SOX Family Members in Pancreatic Islets ...... 84 4.3 SOX4 Knockout on HFD Are Glucose Intolerance and Have Reduced Plasma Insulin .... 86 4.4 SOX4 is Required for β-cell Proliferation ...... 90 4.5 SOX4 Regulates the Cell-Cycle Inhibitor CDKN1A in β-cells to Modulate Proliferation 92 4.6 Discussion ...... 95 5. Mediator Subunit MED15 Is Required for Pancreatic Endocrine Cell Specification and a Functional β-cell State ...... 97 5.1 Introduction ...... 97 5.2 Characterization of MED15 Expression ...... 98 5.3 MED15 is Required for Pancreatic Endocrine Cell Specification ...... 100 5.4 MED15 is Required for Endocrine Specification But Not Pancreatic Progenitor Fate .... 102 5.5 MED15 is Required for β-cell Differentiation Following Endocrine Specification ...... 103 5.6 Loss of MED15 Results in Glucose Intolerance ...... 104 5.7 Identification of Gene Expression Changes in IM15KO ...... 105 5.8 Loss of MED15 Impacts β-cell Metabolism and Mitochondrial Function ...... 107 5.9 Discussion ...... 108 6. Conclusions ...... 114

vii

6.1 Research Summary ...... 114 6.2 Future Directions ...... 121 6.3 Final Thoughts...... 123 Bibliography ...... 124 Appendices ...... 144 Appendix A: IM15KO RNA-Seq Differentially Expressed Genes ...... 144 Appendix B: Taqman Low Density Array Gene Assay IDs ...... 146

viii

List of Tables

Table 1: Human and mouse embryonic pancreas development timeline...... 15 Table 2: Mouse strains ...... 32 Table 3: PCR primers...... 52 Table 4: Antibodies used ...... 53 Table 5: Chromatin immunoprecipitation buffers ...... 55 Table 6: Applied Biosystems, Inc. Taqman® low-density array gene ID ...... 146

List of Figures

Figure 1: Illustration of early pancreas formation...... 8 Figure 2: Diagram of Neurog3 induction and delamination from the epithelium...... 11 Figure 3: Diagram highlighting transcription factor and cell lineage relationships during pancreatogenesis...... 14 Figure 4: Model of hESC directed differentiation adapted from Rezania et al. (2014)...... 16 Figure 5: Illustration of Mediator complex interactions and Mediator complex modules...... 28 Figure 6: Sox4 is expressed during pancreas development and hESC differentiation...... 58 Figure 7: Sox4 is expressed in a subset of trunk epithelial cells...... 59 Figure 8: SOX4 expression becomes restricted to endocrine lineage...... 60 Figure 9: SOX4 protein expression is lost in knockout models...... 61 Figure 10: Endogenous SOX4 is induced by NEUROG3 in the mPAC model of endocrine development...... 63 Figure 11: Validation of Pdx1-Cre mediated recombination with lineage tracing...... 64 Figure 12: Pdx1-Cre; Sox4flox/flox knockout embryos (PS4KO) have defects in endocrine cell genesis...... 65 Figure 13: PS4KO mice have no changes in apoptosis or proliferation at E15.5 or E18.5...... 66 Figure 14: SOX4 directly regulates Neurog3 expression in the mPAC model of endocrine development...... 67 Figure 15: Mouse genomic region containing the Neurog3 gene and its upstream sequences. ... 68 Figure 16: Neurog3-Cre; Sox4flox/flox embryos (ES4KO) have defective endocrine cell formation...... 70 Figure 17: EdU labeling of cells in ES4KO mouse pancreas at E18.5 demonstrated no changes in cell proliferation following Sox4 ablation...... 71 Figure 18: Sox4 null endocrine progenitor cells remain in the endocrine lineage but do not differentiate into islet endocrine cells...... 72 Figure 19: Lineage tracing endocrine progenitor cells demonstrated the presence of hormone negative endocrine specified cells...... 73 Figure 20: Examination of lineage traced endocrine specified cells that express exocrine or ductal markers...... 74 Figure 21: SOX4 potentiates the induction of pro-β-cell genes during endocrine pancreas development...... 75 Figure 22: SOX4 cooperates with NEUROG3 to increase Pax4 and Neurod1 expression...... 76 Figure 23: Flow cytometry analysis of mTmG labeled EMS4KO cells...... 77 Figure 24: Model for SOX4 roles during pancreas development...... 78 Figure 25: Taqman gene expression profiling of Sox genes in isolated human (A) and mouse (B) islets...... 84 Figure 26: Sox4 is the highest expressed member of the SOX class of transcription factors...... 85 Figure 27: Validation of the SOX4 knockout model...... 86

Figure 28: βS4KO mice on chow diet had no significant differences in body weight or glucose tolerance over time...... 87 Figure 29: βS4KO mice on HFD have declining glucose tolerance over time...... 88 Figure 30: βS4KO mice at 30 weeks of age had similar insulin tolerance compared to littermate controls...... 89 Figure 31: βS4KO mice on HFD have reduced plasma insulin following glucose bolus despite functionally unimpaired islets...... 90 Figure 32: βS4KO have reduced β-cell mass...... 91 Figure 33: βS4KO have a reduction in β-cell proliferation rates...... 92 Figure 34: Taqman® low-density array analysis of cell-cycle genes in βS4KO islets following one week of HFD...... 94 Figure 35: Med15 expression in the pancreas...... 99 Figure 36: Validation of Med15 knockout models...... 100 Figure 37: Characterization of PM15KO mice...... 101 Figure 38: PM15KO mice have no differences in progenitor cell population...... 102 Figure 39: PM15KO mice have significantly reduced endocrine progenitor cells...... 103 Figure 40: NM15KO embryonic pancreases have reduced β-cell numbers versus controls...... 104 Figure 41: IM15KO mice are glucose intolerant despite similar number of β-cells...... 106 Figure 42: Analysis of differentially expressed genes in RNA-Seq data obtained from 8 week old IM15KO islets...... 109 Figure 43: Protein validation of IM15KO differentially expressed genes...... 110 Figure 44: Glucose-stimulated insulin secretion and cellular respiration profile of IM15KO islets...... 111 Figure 45: 174 differentially expressed genes from IM15KO RNA-Seq analysis by DESeq2. . 144

List of Abbreviations

CHGA Chomogranin A NM15KO Neurog3-Cre; Med15flox/flox ChIP Chomatin immunoprecipitation ONECUT1 One cut homeobox 1 CoIP Co-immunoprecipitation PAX4 Paired box 4 Cre Cre recombinase PDX1 Pancreatic and duodenal homeobox 1 βS4KO Pdx1-CreER; Sox4flox/flox PM15KO Pdx1-Cre; Med15flox/flox E Embryonic day Pol II RNA polymerase II EMS4KO Neurog3-Cre; mTmG; PPY Pancreatic polypeptide Sox4flox/flox ES4KO Neurog3-Cre; Sox4flox/flox PS4KO Pdx1-Cre; Sox4flox/flox FGF Fibroblast growth factor PTF1A Pancreas transcription factor 1 subunit α FLTP Flattop qPCR Quantitative polymerase chain reaction GCG Glucagon RF Red fluorescence GF Green fluorescence RNA-Seq RNA sequencing (e)GFP (enhanced) green fluorescence RFP Red fluorescence protein protein GHRL Ghrelin RFX6 Regulatory factor X, 6 GSIS Glucose-stimulated insulin SOX SRY-related high mobility group secretion box HA Haemagglutinin SRY Sex-determining region Y hESC Human embryonic stem cell SST Somatostatin HMG High mobility group STZ Streptozotocin IM15KO Ins1-Cre; Med15flox/flox TF Transcription factor INS Insulin TGF-β Transforming growth factor beta MAFA V-maf Avian WT Wild type musculoaponeurotic fibrosarcoma homologue A MED15 Mediator complex subunit 15 MIN6 Mouse insulinoma 6 mPAC Mouse pancreatic ductal adenocarcinoma mTmG Membrane-targeted tdTomato, membrane-targeted eGFP NEUROD1 Neurogenic differentiation 1 NEUROG3 Neurogenin 3 NKX2-2 NK homeobox 2 NKX6-1 NK homeobox 1

xii

Acknowledgements

The time spent during my doctoral degree is a period of my life that I will always cherish. This could not have been possible without the dedication and support of my mentor Dr. Francis Lynn. I am very fortunate to have had Francis take a chance on me as one of his first graduate students. He has guided by example, with hard-work and tenacity, and is one of the most intelligent people I know. Francis, through his genuine passion for science, has fostered a treasured environment for excellence and learning. I am deeply grateful for his patience, despite my stubbornness, and for always challenging me to reach my full potential. The work presented in this thesis could not have happened without the support of many other mentors I have had along my journey. I would like to thank Dr. Pamela Hoodless for her continued support as a mentor, but also for supporting me, during a time when I had lost confidence in myself academically. I would also like to thank Dr. Jim Johnson and Dr. Bruce Verchere for their mentorship both during committee meetings, and outside. I am fortunate to have had their advice on science and life throughout my degree. My degree has been one of the best times of my life because of the people that surrounded me on a daily basis, Nicole, Paul, Thilo, Cuilan, Mei. I also want to thank Sigrid, Stephanie, Jaques, Alex, Samantha, and the many other graduate students who have befriended me along the way. They are some of the smartest people I know, but that is not why they are unforgettable. In the end, I will always value the way they made me feel, as a friend and colleague. Their friendship has gotten me through both good and bad times. I cannot begin to fathom what a future would be like without beer Fridays, three-way conversations about our random dreams and chicken bombings, or our lab hikes. However, I am confident that I will continue to find random skittles well into the future. I am grateful to be part of the CFRI and am thankful for the support from all the labs, core services, and diabetes research groups in Vancouver. I am fortunate to be amongst a group of scientists such as Drs. Lynn, Verchere, Johnson, Kieffer, Lefebvre, and many others who have made a commitment to improve the science community by organizing conferences, workshops, and student research days. This helped me to discover a passion for scientific outreach and student mentorship. I have a special thanks to students Sam, Sara, Vivian, Christina, and Sujan that I co-mentored with Francis. They worked hard to contribute to the success of many projects, and who taught me how to be a better scientist. I would also like to thank my family and friends. Although my parents rarely say it, their actions demonstrate the depth of their love for me. They have nurtured a curiosity for the world and a desire to contribute positively to it. I hope my actions will always make them proud. Kevin, thanks for putting up with me without complaining all these years, we have had some good times. To Andrew and Ryan, your friendships have brought me many years of memories and laughter, and have truly been a foundation for me. Finally, to Nancy, you have always encouraged, listened, and supported me, despite my general grumpiness and occasional melt- down. I once thought I knew what it was to be kind and generous until I met you. I am blessed to have such wonderful people in my life.

xiii

To my parents and grandparents for their love and support.

I appreciate their sacrifices,

so that I would have the opportunity for a better life.

xiv

1. Introduction

1.1 Diabetes Mellitus

Diabetes mellitus is estimated to affect 415 million individuals worldwide as of 2015, projected

to increase to 642 million individuals by 20401. Diabetes mellitus is categorized as a metabolic

disease with symptoms including polyuria, polydipsia, weight loss, and fatigue. A hallmark of

the disease is chronic elevated hyperglycemia that can lead to retinopathy, nephropathy,

neuropathy, and cardiovascular diseases2. In addition to the 5 million diabetes-related deaths

worldwide in 2015, many more individuals are burdened with a reduced quality of life and

complications from the disease. In Canada alone, over 2 million individuals live with diabetes3

with direct drug costs to individuals and indirect costs to the healthcare system through increased

system usage4.

1.2 A Brief History of Diabetes Research

The earliest characterization of a disease with symptoms similar to diabetes mellitus was over

2000 years ago describing polyuria and polydipsia5. Despite the longstanding awareness of the

existence of such a disease, the development of a long-term treatment option would not exist for another two millennia.

The pancreas is comprised of three major tissue types with two major functions. Exocrine cells

of the pancreas are responsible for the release of bicarbonate, digestive enzymes, and lipases into

acini. These are drained by the ductal tissue that channels these products into the duodenum

during food digestion. Finally, endocrine cells occur in clusters known as islets of Langerhans,

comprised of alpha (α), beta (β), delta (δ), and pancreatic polypeptide (PP) cells. Building upon

work demonstrating that a substance in the pancreas was responsible for lowering blood glucose

levels, Banting, Best, Collip, and Macleod would succeed in obtaining relatively pure pancreatic

extracts containing insulin, successfully treating patients with diabetes in 19226. Further work

would go on to show that β-cells are the primary insulin producing cells in the body, cementing

the importance of insulin and these cells in relation to diabetes.

1.3 Systemic Insulin Signaling and Insulin Secretion

Central to diabetes is the dysregulation of blood glucose levels, tightly controlled to levels

between 4-8 mM in humans. Under fasting conditions, β-cells exocytose insulin granules at a

very low oscillatory rate7. Following a meal, elevations in blood glucose levels result in glucose

uptake into β-cells through the glucose transporter GLUT2 in mice and GLUT1 in humans8,9.

These glucose transporters possess a high Km around 11 mM glucose for GLUT2, reflecting the

half-maximal insulin exocytosis rate observed in mouse islets. Within β-cells, glycolysis coupled

with cellular respiration shifts the intracellular ATP/ADP ratio. ATP binding to the ATP

sensitive potassium channels (KATP) leads to channel closure and membrane depolarization. This

causes voltage-gated calcium channel activation and calcium influx into the cell. In turn, this

results in insulin granule exocytosis through calcium-responsive insulin vesicle associated

proteins10. As a peptide hormone, insulin secreted into the bloodstream acts upon target tissues

including skeletal muscle, liver, adipose, and the central nervous system to modulate the

systemic response to elevated glucose levels11. Insulin binding to insulin receptor tyrosine kinase

at the cell-surface leads to phosphorylation of its substrate adaptors such as the IRS family of

proteins. Insulin signaling primarily regulates glucose levels by increasing the surface expression

of glucose transporter GLUT4 at the plasma membrane, leading to increased glucose uptake11.

Clamp studies have also demonstrated that approximately 80% of postprandial glucose uptake is accounted for by skeletal muscles12. In other tissues however, insulin may play additional roles

such as the negative regulation of hepatic gluconeogenesis and the promotion of glycogen synthesis13. Studies utilizing germline knockouts of the insulin receptor, or its receptor substrates

IRS1 and IRS2, displayed insulin resistance and hyperglycemia14–17. These studies demonstrate

that insulin signaling is critical for glucose homeostasis and that loss of insulin signaling leads to

diabetes.

1.4 Diagnosis of Diabetes

Clinically, diabetes mellitus is diagnosed based on blood glucose levels, with a fasting level ≥

7.0 mmol/L or glucose levels >11.1 mmol/L following a two hour oral challenge with 75g

glucose. Glycated hemoglobin (HbA1C) in the blood is also utilized as a marker of chronic

glycaemia, reflecting average blood glucose levels over 60-90 days. This is utilized for

diagnosis, with a cut-off level of ≥6.5% defining individuals with diabetes18. While these tests

are able to identify individuals with diabetes, the aetiologies of diabetes are wide ranging.

1.5 Aetiology of Diabetes Mellitus

Major categories of diabetes include type 1 diabetes, type 2 diabetes, maturity-onset of the young

(MODY), gestational diabetes, and transient & permanent neonatal diabetes19. Type 1 diabetes is

characterized by an autoimmune reaction against β-cells defined by the presence of auto-

antibodies against self-antigens such as insulin, glutamic acid decarboxylase 65 (GAD65), and

islet antigen 2 (IA-2)20. Subsequently, islets undergo infiltration by macrophages and

lymphocytes, leading to immune-mediated destruction of β-cells, leaving behind other endocrine

cells of the islet. This results in systemic insulin deficiency that leads to chronic

hyperglycemia21. Type 1 diabetes has a genetic component, with genome-wide association

studies (GWAS) identifying over 50 loci to date increasing risk for type 1 diabetes22. Several of these loci are in the HLA genomic region, encoding for polymorphic antigen-presenting proteins.

Non-HLA region associated genes include INS, PTPN22, and CTLA-4. Insulin is a primary autoantigen in type 1 diabetes, while PTPN22 and CTLA-4 are believed to prevent T-cell activation. Further, environmental factors such as viral infections are postulated to be triggering factors for type 1 diabetes23.

Type 2 diabetes also has a demonstrated a heritable genetic component, with GWAS studies

identifying multiple loci24. Several associated genes are involved in the cell cycle including

CDKN2A/B, CDC123, and CDKN1C. Interestingly, some genes encoding for TFs found during

β-cell development are associated with type 2 diabetes risk. For example, both HNF1A/B have

been highlighted by GWAS studies, while another SNP was initially linked to CDKAL124,25.

However, later work demonstrated that the SNP loci most likely regulated Sox4 expression. Type

2 diabetes is also associated with obesity, insulin resistance, and insulin deficiency. One unifying

theory postulates that nutrient surplus leads to obesity and insulin resistance. Obesity is observed

in most patients with type 2 diabetes26. In addition, over 90% of individuals with type 2 diabetes

also display insulin resistance in tissues such as the liver and skeletal muscles, where the

majority of glucose uptake occurs27,28. High-fat diet and obesity are linked to changes in adipokine signaling from adipose tissue, now understood to be a major endocrine organ27,29,30.

This may contribute to peripheral insulin resistance, which is associated with compensatory increases in β-cell mass and hyperinsulinemia31. However, this relationship may not be as

straightforward, as hyperinsulinemia has also been demonstrated to precede diet-induced obesity

and changes in circulating fatty acids32. Furthermore, not all obese individuals develop diabetes.

Genetic or other environmental factors may also increase susceptibility to higher glucose and

lipid concentrations, associated with obesity and insulin resistance, and is postulated to

contribute to β-cell failure33. Additionally, β-cell ER-stress driven by glucolipotoxicity and

insulin demand may also drive the cells towards dysfunction or death34,35. Long-term outcomes of these interactions in the setting of insulin resistance may lead to the loss of functional β-cell mass and insulin insufficiency, leading to diabetes.

In MODY and neonatal diabetes, heritable genetic mutations are the primary cause of diabetes.

Both forms are commonly characterized by autosomal dominant mutations in genes that have roles in the insulin secretion pathway, leading to impaired insulin secretion2. Examples include

glucokinase (GCK), glucose transporter SLC2A2 (GLUT2), and potassium voltage-gated channel

(KCNJ11). Mutations for transcription factors (TFs) are prominent amongst currently identified

genes involved in permanent neonatal diabetes and MODY, suggesting their importance the

formation and function of pancreatic β-cells. Examples include PDX1 and PTF1A, with

homozygous loss of function mutations in humans resulting in pancreas agenesis36,37.

NEUROG3, the master regulator of pancreatic endocrine cell development in rodents, was also

reported to result in the loss of enteroendocrine cell resulting in congenital diarrhea. However,

new evidence associates NEUROG3 mutations with neonatal diabetes and pancreatic endocrine

cell development, based on patient case-study and loss of function study in human embryonic

stem cell (hESC) differentiations in-vitro38,39. Finally, many other TFs such as hepatocyte

nuclear factor 4 HNF4A, neuronal differentiation 1 NEUROD1, and NK2 homeobox 2 NKX2-2

have also been identified as risk variants in both MODY and neonatal diabetes19.

1.6 Current Treatments for Diabetes Mellitus

Depending on the aetiology of diabetes, multiple treatment options are available for individuals.

Effective treatment aims to reduce HbA1C levels to <6.5% while reducing both hyper- and

hypoglycemia defined as blood glucose levels >7 mM or <4 mM, respectively2. Exogenous

insulin is an absolute requirement for both type 1 diabetes and specific cases of MODY as β-cell

insulin secretion is lost. Weight loss and increased exercise may be used at first for individuals with impaired glucose tolerance (blood glucose between 6.0-6.5 mmol/L)18. Administration of

metformin, which decreases hepatic gluconeogenesis, is the first line of drug treatment for type 2

diabetes in Canada. A second oral anti-diabetic agent may be included to achieve optimal blood glucose control. Examples include DPP-4 inhibitors and GLP1R agonists; a class of drugs that target the incretin signaling pathway that acts upon β-cells to amplify glucose-stimulated insulin exocytosis40. Sulfonylureas are another class of oral drugs that act on the SUR1 subunit of the

potassium channel to induce closure and insulin secretion. Interestingly, a subset of MODY or

neonatal diabetes with mutations in HNF1A, HNF4A, and KATP channel mutations exhibit good

response to sulfonylurea treatment41. Finally, exogenous insulin will eventually be required for

individuals with type 2 diabetes due to the progressive decline in β-cell mass. There is a unifying

theme of β-cell dysfunction and death in diabetes, leading to insulin insufficiency. Therefore,

strategies to replenish or replace the natural β-cell population offer an attractive alternative to the

daily and lifelong dependency on insulin or drugs.

Islet transplantation represents one cell-based approach for the long-term treatment of diabetes42.

This strategy is limited by the need for immunosuppression, donor tissue supply, and the loss of efficacy of transplanted islets over time43. Alternative treatments in development include:

promoting the survival and replication of pre-existing β-cells44, generating β-cells from

pancreatic and alternative tissues45, as well as expanding and differentiating β-cells from stem-

cell sources46,47. These have been developed based on the understanding of β-cell failure, and

insulin insufficiency; as well as the major mechanisms for β-cell formation and maintenance.

Central to our understanding of these topics is the regulation of transcription: variations in gene

expression are integral to the formation and function of β-cells, as well as the changes that may

occur during the course of diabetes. As the theme of this thesis, a brief overview of

transcriptional regulation during β-cell development and maintenance is explored below.

1.6 Early Pancreas Budding and Development

The pancreas is formed from a dorsal and a ventral bud, which arise from the posterior foregut in the developing embryo as summarized in Fig. 1. Instructive cues from adjacent tissues are required for the specification of pancreatic tissue. Prior to budding of the pancreas, around embryonic day (E)8.5, retinoic acid (RA) signaling provides a posteriorizing signal for the gut tube, and absence of RA leads to anteriorization and loss of the dorsal pancreas48. Studies removing the notochord have also demonstrated the loss of dorsal pancreatic bud development, as activin and FGF2 signaling are required to inhibit SHH to permit bud induction49,50.

Additionally, VEGF signaling by the dorsal aorta has been identified in vitro to be critical for

dorsal pancreatic bud specification51 (Fig. 1B). In contrast, the ventral pancreatic bud receives different cues compared to the dorsal bud. Co-culture experiments in vitro have demonstrated that foregut adjacent to the cardiac mesoderm were induced to a liver cell fate, potentially via

FGF2, while pancreas specification remained the default fate52. In summary, the integration of

these signals results in the combinatorial expression of TFs in the prospective pancreatic region.

However, it is incompletely understood how these signals are coupled to the expression of

pancreas specifying TFs.

Figure 1: Illustration of early pancreas formation. (A) E8.5: Notochord (purple) is required for dorsal pancreas bud induction. Notochord release of activin and FGF2 suppress Shh in the foregut tube and permits pancreas specification. (B) E9.0: Dorsal aorta forms below the notochord separating it from the posterior foregut. VEGF released by the dorsal aorta and vitelline veins contribute to pancreas induction. (C) E9.5: Inductive cues specify cells to become pancreatic progenitors with FGF10 released by the mesenchyme supporting cellular proliferation. The buds are comprised of a stratified epithelium. (D) E12.5: Branching morphogenesis generates a ductal epithelium with tip and trunk cells denoted by unique TF combinations.

Initial specification of the pancreas from the posterior foregut region occurs prior to morphological changes, as suggested by early explant studies53. Combinatorial signaling induces

a specific genetic pattern of TFs54 that define the nascent pancreas by approximately E9.0.

Expression of Hlxb9, Pdx1, Ptf1a, GATA4/6 and Sox9 have been demonstrated to be important

for the proper induction of the pancreas, with gene knockouts displaying pancreas agenesis and

stalled pancreatic progenitor development55–61. Other TFs are also present including Nkx6-1,

Hes1, and FoxA1/2 but are not an absolute requirement for pancreatic bud formation62–64. The

pancreatic buds then expand into the surrounding mesenchyme tissue: growing from a single-

layer epithelium to a transiently stratified bud65. At this time, around E9.5, the dorsal bud is

comprised mainly of pancreatic progenitor cells (Fig 1C). These cells retain the capacity to form

all subsequent lineages of the pancreas: ductal, exocrine, and endocrine. During this time, the

pancreas undergoes primary transition between E9.5-10.5, defined by the first expression of

endocrine and exocrine proteins66. Following initial bud formation, the pancreatic epithelium

proliferates and the stratified cells undergo polarization to form a network of epithelial tubes65,67.

Growth of the epithelium is supported in part by FGF10 signaling from the mesenchyme. Fgf10-/-

mice exhibited reduced pancreatic growth68. FGF10 has been demonstrated to be expressed in

the mesenchyme, as mesenchyme separation experiments demonstrated pancreatic hypoplasia69.

Within the pancreatic buds, TF expression of Sox9 appear to be important for this signaling as it

is required for the expression of the Fgfr2 receptor to mediate FGF signaling70. Finally, HES1

signaling in the pancreatic buds also appeared to be important for early pancreas development,

supporting cellular proliferation. Hes1 loss of function studies led to increased expression of the

cell-cycle inhibitor p57, which reduced cellular proliferation by preventing cell-cycling71.

1.7 Tip and Trunk Formation

As pancreas epithelium continues to expand, the pancreatic epithelium undergoes branching morphogenesis. This process is the remodelling of the epithelial tube network giving rise to an interconnected ductal network or “plexus”65. Proliferation together with continued remodeling resolves into the formation of single lumen branches that have distinct “tip” and “trunk” domains65,72,73. These cellular compartments have a unique signature of TFs with Ptf1a, c-Myc,

and Cpa1 representing the “tip” domain, while Sox9, Onecut-1, and Nkx6-1 defining the “trunk”

domain74–76 (Fig. 1D). Lineage tracing experiments have highlighted that tip cells are highly

proliferative multipotent progenitor cells (MPCs), contributing to epithelial outgrowth, and will

terminally differentiate to exocrine cells by E1475. One the other hand, trunk cells are bipotential. Activation of the master regulator neurogenin 3 (Neurog3) within a subset of trunk- cells will lead to the specification of endocrine cell fate and delamination from the epithelium, while non-induced cells will contribute towards the ductal system of the pancreas77,78. In the

epithelium, transcriptional regulation is important for defining these regions. For example, loss

of function of Nkx6-1 and Nkx6-2 expanded the domain of PTF1A expressing epithelial cells

while gain of function experiments demonstrated increased NEUROG3 positive (+) cells at the

expense of PTF1A expression. Conversely, mis-expression of Ptf1a reduced the number of

NEUROG3+ cells76. These experiments support the role of TFs in establishing tip and trunk

domains, with trunk specification demarcating the first steps of lineage commitment of

pancreatic progenitor cells (Fig. 3).

1.8 Endocrine Cell Specification

As branching morphogenesis continues, through E12.5-E14.5, remodeling of the web-like ductal

plexus combined with branching and outgrowth of the tip regions leads to the three dimensional

arborisation of the organ65,72,73. During the secondary transition of the pancreas, between E13.5-

E16.5, a large wave of exocrine and endocrine differentiation occurs. Although exocrine cells

derive from tip cells, endocrine differentiation begins with the induction of NEUROG3 from the

trunk. This can first be detected in the pancreas as early as E9.5, where rare glucagon- and

insulin-positive cells can also be found54. Within the epithelium, although active Notch signaling

via HES1 is important for early epithelial proliferation, it is believed to prevent expression of

NEUROG3. Observations of mice deficient for the Notch effector Delta-like gene 1 (Dll1)

revealed an increase in endocrine cells79. Stochastic Notch signaling is believed to allow a subset

of trunk cells to activate NEUROG3 expression80,81, and gene dosage studies utilizing Neurog3 heterozygotes or hypomorph alleles have also demonstrated a high-dose requirement for endocrine specification82. Additionally, induction of NEUROG3 also appears to occur rapidly, as

NEUROG3-reporter protein experiments have shown rapid Neurog3 induction followed by

downregulated within 12 hours of initial expression83. Successful NEUROG3 expression is

followed by delamination and endocrine commitment77 (Fig. 2). The majority of NEUROG3

induction and endocrine specification occurs from E13.5-E15.5. Interestingly, inducible

NEUROG3 pulse chase expression experiments demonstrated windows of specific endocrine

subtype competency during pancreas development. Glucagon-positive cells were predominantly

induced at early stages, followed by insulin, somatostatin, and pancreatic polypeptide84,

suggesting additional cues are present in trunk cells84.

Figure 2: Diagram of Neurog3 induction and delamination from the epithelium. Trunk cells within the pancreatic epithelium have stochastic Neurog3 induction due to low local notch signaling. This leads to delamination and activation of endocrine commitment. These cells cluster away from the ductal epithelium and eventually form islets of Langerhans.

1.9 Endocrine Cell Differentiation

Following transient NEUROG3 expression, a sequential cascade of TF expression is necessary to specify individual endocrine cell types (Fig. 3). This cascade is believed to begin with expression of TFs important for permitting the endocrine progenitor cells to differentiate into mono- hormonal cell-types. Insm1, Isl1, NeuroD1, Pax6, and Rfx6 are TFs believed to be direct targets

of NEUROG3 and are expressed downstream85. Knockout studies of Insm1, Isl1, NeuroD1, and

Pax6 displayed reductions of all endocrine cell types. Knockout of Rfx6 had a slightly different

phenotype with retention of endocrine markers chromograinin A (CHGA) and synaptophysin

(SYN), but a loss of hormone staining with the exception of pancreatic polypeptide77,86–93.

Additionally, NKX2-2 and NKX6-1 are early markers of the pancreatic epithelium with

expression starting around E8.7554 that becomes restricted to centrally located epithelial cells.

Following NEUROG3 activation, NKX2-2 is required for the differentiation of alpha, beta, and

PP-cell populations94. Mice lacking Nkx2-2 exhibited normal islet size, although they had

reduced hormone staining and expanded numbers of ghrelin-producing epsilon (ε) cells95.

Finally, the Nkx2-2 knockout had reductions in Pax6, Hlxb9, Pdx1, and Nkx6-1 expression

demonstrating that these factors are induced downstream64,96.

Downstream of the initial group of TFs induced by NEUROG3, additional TF networks produce

finer delineations in endocrine cell subtypes. For instance, Neurog3-Cre; Nkx6-1f/- conditional

mutants had a loss of β-cells while other endocrine cell types increased, without changes in

proliferation or survival. Neurog3-Cre; Nkx6-1 overexpression had the opposite phenotype with

increased β-cell numbers and reductions in other endocrine types97. As another example, TFs that balances the fate of endocrine subtypes, Pax4-/- mice had an increase in α-cell at the expense of

β- and δ-cells98. Complementary experiments with Arx-deficient mice exhibited the opposite phenotype, with an increase in β- and δ-cells while α-cell numbers were reduced99,100. These observations again suggest that TFs act in concert to define specific endocrine subtypes, with

Pax4 postulated to specify a bi-potential endocrine precursor to β or δ-cells. Pax4-deficient pancreas had normal expression of Isl1, Nkx2-2, and Pax6, suggesting that it operates simultaneously or downstream of these factors. However, Pax4 deletion lacked the expression of

Hlxb9 and Pdx1, suggesting that these genes are required further downstream in endocrine cell

development96.

Finally, β-cell specification involves additional TFs such as MafB/A, Pdx1, and Nkx6-1. For instance, MafB is expressed predominantly in α- and β-cells. MafB knockout pancreas exhibited no changes in early TFs such as Pax6, Isl1, and NeuroD1 or in genes related to cell lineage commitment such as Arx and Pax4101,102. However, germline deletion of MafB resulted in

decreased GCG and INS+ cell types. In mice, MafB expression stimulates the expression of

MafA, important for β-cell maturation, insulin gene transcription, and glucose stimulated insulin

responsiveness103–105. Other TFs such as PDX1, NEUROD1, and FOXA2 also appear to regulate

aspects of terminal β-cell differentiation; controlling insulin expression, glucose-stimulated

insulin secretion (GSIS), and insulin granule trafficking55,104,106,107.

In summary, from early patterning to endocrine cell differentiation, external signals from soluble

factors induce combinatorial TF expression to establish the pancreatic domain. As development

proceeds, lineage commitment events increasingly restrict cellular differentiation potential. TFs

regulate gene expression, with Neurog3 and Ptf1a as examples of specific gene-dose

requirements for TFs. Loss of function studies highlight a balance between multiple TFs, with

Nkx2-2 and Rfx6 knockout pancreas displaying altered endocrine cell numbers. Secondly, during

pancreas formation, sequential TF expression is important for endocrine lineage determination,

with loss of specific TFs compromising downstream TF expression and ultimately the

differentiation of cells. Finally, it appears that coordinated TF expression establishes cooperative

and antagonistic TF interactions, helping to setup the correct distribution of cell populations in

the pancreas. These will be major recurring themes for the TFs studied in this thesis.

Figure 3: Diagram highlighting transcription factor and cell lineage relationships during pancreatogenesis. Pancreatic progenitors express a combination of TFs that are instructive for the pancreatic fate. They are highly proliferative and give rise to multipotent tip-cells that have the capacity to form acinar, ductal, or endocrine cells. Tip cells have a unique TF signature and this domain is restricted from trunk-cells through mutual repression via PTF1A and NKX6.1. However, tip cell proliferation combined with tubulogenesis contributes to the trunk cell population as the epithelium expands. Trunk cells are bi-potent and may be fated to either a ductal cell lineage, or the endocrine lineage upon Neurog3 activation. High levels of NEUROG3 activate the expression of TFs required for all endocrine lineages, which then establishes sequential TF activation important for specific endocrine cell fate delineations.

1.10 Human pancreas development

Despite disparities in overall gestation length, islet architecture, and transcription factors

involved, the mouse model has identified many features conserved in humans. Mouse embryonic

pancreas development spans approximately 12 days in-utero, while human pancreas

development begins from around 4-5 weeks post-conception until birth at around 40 weeks of age. Utilizing Carnegie stages (CS) of embryonic development to allow for comparisons between organisms, human pancreas budding occurs around CS13, similar to the E9.5 stage of mouse development (Table 1). TFs such as PDX1 and PTF1A have also been identified in human pancreatic progenitor cells, and clinical studies of individuals with mutations in these genes demonstrated pancreas agenesis, suggesting a similar role in organ specification36,37. However, differences exist, with immunostaining demonstrating SOX9 expression in tip-cells in humans,

versus its absence in mouse pancreatic tip-cells108. Finally, the sequence of TF expression is

similar, with detection of NEUROG3 followed by the observation of β-cells at around 10 weeks

post-conception, or CS21 which corresponds to approximately E15 in mice109. These observations of TF expression patterns in human fetal sections as well in directed hESC

differentiation46,47,110 (Fig. 4), suggest that mouse gene knockout studies of TFs during pancreas

development is useful in modeling humans β-cell development.

Table 1: Human and mouse embryonic pancreas development timeline. Carnegie stages (CS) and their approximate mouse developmental timepoints are shown together with shared features of pancreas development.

Human Mouse development development Pancreatic features CS12 E9.0 Detection of PDX1+ expression in posterior foregut CS13 E9.5-E10.0 Dorsal and ventral pancreas budding CS15 E11 Expansion of pancreatic buds CS20 E14-15 Tip and trunk domains detectable via immunostaining CS21 E15.5 Detection of NEUROG3+ cells and INS+ cells

Figure 4: Model of hESC directed differentiation adapted from Rezania et al. (2014). Summary of the multi-stage differentiation protocol with the number of days in culture as well as the major small molecules added per stage. Markers of differentiation for specific tissues are indicated below the cell-types following each stage, demonstrating a similar sequence of expression compared to mouse models.

1.10 β-cell Maturation and Identity

Following the formation of a nominal β-cell mass during development, several lines of evidence

suggest differences in early postnatal islet function compared to adult β-cells. Rodent and human

studies have demonstrated reduced islet glucose stimulated insulin secretion of β-cells within the

first 21 days following parturition, despite the presence of glucose transporter, insulin, and the

exocytotic machinery111–113. Newborn islets also exhibit higher basal insulin secretion with

reduced β-cell membrane potential and potassium channel depolarization response to

114–116 secretagogues . Other evidence include reduced KATP channel conductance in patch-clamp experiments and lower expression of TFs such as PDX1 and MAFA that are important for insulin expression104,117. The cause for functional maturation has not been fully characterized,

although several studies have suggested that a shift in the nutrient environment may change the

transcriptional program of β-cells118–120. Microarray and gene expression studies have also

revealed increases in glucose metabolic genes such as the high Km hexokinase and PFK1, while

the expression of Ldha and Mct1 are reduced121–123. It is believed that these changes allow β-cells

to better couple glucose influx to ATP generation and cellular depolarization at higher glucose

concentrations. In sum, β-cell status is dynamic following differentiation. In newborns, changes

in the expression of metabolic genes are correlated with improved glucose-stimulated insulin

secretion and a shift in major nutrient composition.

1.11 Regulation of β-cell Identity and Function

Terminally differentiated cells express genes specific to their function and are restricted to one

cell fate under normal conditions. Initially derived from pluripotent cells, they lose the

competency to adopt alternate cell fates even when transferred to conditions instructive for other

cell fates124,125. This increasing loss of differentiation potential suggests the adoption of the

lowest free energy state as the simplest explanation for the control of cell-fate. Hans Spemann

proposed a “fantastical” experiment in 1938 to replace the nucleus of an unfertilized egg with

one from a differentiated cell to address whether cells were truly restricted. The advent of

somatic cell nuclear transfer would follow through with this experiment and demonstrate the

capacity to reprogram a terminally differentiated cell into a pluripotent state126,127. Significant

research in stem-cells and the generation of induced pluripotent cells would show that cell states

require both active positive and negative control for lineage induction as well as maintenance of

a terminal cell state128–130. Similarly for β-cells, evidence point towards both active and passive control of β-cell fate. Chromatin immunoprecipitation and sequencing combined with gene expression data demonstrate passive control via chromatin modifications131,132. Furthermore, loss

of function experiments demonstrated the requirement for active TF expression for adult β-cell

function. Conditional knockout of the Pdx1 gene with Cre recombinase expressed via the rat

insulin promoter resulted in the loss of Glut2, MafA, and Nkx6-1, terminal β-cell markers, while

α-cell genes became mis-expressed133. Other examples of active TF regulation include the Nkx6-

1 knockout where β-cell identity is destabilized through the loss of genes important for β-cell

function134, as well as MafA knockout which resulted in compromised glucose-stimulated insulin

secretion135. In a seminal paper by Talchai et al. (2012), β-cell conditional deletion of FoxO1 in

two mouse models of β-cell stress led to the loss of identity, termed “dedifferentiation”. Despite

the loss of insulin, lineage tracing revealed that these cells mis-expressed genes found during

earlier developmental timeframes, and were able to shift their gene expression profile to that of other endocrine cell types136. In vivo, analysis of the db/db and Glut4-Cre driven insulin receptor

knockout models of diabetes also supported the concept identity loss, demonstrating increases in

CHGA-positive, insulin-negative cells136. Finally, the db/db mouse model was also characterized

to have reductions in Neurod1 and Nkx6-1, and genes important for β-cell function such as Glut2

and the potassium channel subunit Kir6-2137.

Changes to extracellular conditions have also been demonstrated to alter β-cell identity, lending support to the idea of β-cell plasticity. Long-term culture of the betaTC-6 cell-line in 11.1mM glucose led to decreased insulin gene expression and content compared to 0.8 mM glucose culture138. MIN6 cells labelled with a dual-reporter for PDX1 and INS1 also demonstrated

decreased insulin secretion and expression of genes associated with a mature β-cell state upon

addition of activin A139. Additionally, there appear to be more dedifferentiated β-cells in sections in samples from patients with type 2 diabetes, correlating the glucolipotoxic milieu with changes in β-cell identity140. Together, these studies support the role of TFs in reinforcing β-cell function,

while sufficient levels of stress on β-cells is able to destabilize the β-cell state leading to loss of

differentiation, as observed in some models of diabetes.

Active maintenance may confer continuity and stability for β-cell identity, through the action of

TF auto-regulatory networks. During development, expression of Nkx6-1 antagonizes Arx and reinforces β-cell formation. Pdx1 increases Nkx6-1 expression during development55, while

Nkx6-1 prevents Arx repression of Pdx197, permitting continued Pdx1 expression. Continuity of these interactions provides a stable link from developing endocrine cells to mature β-cells.

Further, in β-cell lines, Pdx1 expression also controls Neurod1 and MafA expression, that can positively feedback on Pdx1133,141. Under stress conditions, this may be stabilizing as Pdx1

expression is maintained during the unfolded protein response142. β-cell dedifferentiation may

also be a natural response to stress; Alleviation of glucolipotoxicity or ER-stress can restore β-

cell identity143,144.

Importantly, the β-cell population is heterogeneous in both functional output and gene

expression. Studies have demonstrated varying levels of insulin expression with transitions from

low to high levels of insulin over time130,145. These differential cellular states have been

postulated to vary with environmental conditions or perhaps due to differences in individual β-

cell age. In addition, the marker Fltp has been identified to be enriched in mature β-cells. Cells with low expression of Fltp may represent immature β-cells that have higher proliferation rates within the same islet, providing additional support for β-cell heterogeneity146. Intriguingly, a recent publication has characterized the presence of “hub” β-cells within an islet. These cells

have a pacemaker like function for glucose stimulated insulin secretion. Experimental disruption

of single hub-cells decreased total islet insulin secretion, suggesting functional roles for

heterogeneity within islets147.

1.12 β-cell Differentiation Strategies

The understanding of early patterning has been exploited for the directed differentiation of

pluripotent cells to functional β-cells, ultimately for transplantation and treatment of

diabetes47,148–151. Environmental manipulation coaxes human embryonic stem cells into an

endocrine lineage by mimicking the sequential exposure to signaling molecules a stem cell

receives. For example, the addition of activin to culture mimics initial Nodal signaling, and achieved >80% definitive endoderm specification, as characterized by gene expression149.

Improved protocols now demonstrate the correct sequence of cellular markers with induction of

Neurog3, and finally insulin, MafA, and Nkx6-1 expression. In vitro however, the final cell population still does not represent adult β-cells in terms of functional output. Similar to newborn

β-cells, derived cells show evidence of high basal insulin secretion, and low levels of glucose stimulated insulin secretion46. However, once transplanted into an in-vivo environment, these

cells acquire increased terminal β-cell gene expression as well as functional competency. This

suggests that yet unknown processes continue to shape β-cell differentiation47,150. These may

include exposure to oscillatory glucose concentrations as discussed, circulating factors, or

perhaps changes in architecture151.

Another approach utilizes combinations of TFs found in β-cell development, delivered via

adenovirus to directly reprogram cells into pancreatic β-cells45. In that study, combined MafA,

Neurog3, and Pdx1 delivery was able to induce the expression of several markers of mature β-

cells, with expression of INS and GLUT2. Functionally, following streptozotocin (STZ) ablation

of β-cells, adenoviral transduction of TFs led to a small increase in plasma insulin levels.

Follow-up studies by the Zhou group showed long-term survival of reprogrammed β-cell

clusters. These cells display improved GSIS over several months and increased β-cell terminal

gene expression152. Other studies have also shown partial reprogramming from a wide array of other tissues using the transduction of several TFs153. One caveat of this approach is the correct

viral dose, with hypoglycemic events observed up to 60 days post treatment. An alternative

strategy that may overcome this include modifying the factors to enable reprogramming into

alpha and delta-cells to increase blood glucose and inhibit insulin release, respectively154.

1.13 Maintenance of Adult β-cell Mass

Following differentiation, β-cell numbers in the adult pancreas are determined through a

dynamic equilibrium between proliferation and cellular apoptosis. The initial population is

determined by endocrine cell differentiation during development. Reduced differentiation may

lead to diabetes and failure to thrive as observed in TF knockouts86,94. In the Goto-Kakizaki model of type 2 diabetes, β-cell mass is approximately 30% that of age-matched controls, and is believed to be a primary contributor towards later diabetes development155. Neurog3 reporter

experiments have elucidated that endocrine commitment terminates following birth83, while

lineage tracing experiments in adults have determined that β-cells arise from pre-established β-

cells156,157. In rodents, β-cells appear to have a high proliferation rate in first few weeks

following birth that exponentially decays to less than 1% per day by 3 months of age158. Utilizing

both adult pancreatic sections and pulsed atmospheric radiation decay from atomic bomb testing,

human β-cell proliferation was estimated to occur only within the first 30 years of life159.

However, it is unclear whether proliferation ceases in late-stages of life. Conversely, apoptosis

has also been mathematically modelled and experimentally demonstrated to be high in the weeks

following birth. This was attributed to islet architecture remodelling, and stabilized to a generally

constant rate over time160. Given these variables and the estimated number of pancreatic β-cells

present through adulthood, Finegood et al. estimated individual β-cell lifespan to be between 30-

90 days in rodents160. This has also been challenged as data from Kushner and colleagues using long-term BrDU labeling, have estimated that replication rates in mice approaches zero after

approximately a year, suggesting a long lifespan for β-cells as rodents can live for up to 3 years

of age161.

The concept of β-cell differentiation has also been explored in the adult pancreas. Initial

characterizations noted the close proximity of islets to ductal systems within the pancreas and of

sporadic β-cells within or closely associated with duct, suggesting that a potential adult

multipotent cell population could exist within the pancreas that contributed to the replenishment

of β-cell mass162. This is difficult to consolidate with recent lineage tracing experiments that

have labeled either the ductal or exocrine cellular populations, demonstrating no significant

labeling of endocrine cells following birth74,163. In instances of pancreatic injury utilizing the

pancreatic ductal ligation model, increases in Neurog3 expression have been observed, and

related to increased β-cell proliferation164. However, it remains unclear whether Neurog3

expression is activated in a non-β-cell population or whether increases in Neurog3 expression are

due to de-repression via stress to pre-existing β-cells. Therefore, the current literature strongly

suggests that under normal conditions, at least, β-cell mass in the adult is governed primarily through the interplay between β-cell proliferation and apoptosis.

1.14 Modulators of β-cell Replication

Although β-cell replication rate has an intrinsic cell-cycle component, discussed below,

proliferation can be increased by external stimuli165. Mice treated with pancreatectomy, STZ, and

diphtheria toxin have demonstrated a recovery of β-cell mass following initial ablation166–168.

Elevations in fatty acids and glucose, frequently observed in type 2 diabetes are also able to

change β-cell replication rates depending on context. For example, high-fat diet fed mice are able

to increase their proliferation rates by one week, and β-cell mass by three-weeks of feeding169.

Acute 4 day ex vivo incubation of islets with palmitate or oleate fatty acids was also able to

increase the number of labeled proliferating cells170. However, elevating free fatty acids by

continuous lipid infusion reduced cell proliferation by induction of cell cycle inhibitors171. Thus

it appears that fatty acids may increase proliferation acutely, while chronic levels demonstrate

lipotoxicity; characterized by decreased insulin secretion and β-cell proliferation33. Using β-cell

specific ablation combined with islet transplant, glucose has also been related to increased β-cell

proliferation via Gck activation172. This is supported in part by Gck inhibition and

haploinsufficiency studies demonstrating a reduction in β-cell replication173,174 Finally, as another example of a β-cell proliferative response to stress in vivo, increased insulin resistance

during pregnancy increases maternal β-cell mass. Gene knockout studies point towards a

requirement for prolactin, placental lactogen, and serotonin signaling175,176. Prolactin signaling in

INS-1 cells in vitro leads to nuclear translocation of STAT5177,178, while STAT signaling itself

has been connected to downregulation of menin and cell cycle inhibitor expression in β-cells, decreasing proliferation179–181.

1.15 β-cell Cycle Gene Regulation

Commitment to the endocrine lineage is marked by the upregulation of Cdkn1a (p21) and cell

cycle exit182. Postnatal β-cell replication requires re-entry into the cell cycle, and this may be

mediated by degradation of cell cycle inhibitors and expression of cell-cycle kinases and their

158,183,184 cognate cyclins . Regulation of the G1-S transition appears to be critical and loss of

function studies of CDK4 and their cognate Cyclin Ds demonstrate a severe reduction in β-cell

mass185–187. Cell-cycle inhibitors of the INK and CIP/KIP/WAP families are also expressed in β-

cells184. p57 expression has also been suggested to restrain progenitor cell proliferation,

downregulated after endocrine specification, when its family members p21 and p27 increase71.

Age related loss of proliferative capacity has also been linked to the loss of EZH2 repression of

cell-cycle inhibitors p16 and p19188. As discussed above, changes to cell cycle gene expression

appears to be terminal to intracellular pathways controlling proliferation, and thus represent a

potential therapeutic target. To that end, experiments utilizing adenoviral overexpression of cell

cycle genes in islets have demonstrated increases in β-cell replication44,189. Taken together, the

sum of cell cycle gene expression is permissive for β-cell replication postnatally. However, cell

cycle regulation can be changed in the face of external stress or signaling, although many of the

direct relationships remain to be explored.

1.16 SOX Factors in Pancreas Development

Sry-related HMG-box (SOX) TFs are present throughout the lineage commitment choices during

pancreas development190. Some SOX proteins are required for both initial cell specification, and proliferation of those cells afterwards191–193. Thus, they are of interest for study during a similar

sequence of events for β-cells. Sex-determining region Y (SRY) is the archetypical member of

the SOX TFs, first to be identified, and is required for testis differentiation194. Twenty different

SOX family members have been discovered, based upon their sequence consensus at the high- mobility-group (HMG) DNA binding domain, with approximately 66% consensus at this domain across all SOX proteins195. The DNA binding domain for all SOX proteins have a common

motif: AACAAAG (TTGTTTC), determined based on experimental evidence196,197, although this

sequence can be expanded upon or altered based on cofactor binding198.

Perhaps the best studied functions of SOX factors are during development processes. Loss of function studies of SOX genes demonstrate a requirement for the development of tissues derived

from all germ layers, including the nervous system, skin, bone, muscle, endoderm, liver, and

pancreas199–201. SOX proteins are also clustered into groups A-H depending on the intra-family similarity within the HMG-domain sequence. Members of the same SOX group have also been found to be similarly expression within specific tissues, and have been identified to have functional redundancy202,203. Importantly, SOX factors operate with partner factors and thus their

target genes vary significantly depending on the specific SOX and cofactor proteins expressed

within a cell. As an example of SOX factor exchange during development, SOX2 and OCT4

have a critical partnership in pluripotent stem cells204. However, OCT4 binding to SOX17

modifies the DNA motif recognition of the new complex to activate endoderm targets, mediating

lineage specification205. In another example, SOX2 is expressed and required for neural

progenitor cells, but a switch to SOX C group members drives neuronal differentiation and

maturation192,206,207.

In the context of pancreas development, initial studies demonstrated the presence of multiple

SOX family members present by E11.5 based on in-situ hybridization208. SOX9 is the best- characterized factor, found as early as E9.0 in the pancreatic bud with knockout studies demonstrating a role for SOX9 in the development of the progenitor cell population. In addition,

SOX9 binds and activates the Neurog3 promoter, which is critical for entry to the endocrine lineage for mice and humans58,209. SOX9 expression becomes restricted to epithelial trunk cells

by E15.5 and finally the pancreatic duct lineage in the adult58,59. Interestingly, lineage tracing

with pulsed tamoxifen demonstrated decreasing endocrine cell specification as embryonic development progresses78. As the expression pattern of SOX factors change during lineage

specification, one question of interest was whether any of the other SOX genes expressed in

pancreatic trunk cells could contribute to endocrine specification and maturation.

This was explored in the context of the SoxC family gene Sox4. Sox4 was of particular interest as

it is required for neural differentiation, cells that share many similar features with β-cells.

Moreover, GWAS studies in humans demonstrated an increased risk for diabetes in individuals

with single-nucleotide polymorphisms in what is now considered to be a Sox4 regulatory

enhancer24,210. In mice, germline deletion of Sox4 resulted in embryonic lethality by E14.5 due to

cardiac deformities199. However, an early study utilized pancreatic explants in-vitro to overcome

this, and observed a decrease in the number of endocrine cells after 8 days211. Additionally, a

study involving an ENU mutagenesis forward screen for impaired glucose tolerance identified a

mutation in the Sox4 HMG DNA binding domain212. Sox4 heterozygotes for this mutant allele

had normal β-cell numbers, but exhibited impaired insulin secretion, suggesting that SOX4 may

also be important for β-cell function. Paradoxically, mice were heterozygous for the mutant

allele, but overall Sox4 expression was increased. Remaining questions include whether SOX4

has any roles during β-cell development, which may be addressed by a conditional homozygous

Sox4 knockout. Additional studies regarding Sox4 in pancreatic development demonstrated that

sox4b, the zebrafish ortholog, was required for α-cell differentiation but not β-cell differentiation213. However, several TFs critical for mammalian β-cell development have a different regulatory profile in zebrafish, suggesting that SOX4 may also have variable roles in β- cell differentiation between species214,215. Taken together, it appears that SOX4 may be important

for β-cell differentiation, but whether it is significant in vivo, when it acts upon endocrine cell

differentiation, and whether it is involved in adult β-cell function remains a black box.

1.17 Mediator and Transcriptional Regulation

TFs have been well characterized during pancreatic development, transforming environmental

and intracellular information to gene expression changes. Ultimately, assembly of the RNA

Polymerase II (Pol II) holoenzyme complex on gene promoters initiates transcription216. Basal

transcription requires general TFs such as TF-IIA, -IIB, -IID, -IIE, -IIF and -IIH to bind the

promoter and recruit Pol II217. In contrast, regulated transcription requires controlled Pol II

recruitment to specific genomic loci, controlled by DNA sequence specific TFs. TFs do not bind

Pol II directly, and instead bind intermediaries such as the Mediator complex (Mediator) to

recruit Pol II recruitment to specific regions, followed by pre-initation complex formation.

Moreover, TF-Mediator interactions may increase or decrease Pol II stability during initiation or

elongation to regulate gene transcription218–220.

Mediator has a large and flexible structure that is comprised of 30 unique subunits in mammals221. Crystal structures and cryo-electron microscopy have helped to define four distinct

domains of Mediator: the head, middle, tail, and CDK8 module222. The head module is required

for Pol II binding and together with the middle domain make up the 15 “core” mediator subunits

required for activity, based on yeast experiments222,223. Importantly, Mediator has variability with

its composition; Mediator is able to carry out different functions depending on its constituents.

As one example, the CDK8 module is thought to negatively regulate Mediator interactions, and

CDK8 disengagement alters functionality221. Additionally, the tail region subunits have limited

contact with Pol II but are well situated for protein-protein interactions, with experimental evidence demonstrating importance for TF docking to the Mediator224.

Figure 5: Illustration of Mediator complex interactions and Mediator complex modules. (A) Illustration of Mediator complex interaction with transcription factors at enhancers distal to the promoter, permitting engagement with the RNA Polymerase holoenzyme. (B) The Mediator complex is comprised of four defined modules, each comprised of multiple Mediator subunits. MED15 is situated in the tail module of Mediator.

The major role for Mediator is to integrate regulation from TFs bound to distal enhancers to Pol

II by protein binding225. Specific Mediator subunits have evolutionarily conserved domains that

facilitate their binding to other proteins. Early work demonstrated binding of the MED1 subunit to ligand bound thyroid hormone receptor through an LXXLL motif, followed by the

identification of other nuclear hormone receptors that require this interaction226,227. Loss of

MED1 subunit from Mediator also demonstrated specific nuclear hormone pathway defects,

despite the presence of intact Mediator complex with the majority of subunits present228. Further,

MED1 co-regulates TFs PPARγ and GATA1 to determine adipose and hematopoietic lineages,

respectively. This highlights a role for specific subunits interactions with master regulators of development. Another example is the tail subunit MED15, interacting with regulators of fatty acid metabolism. MED15 binds to SREBP1α through a kinase-inducible domain interacting

(KIX) domain229,230. In C. elegans, the ortholog MDT-15 interacts with SBP-1, the worm

ortholog of SREBP, as well as NHR-49 (HNF4α). MDT-15 knockdown resulted in specific fatty-

acid synthesis defects that may be rescued by exogenous delivery231. The MED15 variant in

yeast and plants also appears to play a role in fatty acid metabolism230,232, suggesting that

specific Mediator subunit-pathway interactions may have ancient evolutionary origins. Although

several subunits now have known roles, many Mediator subunits have no known biological role

in vivo.

There has been no characterization of Mediator function in pancreatic development. This is

perhaps surprising as many TFs have been mapped during pancreatic development, but few have

been associated with Mediator despite its key integrative role for Pol II activation. However, a

few observations suggest that the subunit MED15 may be of interest during β-cell development.

Activin and retinoic acid signaling during early pancreas development are instructive for bud formation. In hESC differentiation, these pathways are also targeted by the addition of soluble factors in order to generate β-like cells. Both TGF and nuclear receptor signaling are also

regulated by MED15 and mediator tail subunits225,233. Experiments in Xenopus have

demonstrated co-immunoprecipitation of MED15 with intracellular Smad complexes important

for TGF-β/Activin/Nodal signaling, as well as recruitment of MED15 to activin targets233.

Furthermore, glucolipotoxicity may stimulate formation of reactive oxygen species (ROS) in β-

cells. As β-cells express limited antioxidant enzymes, they are vulnerable to excess ROS. In this

setting, MED15 may be important for both lipid metabolism and response oxidative stress.

Experiments in worms have identified the MED15 ortholog to have roles in oxidative stress

response as well as fatty acid metabolism231,234. These lines of evidence suggest that the MED15

subunit may have uncharacterized roles in β-cell development and function.

1.18 Thesis Objectives

Rapid advances have been made towards the goal of generating functional β-cells46,47,149,151, and a framework of understanding β-cell gene regulation has been instrumental. While several markers of different stages of β-cell development exist, questions remain regarding: how TF network transitions are coupled during development; what changes must be established during successive lineage decisions; and how terminal differentiation is maintained in β-cell populations. The objective of this thesis is to provide a more comprehensive understanding of transcriptional regulation during β-cell development and adult maintenance with a specific interest in transitions between developmental stages and the target genes that are important for each stage.

Studies of SOX proteins have underscored their impact in regulating sequential lineage decision.

In the pancreas, SOX9 is critical for progenitor cell development. SOX4 is also expressed in the epithelium and germline null pancreas explants revealed an absence of β-cells. Thus it appears that it is indispensable for β-cell development, but it is unknown when SOX4 acts in vivo and in which cells. There is evidence of a SOX4/SOX9 duality in liver duct development235. However, whether it operates in tandem with SOX9 in progenitors or in sequence, replacing SOX9 to initiate endocrine specification is unknown. The overall hypothesis was that expression of SOX4 is necessary for pancreatic β-cell differentiation in vivo. The specific aims were:

1. To determine when and which cells express SOX4 during pancreas development

2. To determine its role during development, when it is required, and what genes it regulates

3. To identify factors that cooperate with SOX4 in the regulation lineage specification

Additionally, SOX4 is expressed in pancreatic islets. Given my interest in cell transitions and

how SOX proteins may assume multiple roles depending on co-factor context, I hypothesized

that SOX4 expression is required for positively maintaining functional β-cell mass. Here, I

explored:

1. Whether SOX4 is necessary for terminally differentiated β-cell function

2. Which genes are regulated by SOX4 in adult β-cells

To further explore elements of gene regulation, we also investigated the potential role of

Mediator subunit MED15 during pancreatic development. Based on previous data that MED15 is

required for pathways important for early pancreas formation, I hypothesized that MED15

expression is permissive for pancreas formation and differentiation. As a new candidate gene in

pancreas development, the aims were to:

1. Characterize the expression profile of MED15 throughout pancreas development

2. Identify if MED15 is required for pancreas formation through loss-of-function studies

3. Determine the TFs that MED15 binds and the targets that are regulated by MED15.

To address these hypotheses, Chapter 3 utilized multiple conditional knockout models of Sox4 to

identify the populations that requires Sox4 during pancreatic development. This was combined

with in vitro experiments to identify SOX4 co-factors and targets of SOX4. In Chapter 4, an inducible Sox4 knockout model was employed to examine the role SOX4 in maintaining

functional β-cell mass. Finally in Chapter 5, the first characterization of MED15 was performed

during pancreatic development and in maturing β-cells. A combination of immunofluorescence

analysis, transcriptome sequencing, and functional assays were used to identify a role for

MED15 in β-cells. Together, these data may provide a better roadmap of TFs and co-regulators

during pancreatic lineage specification and regulation of β-cell state.

2 Materials and Methods

2.1 Chemicals

Chemicals were purchased from Sigma Aldrich (Oakville, ON, Canada) or Thermo Fisher

Scientific (Waltham, MA, USA). Oligonucleotides were purchased from IDT (Table 3;

Coralville, IA, USA). Tissue culture reagents were from Hyclone (Logan, UT, USA) and cultureware from BD-Falcon or Corning (Fisher Scientific; Corning, NY, USA). Secondary antibodies were from Jackson Immunoresearch (Cedarlane Labs; Burlington, ON Canada).

Primary antibodies were purchased from various vendors (Table 4).

2.2 Animals

Mice were housed on a 12-hr light-dark cycle in a controlled climate according to protocols approved by the UBC Animal Care Committee in accordance with guidelines established by the

Canadian Council on Animal Care.

2.3 Diet

Mice were fed ad libitum with standard chow diet 5010 (Lab Diet) unless specified. High-fat diet: Research Diets #D12331; 58% kcal from fat.

2.4 Mouse Strains

Table 2: Mouse strains

Strain Source Pdx1-Cre MMRRC (Davis, CA, USA)236 Neurog3-Cre+ BAC Jackson Laboratory237 (Bar Harbor, ME) Pdx1-CreER MMRRC236 Ins1-Cre Jorge Ferrer Laboratory238 mTmGflox(Gt(ROSA)ACTB-tdTomato-EGFP Jackson Laboratory239 Sox4flox Courtesy Dr. Veronique Lefebvre240

Strain Source MED15flox Med15tm1a(KOMP)Wtsi (KOMP Repository)

Mouse strains used were on the C57BL/6 background, except Pdx1-Cre and derivatives which

were CD1 or mixed background, respectively. PS4KO, ES4KO, βS4KO, PM15KO, EM15KO,

and IM15KO controls were Cre- littermates. EMS4KO control genotypes were Neurog3-Cre+;

mTmGflox/+; Sox4flox/wt.

2.5 Timed Matings

Single male mice were paired with female mice in a cage and vaginal plugs were assessed the following morning. Plugged female mice were set as embryonic day E0.5 at midday of discovery and isolated in a separate cage until time of harvest.

2.6 Tamoxifen Administration

Tamoxifen was dissolved in corn-oil at 60 mg/mL with 2-4 minutes of sonication in a Misonix

3000 sonicator with cup-horn attachment. Six-week-old control and experimental mice were administered 8mg of tamoxifen by oral gavage every other day for three rounds of administration.

2.7 Physiological Measurements

Glucose tolerance test: Mice were shaved the day prior to experiment and fasted for 10 hours during the dark-cycle. On the day of experiment, mice were weighed and fasting blood glucose levels obtained by saphenous vein collection using a OneTouch® UltraMini® glucometer. Two g/kg of 40% w/v D-glucose in water was delivered via intraperitoneal (IP) injection and blood glucose levels determined at 15, 30, 45, 60, and 90 minutes. Blood glucose levels greater than detection limit were reported as the maximum possible value on the glucometer: 33.3 mM. 33

Plasma collection: Blood was collected into microcapillary tubes at fasting and 10 minutes

following IP glucose bolus and immediately transferred to 1.5mL microcentrifuge tubes on ice.

Blood samples were centrifuged immediately following procedure at 5000g for 10 minutes at

4°C and plasma was collected for determination of insulin levels using ELISA (ALPCO

STELLUX®, Salem, NH, USA). Intraperitoneal insulin tolerance test: Performed following a 3-

hour pre-fast during the light-cycle. Mice were weighed and injected with 1U/kg insulin

delivered via intraperitoneal injection. Blood glucose measurements were sampled by saphenous

vein post-insulin injection as above.

2.8 Embryonic Pancreas Dissection

Pregnant dams were euthanized by isofluorane anesthesia followed by cervical dislocation. The

uterus was removed and placed in PBS on ice. Individual embryos were removed with forceps

and scissors from the uterus, umbilical cords detached, and embryos ≥E15.5 were decapitated

following established guidelines. Tail samples were retained for genotyping. Histology:

pancreas, stomach, spleen, and a portion of the duodenum were isolated together from individual

embryos with the aid an Olympus SZX7 (Richmond Hill, ON, Canada) dissecting microscope

with curved and straight forceps and placed into cold 4% paraformaldehyde in PBS for 6 hours

for samples ≤E15.5, and overnight for samples >E15.5. RNA isolation: pancreas only was

3 dissected and placed into 200 µL of TRIzol reagent and homogenized through a 26 /8 gauge needle. Final volume was then adjusted to 1mL of TRIzol reagent and frozen at -80°C.

2.9 Adult Pancreas Collection

Mice were anesthetized with isofluorane followed by cervical dislocation. A horizontal incision was performed below the diaphragm followed by cuts across the diaphragm and laterally towards the axilla to enable access to chest cavity. The right atria was cut with surgical scissors followed

by injection of 5 mL of PBS and then 5 mL of 4% PFA in PBS into the left ventricle at a slow rate. Blanching of the liver and pancreas was used as a marker of successful perfusion, and the pancreas was dissected and placed in 4% PFA in PBS and fixed at 4°C overnight.

2.10 Cell Culture

Cells were cultured in 25 mM glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS (Hyclone), 1% penicillin (100 U/mL; Hyclone), streptomycin (100 ug/mL; Hyclone), and 2 mM L-glutamine (Hyclone). Cells were split 1:5 twice weekly. Viable cell numbers were determined by 1:1 trypan blue staining and counting via haemocytometer.

2.11 Fluorescence Activated Cell Sorting

Mouse E17.5 embryos were harvested and immediately dissected on ice to obtain pancreases that were placed in 2 mL of 0.25% Trypsin for 20 minutes with periodic agitation to homogenize the tissue. 1 mL of cold FBS and 2 mL of cold PBS were added and mixed to stop digestion. The solution was filtered through a 40 μm nylon filter and centrifuged at 200 g for 5 minutes at 4C.

Supernatant was aspirated and the cell pellet was resuspended in 2mL of cold PBS by pipette to wash the cells. Cells were centrifuged again at 200 g for 5 minutes at 4C and supernatant aspirated. Cells were resuspended in 350 μL of 2% FBS in PBS and placed on ice for sorting by

BD FACS Aria II.

2.12 Islet Isolation

Collagenase type XI (6 mL) in Hank’s buffered saline solution (HBSS) at 900 units/mL, 80 mL of ice-cold HBSS with 1 mM CaCl2, and 20 mL of RPMI 1640 supplemented with 10% FBS,

100 U/mL penicillin, 100 μg/mL streptomycin, and 2 mM L-glutamine was prepared per mouse.

Mice were anesthetized by isofluorane, cervically dislocated, and skinned to prevent hair

contamination. A lateral incision was made into the abdominal cavity and the cecum was

displaced to the right of the body. The hepatopancreatic duct was clamped via hemostat proximal

to the duodenum. A 27-30 gauge needle tip cannulated the common bile duct to deliver 3 mL of

collagenase in HBSS. Successful injection resulted in the total inflation of the head, tail and

gastric lobe of the pancreas. The pancreas was removed by surgical scissors and placed into a 50

mL conical tube with 3 mL of collagenase in HBSS on ice until ready to proceed. Collected

pancreases were enzymatically digested by placing the conical tubes in a 37°C water bath for 14

minutes, following by mechanical shaking for 2-5 minutes until a homogenous suspension was achieved. Pancreas suspensions were washed twice with 30 mL of ice-cold HBSS with CaCl2

with supernatant decanted and discarded. The pellet was resuspended in 20 mL of ice-cold HBSS

with CaCl2 and passed through a 70 μm nylon filter to retain islets and allow exocrine tissue to

flow-through to discard. Finally, the filter was inverted and washed with 20 mL of prepared

RPMI 1640 supplemented with penicillin/streptomycin/glutamine to recover islets onto a tissue- culture plate, followed by hand-picking under a dissecting microscope (Olympus SZX16) with

oblique illumination to obtain islets.

2.13 Human Islet Culture

Islets were obtained from the clinical islet core at the University of Alberta or from Dr. Patrick

MacDonald at the University of Alberta. Islets were placed in CMRL 1066 media supplemented

with 10% FBS, 1% penicillin (100 U/mL), streptomycin (100 ug/mL), and 2 mM L-glutamine.

Islets were handpicked under a dissecting microscope with oblique illumination.

2.14 804G ECM Matrix Preparation

804G cells were cultured in 5.5 mM glucose DMEM supplemented with 10% FBS, 100 U/mL

penicillin, 100 μg/mL streptomycin, and 2 mM L-glutamine. 804G ECM matrix was prepared

following 48 hour culture in serum-free 5.5 mM glucose DMEM by collecting the media and centrifuging for 5 minutes at 300g. Supernatant was then 0.2 μm filtered and stored at -20°C.

2.15 Glucose Stimulated Insulin Secretion Assay

One day prior to experiments, 804G matrix was deposited onto 12-well flat-bottom tissue culture

plates. One the day of experiment, 804G matrix coated wells were washed 3 times with sterile

water, and allowed to briefly air-dry. 1 mL of RPMI 1640 with 10% FBS, 100 U/mL penicillin,

100 μg/mL streptomycin, and 2 mM L-glutamine was aliquoted into each well. Islets were

isolated and 50 size-matched islets were deposited per well and allowed to recover overnight. On

the following day, islets were washed in PBS and pre-incubated in 500 μL of Krebs-ringer

bicarbonate HEPES (KRBH; 114 mM NaCl, 20 mM HEPES, 4.7 mM KCl, 2.5 mM CaCl2, 1.2

mM KH2PO4, 1.2 mM MgSO4, 0.2% BSA, pH 7.4), supplemented with 2.8 mM D-glucose for 1

hour. After preincubation, the media was exchanged with 500 μL KRBH with 2.8 mM D-glucose and incubated again for 1 hour for basal secreted insulin levels, and replaced with KRBH supplemented with 16 mM D-glucose or 40 mM KCl for 1 hour for stimulated insulin levels.

Both basal and stimulated insulin media were collected after their respective incubations in a microcentrifuge tube and centrifuged for 5 minutes at 300 g, and supernatant transferred to a clean microcentrifuge tube and stored at -20°C. Finally, 10 islets per well were removed to assess total insulin, removed, and transferred to 500 μL of acid ethanol (1M HCl in 70% ethanol in H2O) and stored at -20°C. Remaining islets were washed with PBS and cell lysate collected

for protein quantification.

2.16 Insulin ELISA

ALPCO STELLUX Chemiluminescent ELISA was utilized according to manufacturer’s

instructions. Plasma insulin samples were used directly. GSIS basal samples were used directly,

16 mM glucose incubation media were diluted 5x, 40 mM KCl incubation media were diluted

20x, and acid ethanol total insulin samples were diluted 200x. Briefly, 5 μL per well of samples

and standards were loaded in two technical replicates, with standards for each separate plate

used. Samples were incubated for 2 hours on an orbital shaker at 850RPM with insulin-antibody

conjugate, washed 6 times, and incubated with substrate for 1 minute. ELISA plates were read on

a POLARstar Omega microplate reader (BMG Labtech, Guelph, ON, Canada) with 1 second

integration time per well.

2.17 Seahorse Metabolism Assay

Day 1: Seahorse XFe96 spheroid microplates were coated with CellTak according to

manufacturer’s instructions, with 30 μL of CellTak working solution (CellTak in 0.1 M

NaHCO3, pH 8.0) per well dispensed by multichannel pipette. Plates were spun down at 200g for

5 minutes and incubated at 37°C for 1 hour in a non-CO2 incubator. CellTak working solution

aspirated and plates were washed 2x with 200 μL of sterile 37°C water with a multichannel

pipette and air-dried. 160 μL of warm RPMI 1640 supplemented with 10% FBS, 100 U/mL

penicillin, 100 μg/mL streptomycin, and 2 mM L-glutamine was added via multichannel pipette to wells. 10 size-matched islets in 15 μL of RPMI 1640 supplemented with 10% FBS, 100 U/mL

penicillin, 100 μg/mL streptomycin, and 2 mM L-glutamine was gently deposited into the centre of the spheroid wells with a canted P20 pipette tip. Islets attached overnight, with 6 technical replicates per animal, and 4 wells left as blanks with 175 μL prepared RPMI. The Seahorse sensor plate was prepared by adding 200 μL of buffer calibrant to the buffer cartridge. Sensor cartridge was placed into the buffer and mixed gently several times and placed into a 37°C non-

CO2 incubator overnight. Day 2: Seahorse media was prepared fresh: Gibco D5030 with 2.8 mM

glucose, 2 mM Glutamine, pH 7.4, and filter sterilized. Islets were washed by removing 150 μL

of media with a multichannel pipette, working in groups. Seahorse media (175 μL) was replaced,

removed again, and 150 μL of Seahorse media was added back for a final volume of 175 μL.

Islets were incubated at 37°C in a non-CO2 incubator for at least 1 hour. Drugs were prepared in

Seahorse media during this time as below:

Condition 16 mM Glucose Oligomycin FCCP AA + Rotenone Stock Conc. 2 M 5 mM 5 mM 10 mM Initial Port Conc. 108.4 mM 18 uM 10 uM 22 uM Dilution 2 mL: 105.6 uL 5.76 uL stock in 3.2 uL stock in 3.52 uL each Example stock in 1894.4 1.6 mL 1.6 mL stock in 1.6 ml uL media Final Well Conc 16 mM 2 uM 1 uM 2uM Well Volume 200 uL 225 uL 250 uL 275 uL

Sensor cartridge was removed from buffer plate and drugs were loaded via port-assist plates. For all wells without any drug conditions, Seahorse media was still included to the empty drug ports to fill all ports. Seahorse machine was run with the following measurement cycles: 5 cycles basal

Seahorse media, 6 cycles 16 mM glucose, 4 cycles of oligomycin, 4 cycles of FCCP, 4 cycles of

AA + Rotenone. Following measurements, islet images were captured on an Olympus inverted microscope and quantified for image area on ImagePro Analyzer (Mediacy, Rockville, MD,

USA). Readings were then normalized to islet area.

2.18 Cell Lysate Collection and Quantitation

Islets or cells were collected by addition of 95°C non-reducing sample buffer (62.5 mM Tris

Buffer pH 6.8, 1 mM sodium vanadate (activated), 1 mM sodium fluoride, 2% w/v SDS, 10%

glycerol). Samples were briefly heated to 95°C and sonicated in microcentrifuge tubes in a

Mysonix 3000 sonicator for 2 minutes (amplitude 80, 30 second pulses). Samples were centrifuged at 10,000 g for 10 minutes and supernatant collected. A fraction of the supernatant

was quantified by MicroBCA or BCA assay according to manufacturer’s instructions. Samples

and BSA standard curves were assessed with two technical replicates of each sample in a

microplate reader after standard incubation periods at 37°C. Samples for western blot were

supplemented with 0.1% bromophenol blue, 2% SDS, 10% glycerol, and 10% β-

mercaptoethanol.

2.19 Western Blot

Up to 50 μg of cell lysate western blot samples were loaded on a 12% SDS/PAGE gel and

separated in running buffer (25 mM Tris-base, 192 mM glycine, 0.1% SDS). Proteins were

transferred to a nitrocellulose or PVDF membrane in transfer buffer (25 mM Tris-base, 192 mM

glycine, 20% v/v methanol, pH 8.3) at 4°C for 2 hours at 30V. Membranes were then washed 2x with tris-buffered saline with Tween-20 (TBST), stained with Ponceau S (0.1% w/v in 5% v/v acetic acid) to validate protein transfer, and washed 2x with TBST again. Membranes were then blocked with 5% w/v milk powder in TBST for 1 hour, and then probed with antibodies in milk solution overnight at 4°C on a rocker. On the following day, membranes were washed 3x 5 minutes with TBST and incubated with appropriate HRP-conjugated secondary antibodies for 1 hour at room temperature. Membranes were washed 3x 5 minutes and incubated in the dark with

Luminata Crescendo HRP-substrate (Millipore, Etobicoke, ON, Canada). Finally, X-ray films were exposed to membranes in a darkroom and developed for appropriate saturation. If necessary, proteins were stripped with Restore solution for 15 minutes at room temperature, washed 2x in TBST, blocked with milk solution, and reprobed again for another target.

Quantification was performed by band densitometry with ImageJ (NIH) and normalized to β- actin or GAPDH loading control.

2.20 Chromatin Immunoprecipitation

All buffer compositions shown below in Table 5. Cells were grown to confluency on a 15cm tissue culture plate. Cells were fixed with addition of crosslinking buffer for 10 minutes at room temperature with gentle rocking. Fixation was stopped with addition of glycine to a final concentration of 125 mM and incubation for 5 minutes with gentle rocking. Cells were then washed 3x with ice-cold PBS and collected by addition of 5 mL of ice-cold PBS with protease inhibitors (Complete EDTA-Free; Roche) and scraping cells to a 15 mL conical tube. Cells were pelleted by centrifugation at 2000 g for 10 minutes at 4°C. Liquids were decanted and pellets were snap frozen in liquid nitrogen and stored at -80°C until desired. Cells were lysed in 20 cell pellet volumes in L1 buffer for 10 minutes at 4°C then pelleted by centrifugation at 3000 g for 10 minutes at 4°C. Pellets were washed with 20 cell pellet volumes of L2 buffer for 10 minutes at room temperature and repelleted as above. Nuclei were resuspended in 5 cell pellet volumes of

L3 buffer in non-stick 2mL flat bottom tubes and sonicated in a Misonix Cup Horn sonicator

(80% amplitude for 24 cycles, 30 second cycles on ice) to obtain chromatin fragments 200-1000 bp in size. Insoluble material was removed by centrifugation at 20,000 g for 10 minutes at 4°C.

Nuclear lysate was adjusted to 1 mL by adding L3 buffer supplemented with 0.3 M NaCl, 1%

Triton X-100, 0.1% deoxycholate. The lysate was pre-cleared with 40 μL of protein A/G

Dynabeads (pre-rinsed in L3+ buffer) for 1 hour with rotation. Afterwards, 5% of the ChIP sample was retained at -20°C as input samples. The remainder of the lysate was split and incubated with 4 μg of antibody or IgG control overnight at 4°C with rotation. Then, 40 μL of pre-rinsed protein A/G Dynabeads was added to the IP samples and incubated overnight at 4°C with rotation. Finally, bound beads were washed twice with low salt buffer, high salt buffer, LiCl buffer, and TE. All washes were performed for 5 minutes at 4°C. Immunoprecipitated materials

were eluted from the beads by addition of 200 μL of elution buffer and incubation at 65°C with continuous vortexing for 1 hour. 150 μL of elution buffer was added to the input samples and processed simultaneously. Crosslinking was reversed by overnight incubation at 65°C with interval mixing for 10 seconds every 20 minutes. DNA was phenol-chloroform purified with an equal volume of phenol/chloroform/isoamyl alcohol (25:24:1) pH 7.8-8.2. Sample tubes were vortexed for 30 seconds and centrifuged at 12,000 g for 10 minutes. Aqueous fraction was transferred to a clean microcentrifuge tube, and samples were phenol-chloroform extracted again and pooled with initial aqueous fraction. Samples were then chloroform-back extracted with addition of an equal volume of chloroform, vortexed for 30 seconds, and centrifuged. Top aqueous fraction was transferred to a clean tube and ethanol precipitation was carried out with addition of NH4OAc to 0.75 M and 1μg of glycogen. 2.5 volumes of 100% ethanol was added

and incubated at -20°C overnight. Samples were centrifuged at 4°C at 20,000 g for 20 minutes

and decanted. Samples were washed twice by 300 μL of 80% ethanol, centrifuged at 4°C at

20,000 g for 10 minutes and decanted. Residual ethanol was removed by quick centrifugation

and pipetting, samples air-dried and resuspended in 50 μL of molecular grade water.

2.21 Gene Expression Studies

RNA isolation: was performed with TRIzol according to manufacturer’s instructions. In brief,

200 uL of chloroform was added to 1 mL of TRIzol sample, vortexed for 10 seconds, and

centrifuged at 12,000 g for 10 minutes. Aqueous top fraction was removed and added to 500 uL

of isopropanol with 1 uL of glycogen (20 mg/mL) to aid in pelleting, vortexed for 10 seconds,

and centrifuged at 12,000 g for 5 minutes. Supernatant was decanted and washed with 1.4 mL of

70% ethanol twice. Following final decanting of ethanol, samples were briefly pulse centrifuged

wand residual ethanol removed by pipette and allowed to air-dry. Samples were then

resuspended in DEPC treated molecular grade water. Samples were then DNAse treated

(TURBO DNA-free; Life Technologies) according to manufacturer’s protocols. Following

DNase clean-up, RNA concentration was determined via Nanodrop Spectrophotometer. cDNA synthesis: 0.5-2 μg of RNA was combined with oligodT (10 μM) and random hexamer primers

(200 mM) and briefly denatured and placed on ice. Superscript II reverse transcriptase (Life

Technologies) was used according to manufacturer’s instructions to synthesize cDNA. Control reverse transcriptase reactions were identical except with the omission of Superscript II.

Reaction was performed in an Eppendorf thermocycler with the following conditions: 15°C for

10 minutes, 25°C for 10 minutes, 37°C for 15 minutes, 42°C for 45 minutes, 50°C for 10 minutes, 55°C for five minutes, and 95°C for 3 seconds. Quantitative polymerase chain reaction:

(qPCR): 40 ng of cDNA was used as template for qPCR using the PrimeTime primer/probesets

(IDT) in a ViiA7 (Life Technologies) real-time PCR machine. Samples were run in three technical replicates on a 384-well plate on “fast” run settings. Primer sequences are indicated in

Table 3. Expression was determined using the ∆∆CT method with glucouronidase B (Gusb) as a reference gene unless otherwise indicated. Taqman low-density array: Experimental was performed according to manufacturer’s instructions. Gene assay IDs available in Appendix B.

Briefly, 200 μg of cDNA from control or experimental islets were combined with 2x Taqman®

Universal Master Mix II, no UNG for a total of 100 μL and loaded into the desired port.

Cartridges were centrifuged, sealed, and prepared as directed. Taqman® arrays were loaded into a ViiA7 real-time PCR machine with the array card adapter and run on “standard” settings.

Primer information is indicated in Table 6. Expression was determined using the ∆∆CT method with β-actin as the reference gene.

2.22 RNA-Sequencing

Following total RNA extraction and DNase treatment, samples were rRNA depleted according to

manufacturer’s instructions utilizing the RiboGone™ Mammalian-Low Input Ribosomal RNA

Removal Kit (Takara Bio USA Inc., Mountain View, CA, USA), Agencourt AMPure XP SPRI beads (Beckman Coulter), and DynaMag-96 side magnet (Thermo Fisher Scientific). rRNA depleted total RNA were assessed for rRNA depletion, RNA integrity, size distribution, and concentration prior to library generation with the Agilent RNA Pico 6000 kit on the Agilent 2100

Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). rRNA depleted RNA passing quality control were used in the SMARTer Stranded RNA-Seq Kit (Takara Bio USA) according to manufacturer’s instructions. Following RNA incubation with stranded N6 primer, buffer, and water, samples were incubated at 94°C for 4 minutes, appropriate for samples with RIN values between 4-7. Following purification of first-strand cDNA, libraries were PCR amplified in accordance with their input RNA amount, and purified. Final adapter-ligated cDNA were quantitated by Qubit® 3.0 dsDNA high-sensitivity assay kit (Thermo Fisher Scientific) together with qPCR amplification using standard curves of a known concentration of adapter-ligated libraries. Sequencing was performed on the NextSeq 500 with the High Output Reagent

Cartridge v2 150 cycles (75bp x 2) (Illumina, San Diego, CA, USA).

2.23 Bioinformatics Analysis

Linux Ubuntu operating system version 14.04 LTS was used for analyses. Fastq files from each sample were concatenated. ENSEMBL Mus musculus genome GRCm38p.4 fastq and gtf annotation file were used for alignments. STAR v2.4.0.1241 was utilized to index the genome

fastq files, align sample transcriptomes against the genome, and output as sorted by coordinate

BAM files. Counts were generated from BAM files with HTSeq242 against annotated genes from

GRCm38p.4. Finally, differential gene analysis was performed between control and

experimental samples using DESeq2243, with an FPKM criteria of ≥5 in two or more samples,

applying a Wald test on P values from genes passing filtering, and adjusted for multiple testing

via Benjamini and Hochberg procedure.

2.24 Tissue Processing for Histology

Following placement into 4% PFA in PBS for the appropriate time frame, samples were washed for 3x10 minutes in PBS. Paraffin processing: Samples were placed into histology cassettes

(Fisher Scientific) with foam inserts and then placed in 50% ethanol overnight followed by 70% ethanol overnight. Pancreases were dehydrated in a gradient for 30 minutes twice in 95% ethanol, three times in 100% ethanol, and twice in xylene. Samples were then transferred into

Paraplast X-TRA® at 65°C in a temperature controlled oven for two 1 hour incubations followed immediately by final paraffin embedding with the Histocentre 2 (Thermo Shandon). Tissue in paraffin were sectioned at 5 μm with the Shandon Finesse microtome (Thermo Shandon) with

Feather R35 microtome blades, collected onto glass coverslips, and allowed to dry overnight at room temperature. Cryo-processing: Samples were placed in 20% sucrose, transferred to 30% sucrose, and finally to a 1:1 mixture of 30% sucrose: Tissue-Tek® O.C.T. (EMS Hatfield PA,

USA) with each step overnight or until samples were no longer floating, whichever was longer.

Samples were then placed into O.C.T. in a plastic-histology cassette and frozen on dry-ice and stored at -80°C. Prior to cryosectioning, samples were defrosted slowly at -20°C overnight sectioned at 10 μm thickness in a cryostat. Sections were quickly warmed to adhere to glass cover slips then immediately stored at -20°C.

2.25 Immunostaining

Paraffin sections were rehydrated in a gradient (2x5 minutes xylene, 3x2 minutes 100% ethanol,

2x2 minutes 95% ethanol, 2x2 minutes 70% ethanol, 2x2 minutes 50% ethanol, 1x2 minutes

H2O). Antigen retrieval was performed with slides in citrate buffer (10 mM sodium citrate,

0.05% Tween-20, pH 6.0) immersed in a water bath, heated to 95°C for 20 minutes. Sections

were allowed to cool to room temperature over 60 minutes and washed with 2x with PBS for 5

minutes. For cryosection immunofluorescence, sections were rehydrated in PBS before

continuing. Sections were blocked for 1 hour with 5% horse serum in PBS followed by

incubation overnight at 4°C with primary antibody. The following day, sections were incubated

for 1 hour at room temperature with appropriate fluorophore conjugated secondary antibodies

and 15 µM DAPI or 0.1µM TOPRO-3 Iodide (Life Technologies) to label DNA. Sections were

washed 3x 10 minutes in PBS, and aqueous mounted with Prolong Diamond with a 1.5x24x60

mm coverslip. Imaging was carried out in the CFRI Diabetes Imaging Core Facility using a

Leica SP5 II, Leica SP8 confocal imaging system, or Olympus BX61 wide-field microscope. For immunohistochemistry, prior to antigen retrieval, slides were peroxidase quenched in 0.3% H2O2

in PBS for 15 minutes. Following primary antibody incubation, VECTASTAIN Elite ABC HRP

kit was utilized according to manufacturer’s recommendations. Slides were washed in PBS 2

times then incubated in 1:200 biotinylated anti-rabbit antibodies in PBS. After washing in PBS 2

times, slides were incubated for 30 minutes in ABC reagent then washed in PBS 2 times.

Sections were developed in Metal Enhanced DAB substrate until desired intensity,

counterstained, and permanently mounted. Imaging was carried out in the CFRI Diabetes

Imaging Core Facility with an Olympus BX61 wide-field microscope.

2.26 TSA Staining

Performed according to manufacturer’s instructions for TSA Plus Fluorescence kit (Perkin-

Elmer, Waltham, MA, USA). Following standard first day immunostaining protocols for paraffin or cryosections, sections were washed 3x for 5 minutes in PBS. Sections were then incubated for

60 minutes at room temperature with HRP labeled secondary antibody diluted 1:200 in TNB buffer (0.1 M Tris-HCl pH 7.5, 0.15 M NaCl, 0.5% Blocking reagent). Slides were washed 3x 5 minutes in TNT buffer (0.1 M Tris-HCl pH 7.5, 0.15 M NaCl, 0.05% v/v Tween-20), and incubated in SA-HRP at 1:100 dilution in TNB buffer at room temperature for 60 minutes.

Sections were washed 3x 5 minutes in TNT buffer, and fluorophore tyramide working solution was added onto each section. Slides were developed for 2-10 minutes as appropriate, quenched in PBS, and washed 3x 5 minutes with TNT buffer, and aqueous mounted in Prolong Diamond.

2.27 EdU Staining

For adult studies EdU labeling was performed by daily IP injections for 1 week of 1 mg EdU.

For embryonic studies, EdU labeling was performed by IP injection of 1 mg EdU, 12-hours prior to sacrifice. Tissues were sectioned and followed standard immunostaining procedures. After antigen retrieval, sections were permeabilized in PBS-Triton (0.3%; 15 min/RT). Sections were then washed in PBS and incubated in 100 mM Tris pH 8.5, 1 mM CuSO4, 30 μM Alexa 594- conjugated Azide (Life Technologies), 100 mM Ascorbate for 1 hour in the dark (Salic and

Mitchison 2008). Sections were then washed with PBS and immunostained with standard protocols.

2.28 TUNEL Staining

In situ Cell Death Detection Kit, Fluorescein (Roche) used according to manufacturer’s recommended protocol. Fixed cryosectioned tissue were permeabilized in 0.2% Triton X-100 in

PBS for 10 minutes and then washed in PBS twice for 5 minutes each. Sections then incubated in

TUNEL reaction mixture for 1 hour in the dark at 37°C in a pre-warmed humidified chamber.

Slides then washed 2 times in PBS for 5 minutes each and followed by immunostaining or mounted for microscope evaluation. Positive control sections created by incubation in DNaseI recombinant for 10 minutes (3 U/mL in 50 mM Tris-HCl pH 7.5) prior to labeling reaction.

2.29 Morphometric Analysis

Serial sections were chosen at 100 µm intervals for E18.5 embryos and 50 µm intervals for

E15.5 embryos and stained. For adult pancreas, 8 total serial sections at 500 µm intervals for each pancreas were stained. The whole section was imaged (Olympus BX61) and tiled with

InVivo software or the cellSens Dimension software (Media Cybernetics, Rockville MD, USA;

Olympus). Cell quantification was performed using custom workflows in CellProfiler (Carpenter et al. 2006) and Image-Pro Analyzer (Media Cybernetics). Counts were normalized to the total

number of pancreatic nuclei of all selected sections or as indicated. For EdU quantification of adult β-cells, serial sections were obtained as above and high magnification images were captured by confocal microscopy on the Leica SP8 platform. All EdU+ INS+ cells were normalized to the total number of INS+ cells quantified.

2.30 Adenovirus Generation and Adenoviral Infection

Sox4 adenovirus was generated using the pAdTrack system as described 244. Neurog3 adenovirus

was as described previously (Gasa et al., 2004; Sabatini et al., 2013). Viruses were amplified in

293 cells and purified with Adeno-X Maxi adenovirus purification kits using the manufacturer’s

protocols (Clontech Laboratories, Mountain View CA, USA). mPAC L20 cells were infected at

an MOI 100:1 for 2 hours. Cells were washed briefly in PBS, media was replaced and they were

cultured for 48 hours.

2.31 Luciferase Assay

mPAC L20 cells were plated at 2e5 cells per well in 24-well plates. The next day cells were

transfected using Lipofectamine 2000 (2ul/µg; Life Technologies) with 1µg of reporter Neurog3-

pFOXluc1 or Neurod1-pGL3 reporter, 50ng of pCMV-renilla luciferase, and 100 ng of expression vectors (mouse Sox4; mouse Neurog3; rat Sox9). Transfections were balanced with empty expression vector. 25 mM glucose DMEM serum-free was used for complexing, and initial incubation with cells for 4 hours. Following incubation, 25 mM glucose DMEM supplemented 15% FBS, 1% penicillin (100 U/mL), streptomycin (100 ug/mL), and 2 mM L- glutamine was used to adjust serum concentrations to 10% FBS and cells were incubated for 48 hours. Luciferase activity was determined utilizing the Dual Luciferase Reporter Assay

(Promega) on the SpectraMaxL plate luminometer with dual injectors (Molecular Devices,

Downingtown PA, USA). Luciferase activity was normalized to renilla luciferase activity.

2.32 Gene Targeting in mPAC-L20 Cells

Transcriptional Activator Like Effector Nucleases (TALEN)s were generated in house using the protocol and TALE toolbox (pTALEN_v2) (246). Guanine binding was encoded by the repeat-

variable diresidue Asn-His (NH) as described 247. TALEN binding sites flanked the stop codon

of the Sox4 gene with the forward TALEN designed to bind to the sequence: 5’-

TCTCTAACCTGGTCTTCACC-3’ and the reverse TALEN to the sequence: 5’-

TGGCCCACCTTCTCCCCGGC-3’. This creates a doubled stranded break which can be

resolved through homology-directed repair with our targeting vector (pSox4HA-2AGLoxPP)

designed with 600 bp homology arms and a hemagglutinin (HA) included at the C-terminus of

Sox4 to replace the stop codon. Homologous recombination with the vector inserts a porcine

teschovirus 2A peptidefused in frame with Sox4HA on its N-terminal side and eGFP on its C-

terminal side, which also included a stop codon. Selection for positively targeted clones was

facilitated by a LoxP flanked PGK-Puro cassette that was situated downstream of the eGFP but

upstream of the 3’ homology arm. Gene targeting was carried out by co-transfecting 1.5 μg of

each TALEN with 3 μg of intact pSox4HA-2AGLoxPP in a 1 well of a 6-well plate with 12 μl of

Lipofectamine 2000 (Life Technologies) in 1 mL total volume overnight. The next day media was changed and 48 hours following the transfection, selection was initiated with 3 ug/mL

puromycin. After 7 days of selection, clones were picked, plated in a 96-well plate and then 96-

well plates were replica plated prior to genotyping. Cells were lysed in QuickExtract DNA

extraction solution using the manufacturer’s protocol (Epicentre Biotechnologies; Madison WI,

USA) and then genotyping carried out using PCR across the homology arms. Sequencing across

homology arms confirmed in-frame knock-in. Approximately 23% of screened clones were properly targeted. The PGK-Puro cassette was excised by infecting the correctly targeted cells with an adenovirus expressing Cre recombinase as described above (Vector Labs; Philadelphia

PA, USA). All experiments reported here were carried out using mPAC-L20-Sox4HA-2AGFP clone A8.

2.33 Human Embryonic Stem Cell Culture

CyT49 human embryonic stem cells (hESCs) 149 were cultured on mitomycin C treated mouse

embryonic fibroblasts using a growth media of DMEM/F12 (Cellgro; Manassas, VA, USA)

containing 10% Xeno-free KnockOut Serum Replacement (Life Technologies), 1% non-essential

amino acids, 1% GlutaMAX, 1% penicillin/streptomycin (pen/strep) (Life Technologies) and 0.1

mM 2-mercaptoethanol (Sigma) supplemented with 10 ng/mL heregulin-1β (Peprotech; Quebec,

Canada) and 10 ng/mL activinA (R&D) that was changed daily. hESCs were passaged twice-

weekly and plated at a density of 0.5x106 or 1x106 per 35-mm or 60-mm dish, respectively. Cells

were differentiated using a modified protocol and adherent cell cultures47. Briefly, to form

definitive endoderm cells were cultured for three days in RPMI (Hyclone), 1% GlutaMAX (Life

Technologies), 0.5% penicillin/streptomycin (Life Technologies). For the first 24 hours media

was supplemented with 100 ng/mL activin A (e-bioscience), 25 ng/mL Wnt3a (R&D), 1:5000 insulin-transferrin-selenium (ITS) (Gibco), followed by 48 hours with 100 ng/mL activin A,

1:5000 ITS and 0.2% dFBS (Life Technologies). To induce primitive gut tube, cells were cultured in RPMI with 0.5% pen/strep, 1% GlutaMAX supplemented for the first 24 hours with

2.5 μM TGF-β RI kinase inhibitor IV (EMD Bioscience; Billerica, MA, USA), 25 ng/mL KGF

(R&D Systems; Minneapolis, MN, USA), 1:1000 ITS, 0.2% dFBS followed by 48 hour culture in the same media without TGF-β RI kinase inhibitor IV. For formation of posterior foregut cells

were cultured for 72 hours in DMEM-High Glucose (Hyclone) with 0.5x B27 (Life

Technologies), 1% GlutaMAX, 0.5% pen/strep and supplemented with 0.25 μM KAAD-

Cyclopamine (Toronto Research Chemicals; Toronto, ON, Canada), 2 nM TTNPB (Sigma), and

50 ng/mL Noggin (R&D Systems). Finally, to produce pancreatic progenitors cells were cultured

for 72 hours in DMEM-HG with 0.5x B27, 1% GlutaMAX, 0.5% pen/strep supplemented with

50 ng/mL Noggin, 50 ng/mL KGF, and 50 ng/mL EGF (R&D Systems).

2.34 Statistics

All statistical analyses were performed with Prism5 (GraphPad Software, La Jolla, CA, USA).

Two-tailed unpaired Student’s t-tests, Mann-Whitney U-tests, one-way ANOVA with post-hoc

Tukey tests, Kruskal-Wallis H tests with post-hoc Dunn’s tests, or multiple t-tests with Holm-

Sidak multiple comparison correction were performed as appropriate, with p≤0.05 considered

significant. Error bars represent SEM.

Table 3: PCR primers

Taqman Primers Gene Probe Primer 1 Primer 2 Dye Gusb TCTAGCTGGAAAT CACCCCTACCAC ACTTTGCCACCC FAM/ZEN/I GTTCACTGCCCTG TTACATCG TCATCC ABLK_FQ Med15 CAACAGCAGCAGC GCATGGCTGTGG CCTGCTGTTGCT FAM/ZEN/I (Mouse) AGCAACAACAA TGTCTA GGAATTG ABLK_FQ Med15 AAACTCAAGAATG CTGTTTGGTCGG CTTCTGGACATT FAM/ZEN/I (Human) ACATGGCGGTGC TGGCA CTGACAGACC ABLK_FQ Sox4 CCGCGCTCCCTTCA TCCTCTCCTGCCT TCGAATTCCCG FAM/ZEN/I (Mouse) GTAGGTG CTTGG GACTATTGC ABLK_FQ Sox9 AGGGTCTCTTCTCG CAAGACTCTGGG GGGCTGGTACT FAM/ZEN/I (Mouse) CTCTCGTTCA CAAGCTC TGTAATCGG ABLK_FQ Sox11 ACCATCCTCTTTAT ACCTGGTGTTCA CACCGTTTCCA FAM/ZEN/I (Mouse) CCTGACCGCC CGTATTGAG ATTTCTCAGC ABLK_FQ Sox12 TTTCACCTACTGAG TCGTCTAGTATC GCCCCAATACC FAM/ZEN/I (Mouse) CCCACCGTC GCCGACC TGATTCCTG ABLK_FQ Sox4 CCGCGCTCCCTTCA TCCTCTCCTGCCT TCGAATTCCCG FAM/ZEN/I (Human) GTAGGTC CTTGG GACTATTGC ABLK_FQ Sox9 TCTGGAGACTTCTG ACTTGCACAACG CTGGTACTTGTA FAM/ZEN/I (Human) AACGAGAGCGA CCGAG ATCCGGGTG ABLK_FQ Sox12 TTTCACCTACTGAG TCGTCTAGTATC GCCCCAATACC FAM/ZEN/I (Human) CCCACCGTC GCCGACC TGATTCCTG ABLK_FQ NeuroD1 TCATGAGTGCCCA TCTTTCGATAGC ATAGTGAAACT FAM/ZEN/I GCTTAATGCCA CATTCGCATC GACGTGCCTC ABLK_FQ Pax4 ACTGCCAATCCCTC ATGTTCCAGTGA CCAGGCAAATT FAM/ZEN/I CTTCCTGTG CACCTCATC CCACATATGAG ABLK_FQ Gcg TGAACACCAAGAG AGCGACTACAGC CTCACATCACT FAM/ZEN/I GAACCGGAAGAA AAATACCTG GGTAAAGGTCC ABLK_FQ Ppy CGCAGAGCAGGGA GGAGAACACAGG CACGGGCTGAA FAM/ZEN/I ATCAAGCCA TGGACTTC GACAAGAG ABLK_FQ Grhl AGAGGACAAGCAG AGAAAGCCCAGC CAACATCGAAG FAM/ZEN/I AAGAGACAGAGGA AGAGAAAG GGAGCATTG ABLK_FQ Sst AGTTCCTGTTTCCC AGACTCCGTCAG GCATCATTCTCT FAM/ZEN/I GGTGGCA TTTCTGC GTCTGGTTGG ABLK_FQ Ins2 CCTCCACCCAGCTC TGATCTACAATG GGCTTCTTCTAC FAM/ZEN/I CAGTTGT CCACGCTTC ACACCCATG ABLK_FG Neurog3 CCTCCCTTTCCACT TCCCCATCCTATC CCTCCACTACCT FAM/ZEN/I AGCACCCAC TGCTCC CCCACTC ABLK_FQ Conventional PCR Primers Gene Primer 1 Primer 2 Primer 3 Neurog3 CCCCAAACCTCCTTC AGCTGGCCGGACAAA (ChIP) ATGCT GG

Conventional PCR Primers Gene Primer 1 Primer 2 Primer 3 NeuroD1 CCTCTGTTATGACAT AGAAAACAAACTCTTC (ChIP) TGTTGTGGGAC ATCCAGTG Pax4 GAAGGAGCTGGAAT AAAGGCTCAGTCACTC (ChIP) GGCCTT AGCC Cre TCCCGCAGAACCTG CCCCAGAAATGCCAG AAGATG ATTACG Pdx1-Cre CGTAGAGAGTCCGC CCCCAGAAATGCCAG GAGCCA ATTACG Pdx1- AACCTGGATAGTGA TTCCATGGAGCGAACG CreER AACAGGGGC ACGAGACC Med15 TAAGGTGCTGTGTGT GTAATGATGAGGCTAA GGTTGGC AAGGGCTG mTmG CTCTGCTGCCTCCTG CGAGGCGGATCACAA TCAATGGGCGGGGGT GCTTCT GCAATA CGTT Sox4 GAAGGAGGCGGAGA CATAGCTCAACACAAA GTAGACGG TGCCAACGC

Table 4: Antibodies used

Primary Host Manufacturer Cat No. Concentration Antibody Species Used SOX4 Rabbit Abcam (Toronto, ON, Ab52043 1:600 with TSA Canada) INS Guinea Pig Linco 4011-01F 1:1000 (P+F) INS Mouse Sigma I2018 1:1000 (P+F) INS Guinea Pig DAKO (Mississauga, A0564 1:500 (P+F) ON, Canada) GCG Mouse Sigma G2654 1:2000 (P+F) GHRL Rabbit Phoenix Pharm Inc H-031-31 1:500 (P) (Burlingame, CA, USA) SST Rat Millipore MAB354 1:100 (P) PP Goat Sigma SAB2500747 1:100 (P) NEUROG3 Mouse Developmental Studies F25A1B3 1:100 (P+F) Hybridoma Bank (Iowa City, IA, USA) SOX9 Rabbit Millipore AB5535 1:500 (P) CHGA Rabbit Thermo Scientific (Lab #RB-9003- 1:400 (P+F) Vision) P1 SYP Mouse Thermo Scientific (Lab MS-1150-S 1:100 (P+F) Vision) AMY Goat Santa Cruz (Dallas, TX, Sc-12821 1:100 (P+F) USA) DBA-Lectin Vector Labs RL-1032 1:200 (P+F)

Primary Host Manufacturer Cat No. Concentration Antibody Species Used Cleaved Rabbit Cell Signaling (Beverly, #9661L 1:100 (P+F) Caspase-3 MA, USA) HA Mouse Cell Signaling 2367 1:100 HA Rabbit Abcam Ab9110 4ug/ChIP GFP Rabbit MBL (Woburn, MA, 598 1:500 (P+F) USA) GAPDH Mouse Sigma G8795 1:2000 (Western) Mouse IgG Donkey Jackson Immuno (West 715-001-003 1:50 (P+F) Grove, PA, USA)

Secondary Target Host Manufacturer Cat No. Concentration Antibody Species Used Cy3 Guinea Pig Donkey Jackson IR 706-166-148 1:200 IgG (H+L) DyLight 488 Guinea Pig Donkey Jackson IR 95746 1:250 IgG (H+L) DyLight 488 Mouse IgG Donkey Jackson IR 86176 1:250 (H+L) DyLight 488 Rabbit IgG Donkey Jackson IR 85834 1:250 (H+L) DyLight 594 Guinea Pig Donkey Jackson IR 82667 1:450 IgG (H+L) DyLight 594 Mouse IgG Donkey Jackson IR 96864 1:450 (H+L) DyLight 594 Rabbit IgG Donkey Jackson IR 84473 1:450 (H+L) DyLight 594 Rat IgG Donkey Jackson IR 87283 1:450 (H+L) DyLight 594 Goat IgG Donkey Jackson IR 95247 1:450 (H+L) Biotin Mouse IgG Goat Vector Labs X0303 1:200 (H+L) Biotin Rabbit IgG Goat Vector Labs BA-1000 1:200 (H+L)

Table 5: Chromatin immunoprecipitation buffers

Crosslinking Buffer (~11x) High salt buffer 100 mM NaCl 500 mM NaCl 1 mM EDTA 0.1% SDS 0.5 mM EGTA 1% Triton X-100 25 mM HEPES-KOH pH 8.0 2 mM EDTA 11% Formaldehyde 20 mM Tris-HCl, pH 8.1 H20 H20 Lysis buffer 1 (20 cell pellet volumes) LiCl buffer 50 mM HEPES-KOH pH 7.5 250 mM LiCl 140 mM NaCl 1% NP-40 1 mM EDTA 1% deoxycholate 10% Glycerol 1 mM EDTA 0.50% NP-40 10 mM Tris-HCl, pH 8.1 0.25% Triton X-100 H20 H20 Add protease inhibitor Lysis buffer 2 (20 cell pellet volumes) TE 200 mM NaCl 1 mM EDTA 1 mM EDTA 10 mM Tris-HCl, pH 8.0 0.5 mM EGTA H20 10 mM Tris pH 8.0 H20 Add protease inhibitor Lysis buffer 3 (4 cell pellet volumes) Elution Buffer 1 mM EDTA 1% SDS 0.5 mM EGTA 100 mM NaHCO3 10 mM Tris pH 8.0 H20 H20 Add protease inhibitor L3+ buffer 300 mM NaCl 0.50% Deoxycholate 0.05% Triton X-100 L3 buffer Low salt buffer 150 mM NaCl 0.1% SDS 1% Triton X-100 2 mM EDTA 20 mM Tris-HCl, pH 8.1 H20

3. SOX4 Is Required for Pancreatic Endocrine Cell Development

3.1 Introduction

The overarching goal of this thesis was to better understand transcriptional regulation during β- cell development. Specific goals were to improve characterization of TFs with respect to their expression, function in pancreatic cells throughout development, and how they act in lineage specification. In Chapter 3, SOX proteins are the focus as they have established roles in cell fate decisions and act to “bridge” initial cell states with subsequent states. This is accomplished by exchanging co-factors to regulate new gene sets. Alternatively, a SOX protein may increase the expression of another SOX family member, which competes with co-factors to drive separate gene targets during specification. In this context, SOX9 and SOX4 are both expressed in the pancreatic epithelium. However, SOX9 expression is downregulated in endocrine progenitor cells, while SOX4 expression becomes restricted to this population. SOX9 is critical for progenitor cell maintenance58,78, while germline Sox4 knockout studies showed that it was critical for α- and β-cell development in pancreatic explants211. This suggests that during pancreatic development, SOX4 may be required in sequence following SOX9 expression for induction of the endocrine lineage. Here, the hypothesis that SOX4 expression is required for endocrine specification in-vivo is tested.

The continued role of SOX4 in endocrine progenitors was also examined as SOX proteins have been identified to couple lineage states and maintain differentiation191. Endocrine differentiation requires inhibition of Notch signaling and transient activation of NEUROG3 in epithelial trunk progenitor cells, and is accompanied by exit from the cell cycle and migration into the mesenchyme63,75,77,81,83,236,248,249. After NEUROG3 induction, downstream regulators of β-cell specification include: PAX4, NEUROD1, RFX6, NKX2.2, NKX6.1 and MAFA among others 56

86,89,91,92,94,98,105. These genes are essential for normal β-cell development, but it is unclear how

endocrine cell fate induction is coupled to the activation of these gene networks. NEUROG3

induction coupled with a co-factor present in both progenitor and endocrine specified cells is one

mechanism to link endocrine cell fate with the expression of downstream genes, as experiments

demonstrate the regulation of Neurod1 in this manner250. However, it is unclear whether any

other TFs bridge the endocrine progenitor fate choice with downstream terminal differentiation.

Improved understanding of this lineage transition will be important to consolidate how TFs

operate in unison during pancreatic development. Studies in this chapter demonstrate that Sox4

may play a bridging role during endocrine cell genesis.

This study expands on these previous observations to define the spatiotemporal importance of

SOX4 during endocrine formation in vivo. SOX4 is demonstrated to be required in a cell-

autonomous capacity for efficient endocrine cell differentiation both prior to and downstream of

NEUROG3. As such, the function of SOX4 during pancreatic endocrine cell formation mirrors

roles of SOX family genes in other systems: to bridge key developmental transitions and couple

fate decisions.

3.2 Sox4 is Dynamically Expressed during Mouse Pancreas Formation

The pattern of Sox4 expression was characterized in vivo. As seen in Fig. 4A, Sox4 is expressed throughout pancreas development with the highest levels at E11.5. A gradual decline in Sox4 message was observed after E13.5; however, significant expression was seen during the secondary transition when the majority of NEUROG3+ endocrine progenitors are derived (E12.5-

14; Fig. 6A). During human embryonic stem cell differentiation, Sox4 is expressed at high levels

in stage 3 and 4 pancreatic progenitors (Fig. 6B; courtesy of N.A.J. Krentz).

Figure 6: Sox4 is expressed during pancreas development and hESC differentiation. qPCR analyses of Sox4 expression in E11.5 to E18.5 mouse pancreas (A) and following directed differentiation of human embryonic stem cells (B). n=4-7.

Because suitable antibodies were not initially available, in situ hybridization was carried out to

localize Sox4 expression (Fig. 7). The epithelium at this stage was branched and tubular with

glucagon positive cells found in proximity to the trunk (dotted area, Fig. 7A). Low expression of

Sox4 mRNA was observed throughout the epithelium (E14.5; Fig. 7B). Occasional cells expressing higher levels of Sox4 (arrowheads, Fig. 7B) were noted to be scattered throughout the epithelium (dotted area, Fig. 7B): an expression pattern reminiscent of the bipotent trunk cells that activate Neurog375,251. Next, immunostaining was used to determine the spatiotemporal expression of SOX4. Early during pancreatic organogenesis (E11.5 & E12.5), SOX4 co- localized with PDX1 and NKX6-1 in the multipotent progenitors and trunk cells respectively

(Figs. 8A-B; Fig. 9A)76. Following the secondary transition, SOX4 was progressively excluded

from pancreatic epithelial cells; however, it remained highly expressed in all NEUROG3-

expressing endocrine progenitors (Figs 8C-D) and in nascent endocrine cells (Fig. 9A). By

E17.5, SOX4 expression is confined largely to endocrine cells (Fig. 9C).

Figure 7: Sox4 is expressed in a subset of trunk epithelial cells. In situ hybridization for Glucagon (Gcg)(A) and Sox4 (B) at E14.5. Sox4 is expressed in a subset of pancreatic epithelial cells (Arrowheads) in a domain that resembles the epithelial trunk (dotted area) with nascent endocrine cells adjacent to this region (Arrows). Inset shows a magnified region of interest. Scale bars=50 μm

To verify that SOX4 was expressed in NEUROG3+ endocrine progenitor cells, the mouse

mPACL20 pancreatic ductal cell line was utilized. This is a method that has previously been

used to model endocrine cell differentiation87,252. Using TALEN-driven homologous recombination in the mPACL20 cells, a hemagglutinin (HA) tag followed by 2A-GFP was

knocked onto Sox4 (mPAC-Sox4-HA; Fig. 10A) and 11 clonal lines were derived with 23%

targeting efficiency. As predicted from the immunostaining results (Fig. 8), when this clonal

mPAC-Sox4-HA cell line was transduced with a Neurog3 adenovirus (Ad-Neurog3)87, marked

increases in endogenous promoter activity led to appreciable HA and GFP that was detectable by

both immunostaining (Fig. 10C) and western blotting (Fig.10D).

Figure 8: SOX4 expression becomes restricted to endocrine lineage. Co-immunostaining of SOX4 (green) (A-D) with: NKX6-1 (red; A; E11.5); PDX1 (red; B; E12.5); NEUROG3 (red; C: E14.5; D: E15.5). SOX4 expression becomes increasingly restricted by E15.5 (D) with a subset encompassing NEUROG3+ cells n=3; scale bars=50 μm.

Figure 9: SOX4 protein expression is lost in knockout models. (A) Representative immunostaining on pancreatic sections for glucagon (GCG; white), insulin (INS; red) and SOX4 protein levels (SOX4; green) demonstrate SOX4 expression in epithelium that is reduced in PS4KO (B). Co-immunostaining for glucagon (GCG; white) and insulin (INS; red) demonstrates SOX4 expression in endocrine cells and adjacent cells resembling epithelial trunk regions (A,C). (C) Co-immunostaining for GCG (white) and INS (red) demonstrates SOX4 expression predominantly in the endocrine compartment (D) E17.5 ES4KO shows reduction in nuclear SOX4 protein. Representative from n=3; scale bars=50 μm.

In support, unmodified mPACL20 cells transduced with Ad-Neurog3 showed a 240-fold increase

in Sox4 message expression (Fig. 10E). These observations demonstrated that SOX4 is widely expressed in the multipotent pancreatic progenitors and that its expression is maintained within the endocrine progenitors. In order to assess whether there is an important cell autonomous role for SOX4 in pancreatic or endocrine cell formation, loss-of-function approaches were used.

3.3 Validation of SOX4 Mouse Models

To determine if SOX4 is important prior to endocrine progenitor formation, mice harboring a floxed Sox4 allele (Sox4flox) were crossed with Pdx1-Cre transgenic mice to generate pancreatic

Sox4 knockout embryos (Pdx1-Cre+/Sox4flox/flox; PS4KO). Lineage analyses indicate that recombination is extremely efficient early during pancreas organogenesis (>95% of epithelial cells at E12.5; Fig. 11A). In agreement, SOX4 immunostaining at E15.5 verified that SOX4 protein expression was significantly reduced compared to controls (Figs. 9A-B). Moreover,

Sox4flox crossed with Neurog3-Cre mice to generate endocrine progenitor Sox4 knockout embryos (Neurog3-Cre+/Sox4flox/flox; ES4KO) also demonstrated reduced SOX4 protein

expression versus controls (Figs. 9C-D).

Figure 10: Endogenous SOX4 is induced by NEUROG3 in the mPAC model of endocrine development. Schematic of Sox4-HA-2A-eGFP knock-in-add-on allele, 3’ red box is stop codon (A). Immunofluorescence analyses show induction of HA-tagged SOX4 and GFP following adenoviral (Ad-)Neurog3 in mPAC-Sox4-HA cells (B) but not in control cells (C).Western blotting of mPAC-Sox4-HA cell lysates showed an increase in HA and GFP immunoreactivity at predicted electrophoretic mobilities (~47 kDa & ~70 kDa+) following Ad-Neurog3 transduction (D). Sox4 expression is increased by Ad-Neurog3 in unmodified mPACL20 cells (E). n=4;* p≤0.05.

Figure 11: Validation of Pdx1-Cre mediated recombination with lineage tracing. (A) Pdx1- Cre mediated pancreatic recombination is assessed at E12.5 by co-immunostaining of floxed mTmG lineage tracing (GFP; green) and pancreatic cells (PDX1; red). Representative from n=3; scale bars=50 μm.

The PS4KO model of Sox4 deletion was first assessed to determine whether Sox4 had functional

roles in pancreas development. Initial quantification of endocrine cell types at E18.5 revealed

that pancreatic deletion of Sox4 significantly reduced insulin+, glucagon+ (INS+, GCG+; Figs.

12A,B,I,J), somatostatin+ (SST+; Figs. 12C,D,L), and pancreatic polypeptide+ (PPY; Figs.

12E,F,M) cells, without changes in ghrelin+ cells (GHRL+; Fig. 12C,D,K). As SOX4 is lost early

in pancreatic progenitor cells, pancreatic sections were assessed at E15.5 to address the

possibility of a defect in initial endocrine cell specification. Bipotent SOX9+ cells and

NEUROG3+ endocrine progenitor cells were quantified. A significant reduction in NEUROG3+

cells, but not SOX9+ cells, was observed (Fig. 12G,H,N,O); suggesting that SOX4 is dispensable

until endocrine differentiation is initiated within the SOX9+ progenitor58,251. To determine

whether apoptosis or proliferation accounted for differences in NEUROG3+ cells, TUNEL and

EdU labeling experiments were performed at E15.5. Our results showed no significant

differences in either cell death (Figs. 13A,B,E-H) or proliferation (Figs. 13C,D).

Figure 12: Pdx1-Cre; Sox4flox/flox knockout embryos (PS4KO) have defects in endocrine cell genesis. Immunohistochemical analyses of insulin (INS; A,B; green; E18.5), glucagon (GCG; A,B; red; E18.5); ghrelin (GHRL; C-D; green; E18.5), somatostatin (SST; C-D; red; E18.5); pancreatic polypeptide (PPY;E-F; green; E18.5), INS (E,F; red; E18.5), SOX9 (G-H; red; E15.5) and NEUROG3 (G-H; green; E15.5) immunoreactive (+) cells in control (A,C,E,G) and PS4KO (B,D,F,H) pancreas. PS4KO (black bars) had significantly reduced numbers of INS+ (I), GCG+ (J), SST+ (L), PPY+ (M) and NEUROG3+ (N) cells but not in GHRL+ (K) or SOX9+ (O) cells compared with controls (white bars). Sox4 expression (P) was significantly reduced at E15.5 in PS4KO pancreas without changes in Sox11 (Q) or Sox12 (R). n=3-4; * p≤0.05, unpaired two- tailed t-tests.; Scale bars=50 µm; nuclei=blue.

Figure 13: PS4KO mice have no changes in apoptosis or proliferation at E15.5 or E18.5. Positive control for TUNEL detection following DNaseI incubation at E15.5 (A) and E18.5 (B). At E12.5, neither the number of proliferating pancreatic epithelial cells (C; Pdx1+EdU+) nor the number of total pancreatic cells (D) were different at E15.5 between control (white bars) and PS4KO (black bars) mice. TUNEL+ cells (green) in E15.5 control (E) versus PS4KO (F). TUNEL (green) and insulin+ (red) immunodetection in control (G) or PS4KO (H) pancreas at E18.5 There were no detectable TUNEL+ cells in embryonic E18.5 pancreases, while insulin expression was reduced. Representative from n=3; unpaired two-tailed t-tests. Scale bars=50 μm.

Despite the marked phenotype and loss of SOX4 protein (Fig. 9B), Sox4 message was significantly reduced but detectable in knockout pancreas by qPCR (Fig. 12P). This may result from a long half-life of Sox4 message or expression from neurons, blood vessels, and mesenchyme that are closely associated with the epithelium. Expression of related SOXC class family members, Sox11 and Sox12, were not significantly different suggesting their upregulation

did not compensate for loss of Sox4 (Figs. 12Q,R). These findings demonstrate that Sox4 is

essential for efficient endocrine cell formation.

3.4 Sox4 Activates the Neurog3 Promoter and Induces its Expression

To investigate whether the cooperative actions of NEUROG3 and SOX4 are important for

endocrine cell genesis, unmodified mPACL20 pancreatic ductal cells were transduced with Ad-

Sox4 and Ad-Neurog3 (Fig. 14A)58,87. Co-transduction of mPACL20 cells resulted in a

significant increase in endogenous Neurog3-expression (Fig. 14A). In order to determine if Sox4 binds and directly regulates Neurog3 in the context of development, the modified mPAC-Sox4-

HA cells were transduced with Ad-Neurog3 to induce endogenous Sox4 (Fig. 10E) and ChIP was carried out using anti-HA antibodies (Fig. 14B). Significant enrichment of SOX4-HA at a conserved region of the Neurog3 promoter containing two partial SOX4 consensus binding sites were observed (Fig. 15). Luciferase assays next demonstrated that SOX4 can transactivate a 2kb fragment of the Neurog3 promoter (Fig. 14C)253. Collectively, these results suggest that Sox4

expression within a subset of bipotent trunk progenitors enhances Neurog3 expression and drives

subsequent endocrine progenitor cell formation.

Figure 14: SOX4 directly regulates Neurog3 expression in the mPAC model of endocrine development. Ad-Neurog3 and Ad-Sox4 cooperatively increase expression from the Neurog3 locus in mPACL20 cells (A). HA-ChIP (B) following Ad-Neurog3 showed HA-tagged SOX4 protein enriched at the Neurog3 promoter. Luciferase assays demonstrate that SOX4 activates a truncated 2kb fragment of the Neurog3 promoter in mPACL20 cells (C). n =3-5; * P≤0.05, one- way ANOVA.

Figure 15: Mouse genomic region containing the Neurog3 gene and its upstream sequences. Two partial SOX4 DNA binding motifs, CTTTG and CTTTGT, are identified in red boxes at - 238 and -134bp upstream of the Neurog3 transcriptional start site, respectively. These motifs are situated in the Neurog3 amplicon utilized for the ChIP assay in Fig. 14. The magnified region of interest contains these DNA binding motifs highlighted in red boxes and also show these regions to be conserved across a number of mammalian species.

3.5 Sox4 Regulates Endocrine Cell Differentiation Downstream of Neurog3

Sox4 expression is maintained at high levels in both mouse and human islets (Fig. 12P); however

Neurog3 expression is not254. In order to test if the maintenance of Sox4 expression downstream

of Neurog3 remains important for islet cell formation, mice harboring Sox4flox were crossed with

the Neurog3-Cre+ transgenic mice (Neurog3-Cre+/Sox4flox/flox; ES4KO). Because reduction of

Sox4 expression would likely occur with some delay following Neurog3-induction in the ES4KO

mice, we hypothesized that no difference in the number of NEUROG3+ cells would be observed.

To test this, we analyzed the number of NEUROG3+ cells in the ES4KO pancreas at E15.5 and

found no difference between genotypes (Fig. 16A-C). Despite this, quantification of endocrine

cells within E18.5 ES4KO pancreas demonstrated significant reductions of INS+, GCG+ (Fig.

16D-G), SST+ (Fig. 16H-I,K) and SOX4+ (Fig. 9D) cells but not PPY+ cells (Fig. 16L-N). This

suggested that in addition to its role in inducing Neurog3 expression in pancreatic progenitors,

SOX4 remains necessary for continued endocrine cell development.

3.6 Loss of Sox4 Does Not Change the Number of NEUROG3 Lineage Cells

Because a significant reduction in endocrine mass was observed in the ES4KO mouse and Sox4

has previously been shown to be important for cell survival and cell proliferation201, the effects

of loss of Sox4 on proliferation (Figs. 17A-C) and apoptosis (Figs. 17D-G) were tested: neither were impacted. Previous studies have suggested “dedifferentiated” endocrine lineage cells maintain expression of the neuroendocrine secretory granule protein chromogranin A

(CHGA)255. Therefore, in order to address the possibility that islet endocrine cells do not fully

differentiate in ES4KO mice, CHGA immunostaining was carried out. Remarkably, despite the reduction in INS+ (Figs. 16D-F, Fig. 18B) and other endocrine cells (Fig. 16) no difference in the

number of CHGA+ cells (Fig. 18A) was observed. Furthermore, co-staining for INS and GCG

with CHGA uncovered a significantly increased number of cluster-forming (dotted area) CHGA- expressing cells that did not co-stain for either of these hormones (Fig. 18C,D). These data suggest that loss of SOX4 in the ES4KO impacts normal islet endocrine cell differentiation or maturation downstream of Neurogenin3.

Figure 16: Neurog3-Cre; Sox4flox/flox embryos (ES4KO) have defective endocrine cell formation. Pancreatic immunostaining for SOX9 (A,B; red) and NEUROG3 (A,B; green) and subsequent morphometric analyses showed no difference in NEUROG3 (C) or SOX9+ cells (not shown) in the ES4KO (c.f. B vs A). Immunostaining for INS+ (D,E; green), GCG+ (D,E; red), GHRL+ (H,I; green), SST+ (H,I; red), PPY+ (L,M; green), INS+ (L,M; red) in control (D,H,L) and ES4KO (E,I,M) E18.5 pancreases. Morphometric analyses showed that ES4KO (black bars) have significantly reduced INS+ (F), GCG+ (G) and SST+ (K) cell populations but normal

GHRL+ (J) and PPY+ (N) cell populations compared with controls (white bars). n=3-4; * p≤0.05, unpaired two-tailed t-tests; Scale bars=50 µm; nuclei=blue.

Figure 17: EdU labeling of cells in ES4KO mouse pancreas at E18.5 demonstrated no changes in cell proliferation following Sox4 ablation. Insulin+ (green) and EdU+ (red) cells in control (A) versus ES4KO (B) pancreas. Proliferating β-cells were detectable as Insulin+ EdU+ cells (arrowheads; A, B) in equal numbers in both control (C; white bars) and ES4KO (C; black bars) pancreas. TUNEL detection in ES4KO mice at E15.5 (D, E) and at E18.5 (F, G) demonstrated no difference between controls (D, F) versus knockout (E, G) mice despite the decrease in insulin staining. Representative from n=3-4, unpaired two-tailed t-test; scale bars=50 μm.

Figure 18: Sox4 null endocrine progenitor cells remain in the endocrine lineage but do not differentiate into islet endocrine cells. Chromogranin+ (CHGA; A), but not INS+ (B), cells are present in normal numbers in the E18.5 ES4KO (black bars). Significant numbers of CHGA +/INS-/GCG- cells can be appreciated in the ES4KO (D; Dotted Area) but not compared to control animals (C; Arrowheads). n=4; * p≤0.05, unpaired two-tailed t-tests; Scale bars=50 µm; nuclei=blue.

To show that these CHGA-expressing, INS- or GCG- cells were derived from the Sox4-null

Neurog3 lineage, mice harboring a ROSA26mT/mG reporter allele239 were crossed with ES4KO

mice (EMS4KO). In this model, CRE expression in endocrine progenitors drives a switch from

red fluorescence (RF) to green fluorescence (GF). Neurog3-lineage GFP+ cells were quantified

and no reduction in the total number of GFP+ cells in the EMS4KO was found (Fig. 19A);

supporting the observation that neither cell proliferation nor death are impacted. Despite a

similar number of pancreatic GFP+ cells, a significant increase in non-insulin/glucagon staining,

GFP+ cells was observed in the EMS4KO (Fig. 19B-D). Notably, no differences in the numbers

of GFP+ cells that expressed the acinar cell marker amylase (Figs. 20A,B,E) or the ductal cell

marker DBA-lectin (Figs. 20C,D,F) were observed. These experiments reveal that Sox4

knockout cells remain present within the pancreas, incapable of differentiating into islet

hormone-expressing endocrine cells and that they persist in a state not normally present in the pancreas, with CHGA+ and synaptophysin+ (not shown) but no islet hormone expression.

Figure 19: Lineage tracing endocrine progenitor cells demonstrated the presence of hormone negative endocrine specified cells. There is no change in the total number of labeled endocrine lineage GFP+ cells in the E18.5 ES4KO pancreas (A). However, significantly increased numbers of GFP+ cells that did not co-stain for either INS or GCG were observed (B- D; Dotted Area) compared to littermate controls (white bars; g; Arrowheads). n=3-5; * p ≤ 0.05, unpaired two-tailed t-tests; Scale bars=50 µm; nuclei=blue.

3.7 Sox4 Regulates Factors Downstream of Neurogenin3

As lineage tracing revealed that Sox4-null, NEUROG3 lineage cells remained fated to a neuroendocrine lineage, the possibility that SOX4 regulated genes important for β-cell differentiation downstream of NEUROG3 was explored. To that end, qPCR expression analyses were carried out at E17.5 (Figs. 21A-J). The expression profiles of the endocrine hormones

(Figs. 21A-E) mirrored tissue morphometry results (Fig. 16). Importantly, gene expression levels

were significantly reduced for TFs Neurod1 and Pax4, known to be downstream of endocrine cell specification (Figs. 21F,G).

Figure 20: Examination of lineage traced endocrine specified cells that express exocrine or ductal markers. Quantification of lineage traced endocrine progenitor cells (A,B; GFP; green) that have exocrine gene expression (A,B; amylase; red) at E18.5 demonstrate a low level (<1%) of GFP+/amylase+ cells that was not significantly different between control (A,E; white bars) and EMS4KO mice (B,E; black bars). Endocrine progenitor cells (C,D; GFP; green) that express the ductal marker DBA-Lectin (C,D; red) demonstrated no significant differences between control (C,F; white bars) and EMS4KO (D,F; black bars) pancreas. n=3-4, unpaired two-tailed t- tests; scale bars=50 μm.

To test whether SOX4 binding to the Pax4 and Neurod1 promoters was important for β-cell formation, mPACL20 cells were transduced with Ad-Neurog3 and/or Ad-Sox4, and both Pax4

(Fig. 22A) and Neurod1 (Fig. 22B) expression was assessed using qPCR. Remarkably, in this system the combination of both Ad-Neurog3 and Ad-Sox4 induced Pax4 expression ~200-fold

and Neurod1 expression ~10-fold compared to Neurog3 alone. Neurod1 promoter-driven luciferase transcription256 was significantly induced with co-transfection of Sox4 (Fig. 22C). HA-

ChIP on Ad-Neurog3 transduced mPAC-Sox4-HA cells demonstrated interaction of SOX4-HA at potential binding sites within both the Pax4 and Neurod1 (Fig. 22D) promoters: providing further evidence for the direct regulation of these β-cell transcription factors by SOX4.

Figure 21: SOX4 potentiates the induction of pro-β-cell genes during endocrine pancreas development. Ins2 (A), Gcg (B), Sst (D), Ppy (E), Pax4 (F), Neurod1 (G), and Sox4 (H) message levels were significantly reduced, while Ghrl (C), Sox11 (I), and Sox12 (J) were not in E17.5 ES4KO (black bars). n=5, * p ≤ 0.05, unpaired two-tailed t-tests.

In order to understand if Pax4 and Neurod1 are SOX4 targets in vivo, fluorescence-activated cell sorter (FACS) purification of “newborn” Neurog3-lineage cells was carried out163,239 (Figs.

23A-B) using the Neurog3-Cre; mTmG mouse. We reasoned that newly-born cells would possess both RF and GF due to protein stability. FACS purification of these cells demonstrated significant enrichment of both Neurog3 and Sox4, indicating that Sox4 is indeed enriched in the

newborn endocrine progenitor pool in vivo (Fig. 23C,D). In addition, significant reductions in expression of Pax4, NeuroD1 and Ins2 (but not Arx) were observed in the Neurog3-lineage cells from the EMS4KO (Fig.23E-H). The previous demonstrations that Pax4 and NeuroD1 are directly downstream of Neurog3 and the observation that expression of these factors was reduced in the newborn (RF+GF) endocrine cells that expressed normal Neurog3 levels (Fig.23D), lends credence to them being direct, biologically relevant, SOX4 targets in vivo.

Figure 22: SOX4 cooperates with NEUROG3 to increase Pax4 and Neurod1 expression. Pax4 (A) and Neurod1 (B) expression was induced following combined Ad-Neurog3 and Ad- Sox4 in mPACL20 cells. SOX4 significantly induced Neurod1 promoter driven luciferase activity in mPACL20 cells (C). ChIP demonstrated that HA-tagged SOX4 was enriched at the Pax4 and Neurod1 promoters in the Ad-Neurog3 transduced mPAC-Sox4-HA cells (D; n=3). n=4-6; * p≤0.05, one-way ANOVA.

Figure 23: Flow cytometry analysis of mTmG labeled EMS4KO cells. Representative flow cytometry analyses from dissociated E17.5 mTmG+ (A) and Neurog3-Cre; mTmG+ (B) pancreas demonstrating that distinct populations of cells can be separated. [A] Tomato+ non- endocrine cells; [B]: Tomato+GFP+ newborn endocrine cells; [C]: GFP+ enriched maturing endocrine cells. qPCR analyses of Sox4 (C); Neurog3 (D); Neurod1 (E); Pax4 (F); Ins2 (G); and Arx (H) expression in newborn endocrine cells from E17.5 control and EMS4KO mice. (I) Stored cells as a percentage of total input pancreatic cells in 50,000 analysed cells show no differences in populations between control and EMS4KO samples. n=3-5; p≤0.05, one-way ANOVA.

3.8 Discussion

Many of the developmental transitions that underlie formation of β-cells are marked by expression of unique SOX proteins. In embryonic stem cells, SOX2 helps maintain the network

of factors that are important for the pluripotent state204. SOX factors remain important downstream of SOX2 during definitive endoderm formation, when OCT4 complexes with

SOX17 to regulate endoderm induction205. SOX17 then likely interacts with undefined factors

until expression wanes in the pancreatic buds at the time when SOX9 is induced257. SOX9

establishes the gene expression program necessary for pancreatic budding and growth and

remains important in the formation of the endocrine progenitor cells258; however, it is silenced in

the endocrine lineage.

This study provides the first evidence of a cell autonomous role for SOX4 during normal endocrine cell formation. Based on our data, we propose a model (Fig. 24) where SOX4 within a subset of multipotent epithelial progenitor cells stabilizes NEUROG3 transcription and endocrine progenitor fate. SOX4 then couples endocrine progenitor cell differentiation with subsequent development through cooperation with NEUROG3.

Figure 24: Model for SOX4 roles during pancreas development. Sox4-expressing bi-potent progenitor cells (red) activate high levels of Neurog3 expression (yellow). Sox4 and Neurog3 cooperate following Neurog3 activation to increase Neurog3 levels and also to increase expression of downstream targets Pax4 and Neurod1, genes necessary for mature endocrine cell formation (light and dark blue).

SOX4 has been implicated in development of a wide array of mammalian tissues. Previous

pancreatic studies utilized germline knockout models and were unable to definitively show that cell autonomous actions of SOX4 in the pancreatic epithelium were necessary208,211,212.

Furthermore, the early studies did not uncover when SOX4 becomes important for islet cell

formation or how it functions to regulate islet cell genesis. Finally, these previous studies did not

attempt to assess whether SOX4 expression in the exocrine progenitors is important for their

differentiation. Studies in Chapter 3 address all of these previous questions and for the first time

shows that SOX4 expression within the trunk cells is necessary for endocrine cell development,

and that loss of Sox4 impacts both formation of progenitors and terminal differentiation of all

islet cells but not the exocrine tissue.

Of note zebrafish Sox4b, an ortholog of Sox4, was transiently expressed in endocrine precursor

cells, and knockdown of Sox4b uncovered a requirement for α-cell differentiation. Other

endocrine cell types, however, were unchanged213. The differences in observed phenotypes

between our study and this former one may be attributed in part to unique endocrine

differentiation pathways present in fish and mice or to the differences in knockout/knockdown

approaches used.

In the multipotent pancreatic progenitor cells, NEUROG3 expression is repressed through active

Notch signaling79,80. Endocrine cell differentiation requires loss of Notch signaling and binding

of a number of transcription factors to the Neurog3 promoter including: , ONECUT-1, TCF2,

FOXA2 and NEUROG3 itself58,63,79,82,88,259. This study adds to the cadre of factors that regulate

Neurog3 by demonstrating that SOX4 participates in this Neurog3 regulatory network.

SOX4 remains important for endocrine differentiation downstream of its initial role in activating

Neurog3 expression. SOX4 may act to stabilize endocrine differentiation following a transient

NEUROG3 induction. Pax4 and Neurod1, which lie directly downstream of NEUROG3 and are

essential during β-cell formation98,260, were significantly reduced in the ES4KO pancreas. Null

mutation of Neurod1 leads to reductions in α-, β-, and δ-cells86 while pancreatic Pax4 knockouts

had reductions of both β- and δ-cells while other endocrine cells including GHRL+ ε-cells were

increased95,98,261. Notably, a dramatic increase in ε-cells was not observed following loss of Sox4,

suggesting that residual Pax4 expression is sufficient to suppress ε-cell formation. Lineage

tracing Sox4 knockout cells provided evidence that NEUROG3+ cell induction was intact and

that these cells remained present within proto-islet cell clusters at E18.5. Interestingly, these cells

were devoid of more terminal gene expression of insulin and glucagon. Evidence from Pax4,

Nkx6-1, and Pax6 knockout models demonstrate a role for TFs in repressing alternate endocrine

cell states. One possibility for the fate of Sox4 knockout cells could be differentiation towards

another enteroendocrine cell type. As δ-cells were significantly reduced in the ES4KO model

while ε and PP-cell numbers were similar, trans-differentiation of α and β-cells to other pancreatic endocrine cell-types is unlikely, although cannot be ruled out based on current data.

Interestingly, lineage tracing of NEUROG3+ cells have provided evidence of gastrin+ cells during development262. Follow-up experiments could address whether Sox4 knockout cells

assume a different cell fate due to the perturbed TF network, or whether these cells have stalled

in their development and occupy a dysfunctional partial neuroendocrine cell state.

Together, these studies helped to address the major points of focus for thesis, and tested the

hypothesis that SOX4 expression is required for pancreatic β-cell differentiation in-vivo.

Increased spatiotemporal resolution was obtained for when SOX4 is required in vivo, and both

binding partners and targets of SOX4 were identified. Importantly, we demonstrated that the

continuity of SOX4 expression from pancreatic progenitor cells to endocrine specified cells was

required to bridge this cell fate transition. Not only does SOX4 induce expression of the master

regulator NEUROG3, it is required as a co-factor to induce downstream TF expression for lineage progression. In conclusion, this work improves understanding of how endocrine fate progression and the related TF networks are regulated. Future studies to understand how SOX4 is induced in endocrine progenitors may be of benefit for protocols of directed differentiation of

β-cells from hESCs.

4. SOX4 Is Required for the Maintenance of Adult β-cell Population via Proliferation

4.1 Introduction

Studies within chapter 3 demonstrated that Sox4 is the predominant Sox family member involved

in the specification and differentiation of β-cells211,263. Here, Sox4 expression was demonstrated

to be maintained in mature islets, and is the highest expressed member of the SOX C group (Fig.

25). Interestingly, GWAS studies on diabetes associated risk variants identified single-nucleotide

polymorphisms (SNPs) located in the Cdkal1 regulatory region, later understood to be Sox4

regulatory elements24,25,210. To add to this, studies have demonstrated that a heterozygous mutant

Sox4 allele derived from ENU mutagenesis, impaired insulin secretion in heterozygous

offspring212,264. These mice however were germline mutants, and further assessment found

double the normal expression of SOX4. Thus, there has yet to be an in vivo loss of function study

to determine the role of SOX4 in adult in β-cells and understand its association with type 2

diabetes.

During development, β-cells are generated through differentiation; however, following birth the

β-cell population is maintained by cellular replication156,157,160,265. Experimental evidence has

demonstrated that several cyclins alone or in combination with other cell cycle genes are

sufficient to drive β-cell proliferation, while loss of Ccnd2 leads to a reduction in β-cell mass44.

Importantly, cell-cycle inhibitors including the INK and ARF family of genes have been

implicated in the negative control of β-cell proliferation, with GWAS studies identifying diabetes

risk associated SNPs in cell-cycle genes such as Cdkn2a, Cdkn2b, and Cdkn1c24,266. Although β- cell replication occurs at a slow-rate, proliferation can be increased following conditions such as high-fat diet, obesity, and pregnancy175,267,268, and these have been correlated to changes in cell- cycle genes. However, there have been few direct regulators of cell-cycle genes defined in β-

cells. As the SOX4 co-factor NEUROG3 is downregulated in β-cells, partner with other co-

factors may allow for SOX4 to regulate new processes in the adult β-cell state. SOX4 may be a

regulator of cell-cycle genes in β-cells, as it is believed to be an oncogene269–271. Further, previous study of Sox4 hypomorphic mice identified reduced proliferation of skin stem cells, with deregulation of cell-cycle genes. Further, these hypomorphic mice were also associated with accelerated aging272. This observation was also of interest as labeling experiments in β-cells have shown diminishing proliferation rates with age157,159,273, while other studies have correlated this

changes in the expression of cell-cycle genes132,188. Therefore, we hypothesized that continued

SOX4 expression in adult β-cells is required for proliferation and maintenance of functional β- cell mass.

In chapter 4, the role of Sox4 in the maintenance of functional β-cell mass is explored. Here,

Sox4 is demonstrated to be the most highly expressed SOX family member in adult β-cells. Sox4 expression was found to be reduced in samples from humans with type 2 diabetes, correlating with GWAS studies that highlight Sox4 association to type 2 diabetes risk. In addition, continued

Sox4 expression is essential for the high-fat diet stimulated expansion of β-cells mass in mice, partly through regulating Cdkn1a expression. In sum these studies demonstrate that changes in

Sox4 expression may increase diabetes risk through impaired normal β-cell replication.

4.2 SOX4 Expression is the Highest of All SOX Family Members in Pancreatic Islets

We sought to understand whether Sox4 remained a critical factor in β-cells, and if it was due to the regulation of different targets. Sox4 expression was assessed in both mouse and human islets, and was the highest expressed SOXC group member, while Sox9 expressed was low (Fig. 26).

Published RNA-Seq data from both humans and mice132,145,274–277 corroborated this finding, and

confirmed that Sox4 expression is the highest of all family members in mice and humans. In

human type 2 diabetic islet samples, Sox4 expression is also diminished versus control samples

(Fig. 26B), a setting correlated with reduced proliferation and cell-cycle gene associated risk.

Figure 25: Taqman gene expression profiling of Sox genes in isolated human (A) and mouse (B) islets. Expression profiling in islets revealed that Sox4 is the most highly expressed SOXC family member, in addition to detectable expression of Sox11, Sox12, and the SOXE family member Sox9. n=3.

Figure 26: Sox4 is the highest expressed member of the SOX class of transcription factors. (A) Averaged Log2 fold FPKM values of SOX family genes obtained from published RNA-Seq studies in isolated human or mouse β-cells normalized to Gusb expression within each study, normalized to a scale from 0-1. (B) Taqman® qPCR analysis on isolated human islets showed decreased Sox4 expression from individuals with type 2 diabetes (T2D) versus controls (CT). n=6, * = p≤0.05, unpaired Mann-Whitney U-test.

To understand how Sox4 contributes to a functional adult β-cell state without its confounding impact during development, a tamoxifen-inducible knockout approach was utilized263. Pdx1-

CreER+/0; Sox4flox/flox mice and littermate Sox4flox/flox mice were administered tamoxifen at 6

weeks of age to generate knockout (βS4KO) or control mice, respectively (Fig 27A). Tamoxifen

delivery led to a significant decrease in islet Sox4 expression by 8 weeks of age (Fig. 27B). In

addition, immunohistochemical staining for SOX4 protein was performed on paraffin sections

taken from mice at 30 week endpoint. This demonstrated strong nuclear SOX4 in control mice

(Fig. 27C) but not in βS4KO tissue (Fig. 27D), indicating effective knockout without recovery

over the experimental time-course.

Figure 27: Validation of the SOX4 knockout model. (A) Schematic of longitudinal mouse study: littermate control or knockout mice received 8mg of tamoxifen in corn oil (60mg/mL) at 6 weeks, every other day. High-fat diet commenced at 7 weeks of age, with weekly body weight, fasting glucose, and bi-weekly IPGTT assessments until endpoint at 30 weeks of age. (B) Taqman® qPCR analysis on 8 week old isolated islets demonstrated a significant decrease in knockout (βS4KO, KO) islets versus controls. (C,D) Immunohistochemistry of 30 week old islets from control (C) demonstrated nuclear protein SOX4 in the islets that is reduced in knockout islets (D). n=4, * = p≤0.05, unpaired Mann-Whitney U-test.

4.3 SOX4 Knockout on HFD Are Glucose Intolerance and Have Reduced Plasma Insulin

To ascertain whether SOX4 is required for normal β-cell function and ultimately glucose

homeostasis, mice were monitored for body weight weekly and glucose tolerance tests were

performed bi-weekly. Chow fed mice exhibited no significant differences in their body weight or

glucose tolerance up to the endpoint at 30 weeks age (Fig. 28). HFD was fed to the mice to 86

stimulate the expansion of β-cell mass and to model the prediabetes state in humans. HFD-fed

control and βS4KO mice gained weight at the same rate (Fig. 29A). Despite this, βS4KO mice

on HFD demonstrated worsening glucose tolerance over time and significant difference in

fasting blood glucose by 30 weeks of age (Figs. 29B-F). Insulin tolerance tests were also carried

out on HFD-fed mice to assess whether differences in insulin resistance contributed to glucose intolerance. Blood glucose levels after a three hour fast were elevated in βS4KO mice versus controls, similar to the increased fasting blood glucose levels observed in IPGTT results from 30 week old mice. However, no significant differences in glucose disposal following IP insulin injection between control and βS4KO mice were observed (Fig. 30).

Figure 28: βS4KO mice on chow diet had no significant differences in body weight or glucose tolerance over time. (A) Weekly body weight measurements were similar between littermate controls (Control) and βS4KO (Knockout) mice. (B-D) Bi-weekly IPGTT assays at 10,

20 and 30 weeks respectively with no significant differences between littermate control and βS4KO mice. (E) Area under the curve plotted against time revealed no significant differences between control and βS4KO mice. n=3-5. (A-D) Two-way repeated measures ANOVA. (E) Multiple t-test with Holm-Sidak multiple comparison correction.

Figure 29: βS4KO mice on HFD have declining glucose tolerance over time. (A) Weekly body weight measurements were similar between HFD-fed littermate controls (Control) and βS4KO mice (Knockout). (B-E) Bi-weekly IPGTT assays of HFD-fed mice demonstrated worsening glucose tolerance results over time with the 30-week time point presenting impaired fasting glucose levels. (F) Area under the curve plotted against time showed separating IPGTT scores between HFD-fed control and βS4KO mice. n=7-10, * = p≤0.05 (A-E) Two-way repeated measures ANOVA. (F) Multiple t-tests with Holm-Sidak multiple comparison correction.

Figure 30: βS4KO mice at 30 weeks of age had similar insulin tolerance compared to littermate controls. (A) 30 week old βS4KO (Knockout) mice had elevated fasting blood glucose levels versus littermate controls (Control). Following insulin administration, blood glucose levels declined at a similar rate compared to littermate controls. (B) Data graphed as fold over basal blood glucose levels, demonstrating no significant differences between groups. n=3-9. Unpaired t-test on area under curve (B).

To understand how β-cell function was impaired in the HFD condition βS4KO, plasma insulin levels were assessed following glucose challenge. This established that βS4KO had significantly reduced stimulated plasma insulin levels (Fig. 31A). Basal plasma insulin in βS4KO trended higher than in controls, which might suggest mis-regulated insulin secretion when SOX4 is deleted. However, this difference was not significant and not observed in static glucose stimulated insulin secretion (GSIS) assay on isolated islets (Fig. 31C). Reduced β-cell insulin secretion may be due to an altered β-cell state as SOX4 has been shown, in development, to impact β-cell differentiation and the expression of Neurod1, a TF important for mature β-cell function and insulin transcription278. Therefore, GSIS assay was performed on size-matched

islets isolated from 30 week old mice on HFD. No significant differences were observed in either

insulin content (Fig. 31B) or glucose- or KCl-stimulated insulin secretion (Fig 31C). This

demonstrated that the function of β-cells in βS4KO mice were normal when accounting for the

number of β-cells. Taken together, these data demonstrated that SOX4 is required for β-cell

compensation to excess nutrients; however, loss of SOX4 did not functionally impair the β-cells

that remained. This suggested that reductions in serum insulin may be due to a decrease in the

number of functional β-cells present in βS4KO mice.

Figure 31: βS4KO mice on HFD have reduced plasma insulin following glucose bolus despite functionally unimpaired islets. (A) Plasma insulin levels show a significant reduction in βS4KO mice following glucose challenge. (B) Isolated islets from 30 week old mice for GSIS have no significant differences in insulin content (C) nor significant differences in function following glucose stimulation (C). n=4, *=p≤0.05. Friedman’s two-way ANOVA (A,C). Unpaired t-test (B).

4.4 SOX4 is Required for β-cell Proliferation

In order to understand if the impaired glucose tolerance was driven by reductions in β-cell mass,

β-cell mass was quantified at 30 weeks of age, and this demonstrated a 64.8% reduction in the number of pancreatic INS+ cells in βS4KO versus controls (Figs. 32A-C). No significant difference in the number of GCG+ cells (Fig. 32D) was observed. As proliferation is essential for maintaining β-cell numbers in adults, replication was quantified utilizing daily EdU injections

for one week following tamoxifen administration, from 7-8 weeks of age. This coincided with

the first week of our high-fat diet feeding regimen: an experimental paradigm that has been well

characterized169. After 1 week of EdU injections and HFD, 7% of total β-cells were EdU labelled

(Figs. 33A,C), similar to published cumulative labeling results in β-cells157. Importantly, the

βS4KO showed a 38.6% reduction in the number of EdU+ INS+ cells (Figs. 33B,C), although at

this point β-cell mass did not differ between genotypes (Fig. 33D), similar to previous observations169. These studies demonstrated that SOX4 is essential for β-cell replication

stimulated by HFD-feeding.

Figure 32: βS4KO have reduced β-cell mass. (A-B) Representative images from tiled acquisition of (A) control (CT) or (B) βS4KO (KO) Sections of pancreas from 30 week old mice with immunostaining for glucagon (red), insulin (green), or DNA (blue). Quantification of pancreatic sections demonstrated a 65% reduction in insulin+ cells (C) but not glucagon+ cells (D) in 30 week old pancreases. n=4, * = p≤0.05. Unpaired two-tailed t-test.

Figure 33: βS4KO have a reduction in β-cell proliferation rates. (A-B) Representative images from (A) control (CT) or (B) βS4KO (KO) 30 week pancreas labelled with EdU for 1 week in-vivo prior to harvest. Sections were immunostained for EdU (A”; B”; red), insulin (A’; B’; green), or DNA (A; B; blue). (C) Quantification of EdU labelled insulin+ cells in representative sections reveal a 39% decrease in labelled insulin+ cells; while (D) counted insulin+ cells remain unchanged at 8 weeks of age. n=4, * = p≤0.05. Unpaired two-tailed t-test.

4.5 SOX4 Regulates the Cell-Cycle Inhibitor CDKN1A in β-cells to Modulate Proliferation

As SOX4 is a transcription factor, we utilized Taqman® low-density arrays to identify SOX4

target genes important for regulating the β-cell cycle. Importantly, these studies were carried out following 1 week on HFD prior to reductions in β-cell mass. Amongst an array of cell-cycle genes, only Cdkn1a (p21) and Mcm10 had significant 2-fold differential expression (Fig. 34A).

However, Cdkn1a expression has one of the highest fold β-actin levels in the panel of genes

assessed, roughly 69x greater than Mcm10 (Figs. 34B,C). This suggested that this highly

expressed gene may be responsible, in part, for the reductions in β-cell proliferation in βS4KO.

CDKN1A has been previously characterized to be associated with β-cell cycle arrest to drive

reductions in β-cell proliferation182. In contrast, MCM10 is an S-phase scaffold protein without its own activity, expressed at extremely low relative abundance levels (Fig. 34C). Further analysis of the Cdkn1a promoter identified three canonical SOX4 DNA binding motifs

[TTGTTT] at -2374bp, -1790bp, and -692bp upstream of the transcriptional start site. A previous study utilized deletion constructs of the human CDKN1A promoter, identifying a domain required by SOX9 to regulate transcription. This is supportive as SOX factors share a common motif due to HMG domain conservation, while independent EMSA and ChIP-Seq experiments on SOX4 and SOX9 have also mapped binding to this common motif193,196,279. Together, these

studies reveal a previously unappreciated role for SOX4 in promoting β-cell proliferation, in part

through the altered expression of the cell-cycle inhibitor p21.

Figure 34: Taqman® low-density array analysis of cell-cycle genes in βS4KO islets following one week of HFD. (A) Taqman® low-density array expression analysis of 8 week old islets, following taxmoifen administration at 6 weeks of age, and 1 week of HFD from 7-8 weeks of age. Blue lines represent 0.5 and 2 fold control expression levels, red line represents control expression levels. Only Cdkn1a and Mcm10 are significantly different. (B) Cdkn1a has the highest fold difference in knockout (KO) islets versus littermate controls (CT) and the greatest expression levels fold β-actin. (C) Mcm10 is differentially expressed, although at ~69x less than Cdkn1a. n=3, * = p≤0.05. Unpaired two-tailed t-test.

4.6 Discussion

Replenishment of β-cell numbers relies on cell-division rather than differentiation83. Mirroring

the neuronal system280, the studies in chapter 3-4 establish pancreatic SOX4 as a factor that

initially specifies a terminally differentiated β-cell state263, and then supports β-cell proliferation.

This is in agreement with experimental evidence that SOX4 can act as an oncogene269 as well as

regulate cell cycle genes and proliferation272,281.

HFD-fed βS4KO mice exhibited glucose intolerance that was attributed to reduced plasma

insulin levels. However, once matched islets were utilized in a static GSIS assay, no significant

differences were discovered in either total insulin content or insulin secretion following 16mM

glucose or KCl challenge. This suggested that β-cell function is unaffected when islet size is

matched, and that reduced serum insulin was due to altered β-cell number in βS4KO mice.

Furthermore, while SOX4 is important for embryonic β-cell differentiation, these data suggest against a requirement in the maintenance of a functional β-cell state. Quantification of immunostained pancreatic sections supported these observations as βS4KO pancreatic sections had diminished β-cell numbers.

Here, HFD feeding of the βS4KO resulted in glucose intolerance that was caused by impaired β- cell replication. To estimate the impact of a 38.6% reduction in proliferation over a period of 30 weeks, we used the proliferation rate derived from a one week EdU labeling period, together with a simple mathematical model of compounding over 24 weeks: Cellsfinal = Cellsinitial [1 +

Time(weeks) (Proliferationknockout – Proliferationcontrol)] . Assuming a steady state of β-cell attrition and

proliferation, this approach estimates a 48.7% decrease over 30 weeks, roughly accounting for

our observed 64.8% reduction in β-cell mass. In fact, previous estimates suggest that β-cell

proliferation rate is significantly higher at birth with an exponential decay157,160. This suggests

that our estimation is conservative and our β-cell quantification more accurately reflects real- world differences. Additionally, TUNEL staining did not show significant levels in our model suggesting changes in cell survival do not affect β-cell mass in the βS4KO.

The screen of potential target genes identified Cdkn1a as the best candidate as it was amongst the most differentially expressed genes with a high abundance level. Cdkn1a arrests cell G1-S

transition and has been correlated with endocrine specification and exit from the cell cycle182.

Mcm10 may also be expressed during late G1 phase, with its increases reflecting a stalled S- phased entry282. Germline Cdkn1a gene knockout does not lead to increased proliferation due to

redundancy with other family members and possible developmental compensation283. However, recent evidence suggests that CDKN1A inhibits β-cell cell-cycle progression, and targeting the protein by small molecule inhibitors improves human islet proliferation284. Overall, the evidence

points towards SOX4 as a positive modulator of proliferation through downregulating p21

expression that is present in adult β-cells.

In summary, chapter 4 explored the role of SOX4 in the adult β-cell state, as data highlighted the presence of SOX4 at high levels relative to its family members. Here, a loss of function approach determined that Sox4 expression is important following β-cell differentiation to maintain effective cellular replication rates in the face of metabolic stress, through modulating the expression of Cdkn1a. This demonstrates that SOX4 has unique roles in both developmental and adult β-cell states. The finding that β-cell replication is reduced in a nutrient surplus setting may help to explain the identified risk associations with Sox4, as obesity and nutrient excess have a high co-incidence rate with type 2 diabetes.

5. Mediator Subunit MED15 Is Required for Pancreatic Endocrine Cell Specification and a Functional β-cell State

5.1 Introduction

Some poorly described aspects of transcriptional regulation in β-cells are how TFs are coupled to

external signals and how TFs in β-cells specifically interface with Pol II to regulate gene

expression. For instance, receptor and gene deletion studies have identified activin and FGF

signaling to be required for pancreatic bud induction and expansion, respectively68,285. However,

the intracellular TFs that ultimately regulate these signals have not been characterized, nor the

changes in gene expression that ensue to specify pancreatic tissue. Following pancreas

specification, single-gene deletion studies have identified TFs important for proper pancreatic

formation and cell fate decisions, as described in chapter 2. Although some studies have

demonstrated cooperative TF binding256,286,287, the interactions for many TFs remain unknown.

Finally, after TF binding to their cognate DNA domain, the terminal interactions with Pol II to

regulate transcription are undefined. Here in chapter 5, some of these questions begin to be

explored in the context of Mediator proteins.

Mediator is an important complex that regulates Pol II assembly, elongation, and termination of

transcription221. This is highlighted by Mediator disruption in yeast, where transcription levels

decreased to virtually undetectable levels218. One of the major functions of Mediator is believed

to be a TF co-regulator, allowing TFs at enhancer sites to dock to the Pol II holoenzyme, and

thereby facilitating gene regulation. This is supported by numerous TF-Mediator studies as well

as ChIP-ChIP genome occupancy experiments that have identified Mediator presence in gene

enhancer locations288,289. Interestingly, single Mediator subunit deletions have identified interactions with transcriptional master regulators of specific lineages219,227,228. For example,

MED1 and MED14 are required for PPARγ recruitment while MED23 is important for ELK1

binding, in adipogenesis219. These and other studies have led to the concept that individual

Mediator subunits facilitate specific signaling pathways or lineage decisions225.

The Mediator subunit, MED15 was of particular interest with regards to pancreas biology.

MED15 is a component of the Mediator tail-domain, a region that cryo-EM experiments have

predicted to be suitable for TF-docking224. Previous studies have demonstrated that MED15 is

required for TGF-β signaling as well as nuclear hormone signaling231,233, pathways which are

important for early pancreas development231,233. Further, MED15 orthologs in yeast, plants, and

worms, are required for fatty acid metabolism and response to stress231,232,290, with potential

implications for maintenance of a mature β-cell state. The role of Mediator has not been

described for β-cell biology, and conversely, many subunits do not have well-established in vivo biological functions. Here, we demonstrate the first known roles for MED15 in mice using a loss-of-function approach. MED15 is required for commitment to the pancreatic endocrine lineage; however the number of pancreatic progenitors and total pancreatic cell numbers are unaffected in the knockout. Finally, loss of MED15 in adult β-cells also alters gene expression leading to impaired β-cell function and ultimately, diabetes.

5.2 Characterization of MED15 Expression

Med15 expression in the pancreas was characterized by Taqman® real-time PCR analysis (Fig.

35A). Med15 mRNA expression was high at E12.5, immediately prior to the secondary transition when a large number of endocrine cells are specified. Med15 expression is reduced from E13.5-

E16.5 relative to other developmental time points, and becomes enriched in adult islets. This was validated at the protein level by western blot in the MIN6 β-cell line and in both mouse and human islets (Fig. 35B).

Figure 35: Med15 expression in the pancreas. (A) Taqman® qPCR analysis of isolated pancreases from E12.5 – E18.5 and 8 week isolated islets demonstrated the presence of Med15 mRNA. (B) Representative western blot against MED15 established the presence of MED15 protein in the MIN6 β-cell line as well as mouse and human islets. n=3.

To determine when MED15 is required during β-cell development, mice harbouring Pdx1-Cre,

Neurog3-Cre, or Ins1-Cre were crossed with Med15flox/flox to generate conditional knockouts:

PM15KO, NM15KO, and IM15KO, respectively (Figs. 36A). The combined use of these models allowed for the identification of the specific lineage that required MED15. Western blot analysis of MED15 demonstrated detection of an expected 85kDa protein, predicted to be MED15, in

PM15KO E15.5 tissue (Fig. 36B). This band was lost in the PM15KO knockout embryos, while additional immunostaining demonstrated reduction of MED15 protein in the models used (Figs.

36C-F). These data support the expression of MED15 in the pancreas, in a subset of pancreatic cells.

Figure 36: Validation of Med15 knockout models. (A) A timeline demonstrating the cell lineages affected by each transgenic model. Pdx1-Cre results in whole pancreas knockout of Med15 impacting progenitor cells and all derivatives. Neurog3-Cre deletes Med15 in endocrine progenitors while Ins1-Cre is expressed in maturing β-cells only. (B) Western blot for MED15 showed reduced MED15 in PM15KO E15.5 pancreas versus wild-type with GAPDH loading control shown below. Immunostaining for MED15 (green) and INS (red) demonstrated loss of MED15 protein in E18.5 PM15KO (C-D) or 8-week old IM15KO (E-F) pancreas. n=3.Scale bars=100 μm.

5.3 MED15 is Required for Pancreatic Endocrine Cell Specification

Initial assessment of PM15KO mice revealed that knockout genotypes were born at non-

Mendelian ratios (Fig. 37A). This was reminiscent of severe TF knockout phenotypes such as pancreatic Pdx1 and Nkx2-2 knockouts that demonstrated a failure to thrive postnatally60,94.

Whole-mount examination of E18.5 pancreases from PM15KO mice exhibited no gross

100

differences in overall pancreas morphology (Fig. 37B). However, immunofluorescence analyses

of pancreatic sections revealed a severe reduction in islet size and β-cell mass (Figs. 37C,D).

This is reflected in the significant glucose intolerance observed in 6 week old PM15KO mice

(Fig. 37E), suggesting defects in functional β-cell mass.

Figure 37: Characterization of PM15KO mice. (A) Initial characterization of PM15KO mice from three breeding pairs demonstrated lower than predicted Mendelian ratio of PM15KO genotype survivors. (B) Whole-mount micrograph of the pancreas showed no gross diferences compared to wild-type at E18.5 (C) Representative immunostaining of pancreatic sections from 6-week old mice for insulin (INS; green) and glucagon (GCG; red) cells versus (D) PM15KO revealed decreased numbers of INS and GCG+ cells. (E) IPGTT of 6-week old mice demonstrated that PM15KO mice were glucose intolerance, with blood glucose reaching limits of detection (33.3 mmol/L). n=3-7, p≤0.05. Unpaired Mann-Whitney U test on AUC (E).

101

5.4 MED15 is Required for Endocrine Specification But Not Pancreatic Progenitor Fate

As the preliminary results pointed towards a defect in cell differentiation, the pancreatic progenitor cell population was assessed by immunostaining at E12.5, as Pdx1-Cre recombination occurs by this time (Fig. 11). PDX1 or SOX9+ are markers of pancreatic epithelial cells at this stage, and the number of these cells was not significantly changed in PM15KO pancreases (Figs.

38A,B). Further, NKX6-1+ trunk-cell or CPA1+ tip-cell numbers were also unchanged (Figs.

38C,D). This demonstrated that pancreatic deletion of Med15 does not alter the number of pancreatic progenitor cells during early pancreatogenesis.

Figure 38: PM15KO mice have no differences in progenitor cell population. CellProfiler quantification of immunofluorescent micrographs from E12.5 pancreatic sections demonstrated no change in cell numbers stained for (A) PDX1+ (B) SOX9+ to label pancreatic cells, as well as (C) NKX6-1+ and (D) CPA1+ to label trunk and tip cells, respectively. n=3-4. Unpaired two- tailed t-test.

Next, pancreatic sections were examined by immunostaining at E15.5, a period of high

NEUROG3 induction and endocrine specification. PM15KO mice exhibited severe reductions in

NEUROG3+ cell numbers (Figs. 39A-C). However, the total number of SOX9+ cells marking the trunk-cells were unchanged, nor were the total number of pancreatic cells different (Figs.

39A,B,D,E). In support of these findings, the number of INS+ cells were reduced by E18.5 (Figs.

102

39F-H), while there were no changes in total number of pancreatic cells (Figs. 39F,G,I). Taken

together, these data support a requirement for MED15 in endocrine lineage specification from bi-

potent trunk cells.

Figure 39: PM15KO mice have significantly reduced endocrine progenitor cells. (A,B) Representative immunostaining of pancreatic sections from E15.5 control (A) or PM15KO (B) pancreas demonstrated a reduction in NEUROG3 (NGN3; green) but not SOX9 (SOX9; red) cells in PM15KO (B) sections, as quantified by Cellprofiler (C,D). Total pancreatic cell numbers are unchanged (E). (F-G) Representative immunostaining of E18.5 pancreas for insulin (INS; green) quantified by Cellprofiler demonstrated a reduction in INS+ cells (H) but not total pancreatic cells (I). n=3-4. p≤0.05. Unpaired two-tailed t-tests.

5.5 MED15 is Required for β-cell Differentiation Following Endocrine Specification

As MED15 is required for NEUROG3 induction, PM15KO would mask a role for MED15 in β- cell differentiation following NEUROG3-induced endocrine specification. Therefore, we utilized

NM15KO mice to address whether MED15 is necessary for β-cell differentiation downstream of

NEUROG3 induction. As shown in chapter 3, we postulated that loss of Med15 expression with

NM15KO occurs with a time-delay and would allow for the high induction of NEUROG3 to specify endocrine fate. This was validated at E15.5 with immunostaining demonstrating no

103

statistical difference in the number of NGN3+ cells nor SOX9+ cells (Figs. 40A,B), as expected

from previous data (Fig. 16). However, E18.5 NM15KO pancreas sections, demonstrated a

reduction in the number of INS+ cells, but not GCG+ cells or total pancreatic cell number (Figs.

40C-E). These data suggest that in addition to a requirement for endocrine specification, MED15 continues to be required for β-cell development.

Figure 40: NM15KO embryonic pancreases have reduced β-cell numbers versus controls. Cellprofiler quantification of E15.5 paraffin sections immunostained for (A) NEUROG3+ and (B) SOX9+ showed no difference in PM15KO versus controls. Quantification of E18.5 sections displayed differences in (C) INS+ cell numbers but no significant differences in (D) GCG+ cell number or (E) total pancreatic cell number. n=3-4. p≤0.05. Unpaired two-tailed t-tests.

5.6 Loss of MED15 Results in Glucose Intolerance

Interestingly, Med15 expression is at a relative maximum in E18.5 and isolated islets from 8 week old mice, versus earlier pancreatic stages (Fig. 35A). In addition, several TFs important for early pancreas development are active in the maintenance of functional β-cells, such as PDX1,

NEUROD1, and NKX6-1. TFs require Mediator complex to engage Pol II, and as MED15 was determined to be important during embryonic pancreas development, it may continue to be required for the function of TFs involved in maintaining a functional β-cell state. To explore this,

IM15KO mice were utilized, where Med15 is conditionally deleted from maturing β-cells that

begin to express Insulin I (Ins1; Fig. 36A).

104

IM15KO mice were born at Mendelian ratios with no significant differences in body weight at 8

weeks of age (Fig. 41A). Following an IP glucose bolus, IM15KO exhibited severe glucose intolerance (Fig. 41B). Blood samples collected prior to glucose injection and 10 minutes after, demonstrated an impaired increase in plasma insulin levels (Fig. 41C) suggesting impaired functional β-cell mass. To investigate further, eight-week old pancreases were then paraffin embedded and sectioned. Quantification of INS+ and GCG+ immunostained cells within pancreatic sections demonstrated no differences in IM15KO pancreas versus controls, nor were total pancreatic cells significantly different (Figs. 41D-G). Collectively, these data demonstrated that conditional knockout of Med15 in maturing β-cells resulted in compromised β-cell function that was not due to a reduction in overall β-cell number.

5.7 Identification of Gene Expression Changes in IM15KO

As MED15 is a TF co-regulator, gene expression changes in the IM15KO were assessed in order to understand the underlying changes that contributed to the functional impairment of β-cells.

Islets isolated from 8 week old mice were processed for high-throughput RNA sequencing

(RNA-Seq). The dataset generated was validated, and unbiased k-means clustering was able to independently segregated 1M15KO and control islets (Figs. 45; Appendix A; page 120-121).

After multiple comparison correction, 174 genes were identified that displayed significant differences in gene expression in IM15KO islets, with the majority downregulated in IM15KO islets compared to controls (Fig. 45; Appendix A; page 120-121). Gene ontology and KEGG analyses of these genes revealed enrichment in pathways related to insulin exocytosis, protein processing, and diabetes (Figs.42A,B; Figs. 43A-E).

105

Figure 41: IM15KO mice are glucose intolerant despite similar number of β-cells. (A) IM15KO mice at 8 weeks of age have no differences in body weight versus control mice (CTRL). (B) IM15KO mice at 8 weeks of age demonstrated glucose intolerance (n=13-16) and (C) were unable to increase blood insulin concentration following glucose bolus as measured by plasma insulin levels 10 minutes post IP injection (n=10). (D-F) Quantification of pancreas sections from IM15KO showed no differences in (D) INS+ cells (E) GCG+ cells or (F) total pancreatic cells at 8 weeks of age. (G,H) Representative images of islets from 8 week sections stained for insulin (INS; green) and glucagon (GCG; red) from control (G) or IM15KO (H) mice. n=4. *=p≤0.05. (B) Two-way ANOVA with Sidak multiple-comparison test. (C) Two-way ANOVA (A,D-F) Unpaired two-tailed t-tests.

106

Notably, MED15 deletion phenocopied the glucose intolerance observed in adult mice with conditional deletion of NKX6-1 from β-cells134. Cross-examination of the IM15KO RNA-Seq dataset with NKX6-1 RNA-Seq and Chromatin-immunoprecipitation-Sequencing (ChIP-Seq) data demonstrated overlapping gene targets that were significant upon Fisher-exact analysis (Fig.

42C). Finally, immunostaining was performed to verify the presence of NKX6-1 and differential expression of β-cell genes (MAFA, UCN3, GLUT2) that are known markers of β-cell function

(Fig. 43F-I). Immunostaining of these proteins reflected the mRNA expression level differences.

These data strongly suggest that MED15 is an important co-regulator for NKX6-1 in maintaining a functional β-cell state.

5.8 Loss of MED15 Impacts β-cell Metabolism and Mitochondrial Function

As data demonstrated functional β-cell impairment and downregulation of multiple genes involved in glucose metabolism, islets were examined ex-vivo by GSIS. Here, islets from 8 week

IM15KO showed decreased insulin exocytosis following stimulation by 16 mM glucose, but not with 40 mM KCl (Fig. 44A). Total insulin content was similar in IM15KO versus controls (Fig.

44B). This demonstrated that MED15 is required for GSIS. Furthermore, as KCl-stimulated insulin secretion was similar in IM15KO, this implied that loss of Med15 impaired β-cell function upstream of the KATP channel, perhaps by modulating metabolism. Therefore, cellular respiration was assessed by Seahorse XF mitochondrial stress assay. Here, IM15KO islets exhibited reduced oxygen consumption rates (OCR) during basal glucose (2.8 mM) and high glucose (16 mM) measurements, and one data point following administration of the uncoupling agent FCCP (Fig. 44C). These results suggest that loss of MED15 led to reductions in glucose- stimulated cellular respiration and ATP generation, important for glucose-stimulated insulin exocytosis. Further, mild mitochondrial impairment was observed as FCCP decoupled OCR had

107

a minor impact, likely contributing to the reduced OCR during 2.8 mM and 16 mM glucose

conditions. Together, these experiments demonstrate that MED15 is a TF co-regulator of genes

important primarily for glucose metabolism and/or mitochondrial function.

5.9 Discussion

Characterization of Mediator subunits have led to the finding that specific subunits interact with

transcriptional master regulators of cell fate. Genome-wide occupancy experiments have

identified Mediator at OCT4, SOX2, and NANOG sites, postulated to form super-enhancers to regulate ES cell pluripotency291,292. Subunit specific roles include MED1 interaction with PPARγ

to regulate the development of adipose tissue228, and with GATA1 in the regulation of

haematopoiesis293. Other subunits such as MED12, MED13, and MED23 have also been

described in chondrogenesis, smooth muscle development, and adipogenesis. Interestingly,

Mediator subunits MED12 and MED25 have also been demonstrated to interact with TFs known

to be important for pancreas development, such as SOX9294,295. In zebrafish, mot gene encodes for Med12 and mutation studies have established a role in the development of a sub-set of neurons. However, despite recent progress, many Mediator subunits lack known roles in mammalian development.

108

Figure 42: Analysis of differentially expressed genes in RNA-Seq data obtained from 8 week old IM15KO islets. (A) GO enrichment of genes in biological processes regulating insulin exocytosis and processing were identified. (B) KEGG analysis of biological pathways demonstrated enrichment for processes in MODY and type 2 diabetes. (C) Overlap of IM15KO differentially expressed genes with a ChIP-Seq dataset of genes identified within 10kb of NKX6- 1 bound chromatin (SRP015805; 134); n=5. (C) Fisher-exact test; p≤0.05.

109

Figure 43: Protein validation of IM15KO differentially expressed genes. (A-E) RNA-Seq gene expression data showed no change in (A) Nkx6-1. However, NKX6-1 targets (B) Pdx1, (C) Glut2, (D) MafA, and (E) Ucn3 have reduced expression. FPKM: fragments per kilobase of transcript per million mapped reads. * = p≤0.05. Significance obtained from DESeq2 analysis: wald-test with B-H multiple-test correction. Representative immunostaining in 8 week pancreatic sections reflect RNA expression data with no changes in NKX6-1 (F,F’; green) or PDX1 (F-I, F’-I’; red) staining. GLUT2 (G,G’; green), MAFA (H,H’; green), and UCN3 (I,I’; green) were reduced in matched conditions. n=3. Scale bars=50 μm.

110

Figure 44: Glucose-stimulated insulin secretion and cellular respiration profile of IM15KO islets. (A) GSIS on 50 matched islets per well, non-serial measurements from 2.8 mM basal glucose to either 16 mM glucose or 40 mM KCl. IM15KO had reduced insulin exocytosis at 16mM glucose but not at 40 mM KCl compared to controls (n=5-6). (B) Total insulin content from islets was similar. (C) Seahorse XF Mito stress kit demonstrated reduced cellular respiration at 2.8 mM glucose (2.8 mM), 16 mM glucose (16 mM), and FCCP. Olig: Oligomycin, AA+R: Antimycin + Rotenone. n=6, p≤0.05. (A,C) Two-way ANOVA (B) Mann- Whitney U test.

This is the first study to demonstrate a role for the MED15 subunit in mammals. Loss of MED15 in pancreatic progenitors led to significant reductions in the master regulator NEUROG3.

However, pancreatic progenitors were unaffected with similar cell numbers and expression of both PDX1 and SOX9 intact. This suggests that Mediator complex is able to function despite the lack of MED15 in pancreatic progenitors. Additionally, MED15 also appears to be a required co- regulator of the pancreatic endocrine cell lineage, in line with observations that Mediator subunits interact with specific pathways. Conditional deletion of Med15 in endocrine progenitor also displayed defects in β-cell development, while loss of Med15 in maturing β-cells impairs β-

cell function in the adult. Multiple roles for MED15 may be conferred by the presence of a KIX

domain, an important protein-protein interaction site that has known interactions with multiple

TFs229. Additionally, MED15 contains conserved intrinsically disordered regions throughout the

111

protein, proposed to be important for protein-protein interactions296,297. Therefore, similar to

other mediator factors, MED15 likely engages multiple TFs through the course of

development219.

Expression data from RNA-Seq of IM15KO offers a potential MED15 interacting candidate that

is able to explain the different phenotypes observed in the mouse models used, NKX6-1. Maike

Sander’s group134 demonstrated that Nkx6-1 gene deletion in 4 week old mice led to glucose

intolerance and reduced plasma insulin levels, features shared by the IM15KO mice. There was

also significant overlap of genes with altered expression in both mouse models, regulating genes

important for glucose metabolism, insulin exocytosis, and β-cell maturity. The pattern of NKX6-

1 expression in pancreatic progenitor and endocrine progenitors also coincides with MED15

expression during pancreas development, suggesting that it is a plausible interacting factor54,97.

Importantly, Nkx6-1-/- mutants also demonstrate a similar 90.7% reduction of NGN3+ cells in

pancreatic sections from E12.5, although this had recovered to wild-type levels by E14.576. This

was postulated to be due a residual population of SOX9+PDX1+ trunk-like progenitor cells that

may contribute to NGN3 induction. One possibility is that MED15 is a central integrator of

multiple TFs present in pancreatic progenitor cells. Therefore, loss of MED15 as a lineage

regulator would likely prevent even the residual SOX9+PDX1+ trunk-like cells from adopting an

endocrine cell fate, although this remains to be examined. Finally, Ngn3-Cre mediated Nkx6-1

deletion in endocrine progenitors also resulted in a specific decrease in β-cell numbers, mirroring

our MED15 results. As MED15 phenocopies NKX6-1 at multiple stages of β-cell development,

MED15 appears to be a good candidate as an NKX6-1 co-regulator of transcription.

Alternatively, MED15 may be interacting with other factors such as the HNF family members during β-cell development. There is evidence that HNF1α is expressed in developing pancreatic

112

epithelium, and is regulated by NKX6-1298,299. Loss of HNF1α also leads to downregulation of

similar genes observed in the RNA-Seq dataset300, supporting the data that it is regulated

downstream of NKX6-1. Finally, worm MDT-15 has been demonstrated to interact with a

related family member, NHR-49 (HNF4α)231. Therefore, MED15 may be able to interact with

HNF proteins in the pancreas, to regulate a gene set that overlaps the differentially expressed

genes in the NKX6-1 knockout. This remains a hypothesis to be experimentally validated.

Finally, loss of MED15 in maturing β-cells specifically impacted glucose-stimulated insulin

exocytosis. Despite the downregulation of genes known to be involved in insulin biosynthesis

and exocytosis such as Ero1lb, Sytl4, Pdi, and MafA, this study demonstrated that IM15KO retained a similar capacity as control mice to secrete insulin following KCl-stimulated depolarization. Therefore, in the context of Med15 deletion, it appears that glucose metabolism and downstream cellular respiration is the basis of the observed phenotype. This may be in part

due to the downregulation of Slc2a2, encoding for the GLUT2 transporter as well as

mitochondrial genes such as Cox6a2, one of the most differentially expressed genes in the

IM15KO RNA-Seq dataset. Overall, MED15 appears to be indispensable for β-cell development

and further examination may provide insight to identify essential TF partners and networks.

113

6. Conclusions

6.1 Research Summary

Research into the transcriptional regulation of β-cell development has continued to yield novel insights, yet has also showed an increasingly complex landscape of factors and interactions.

Efforts are underway to leverage current understanding of β-cell signaling and TF dynamics to

generate β-cells from ES-cells46,47, alternative tissues153,154, or accelerate existing β-cell survival and proliferation301,302. An improved map of β-cell TF networks could offer a clearer

understanding of the underlying changes that occur during lineage specification. Moreover, it would help to elucidate the reinforced TF networks during the terminal β-cell state. As there is evidence of a destabilized β-cell state in type 2 diabetes, this may offer insight on how to restore a functional β-cell state. Toward these ends, this thesis embarked upon the functional characterization of Sox4 and Med15.

One of the earliest loss-of-function studies of Sox4 highlighted a role for this TF in lineage progression, as Sox4-/- cells had a block at the pro-B-cell stage199. Ongoing studies have

demonstrated additional cell fate specification roles for Sox4 in other tissues such as heart, brain,

and bone201. In the pancreas, the earliest studies demonstrated that pancreas explants from Sox4-/-

mice had severely reduced numbers of β-cells211. Additional characterization in zebrafish

revealed that Sox4b ortholog in fish had important roles in the regulation of pancreatic endocrine

development303, while ENU mutagenesis studies revealed a role in insulin secretion212. These

initial studies suggested that Sox4 may be important in β-cell development, although significant

questions remained:

1. What is the spatiotemporal pattern of SOX4 in vivo?

2. Which cells during pancreatic development require SOX4 and what is its role?

114

3. As a co-activator, what are the co-factors for SOX4 during pancreatic development?

4. What are the targets of SOX4 and how are they regulated by SOX4?

To address these questions, multiple mouse models combined with lineage tracing were utilized

to study Sox4 both during development and in adult β-cells in vivo in chapter 3 and 4. Previous

studies identified Sox4 message within the pancreatic epithelium. Here, the first protein level

validation of SOX4 expression was performed revealing expression within pancreatic progenitor

cells that become restricted to NEUROG3+ cells and later, endocrine cells. In addition, SOX4

was shown to be important for the induction of NEUROG3, and therefore necessary for

endocrine cell specification. In this context, in vitro experiments demonstrated that SOX4 acted

in a cooperative manner with NEUROG3 to induce NEUROG3 expression, although other co-

factors may exist. As high levels of NEUROG3 need to be attained for endocrine cell

specification, initial changes in NEUROG3 due to stochastic differences in Notch signaling may

be amplified through a positive regulatory loop that includes SOX4 and FOXA2286.

Next, as SOX4 becomes enriched in maturing endocrine cells, studies were performed with

Neurog3-Cre to determine if SOX4 continued to promote endocrine development. These

experiments determined that loss of SOX4 led to reduced β-cell numbers. One role for SOX4 in endocrine cells was identified with experiments demonstrating that SOX4 cooperates with

NEUROG3 in the promoter binding and induction of Pax4 and Neurod1. This is similar to the interactions between NEUROG3 and NKX2-2 in regulating TF expression following NEUROG3 induction256. These experiments provide evidence of pleiotropy for Sox4, suggesting that multi-

lineage analysis is important to dissect specific roles during pancreatogenesis. Finally, lineage

tracing of Sox4 deletion following endocrine program activation revealed that a subset of these

cells were devoid of islet hormone staining, although they were positive for the neuroendocrine

115

marker CHGA. This suggested that while these cells were able to initiate endocrine

differentiation, it is incomplete. This may be a potential interim cell state observed in previous

single TF knockout studies where endocrine cell numbers are reduced, although further

experiments would be necessary to establish this. The fate of these cells also remains to be

understood as they appear to persist in E18.5 tissue sections. One potential outcome could

involve apoptosis of these cells shortly following birth, as observed in Neurod1 knockout mice86,

since loss of SOX4 reduced Neurod1 expression. Alternatively, these cells may default to another transient endocrine population found only during development, such as Gastrin+ cells, although these hypotheses remain to be formally tested. Taken together, these studies provide an improved understanding of how SOX4 is required for endocrine development, and the initial fate of cells that have compromised endocrine development.

Experiments presented in Chapter 4 utilized the Pdx1-CreER; Sox4flox/flox loss of function model

to determine if SOX4 bridged the transition to a mature β-cell state. Loss of Sox4 at 6 weeks of

age resulted in no differences in body-weight for either HFD or chow-fed mice. However, HFD mice displayed worsening glucose tolerance over time. Additional experiments identified no differences in glucose-stimulated insulin secretion from size matched HFD-fed mouse islets, nor altered insulin sensitivity. These findings were consolidated upon examination of β-cell number,

with decreased β-cell mass in HFD-fed mouse pancreatic sections. As β-cell mass is maintained via replication, one week of EdU was administered to 7-week old mice to label dividing β-cells.

As predicted, loss of SOX4 decreased β-cell replication, with changes in cell-cycle gene Cdkn1a.

Previous research has demonstrated that a Sox4 loss of function mutant allele impacts stimulated insulin secretion. However, the experiments presented in chapter 4 suggest that loss of Sox4 leads to no differences in insulin secretion, with a primary defect of β-cell replication. This may

116

be due to differences in the models utilized. Previous studies used heterozygous mice carrying a loss of function Sox4 allele derived from ENU mutagenesis. These mice may also harbour undetected mutations in other loci. Further, the studies were paradoxically marked by 100% increased expression of Sox4212,264. One speculation may be that this is a form of developmental compensation for the reduced gene dose of the functional allele. The authors argue that impaired insulin secretion is due to overexpression of Sox4, driving an increase in Stxbp6 expression observed in their mouse model. Stxbp6 is a SNARE protein suggested to be involved with granule exocytosis, as their experiments observed reduced granule fusion. Further, in vitro experiments with overexpression of Sox4 also increased the expression of Stxbp6.

In contrast, the studies in chapter 4 utilized Sox4 deletion that led to decreased expression, reflecting the levels of Sox4 expression observed from human patients with type 2 diabetes. The reduction of cellular proliferation is in concordance with observations that Sox4 hypomorphic alleles result in decreased proliferation and accelerated aging272. Moreover, this was observed only in HFD conditions that simulate a nutrient surplus pre-diabetes state. As obesity or high fat diet produces a setting that results in increased β-cell proliferation169,268, this setting may be sensitive to loss of Sox4. Further, Sox4 has been identified as a positive regulator of the polycomb repressive complex gene Ezh2 in epithelial cells. Interestingly, studies in the pancreas have suggested that loss of Ezh2 expression leads to de-repression of cell-cycle inhibitors of the

INK/ARF family leading to age related reductions in β-cell proliferation. A clear relationship between Sox4-Ezh2-p16 will require additional experiments. Finally, given previous studies identifying Sox4 as a type 2 diabetes related risk-factor24,25,210, the work encompassed in this chapter offers a potential mechanism of how Sox4 impacts β-cell development and function.

117

Although experiments described in chapters 3 and 4 have identified key genes regulated by Sox4 several observations were of note. Specifically, loss of Sox4 leads to the presence of cells that have specified to an endocrine lineage, yet do not show detectable levels of hormone. It is unknown whether these “incompletely” differentiated cells resolve to a specific fate during postnatal maturation, or whether they have reprogrammed to an alternative endocrine fate, such as gastrin+ cells, found within embryonic but not adult pancreas262. As these cells are a subset of

endocrine cells in the Sox4 knockout, FACS sorting followed by single-cell RNA-Seq may offer

improved characterization of these cells. In adult islets, SOX4 appears to partner with yet

unidentified factors to regulate cell replication. While studies here have shown an increase in

Cdkn1a, direct regulation of this gene and others by SOX4 is unknown. The ability to isolate

islets allows for future ChIP-Seq and RNA-Seq experiments to identify bona-fide targets with

SOX4 occupancy. Further, as SOX4 upregulates Ezh2 in epithelial cells, their relationship in the pancreas is unknown. It would be of interest to determine whether SOX4 expression level diminishes with age, leading to loss of Ezh2 repression of cell-cycle inhibitors.

A key component of regulated transcription following TF-enhancer binding is the engagement of

Mediator. Indeed, in yeast, there is an absolute requirement for Mediator complex in transcriptional activation218,304. As described in chapter 1, there is now accumulating evidence

that Mediator complex is comprised of variable subunit composition224,305,306. This is important

as subunit knockouts have illustrated that different TFs and signaling pathways activate specific subunits to coordinate gene responses. This in part may be due to the abundance of intrinsically disordered regions within the mediator complex, allowing for induced-fit binding of multiple

transcription factors307. Curiously, subunit composition appears to be reduced in differentiated

liver cells308, postulated to be another mechanism for the reinforcement of a terminally

118

differentiated state. Thus while many TFs have been identified in pancreas development, their

interaction with Mediator has not been previously shown, nor has a specific subunit been linked

to the distinct cell fate decisions during pancreas development. This thesis now contributes to our

understanding of TF regulation by identifying MED15 as a key mediator subunit that coordinates

pancreatic endocrine development.

Experiments presented in chapter 5 offer the first characterization of MED15 in the pancreas.

MED15 message and protein were detected in the pancreas throughout development, localized to

endocrine cells by E18.5. MED15 knockout mice exhibited severe glucose intolerance and frank

diabetes. Loss of MED15 within pancreatic progenitors led to >95% reduction of the master

regulator NEUROG3 by E15.5, demonstrating a requirement for MED15 in endocrine lineage

induction. However, as MED15 is a co-regulator, its binding partners in this context are

unknown. NKX6-1 is hypothesized to be a likely co-factor of MED15 due to multiple similar

phenotypes in NKX6-1 knockouts in development and adult β-cells. An alternate hypothesis could be that MED15 binds to unique TFs at different stages of cell development. MED1 exhibits this characteristic as it is involved in the specification of multiple tissues, segregated spatiotemporally219. MED15 may also be suitable to bind multiple co-factors as it contains intrinsically disordered domains, which can mediate varied protein-protein binding interactions.

Further, these domains are conserved in MED15 as opposed to other Mediator subunits297,

perhaps as they are important for MED15 specific protein-protein interactions. These

hypotheses remain to be assessed.

Furthermore, the signaling cascades that require MED15 expression remain undefined during

pancreatic development. Some hints towards these cascades may be from previous MED15

studies demonstrating roles in nuclear hormone receptor, TGF-β, and fatty acid signaling230–232.

119

Retinoic acid signaling is important for early pancreas formation48, and MED15 may be a co-

regulator of the intracellular nuclear receptors of retinoic acid, a hypothesis that remains to be

tested. In addition, while MED15 is important in Xenopus TGF-β signaling, whether MED15

plays a similar role during pancreas formation remains to be seen. Activin signaling is required

dorsal pancreas specification285, however the Pdx1-Cre; Med15 deletion in the pancreatic

progenitors did not alter the number of pancreatic cells by E12.5. This may be due to the timing

of the Med15 deletion which occurs following initial bud specification. However, loss of Gdf11,

a gene encoding a TGF-β ligand led to a 3-fold expansion of NEUROG3 by E15.5309, as opposed

to the severe reduction in NEUROG3 cells observed in the Med15 knockout mice. This suggests

that MED15 may have alternate functions in mammalian β-cell development. Finally, Neurog3-

Cre mediated loss within specified endocrine cells also led to a decrease in β-cells. Despite the reductions in endocrine cells, neither pancreatic progenitors nor total pancreatic cell numbers were affected in either model, suggesting that MED15 has endocrine lineage specific roles.

Additional work utilizing Med15 deletion in maturing β-cells with Ins1-Cre revealed a requirement for MED15 for β-cell functionality. Intriguingly, Med15 deletion in this context led to a significant decrease in only 174 genes, many of which have been demonstrated to be related to β-cell function. This again supports a highly specific role for MED15 in the regulation of an endocrine cell state as opposed to being a general docking subunit in the Mediator complex. Loss of MED15 in maturing β-cells did not decrease β-cell mass, yet compromised glucose metabolism. Interestingly, despite the downregulation of numerous genes, insulin protein expression and secretion following KCl appeared to be intact. These results suggest that decreased cellular respiration is the principal cause for compromised β-cell function. Rescue experiments to restore cellular respiration in the setting of an altered β-cell state would be

120

interesting to identify whether it is sufficient to restore β-cell functionality despite perturbed

identity. In summary, these experiments represent the first characterization of MED15 in mammals. It is a critical TF co-regulator important for the specification of an endocrine cell state, and has potential for further studies to identify binding partners and its role in adult β-cell

state.

6.2 Future Directions

The data presented in this thesis show key roles for both SOX4 and MED15 in β-cell

development. As these are some of the first studies to characterize these proteins and their roles,

future research avenues remain diverse. For example, all SOXC family members are present

within the pancreas. Other studies of pancreatic factors such as NKX6-1 and NKX6-2

demonstrate the potential for functional redundancy310, and triple-SOXC family knockouts have

shown overlapping roles of SOXC family members during development201. Thus while beyond the original scope of this thesis, use of combined of SOXC gene deletion models could determine

whether compensation occurs to allow for partial NEUROG3 induction during development, or

β-cell proliferation in adult cells. Other avenues of research include further assessment of the

SOX4 knockout. In adults, rescue experiments in ex vivo islets with siRNAs targeted against

Cdkn1a or other cell-cycle genes would help determine the critical factor downstream of SOX4.

Additionally, expression analysis of hESC differentiation noted increased expression of SOX4 as

differentiation progressed towards pancreatic progenitor like cells. Several TFs identified in

mouse β-cell function do not have identical expression or roles in human β-cells, for example

MAFB311. shRNA knockdown of Sox4 in human islets and additional work in hES differentiation

could determine whether SOX4 has conserved functions. Previous GWAS studies pointed

121

towards a likely role for SOX4 in humans, as studies demonstrated a modest increase in odds- ratio for a Cdkal1 SNP, now associated with a Sox4 enhancer.

Future studies for MED15 will include identification of MED15 binding co-factors during pancreas development, using strategies such as co-immunoprecipitation experiments. This may be followed by ChIP experiments at the NEUROG3 promoter to determine co-enrichment of

MED15 and its binding partners. This may help to identify upstream signaling cascades important for NEUROG3 induction. Initial metabolic studies have also identified glucose metabolism as a key defect in Med15 β-cell knockouts. As GLUT2 expression is reduced, future experiments assessing glucose import and rescue with anaplerotic intermediates may pinpoint the root cause of β-cell dysfunction in Med15 knockouts. Moreover, in our Ins1-Cre model, recombination of Med15 floxed allele is detectable since the time of birth238. As β-cells exhibit

postnatal maturation, one question is whether loss of Med15 in β-cells after they have reached

functional maturity, would alter function and gene expression? Finally, Mediator subunits are

found in regions of super-enhancers, defined as enhancer regions that are enriched for master

regulator TF and Mediator binding. These regions are associated with genes that are critical for

cell identity. As MED15 appears to be indispensable for β-cell function, it may serve as the integrative hub for β-cell TFs to establish β-cell identity. Future ChIP-Seq experiments could identify MED15 DNA binding regions for the differentially expressed genes identified in

Chapter 5. This may be followed by chromatin conformation capture with genome sequencing to identify enhancers, and their inferred cognate TFs, that operate in a network to regulate genes important for a functional β-cell state.

Although the studies in chapter 5 were limited to the MED15 subunit, other subunits are likely important for pancreatic progenitor specification as well as acinar and ductal development, as

122

these tissues are unaffected in Med15 knockouts. In addition, MED15 expression appears to be restricted to endocrine cells by E18.5 (Fig. 36). As anecdotal evidence exists for restrictions in

Mediator subunit expression with increasing cellular differentiation, a profile of Mediator subunits during differentiation may be informative as a marker for maturation into a functional β- cell state. However, how mediator subunit expression is regulated in the pancreas is unknown.

Identification of non-endocrine Mediators or factors that regulate Mediator subunit expression may allow for the development of small molecule drugs to prohibit the function or expression of

Mediator subunits required for non-endocrine development, curtailing the generation of unwanted cell-types during ES differentiation. Finally, MED15 sequence exhibits modest conservation between mouse and their human counterparts, and an examination of MED15 in human β-cells will be critical.

6.3 Final Thoughts

Research in the β-cell development field continues to progress rapidly, with the identification of numerous pathways involved in both β-cell development and pathogenesis of type 2 diabetes.

Phase I/II FDA clinical trial is now underway for transplantation of differentiated hESCs

(ClinicalTrials.gov ID: NCT02239354), representing one of the first potential long-term treatment options for millions of patients with diabetes. Continued research into β-cell transcription provides the foundation for this and other strategies. Identification of SOX4 and

MED15, critical TFs for β-cell development and function, opens up new nodes in the β-cell

landscape regulatory landscape. It is the hope that the work encompassed within this thesis

contributes to the understanding of β-cell formation, failure, and ultimately the development of

effective treatments for patients who live with diabetes.

123

Bibliography

1. International Diabetes Federation. IDF Diabetes Atlas, 7 edition. International Diabetes Federation (2015). doi:2-930229-80-2 2. American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 32 Suppl 1, S62-7 (2009). 3. Diabetes, by age group and sex (Number of persons). Statistics Canada (2014). Available at: http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/health53a-eng.htm. 4. Government of Canada, P. H. A. of C. Diabetes in Canada: Facts and figures from a public health perspective - Public Health Agency of Canada. 5. Eknoyan, G. & Nagy, J. A history of diabetes mellitus or how a disease of the kidneys evolved into a kidney disease. Adv. Chronic Kidney Dis. 12, 223–9 (2005). 6. King, K. M. & Rubin, G. A history of diabetes: from antiquity to discovering insulin. Br. J. Nurs. 12, 1091–5 (2003). 7. Rorsman, P. & Braun, M. Regulation of insulin secretion in human pancreatic islets. Annu. Rev. Physiol. 75, 155–79 (2013). 8. MacDonald, P. E., Joseph, J. W. & Rorsman, P. Glucose-sensing mechanisms in pancreatic beta-cells. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 360, 2211–25 (2005). 9. De Vos, A. et al. Human and rat beta cells differ in glucose transporter but not in glucokinase gene expression. J. Clin. Invest. 96, 2489–95 (1995). 10. Fu, Z., Gilbert, E. R. & Liu, D. Regulation of insulin synthesis and secretion and pancreatic Beta-cell dysfunction in diabetes. Curr. Diabetes Rev. 9, 25–53 (2013). 11. Taniguchi, C. M., Emanuelli, B. & Kahn, C. R. Critical nodes in signalling pathways: insights into insulin action. Nat. Rev. Mol. Cell Biol. 7, 85–96 (2006). 12. Cersosimo, E. et al. The effect of graded doses of insulin on peripheral glucose uptake and lactate release in cancer cachexia. Surgery 109, 459–67 (1991). 13. Saltiel, A. R. & Kahn, C. R. Insulin signalling and the regulation of glucose and lipid metabolism. Nature 414, 799–806 (2001). 14. Araki, E. et al. Alternative pathway of insulin signalling in mice with targeted disruption of the IRS-1 gene. Nature 372, 186–90 (1994). 15. Tamemoto, H. et al. Insulin resistance and growth retardation in mice lacking insulin receptor substrate-1. Nature 372, 182–6 (1994). 16. Withers, D. J. et al. Disruption of IRS-2 causes type 2 diabetes in mice. Nature 391, 900– 4 (1998).

124

17. Kido, Y. et al. Tissue-specific insulin resistance in mice with mutations in the insulin receptor, IRS-1, and IRS-2. J. Clin. Invest. 105, 199–205 (2000). 18. American Diabetes Association. Standards of medical care in diabetes--2013. Diabetes Care 36 Suppl 1, S11-66 (2013). 19. Ashcroft, F. M. & Rorsman, P. Diabetes mellitus and the β cell: the last ten years. Cell 148, 1160–71 (2012). 20. Pihoker, C., Gilliam, L. K., Hampe, C. S. & Lernmark, A. Autoantibodies in diabetes. Diabetes 54 Suppl 2, S52-61 (2005). 21. van Belle, T. L., Coppieters, K. T. & von Herrath, M. G. Type 1 diabetes: etiology, immunology, and therapeutic strategies. Physiol. Rev. 91, 79–118 (2011). 22. Barrett, J. C. et al. Genome-wide association study and meta-analysis find that over 40 loci affect risk of type 1 diabetes. Nat. Genet. 41, 703–7 (2009). 23. Rodriguez-Calvo, T., Sabouri, S., Anquetil, F. & von Herrath, M. G. The viral paradigm in type 1 diabetes: Who are the main suspects? Autoimmun. Rev. 15, 964–9 (2016). 24. Scott, L. J. et al. A genome-wide association study of type 2 diabetes in Finns detects multiple susceptibility variants. Science 316, 1341–5 (2007). 25. Steinthorsdottir, V. et al. A variant in CDKAL1 influences insulin response and risk of type 2 diabetes. Nat. Genet. 39, 770–5 (2007). 26. Eckel, R. H. et al. Obesity and type 2 diabetes: what can be unified and what needs to be individualized? Diabetes Care 34, 1424–30 (2011). 27. Kahn, S. E., Hull, R. L. & Utzschneider, K. M. Mechanisms linking obesity to insulin resistance and type 2 diabetes. Nature 444, 840–6 (2006). 28. Weyer, C., Bogardus, C., Mott, D. M. & Pratley, R. E. The natural history of insulin secretory dysfunction and insulin resistance in the pathogenesis of type 2 diabetes mellitus. J. Clin. Invest. 104, 787–94 (1999). 29. Matsubara, T. et al. PGRN is a key adipokine mediating high fat diet-induced insulin resistance and obesity through IL-6 in adipose tissue. Cell Metab. 15, 38–50 (2012). 30. Sell, H., Habich, C. & Eckel, J. Adaptive immunity in obesity and insulin resistance. Nat. Rev. Endocrinol. 8, 709–16 (2012). 31. Weir, G. C. & Bonner-Weir, S. Five stages of evolving beta-cell dysfunction during progression to diabetes. Diabetes 53 Suppl 3, S16-21 (2004). 32. Mehran, A. E. et al. Hyperinsulinemia drives diet-induced obesity independently of brain insulin production. Cell Metab. 16, 723–37 (2012). 33. Poitout, V. & Robertson, R. P. Glucolipotoxicity: Fuel excess and ??-cell dysfunction. Endocrine Reviews 29, 351–366 (2008).

125

34. Eizirik, D. L., Cardozo, A. K. & Cnop, M. The role for endoplasmic reticulum stress in diabetes mellitus. Endocr. Rev. 29, 42–61 (2008). 35. Laybutt, D. R. et al. Endoplasmic reticulum stress contributes to beta cell apoptosis in type 2 diabetes. Diabetologia 50, 752–63 (2007). 36. Stoffers, D. A., Zinkin, N. T., Stanojevic, V., Clarke, W. L. & Habener, J. F. Pancreatic agenesis attributable to a single nucleotide deletion in the human IPF1 gene coding sequence. Nat. Genet. 15, 106–10 (1997). 37. Sellick, G. S. et al. Mutations in PTF1A cause pancreatic and cerebellar agenesis. Nat. Genet. 36, 1301–5 (2004). 38. Rubio-Cabezas, O. et al. Permanent Neonatal Diabetes and Enteric Anendocrinosis Associated With Biallelic Mutations in NEUROG3. Diabetes 60, 1349–53 (2011). 39. McGrath, P. S., Watson, C. L., Ingram, C., Helmrath, M. A. & Wells, J. M. The Basic Helix-Loop-Helix Transcription Factor NEUROG3 Is Required for Development of the Human Endocrine Pancreas. Diabetes 64, 2497–505 (2015). 40. Drucker, D. J. Incretin action in the pancreas: potential promise, possible perils, and pathological pitfalls. Diabetes 62, 3316–23 (2013). 41. Proks, P., Reimann, F., Green, N., Gribble, F. & Ashcroft, F. Sulfonylurea stimulation of insulin secretion. Diabetes 51 Suppl 3, S368-76 (2002). 42. Shapiro, A. M. et al. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N. Engl. J. Med. 343, 230–8 (2000). 43. McCall, M. & Shapiro, a M. J. Update on islet transplantation. Cold Spring Harb. Perspect. Med. 2, a007823 (2012). 44. Fiaschi-taesch, N. M. et al. Induction of Human Beta-Cell Proliferation and Engraftment Using a Single G1/S Regulatory Molecule, cdk6. Diabetes 59, 1926–36 (2010). 45. Zhou, Q., Brown, J., Kanarek, A., Rajagopal, J. & Melton, D. A. In vivo reprogramming of adult pancreatic exocrine cells to beta-cells. Nature 455, 627–32 (2008). 46. Rezania, A. et al. Reversal of diabetes with insulin-producing cells derived in vitro from human pluripotent stem cells. Nat. Biotechnol. 32, 1121–33 (2014). 47. Schulz, T. C. et al. A scalable system for production of functional pancreatic progenitors from human embryonic stem cells. PLoS One 7, e37004 (2012). 48. Martín, M. et al. Dorsal pancreas agenesis in retinoic acid-deficient Raldh2 mutant mice. Dev. Biol. 284, 399–411 (2005). 49. Hebrok, M., Kim, S. K., St Jacques, B., McMahon, A. P. & Melton, D. A. Regulation of pancreas development by hedgehog signaling. Development 127, 4905–13 (2000).

126

50. Kim, S. K. & Melton, D. A. Pancreas development is promoted by cyclopamine, a hedgehog signaling inhibitor. Proc. Natl. Acad. Sci. U. S. A. 95, 13036–41 (1998). 51. Lammert, E., Cleaver, O. & Melton, D. Induction of pancreatic differentiation by signals from blood vessels. Science 294, 564–7 (2001). 52. Deutsch, G., Jung, J., Zheng, M., Lóra, J. & Zaret, K. S. A bipotential precursor population for pancreas and liver within the embryonic endoderm. Development 128, 871– 81 (2001). 53. Wessells, N. K. & Cohen, J. I. I. Early Pancreas Organogenesis : and Mass Effects. 270, 237–270 (1967). 54. Jørgensen, M. C. et al. An illustrated review of early pancreas development in the mouse. Endocr. Rev. 28, 685–705 (2007). 55. Ahlgren, U., Jonsson, J., Jonsson, L., Simu, K. & Edlund, H. beta-cell-specific inactivation of the mouse Ipf1/Pdx1 gene results in loss of the beta-cell phenotype and maturity onset diabetes. Genes Dev. 12, 1763–8 (1998). 56. Ohlsson, H., Karlsson, K. & Edlund, T. IPF1, a homeodomain-containing transactivator of the insulin gene. EMBO J. 12, 4251–9 (1993). 57. Kawaguchi, Y. et al. The role of the transcriptional regulator Ptf1a in converting intestinal to pancreatic progenitors. Nat. Genet. 32, 128–34 (2002). 58. Lynn, F. C. et al. Sox9 coordinates a transcriptional network in pancreatic progenitor cells. Proc. Natl. Acad. Sci. U. S. A. 104, 10500–5 (2007). 59. Seymour, P. A. et al. SOX9 is required for maintenance of the pancreatic progenitor cell pool. Proc. Natl. Acad. Sci. U. S. A. 104, 1865–70 (2007). 60. Harrison, K. a, Thaler, J., Pfaff, S. L., Gu, H. & Kehrl, J. H. Pancreas dorsal lobe agenesis and abnormal islets of Langerhans in Hlxb9-deficient mice. Nat. Genet. 23, 71–5 (1999). 61. Xuan, S. et al. Pancreas-specific deletion of mouse Gata4 and Gata6 causes pancreatic agenesis. J. Clin. Invest. 122, 3516–28 (2012). 62. Gao, N. et al. Dynamic regulation of Pdx1 enhancers by Foxa1 and Foxa2 is essential for pancreas development. Genes Dev. 22, 3435–48 (2008). 63. Jensen, J. et al. Control of endodermal endocrine development by Hes-1. Nat. Genet. 24, 36–44 (2000). 64. Sander, M. et al. Homeobox gene Nkx6.1 lies downstream of Nkx2.2 in the major pathway of beta-cell formation in the pancreas. Development 127, 5533–40 (2000). 65. Villasenor, A., Chong, D. C., Henkemeyer, M. & Cleaver, O. Epithelial dynamics of pancreatic branching morphogenesis. Development 137, 4295–305 (2010). 66. Rutter, W. J. et al. Regulation of specific protein synthesis in cytodifferentiation. J. Cell.

127

Physiol. 72, Suppl 1:1-18 (1968). 67. Hick, A.-C. et al. Mechanism of primitive duct formation in the pancreas and submandibular glands: a role for SDF-1. BMC Dev. Biol. 9, 66 (2009). 68. Bhushan, a et al. Fgf10 is essential for maintaining the proliferative capacity of epithelial progenitor cells during early pancreatic organogenesis. Development 128, 5109–17 (2001). 69. Landsman, L. et al. Pancreatic mesenchyme regulates epithelial organogenesis throughout development. PLoS Biol. 9, e1001143 (2011). 70. Seymour, P. A. et al. A Sox9/Fgf feed-forward loop maintains pancreatic organ identity. Development 139, 3363–72 (2012). 71. Georgia, S., Soliz, R., Li, M., Zhang, P. & Bhushan, A. p57 and Hes1 coordinate cell cycle exit with self-renewal of pancreatic progenitors. Dev. Biol. 298, 22–31 (2006). 72. Bankaitis, E. D., Bechard, M. E. & Wright, C. V. E. E. Feedback control of growth, differentiation, and morphogenesis of pancreatic endocrine progenitors in an epithelial plexus niche. Genes Dev. 29, 2203–16 (2015). 73. Kesavan, G. et al. Cdc42-mediated tubulogenesis controls cell specification. Cell 139, 791–801 (2009). 74. Solar, M. et al. Pancreatic exocrine duct cells give rise to insulin-producing beta cells during embryogenesis but not after birth. Dev. Cell 17, 849–60 (2009). 75. Zhou, Q. et al. A multipotent progenitor domain guides pancreatic organogenesis. Dev. Cell 13, 103–14 (2007). 76. Schaffer, A. E., Freude, K. K., Nelson, S. B. & Sander, M. Nkx6 transcription factors and Ptf1a function as antagonistic lineage determinants in multipotent pancreatic progenitors. Dev. Cell 18, 1022–9 (2010). 77. Gradwohl, G., Dierich, A., LeMeur, M. & Guillemot, F. neurogenin3 is required for the development of the four endocrine cell lineages of the pancreas. Proc. Natl. Acad. Sci. U. S. A. 97, 1607–11 (2000). 78. Kopp, J. L. et al. Sox9+ ductal cells are multipotent progenitors throughout development but do not produce new endocrine cells in the normal or injured adult pancreas. Development 138, 653–65 (2011). 79. Apelqvist, A. et al. Notch signalling controls pancreatic cell differentiation. Nature 400, 877–81 (1999). 80. Qu, X. et al. Notch-mediated post-translational control of Ngn3 protein stability regulates pancreatic patterning and cell fate commitment. Dev. Biol. 376, 1–12 (2013). 81. Murtaugh, L. C., Stanger, B. Z., Kwan, K. M. & Melton, D. A. Notch signaling controls multiple steps of pancreatic differentiation. Proc. Natl. Acad. Sci. U. S. A. 100, 14920–5 128

(2003). 82. Wang, S. et al. Sustained Neurog3 expression in hormone-expressing islet cells is required for endocrine maturation and function. Proc. Natl. Acad. Sci. U. S. A. 106, 9715–20 (2009). 83. Miyatsuka, T., Li, Z. & German, M. S. Chronology of islet differentiation revealed by temporal cell labeling. Diabetes 58, 1863–8 (2009). 84. Johansson, K. A. et al. Temporal control of neurogenin3 activity in pancreas progenitors reveals competence windows for the generation of different endocrine cell types. Dev. Cell 12, 457–65 (2007). 85. Pan, F. C. & Wright, C. Pancreas organogenesis: from bud to plexus to gland. Dev. Dyn. 240, 530–65 (2011). 86. Naya, F. J. et al. Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in BETA2/neuroD-deficient mice. Genes Dev. 11, 2323– 34 (1997). 87. Gasa, R. et al. Proendocrine genes coordinate the pancreatic islet differentiation program in vitro. Proc. Natl. Acad. Sci. U. S. A. 101, 13245–50 (2004). 88. Schwitzgebel, V. M. et al. Expression of neurogenin3 reveals an islet cell precursor population in the pancreas. Development 127, 3533–42 (2000). 89. Smith, S. B. et al. Rfx6 directs islet formation and insulin production in mice and humans. Nature 463, 775–80 (2010). 90. St-Onge, L., Sosa-Pineda, B., Chowdhury, K., Mansouri, A. & Gruss, P. Pax6 is required for differentiation of glucagon-producing alpha-cells in mouse pancreas. Nature 387, 406– 9 (1997). 91. Sander, M. et al. Genetic analysis reveals that PAX6 is required for normal transcription of pancreatic hormone genes and islet development. Genes Dev. 11, 1662–73 (1997). 92. Ahlgren, U., Pfaff, S. L., Jessell, T. M., Edlund, T. & Edlund, H. Independent requirement for ISL1 in formation of pancreatic mesenchyme and islet cells. Nature 385, 257–60 (1997). 93. Marsich, E., Vetere, A., Di Piazza, M., Tell, G. & Paoletti, S. The PAX6 gene is activated by the basic helix-loop-helix transcription factor NeuroD/BETA2. Biochem. J. 376, 707– 15 (2003). 94. Sussel, L. et al. Mice lacking the homeodomain transcription factor Nkx2.2 have diabetes due to arrested differentiation of pancreatic beta cells. Development 125, 2213–21 (1998). 95. Prado, C. L., Pugh-Bernard, A. E., Elghazi, L., Sosa-Pineda, B. & Sussel, L. Ghrelin cells replace insulin-producing beta cells in two mouse models of pancreas development. Proc. Natl. Acad. Sci. U. S. A. 101, 2924–9 (2004).

129

96. Wang, J. et al. The concerted activities of Pax4 and Nkx2.2 are essential to initiate pancreatic beta-cell differentiation. Dev. Biol. 266, 178–89 (2004). 97. Schaffer, A. E. et al. Nkx6.1 controls a gene regulatory network required for establishing and maintaining pancreatic Beta cell identity. PLoS Genet. 9, e1003274 (2013). 98. Sosa-Pineda, B., Chowdhury, K., Torres, M., Oliver, G. & Gruss, P. The Pax4 gene is essential for differentiation of insulin-producing beta cells in the mammalian pancreas. Nature 386, 399–402 (1997). 99. Collombat, P. et al. Opposing actions of Arx and Pax4 in endocrine pancreas development. Genes Dev. 17, 2591–603 (2003). 100. Collombat, P. et al. The simultaneous loss of Arx and Pax4 genes promotes a somatostatin-producing cell fate specification at the expense of the alpha- and beta-cell lineages in the mouse endocrine pancreas. Development 132, 2969–80 (2005). 101. Artner, I. et al. MafB is required for islet beta cell maturation. Proc. Natl. Acad. Sci. U. S. A. 104, 3853–8 (2007). 102. Artner, I. et al. MafA and MafB regulate genes critical to beta-cells in a unique temporal manner. Diabetes 59, 2530–9 (2010). 103. Nishimura, W. et al. A switch from MafB to MafA expression accompanies differentiation to pancreatic beta-cells. Dev. Biol. 293, 526–39 (2006). 104. Wang, H., Brun, T., Kataoka, K., Sharma, a J. & Wollheim, C. B. MAFA controls genes implicated in insulin biosynthesis and secretion. Diabetologia 50, 348–58 (2007). 105. Matsuoka, T. et al. Members of the large Maf transcription family regulate insulin gene transcription in islet beta cells. Mol. Cell. Biol. 23, 6049–62 (2003). 106. Sharma, A. et al. The NeuroD1/BETA2 sequences essential for insulin gene transcription colocalize with those necessary for neurogenesis and p300/CREB binding protein binding. Mol. Cell. Biol. 19, 704–13 (1999). 107. Gao, N. et al. Foxa2 controls vesicle docking and insulin secretion in mature Beta cells. Cell Metab. 6, 267–79 (2007). 108. Jennings, R. E., Berry, A. A., Strutt, J. P., Gerrard, D. T. & Hanley, N. A. Human pancreas development. Development 142, 3126–37 (2015). 109. Jennings, R. E. et al. Development of the human pancreas from foregut to endocrine commitment. Diabetes 62, 3514–22 (2013). 110. Pagliuca, F. W. et al. Generation of Functional Human Pancreatic β Cells In Vitro. Cell 159, 428–439 (2014). 111. Bliss, C. R. & Sharp, G. W. Glucose-induced insulin release in islets of young rats: time- dependent potentiation and effects of 2-bromostearate. Am. J. Physiol. 263, E890-6 (1992). 130

112. Grasso, S., Messina, A., Saporito, N. & Reitano, G. Serum-insulin response to glucose and aminoacids in the premature infant. Lancet (London, England) 2, 755–6 (1968). 113. Hole, R. L., Pian-Smith, M. C. & Sharp, G. W. Development of the biphasic response to glucose in fetal and neonatal rat pancreas. Am. J. Physiol. 254, E167-74 (1988). 114. Otonkoski, T., Andersson, S., Knip, M. & Simell, O. Maturation of insulin response to glucose during human fetal and neonatal development. Studies with perifusion of pancreatic isletlike cell clusters. Diabetes 37, 286–91 (1988). 115. Rozzo, A., Meneghel-Rozzo, T., Delakorda, S. L., Yang, S.-B. & Rupnik, M. Exocytosis of insulin: in vivo maturation of mouse endocrine pancreas. Ann. N. Y. Acad. Sci. 1152, 53–62 (2009). 116. Rorsman, P. et al. Failure of glucose to elicit a normal secretory response in fetal pancreatic beta cells results from glucose insensitivity of the ATP-regulated K+ channels. Proc. Natl. Acad. Sci. U. S. A. 86, 4505–9 (1989). 117. Gauthier, B. R. et al. PDX1 deficiency causes mitochondrial dysfunction and defective insulin secretion through TFAM suppression. Cell Metab. 10, 110–8 (2009). 118. Prentki, M., Matschinsky, F. M. & Madiraju, S. R. M. Metabolic signaling in fuel-induced insulin secretion. Cell Metab. 18, 162–85 (2013). 119. Rutter, G. A., Pullen, T. J., Hodson, D. J. & Martinez-Sanchez, A. Pancreatic β-cell identity, glucose sensing and the control of insulin secretion. Biochem. J. 466, 203–18 (2015). 120. Stolovich-Rain, M. et al. Weaning triggers a maturation step of pancreatic β cells. Dev. Cell 32, 535–45 (2015). 121. Thorrez, L. et al. Tissue-specific disallowance of housekeeping genes: the other face of cell differentiation. Genome Res. 21, 95–105 (2011). 122. Jermendy, A. et al. Rat neonatal beta cells lack the specialised metabolic phenotype of mature beta cells. Diabetologia 54, 594–604 (2011). 123. Lemaire, K., Thorrez, L. & Schuit, F. Disallowed and Allowed Gene Expression: Two Faces of Mature Islet Beta Cells. Annu. Rev. Nutr. 36, 45–71 (2016). 124. Blau, H. M. Differentiation requires continuous active control. Annu. Rev. Biochem. 61, 1213–30 (1992). 125. Holmberg, J. & Perlmann, T. Maintaining differentiated cellular identity. Nat. Rev. Genet. 13, 429–39 (2012). 126. Briggs, R. & King, T. J. Transplantation of Living Nuclei From Blastula Cells into Enucleated Frogs’ Eggs. Proc. Natl. Acad. Sci. U. S. A. 38, 455–63 (1952). 127. Hochedlinger, K. & Jaenisch, R. Monoclonal mice generated by nuclear transfer from mature B and T donor cells. Nature 415, 1035–8 (2002). 131

128. Takahashi, K. & Yamanaka, S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126, 663–76 (2006). 129. Hobert, O. Regulation of terminal differentiation programs in the nervous system. Annu. Rev. Cell Dev. Biol. 27, 681–96 (2011). 130. Szabat, M. et al. Maintenance of β-cell maturity and plasticity in the adult pancreas: developmental biology concepts in adult physiology. Diabetes 61, 1365–71 (2012). 131. Tennant, B. R. et al. Identification and analysis of murine pancreatic islet enhancers. Diabetologia 56, 542–52 (2013). 132. Avrahami, D. et al. Aging-Dependent Demethylation of Regulatory Elements Correlates with Chromatin State and Improved β Cell Function. Cell Metab. 22, 619–32 (2015). 133. Gao, T. et al. Pdx1 maintains β cell identity and function by repressing an α cell program. Cell Metab. 19, 259–71 (2014). 134. Taylor, B. L., Liu, F.-F. & Sander, M. Nkx6.1 is essential for maintaining the functional state of pancreatic beta cells. Cell Rep. 4, 1262–75 (2013). 135. Hang, Y. et al. The MafA transcription factor becomes essential to islet β-cells soon after birth. Diabetes 63, 1994–2005 (2014). 136. Talchai, C., Xuan, S., Lin, H. V., Sussel, L. & Accili, D. Pancreatic β cell dedifferentiation as a mechanism of diabetic β cell failure. Cell 150, 1223–34 (2012). 137. Kjørholt, C., Akerfeldt, M. C., Biden, T. J. & Laybutt, D. R. Chronic hyperglycemia, independent of plasma lipid levels, is sufficient for the loss of beta-cell differentiation and secretory function in the db/db mouse model of diabetes. Diabetes 54, 2755–63 (2005). 138. Poitout, V., Olson, L. K. & Robertson, R. P. Chronic exposure of betaTC-6 cells to supraphysiologic concentrations of glucose decreases binding of the RIPE3b1 insulin gene transcription activator. J. Clin. Invest. 97, 1041–6 (1996). 139. Szabat, M., Johnson, J. D. & Piret, J. M. Reciprocal modulation of adult beta cell maturity by activin A and follistatin. Diabetologia 53, 1680–9 (2010). 140. Cinti, F. et al. Evidence of β-Cell Dedifferentiation in Human Type 2 Diabetes. J. Clin. Endocrinol. Metab. 101, 1044–54 (2016). 141. Vanhoose, A. M. et al. MafA and MafB regulate Pdx1 transcription through the Area II control region in pancreatic beta cells. J. Biol. Chem. 283, 22612–9 (2008). 142. Templin, A. T., Maier, B., Tersey, S. A., Hatanaka, M. & Mirmira, R. G. Maintenance of Pdx1 mRNA translation in islet β-cells during the unfolded protein response. Mol. Endocrinol. 28, 1820–30 (2014). 143. Szabat, M. et al. Reduced Insulin Production Relieves Endoplasmic Reticulum Stress and Induces β Cell Proliferation. Cell Metab. 23, 179–93 (2016).

132

144. Bensellam, M., Laybutt, D. R. & Jonas, J.-C. The molecular mechanisms of pancreatic β- cell glucotoxicity: recent findings and future research directions. Mol. Cell. Endocrinol. 364, 1–27 (2012). 145. Wang, Y. J. et al. Single-Cell Transcriptomics of the Human Endocrine Pancreas. Diabetes 65, 3028–38 (2016). 146. Bader, E. et al. Identification of proliferative and mature β-cells in the islets of Langerhans. Nature 535, 430–4 (2016). 147. Johnston, N. R. et al. Beta Cell Hubs Dictate Pancreatic Islet Responses to Glucose. Cell Metab. 24, 389–401 (2016). 148. Kroon, E. et al. Pancreatic endoderm derived from human embryonic stem cells generates glucose-responsive insulin-secreting cells in vivo. Nat. Biotechnol. 26, 443–52 (2008). 149. D’Amour, K. a et al. Efficient differentiation of human embryonic stem cells to definitive endoderm. Nat. Biotechnol. 23, 1534–41 (2005). 150. Rezania, A. et al. Maturation of human embryonic stem cell-derived pancreatic progenitors into functional islets capable of treating pre-existing diabetes in mice. Diabetes 61, 2016–29 (2012). 151. Johnson, J. D. The quest to make fully functional human pancreatic beta cells from embryonic stem cells: climbing a mountain in the clouds. Diabetologia 59, 2047–57 (2016). 152. Li, W. et al. Long-term persistence and development of induced pancreatic beta cells generated by lineage conversion of acinar cells. Nat. Biotechnol. 32, 1223–30 (2014). 153. Wei, R. & Hong, T. Lineage Reprogramming: A Promising Road for Pancreatic β Cell Regeneration. Trends Endocrinol. Metab. 27, 163–76 (2016). 154. Li, W. et al. In vivo reprogramming of pancreatic acinar cells to three islet endocrine subtypes. Elife 3, e01846 (2014). 155. Movassat, J., Saulnier, C., Serradas, P. & Portha, B. Impaired development of pancreatic beta-cell mass is a primary event during the progression to diabetes in the GK rat. Diabetologia 40, 916–25 (1997). 156. Dor, Y., Brown, J., Martinez, O. I. & Melton, D. A. Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature 429, 41–6 (2004). 157. Teta, M., Rankin, M. M., Long, S. Y., Stein, G. M. & Kushner, J. A. Growth and regeneration of adult beta cells does not involve specialized progenitors. Dev. Cell 12, 817–26 (2007). 158. Kushner, J. A. Beta-cell growth: an unusual paradigm of organogenesis that is cyclin D2/Cdk4 dependent. Cell Cycle 5, 234–7 (2006). 159. Perl, S. et al. Significant human beta-cell turnover is limited to the first three decades of

133

life as determined by in vivo thymidine analog incorporation and radiocarbon dating. J. Clin. Endocrinol. Metab. 95, E234-9 (2010). 160. Finegood, D. T., Scaglia, L. & Bonner-Weir, S. Dynamics of beta-cell mass in the growing rat pancreas. Estimation with a simple mathematical model. Diabetes 44, 249–56 (1995). 161. Teta, M., Long, S. Y., Wartschow, L. M., Rankin, M. M. & Kushner, J. A. Very slow turnover of beta-cells in aged adult mice. Diabetes 54, 2557–67 (2005). 162. Bonner-Weir, S. et al. Transdifferentiation of pancreatic ductal cells to endocrine beta- cells. Biochem. Soc. Trans. 36, 353–6 (2008). 163. Xiao, X. et al. No evidence for β cell neogenesis in murine adult pancreas. J. Clin. Invest. 123, 2207–17 (2013). 164. Van de Casteele, M. et al. Neurogenin 3+ cells contribute to β-cell neogenesis and proliferation in injured adult mouse pancreas. Cell Death Dis. 4, e523 (2013). 165. Heit, J. J., Karnik, S. K. & Kim, S. K. Intrinsic regulators of pancreatic beta-cell proliferation. Annu. Rev. Cell Dev. Biol. 22, 311–38 (2006). 166. Nir, T., Melton, D. A. & Dor, Y. Recovery from diabetes in mice by beta cell regeneration. J. Clin. Invest. 117, 2553–61 (2007). 167. Rankin, M. M. & Kushner, J. A. Adaptive beta-cell proliferation is severely restricted with advanced age. Diabetes 58, 1365–72 (2009). 168. Kataoka, M. et al. Recovery from diabetes in neonatal mice after a low-dose streptozotocin treatment. Biochem. Biophys. Res. Commun. 430, 1103–8 (2013). 169. Mosser, R. E. et al. High-fat diet-induced β-cell proliferation occurs prior to insulin resistance in C57Bl/6J male mice. Am. J. Physiol. Endocrinol. Metab. 308, E573-82 (2015). 170. Brelje, T. C., Bhagroo, N. V, Stout, L. E. & Sorenson, R. L. Beneficial effects of lipids and prolactin on insulin secretion and beta-cell proliferation: a role for lipids in the adaptation of islets to pregnancy. J. Endocrinol. 197, 265–76 (2008). 171. Pascoe, J. et al. Free fatty acids block glucose-induced β-cell proliferation in mice by inducing cell cycle inhibitors p16 and p18. Diabetes 61, 632–41 (2012). 172. Porat, S. et al. Control of pancreatic β cell regeneration by glucose metabolism. Cell Metab. 13, 440–9 (2011). 173. Terauchi, Y. et al. Glucokinase and IRS-2 are required for compensatory beta cell hyperplasia in response to high-fat diet-induced insulin resistance. J. Clin. Invest. 117, 246–57 (2007). 174. Nakamura, A. et al. Impact of small-molecule glucokinase activator on glucose metabolism and beta-cell mass. Endocrinology 150, 1147–54 (2009). 134

175. Kim, H. et al. Serotonin regulates pancreatic beta cell mass during pregnancy. Nat. Med. 16, 804–8 (2010). 176. Freemark, M. et al. Targeted deletion of the PRL receptor: effects on islet development, insulin production, and glucose tolerance. Endocrinology 143, 1378–85 (2002). 177. Stout, L. E., Svensson, A. M. & Sorenson, R. L. Prolactin regulation of islet-derived INS- 1 cells: characteristics and immunocytochemical analysis of STAT5 translocation. Endocrinology 138, 1592–603 (1997). 178. Brelje, T. C., Stout, L. E., Bhagroo, N. V. & Sorenson, R. L. Distinctive roles for prolactin and growth hormone in the activation of signal transducer and activator of transcription 5 in pancreatic islets of langerhans. Endocrinology 145, 4162–75 (2004). 179. Karnik, S. K. et al. Menin controls growth of pancreatic beta-cells in pregnant mice and promotes gestational diabetes mellitus. Science 318, 806–9 (2007). 180. Chen, H. et al. Augmented Stat5 Signaling Bypasses Multiple Impediments to Lactogen- Mediated Proliferation in Human β-Cells. Diabetes 64, 3784–97 (2015). 181. Chamberlain, C. E. et al. Menin determines K-RAS proliferative outputs in endocrine cells. J. Clin. Invest. 124, 4093–101 (2014). 182. Miyatsuka, T., Kosaka, Y., Kim, H. & German, M. S. Neurogenin3 inhibits proliferation in endocrine progenitors by inducing Cdkn1a. Proc. Natl. Acad. Sci. U. S. A. 108, 185–90 (2011). 183. Zhong, L. et al. Essential role of Skp2-mediated p27 degradation in growth and adaptive expansion of pancreatic beta cells. J. Clin. Invest. 117, 2869–76 (2007). 184. Cozar-Castellano, I. et al. Molecular control of cell cycle progression in the pancreatic beta-cell. Endocr. Rev. 27, 356–70 (2006). 185. Rane, S. G. et al. Loss of Cdk4 expression causes insulin-deficient diabetes and Cdk4 activation results in beta-islet cell hyperplasia. Nat. Genet. 22, 44–52 (1999). 186. Kushner, J. A. et al. Cyclins D2 and D1 are essential for postnatal pancreatic beta-cell growth. Mol. Cell. Biol. 25, 3752–62 (2005). 187. Georgia, S. & Bhushan, A. Beta cell replication is the primary mechanism for maintaining postnatal beta cell mass. J. Clin. Invest. 114, 963–8 (2004). 188. Chen, H. et al. Polycomb protein Ezh2 regulates pancreatic beta-cell Ink4a/Arf expression and regeneration in diabetes mellitus. Genes Dev. 23, 975–85 (2009). 189. Cozar-Castellano, I., Takane, K. K., Bottino, R., Balamurugan, A. N. & Stewart, A. F. Induction of beta-cell proliferation and retinoblastoma protein phosphorylation in rat and human islets using adenovirus-mediated transfer of cyclin-dependent kinase-4 and cyclin D1. Diabetes 53, 149–59 (2004). 190. Wegner, M. All purpose Sox: The many roles of Sox proteins in gene expression. Int. J.

135

Biochem. Cell Biol. 42, 381–90 (2010). 191. Potzner, M. R. et al. Sequential requirement of Sox4 and Sox11 during development of the sympathetic nervous system. Development 137, 775–84 (2010). 192. Bergsland, M. et al. Sequentially acting Sox transcription factors in neural lineage development. Genes Dev. 25, 2453–64 (2011). 193. Garside, V. C. et al. SOX9 modulates the expression of key transcription factors required for heart valve development. Development 142, 4340–50 (2015). 194. Koopman, P., Münsterberg, A., Capel, B., Vivian, N. & Lovell-Badge, R. Expression of a candidate sex-determining gene during mouse testis differentiation. Nature 348, 450–2 (1990). 195. Kamachi, Y. & Kondoh, H. Sox proteins: regulators of cell fate specification and differentiation. Development 140, 4129–44 (2013). 196. Jauch, R., Ng, C. K. L., Narasimhan, K. & Kolatkar, P. R. The crystal structure of the Sox4 HMG domain-DNA complex suggests a mechanism for positional interdependence in DNA recognition. Biochem. J. 443, 39–47 (2012). 197. Badis, G. et al. Diversity and complexity in DNA recognition by transcription factors. Science 324, 1720–3 (2009). 198. Ng, C. K. L. et al. Deciphering the Sox-Oct partner code by quantitative cooperativity measurements. Nucleic Acids Res. 40, 4933–41 (2012). 199. Schilham, M. W. et al. Defects in cardiac outflow tract formation and pro-B-lymphocyte expansion in mice lacking Sox-4. Nature 380, 711–4 (1996). 200. Sarkar, A. & Hochedlinger, K. The sox family of transcription factors: versatile regulators of stem and progenitor cell fate. Cell Stem Cell 12, 15–30 (2013). 201. Bhattaram, P. et al. Organogenesis relies on SoxC transcription factors for the survival of neural and mesenchymal progenitors. Nat. Commun. 1, 9 (2010). 202. Dy, P. et al. The three SoxC proteins--Sox4, Sox11 and Sox12--exhibit overlapping expression patterns and molecular properties. Nucleic Acids Res. 36, 3101–17 (2008). 203. Jiang, Y. et al. Transcription factors SOX4 and SOX11 function redundantly to regulate the development of mouse retinal ganglion cells. J. Biol. Chem. 288, 18429–38 (2013). 204. Boyer, L. A. et al. Core transcriptional regulatory circuitry in human embryonic stem cells. Cell 122, 947–56 (2005). 205. Aksoy, I. et al. Oct4 switches partnering from Sox2 to Sox17 to reinterpret the enhancer code and specify endoderm. EMBO J. 32, 938–53 (2013). 206. Wegner, M. SOX after SOX: SOXession regulates neurogenesis. Genes Dev. 25, 2423–8 (2011).

136

207. Cheung, M., Abu-Elmagd, M., Clevers, H. & Scotting, P. J. Roles of Sox4 in central nervous system development. Brain Res. Mol. Brain Res. 79, 180–91 (2000). 208. Lioubinski, O., Müller, M., Wegner, M. & Sander, M. Expression of Sox transcription factors in the developing mouse pancreas. Dev. Dyn. 227, 402–8 (2003). 209. McDonald, E. et al. SOX9 regulates endocrine cell differentiation during human fetal pancreas development. Int. J. Biochem. Cell Biol. 44, 72–83 (2012). 210. Ragvin, A. et al. Long-range gene regulation links genomic type 2 diabetes and obesity risk regions to HHEX, SOX4, and IRX3. Proc. Natl. Acad. Sci. U. S. A. 107, 775–80 (2010). 211. Wilson, M. E. et al. The HMG box transcription factor Sox4 contributes to the development of the endocrine pancreas. Diabetes 54, 3402–9 (2005). 212. Goldsworthy, M. et al. Role of the transcription factor sox4 in insulin secretion and impaired glucose tolerance. Diabetes 57, 2234–44 (2008). 213. Mavropoulos, A. et al. sox4b is a key player of pancreatic alpha cell differentiation in zebrafish. Dev. Biol. 285, 211–23 (2005). 214. Flasse, L. C. et al. Ascl1b and Neurod1, instead of Neurog3, control pancreatic endocrine cell fate in zebrafish. BMC Biol. 11, 78 (2013). 215. Lin, J. W. et al. Differential requirement for ptf1a in endocrine and exocrine lineages of developing zebrafish pancreas. Dev. Biol. 270, 474–86 (2004). 216. Luse, D. S. The RNA polymerase II preinitiation complex. Through what pathway is the complex assembled? Transcription 5, e27050 (2014). 217. Duttke, S. H. C. Evolution and diversification of the basal transcription machinery. Trends Biochem. Sci. 40, 127–9 (2015). 218. Kornberg, R. D. Mediator and the mechanism of transcriptional activation. Trends Biochem. Sci. 30, 235–9 (2005). 219. Yin, J. & Wang, G. The Mediator complex: a master coordinator of transcription and cell lineage development. Development 141, 977–87 (2014). 220. Sainsbury, S., Bernecky, C. & Cramer, P. Structural basis of transcription initiation by RNA polymerase II. Nat. Rev. Mol. Cell Biol. 16, 129–43 (2015). 221. Allen, B. L. & Taatjes, D. J. The Mediator complex: a central integrator of transcription. Nat. Rev. Mol. Cell Biol. 16, 155–66 (2015). 222. Plaschka, C. et al. Architecture of the RNA polymerase II-Mediator core initiation complex. Nature 518, 376–80 (2015). 223. Cai, G. et al. Interaction of the mediator head module with RNA polymerase II. Structure 20, 899–910 (2012).

137

224. Robinson, P. J. et al. Structure of a Complete Mediator-RNA Polymerase II Pre-Initiation Complex. Cell 166, 1411–1422.e16 (2016). 225. Malik, S. & Roeder, R. G. The metazoan Mediator co-activator complex as an integrative hub for transcriptional regulation. Nat. Rev. Genet. 11, 761–72 (2010). 226. Ren, Y. et al. Specific structural motifs determine TRAP220 interactions with nuclear hormone receptors. Mol. Cell. Biol. 20, 5433–46 (2000). 227. Yuan, C. X., Ito, M., Fondell, J. D., Fu, Z. Y. & Roeder, R. G. The TRAP220 component of a thyroid hormone receptor- associated protein (TRAP) coactivator complex interacts directly with nuclear receptors in a ligand-dependent fashion. Proc. Natl. Acad. Sci. U. S. A. 95, 7939–44 (1998). 228. Ge, K. et al. Transcription coactivator TRAP220 is required for PPAR gamma 2- stimulated adipogenesis. Nature 417, 563–7 (2002). 229. Thakur, J. K., Yadav, A. & Yadav, G. Molecular recognition by the KIX domain and its role in gene regulation. Nucleic Acids Res. 42, 2112–25 (2014). 230. Yang, F. et al. An ARC/Mediator subunit required for SREBP control of cholesterol and lipid homeostasis. Nature 442, 700–4 (2006). 231. Taubert, S., Van Gilst, M. R., Hansen, M. & Yamamoto, K. R. A Mediator subunit, MDT- 15, integrates regulation of fatty acid metabolism by NHR-49-dependent and -independent pathways in C. elegans. Genes Dev. 20, 1137–49 (2006). 232. Kim, M. J., Jang, I.-C. & Chua, N.-H. The Mediator Complex MED15 Subunit Mediates Activation of Downstream Lipid-Related Genes by the WRINKLED1 Transcription Factor. Plant Physiol. 171, 1951–64 (2016). 233. Kato, Y., Habas, R., Katsuyama, Y., Näär, A. M. & He, X. A component of the ARC/Mediator complex required for TGF beta/Nodal signalling. Nature 418, 641–6 (2002). 234. Goh, G. Y. S. et al. The conserved Mediator subunit MDT-15 is required for oxidative stress responses in Caenorhabditis elegans. Aging Cell 13, 70–9 (2014). 235. Poncy, A. et al. Transcription factors SOX4 and SOX9 cooperatively control development of bile ducts. Dev. Biol. 404, 136–48 (2015). 236. Gu, G., Dubauskaite, J. & Melton, D. A. Direct evidence for the pancreatic lineage: NGN3+ cells are islet progenitors and are distinct from duct progenitors. Development 129, 2447–57 (2002). 237. Schonhoff, S. E., Giel-Moloney, M. & Leiter, A. B. Neurogenin 3-expressing progenitor cells in the gastrointestinal tract differentiate into both endocrine and non-endocrine cell types. Dev. Biol. 270, 443–54 (2004). 238. Thorens, B. et al. Ins1(Cre) knock-in mice for beta cell-specific gene recombination.

138

Diabetologia 58, 558–65 (2015). 239. Muzumdar, M. D., Tasic, B., Miyamichi, K., Li, L. & Luo, L. A global double-fluorescent Cre reporter mouse. Genesis 45, 593–605 (2007). 240. Penzo-Méndez, A., Dy, P., Pallavi, B. & Lefebvre, V. Generation of mice harboring a Sox4 conditional null allele. Genesis 45, 776–80 (2007). 241. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013). 242. Anders, S., Pyl, P. T. & Huber, W. HTSeq--a Python framework to work with high- throughput sequencing data. Bioinformatics 31, 166–9 (2015). 243. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014). 244. Luo, J. et al. A protocol for rapid generation of recombinant adenoviruses using the AdEasy system. Nat. Protoc. 2, 1236–47 (2007). 245. Sabatini, P. V et al. Npas4 is a novel activity-regulated cytoprotective factor in pancreatic β-cells. Diabetes 62, 2808–20 (2013). 246. Sanjana, N. E. et al. A transcription activator-like effector toolbox for genome engineering. Nat. Protoc. 7, 171–92 (2012). 247. Cong, L., Zhou, R., Kuo, Y.-C., Cunniff, M. & Zhang, F. Comprehensive interrogation of natural TALE DNA-binding modules and transcriptional repressor domains. Nat. Commun. 3, 968 (2012). 248. Gouzi, M., Kim, Y. H., Katsumoto, K., Johansson, K. & Grapin-Botton, A. Neurogenin3 initiates stepwise delamination of differentiating endocrine cells during pancreas development. Dev. Dyn. 240, 589–604 (2011). 249. Kopp, J. L. et al. Progenitor cell domains in the developing and adult pancreas. Cell Cycle 10, 1921–7 (2011). 250. Mastracci, T. L., Anderson, K. R., Papizan, J. B. & Sussel, L. Regulation of Neurod1 contributes to the lineage potential of Neurogenin3+ endocrine precursor cells in the pancreas. PLoS Genet. 9, e1003278 (2013). 251. Seymour, P. A. et al. A dosage-dependent requirement for Sox9 in pancreatic endocrine cell formation. Dev. Biol. 323, 19–30 (2008). 252. Yoshida, T. & Hanahan, D. Murine pancreatic ductal adenocarcinoma produced by in vitro transduction of polyoma middle T oncogene into the islets of Langerhans. Am. J. Pathol. 145, 671–84 (1994). 253. Lee, J. C. et al. Regulation of the pancreatic pro-endocrine gene neurogenin3. Diabetes 50, 928–36 (2001).

139

254. Rukstalis, J. M. & Habener, J. F. Neurogenin3: a master regulator of pancreatic islet differentiation and regeneration. Islets 1, 177–84 (2009). 255. Talchai, C., Xuan, S., Kitamura, T., DePinho, R. A. & Accili, D. Generation of functional insulin-producing cells in the gut by Foxo1 ablation. Nat. Genet. 44, 406–12, S1 (2012). 256. Anderson, K. R. et al. Cooperative transcriptional regulation of the essential pancreatic islet gene NeuroD1 (beta2) by Nkx2.2 and neurogenin 3. J. Biol. Chem. 284, 31236–48 (2009). 257. Choi, E. et al. Dual lineage-specific expression of Sox17 during mouse embryogenesis. Stem Cells 30, 2297–308 (2012). 258. Lynn, F. C. et al. MicroRNA expression is required for pancreatic islet cell genesis in the mouse. Diabetes 56, 2938–45 (2007). 259. Lee, C., Sund, N. & Vatamaniuk, M. Foxa2 controls Pdx1 gene expression in pancreatic β-cells in vivo. Diabetes 51, (2002). 260. Smith, S. B. et al. Neurogenin3 and hepatic nuclear factor 1 cooperate in activating pancreatic expression of Pax4. J. Biol. Chem. 278, 38254–9 (2003). 261. Wang, Q. et al. Ghrelin is a novel target of Pax4 in endocrine progenitors of the pancreas and duodenum. Dev. Dyn. 237, 51–61 (2008). 262. Suissa, Y. et al. Gastrin: a distinct fate of neurogenin3 positive progenitor cells in the embryonic pancreas. PLoS One 8, e70397 (2013). 263. Xu, E. E. et al. SOX4 cooperates with neurogenin 3 to regulate endocrine pancreas formation in mouse models. Diabetologia 58, 1013–23 (2015). 264. Collins, S. C. et al. Increased Expression of the Diabetes Gene SOX4 Reduces Insulin Secretion by Impaired Fusion Pore Expansion. Diabetes 65, 1952–61 (2016). 265. Piper, K. et al. Beta cell differentiation during early human pancreas development. J. Endocrinol. 181, 11–23 (2004). 266. Kerns, S. L. et al. A novel variant in CDKN1C is associated with intrauterine growth restriction, short stature, and early-adulthood-onset diabetes. J. Clin. Endocrinol. Metab. 99, E2117-22 (2014). 267. Butler, A. E. et al. Adaptive changes in pancreatic beta cell fractional area and beta cell turnover in human pregnancy. Diabetologia 53, 2167–76 (2010). 268. Cox, A. R. et al. Extreme obesity induces massive beta cell expansion in mice through self-renewal and does not alter the beta cell lineage. Diabetologia 59, 1231–41 (2016). 269. Vervoort, S. J., van Boxtel, R. & Coffer, P. J. The role of SRY-related HMG box transcription factor 4 (SOX4) in tumorigenesis and metastasis: friend or foe? Oncogene 32, 3397–409 (2013).

140

270. Jafarnejad, S. M., Wani, A. A., Martinka, M. & Li, G. Prognostic significance of Sox4 expression in human cutaneous melanoma and its role in cell migration and invasion. Am. J. Pathol. 177, 2741–52 (2010). 271. Penzo-Méndez, A. I. Critical roles for SoxC transcription factors in development and cancer. Int. J. Biochem. Cell Biol. 42, 425–8 (2010). 272. Foronda, M. et al. Sox4 links tumor suppression to accelerated aging in mice by modulating stem cell activation. Cell Rep. 8, 487–500 (2014). 273. Cnop, M. et al. The long lifespan and low turnover of human islet beta cells estimated by mathematical modelling of lipofuscin accumulation. Diabetologia 53, 321–30 (2010). 274. Friedman-Mazursky, O., Elkon, R. & Efrat, S. Redifferentiation of expanded human islet β cells by inhibition of ARX. Sci. Rep. 6, 20698 (2016). 275. Li, J. et al. Single-cell transcriptomes reveal characteristic features of human pancreatic islet cell types. EMBO Rep. 17, 178–87 (2016). 276. Ku, G. M. et al. Research resource: RNA-Seq reveals unique features of the pancreatic β- cell transcriptome. Mol. Endocrinol. 26, 1783–92 (2012). 277. van der Meulen, T. et al. Urocortin3 mediates somatostatin-dependent negative feedback control of insulin secretion. Nat. Med. 21, 769–76 (2015). 278. Gu, C. et al. Pancreatic beta cells require NeuroD to achieve and maintain functional maturity. Cell Metab. 11, 298–310 (2010). 279. Liu, C.-F. & Lefebvre, V. The transcription factors SOX9 and SOX5/SOX6 cooperate genome-wide through super-enhancers to drive chondrogenesis. Nucleic Acids Res. 43, 8183–203 (2015). 280. Mu, L. et al. SoxC transcription factors are required for neuronal differentiation in adult hippocampal neurogenesis. J. Neurosci. 32, 3067–80 (2012). 281. Feng, W. et al. The chromatin remodeler CHD7 regulates adult neurogenesis via activation of SoxC transcription factors. Cell Stem Cell 13, 62–72 (2013). 282. Thu, Y. M. & Bielinsky, A.-K. Enigmatic roles of Mcm10 in DNA replication. Trends Biochem. Sci. 38, 184–94 (2013). 283. Cozar-Castellano, I., Haught, M. & Stewart, A. F. The cell cycle inhibitory protein p21cip is not essential for maintaining beta-cell cycle arrest or beta-cell function in vivo. Diabetes 55, 3271–8 (2006). 284. Robitaille, K. et al. High-throughput Functional Genomics Identifies Regulators of Primary Human Beta Cell Proliferation. J. Biol. Chem. 291, 4614–25 (2016). 285. Kim, S. K. et al. Activin receptor patterning of foregut organogenesis. Genes Dev. 14, 1866–71 (2000).

141

286. Ejarque, M. et al. Neurogenin3 cooperates with Foxa2 to autoactivate its own expression. J. Biol. Chem. 288, 11705–17 (2013). 287. Miyatsuka, T. et al. Ptf1a and RBP-J cooperate in activating Pdx1 gene expression through binding to Area III. Biochem. Biophys. Res. Commun. 362, 905–9 (2007). 288. Andrau, J.-C. et al. Genome-wide location of the coactivator mediator: Binding without activation and transient Cdk8 interaction on DNA. Mol. Cell 22, 179–92 (2006). 289. Borggrefe, T. & Yue, X. Interactions between subunits of the Mediator complex with gene-specific transcription factors. Semin. Cell Dev. Biol. 22, 759–68 (2011). 290. Thakur, J. K. et al. Mediator subunit Gal11p/MED15 is required for fatty acid-dependent gene activation by yeast transcription factor Oaf1p. J. Biol. Chem. 284, 4422–8 (2009). 291. Whyte, W. A. et al. Master transcription factors and mediator establish super-enhancers at key cell identity genes. Cell 153, 307–19 (2013). 292. Kagey, M. H. et al. Mediator and cohesin connect gene expression and chromatin architecture. Nature 467, 430–5 (2010). 293. Stumpf, M. et al. The mediator complex functions as a coactivator for GATA-1 in erythropoiesis via subunit Med1/TRAP220. Proc. Natl. Acad. Sci. U. S. A. 103, 18504–9 (2006). 294. Nakamura, Y. et al. Wwp2 is essential for palatogenesis mediated by the interaction between Sox9 and mediator subunit 25. Nat. Commun. 2, 251 (2011). 295. Zhou, R. et al. SOX9 interacts with a component of the human thyroid hormone receptor- associated protein complex. Nucleic Acids Res. 30, 3245–52 (2002). 296. Tompa, P. Intrinsically disordered proteins: a 10-year recap. Trends Biochem. Sci. 37, 509–16 (2012). 297. Nagulapalli, M., Maji, S., Dwivedi, N., Dahiya, P. & Thakur, J. K. Evolution of disorder in Mediator complex and its functional relevance. Nucleic Acids Res. 44, 1591–612 (2016). 298. Donelan, W., Koya, V., Li, S.-W. & Yang, L.-J. Distinct regulation of hepatic nuclear factor 1alpha by NKX6.1 in pancreatic beta cells. J. Biol. Chem. 285, 12181–9 (2010). 299. Nammo, T. et al. Expression profile of MODY3/HNF-1alpha protein in the developing mouse pancreas. Diabetologia 45, 1142–53 (2002). 300. Shih, D. Q. et al. Loss of HNF-1alpha function in mice leads to abnormal expression of genes involved in pancreatic islet development and metabolism. Diabetes 50, 2472–80 (2001). 301. Takane, K. K., Kleinberger, J. W., Salim, F. G., Fiaschi-Taesch, N. M. & Stewart, A. F. Regulated and reversible induction of adult human β-cell replication. Diabetes 61, 418–24 (2012). 142

302. Wang, P. et al. A high-throughput chemical screen reveals that harmine-mediated inhibition of DYRK1A increases human pancreatic beta cell replication. Nat. Med. 21, 383–8 (2015). 303. Quiroz, Y. et al. The HMG-box transcription factor Sox4b is required for pituitary expression of gata2a and specification of thyrotrope and gonadotrope cells in zebrafish. Mol. Endocrinol. 26, 1014–27 (2012). 304. Spåhr, H. et al. Analysis of Schizosaccharomyces pombe mediator reveals a set of essential subunits conserved between yeast and metazoan cells. Proc. Natl. Acad. Sci. U. S. A. 98, 11985–90 (2001). 305. Malik, S. et al. Structural and functional organization of TRAP220, the TRAP/mediator subunit that is targeted by nuclear receptors. Mol. Cell. Biol. 24, 8244–54 (2004). 306. Tsai, K.-L. et al. Subunit architecture and functional modular rearrangements of the transcriptional mediator complex. Cell 157, 1430–44 (2014). 307. Tóth-Petróczy, A. et al. Malleable machines in transcription regulation: the mediator complex. PLoS Comput. Biol. 4, e1000243 (2008). 308. D’Alessio, J. A., Ng, R., Willenbring, H. & Tjian, R. Core promoter recognition complex changes accompany liver development. Proc. Natl. Acad. Sci. U. S. A. 108, 3906–11 (2011). 309. Harmon, E. B. et al. GDF11 modulates NGN3+ islet progenitor cell number and promotes beta-cell differentiation in pancreas development. Development 131, 6163–74 (2004). 310. Nelson, S. B., Schaffer, A. E. & Sander, M. The transcription factors Nkx6.1 and Nkx6.2 possess equivalent activities in promoting beta-cell fate specification in Pdx1+ pancreatic progenitor cells. Development 134, 2491–500 (2007). 311. Hang, Y. & Stein, R. MafA and MafB activity in pancreatic β cells. Trends Endocrinol. Metab. 22, 364–73 (2011).

143

Appendices

Appendix A: IM15KO RNA-Seq Differentially Expressed Genes

Figure 45: 174 differentially expressed genes from IM15KO RNA- Seq analysis by DESeq2. (Left) Table of differentially expressed genes organized in columns representing: gene name (genes), mean FPKM (mean), Log2 fold difference of control over IM15KO FPKM values (change), non- adjusted p-value (pvalue), and B-H multiple comparison correction adjusted p-values (padj). Genes organized by padj value. (Right) R module “pheatmap” heatmap plot of differentially expressed genes by their Log2 FPKM values. Data organized by k- means clustering of samples on the x-axis and by genes on the y-axis. Legend depicts scale of FPKM data represented on the graph.

144

145

Appendix B: Taqman Low Density Array Gene Assay IDs Table 6: Applied Biosystems, Inc. Taqman® low-density array gene ID

Taqman Low Density Array Gene Assay ID: Mm02619580_g1 Mm00487804_m1 Mm00437336_m1 Mm01617993_gH Mm00599749_m1 Mm00456190_m1 Mm00437347_m1 Mm00492069_m1 Mm03053308_g1 Mm00485408_m1 Mm00437350_m1 Mm00516196_m1 Mm00437967_m1 Mm00440464_m1 Mm00437353_m1 Mm01138833_m1 Mm00432284_m1 Mm00448100_g1 Mm01301717_m1 Mm00514160_m1 Mm01195085_m1 Mm00465434_m1 Mm00460518_m1 Mm00446953_m1 Mm00432337_m1 Mm00440924_g1 Mm01259683_g1 Mm00497217_m1 Mm00438064_m1 Mm00772799_m1 Mm00731595_gH Mm00786984_s1 Mm03053893_gH Mm01317678_m1 Mm00845209_s1 Mm00712529_m1 Mm01171453_m1 Mm00479224_m1 Mm00463644_m1 Mm00801867_m1 Hs99999901_s1 Mm00487888_m1 Mm00435565_m1 Mm00725863_s1 Mm00432359_m1 Mm00485474_m1 Mm00486320_s1 Mm00484848_m1 Mm00438070_m1 Mm00485586_m1 Mm01281943_s1 Mm00487656_m1 Mm00486880_m1 Mm00803251_m1 Mm02529758_s1 Mm00484963_m1 Mm00772471_m1 Mm00495559_m1 Mm00521599_m1 Mm01278606_m1 Mm00438128_m1 Mm00490624_m1 Mm01243796_g1 Mm00501741_m1 Mm00726334_s1 Mm01731287_m1 Mm02619580_g1 Mm00487804_m1 Mm00432437_m1 Mm00494175_m1 Mm00599749_m1 Mm00456190_m1 Mm00438163_m1 Mm01331624_m1 Mm03053308_g1 Mm00485408_m1 Mm00432448_m1 Mm00545877_m1 Mm00437967_m1 Mm00440464_m1 Mm00438170_m1 Mm01299062_g1 Mm00432284_m1 Mm00448100_g1 Mm01257348_m1 Mm01265783_m1 Mm01195085_m1 Mm00465434_m1 Mm00483241_m1 Mm01342435_m1 Mm00432337_m1 Mm00440924_g1 Mm00466006_m1 Mm00483039_m1 Mm00438064_m1 Mm00772799_m1 Mm01617993_gH Mm01322146_m1 Mm03053893_gH Mm01317678_m1 Mm00492069_m1 Mm00443800_m1 Mm01171453_m1 Mm00479224_m1 Mm00516196_m1 Mm00452114_m1 Hs99999901_s1 Mm00487888_m1 Mm01138833_m1 Mm00438851_m1 Mm00432359_m1 Mm00485474_m1 Mm00514160_m1 Mm00445405_s1 Mm00438070_m1 Mm00485586_m1 Mm00446953_m1 Mm00433382_m1 Mm00486880_m1 Mm00803251_m1 Mm00497217_m1 Mm03053560_s1 Mm00772471_m1 Mm00495559_m1 Mm00786984_s1 Mm00433409_s1 Mm00438128_m1 Mm00490624_m1 Mm00712529_m1 Mm01719732_m1 Mm00726334_s1 Mm01731287_m1 Mm00801867_m1 Mm00444911_m1 Mm00432437_m1 Mm00494175_m1 Mm00725863_s1 Mm00550265_m1 Mm00438163_m1 Mm01331624_m1 Mm00484848_m1 Mm00455902_s1 Mm00432448_m1 Mm00545877_m1 Mm00487656_m1 Mm00493445_m1 Mm00438170_m1 Mm01299062_g1 Mm00484963_m1 Mm00501505_m1 Mm01257348_m1 Mm01265783_m1 Mm01278606_m1 Mm00437328_m1 Mm00483241_m1 Mm01342435_m1

146

Taqman Low Density Array Gene Assay ID: Mm00501741_m1 Mm00437330_m1 Mm00466006_m1 Mm00483039_m1 Mm03053893_gH Mm00485474_m1 Mm00521599_m1 Mm00456190_m1 Mm01171453_m1 Mm00485586_m1 Mm01243796_g1 Mm00485408_m1 Hs99999901_s1 Mm00803251_m1 Mm02619580_g1 Mm00440464_m1 Mm00432359_m1 Mm00495559_m1 Mm00599749_m1 Mm00448100_g1 Mm00438070_m1 Mm00490624_m1 Mm03053308_g1 Mm00465434_m1 Mm00486880_m1 Mm01731287_m1 Mm00437967_m1 Mm00440924_g1 Mm00772471_m1 Mm00494175_m1 Mm00432284_m1 Mm00772799_m1 Mm00438128_m1 Mm01331624_m1 Mm01195085_m1 Mm01317678_m1 Mm00726334_s1 Mm00545877_m1 Mm00432337_m1 Mm00479224_m1 Mm00432437_m1 Mm01299062_g1 Mm00438064_m1 Mm00487888_m1 Mm00438163_m1 Mm01265783_m1 Mm03053893_gH Mm00485474_m1 Mm00432448_m1 Mm01342435_m1 Mm01171453_m1 Mm00485586_m1 Mm00438170_m1 Mm00483039_m1 Hs99999901_s1 Mm00803251_m1 Mm01257348_m1 Mm01322146_m1 Mm00432359_m1 Mm00495559_m1 Mm00483241_m1 Mm00443800_m1 Mm00438070_m1 Mm00490624_m1 Mm00466006_m1 Mm00452114_m1 Mm00486880_m1 Mm01731287_m1 Mm01617993_gH Mm00438851_m1 Mm00772471_m1 Mm00494175_m1 Mm00492069_m1 Mm00445405_s1 Mm00438128_m1 Mm01331624_m1 Mm00516196_m1 Mm00433382_m1 Mm00726334_s1 Mm00545877_m1 Mm01138833_m1 Mm03053560_s1 Mm00432437_m1 Mm01299062_g1 Mm00514160_m1 Mm00433409_s1 Mm00438163_m1 Mm01265783_m1 Mm00446953_m1 Mm01719732_m1 Mm00432448_m1 Mm01342435_m1 Mm00497217_m1 Mm00444911_m1 Mm00438170_m1 Mm00483039_m1 Mm00786984_s1 Mm00550265_m1 Mm01257348_m1 Mm01322146_m1 Mm00712529_m1 Mm00455902_s1 Mm00483241_m1 Mm00443800_m1 Mm00801867_m1 Mm00493445_m1 Mm00466006_m1 Mm00452114_m1 Mm00725863_s1 Mm00501505_m1 Mm01617993_gH Mm00438851_m1 Mm00484848_m1 Mm00437328_m1 Mm00492069_m1 Mm00445405_s1 Mm00487656_m1 Mm00437330_m1 Mm00516196_m1 Mm00433382_m1 Mm00484963_m1 Mm00437336_m1 Mm01138833_m1 Mm03053560_s1 Mm01278606_m1 Mm00437347_m1 Mm00514160_m1 Mm00433409_s1 Mm00501741_m1 Mm00437350_m1 Mm00446953_m1 Mm01719732_m1 Mm00487804_m1 Mm00437353_m1 Mm00497217_m1 Mm00444911_m1 Mm00456190_m1 Mm01301717_m1 Mm00786984_s1 Mm00550265_m1 Mm00485408_m1 Mm00460518_m1 Mm00712529_m1 Mm00455902_s1 Mm00440464_m1 Mm01259683_g1 Mm00801867_m1 Mm00493445_m1 Mm00448100_g1 Mm00731595_gH Mm00725863_s1 Mm00501505_m1 Mm00465434_m1 Mm00845209_s1 Mm00484848_m1 Mm00437328_m1 Mm00440924_g1 Mm00463644_m1 Mm00487656_m1 Mm00437330_m1 Mm00772799_m1 Mm00435565_m1 Mm00484963_m1 Mm00437336_m1 Mm01317678_m1 Mm00486320_s1 Mm01278606_m1 Mm00437347_m1

147

Taqman Low Density Array Gene Assay ID: Mm00479224_m1 Mm01281943_s1 Mm00501741_m1 Mm00437350_m1 Mm00487888_m1 Mm02529758_s1 Mm00487804_m1 Mm00437353_m1 Mm01322146_m1 Mm00445405_s1 Mm01719732_m1 Mm00493445_m1 Mm00443800_m1 Mm00433382_m1 Mm00444911_m1 Mm00501505_m1 Mm00452114_m1 Mm03053560_s1 Mm00550265_m1 Mm00437328_m1 Mm00438851_m1 Mm00433409_s1 Mm00455902_s1 Mm00437330_m1 Mm00437336_m1 Mm00437350_m1 Mm01301717_m1 Mm01259683_g1 Mm00437347_m1 Mm00437353_m1 Mm00460518_m1 Mm00731595_gH Mm00845209_s1 Mm00521599_m1 Mm00437967_m1 Mm01301717_m1 Mm00463644_m1 Mm01243796_g1 Mm00432284_m1 Mm00460518_m1 Mm00435565_m1 Mm02619580_g1 Mm01195085_m1 Mm01259683_g1 Mm00486320_s1 Mm00599749_m1 Mm00432337_m1 Mm00731595_gH Mm01281943_s1 Mm03053308_g1 Mm00438064_m1 Mm00845209_s1 Mm02529758_s1 Mm00463644_m1 Mm00463644_m1 Mm00435565_m1 Mm00486320_s1 Mm02529758_s1 Mm00521599_m1 Mm01243796_g1 Mm01281943_s1

148