Dissertation

Home , ANP32D, Apiin, ARHGEF1, C20orf27, DNA polymerase eta, GTF2E1

The Immune-modulatory and Anti-carcinogenic Mechanisms of the Flavonoid Apigenin

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Daniel A. Arango Tamayo, B.S.

Graduate Program of Molecular Cellular and Developmental Biology

The Ohio State University

2015

Dissertation Committee:

Dr. Andrea I. Doseff, Advisor

Dr. Erich Grotewold, Co-Advisor

Dr. Denis Guttridge

Dr. Tsonwin Hai

Dr. Dawn Chandler

Daniel Arango

2015

Abstract

Dietary phytochemicals provide health benefits against several cancers and inflammatory diseases. Flavonoids are amongst the most abundant dietary phytochemicals emerging as key anti-carcinogenic and anti-inflammatory molecules. Yet, the mechanisms underlying their anti-cancer and anti-inflammatory activities are poorly defined. The goal of this project was to study the immune-modulatory and anti- carcinogenic mechanisms of the flavonoid apigenin. I investigated the modes of action of apigenin using different model systems including a monocytic leukemia cell line, breast cancer cell lines, macrophages and mouse models of inflammation and breast cancer development. In monocytic leukemia, I found that apigenin induces DNA strand breaks leading to the activation of a DNA damage response pathway that results in cell cycle arrest and induction of apoptosis. Using mouse models of inflammation, I showed that apigenin reduces lipopolysaccharide (LPS)-induced lethality by inhibiting the activity of the transcription factor NF-κB and the expression of the pro-inflammatory molecules miR-155 and TNFα. I established, using a pre-clinical mouse model of breast cancer development, that the immune-modulatory and anti-carcinogenic activities of apigenin work in concert to delay breast tumor progression and metastasis by dually acting on malignant and immune cells. My results show that apigenin induces apoptosis

and inhibits proliferation in breast tumors as well as halts macrophages infiltration in the tumor microenvironment by reducing the expression of NF-κB-dependent chemokines and promoting apoptosis in blood monocytes, the macrophage progenitors. Moreover, I implemented the use of a newly formulated celery-based apigenin-rich diet in mouse models of inflammation and breast cancer demonstrating that this diet, as well as apigenin, have anti-inflammatory and anti-carcinogenic activities by immune-modulating monocytes and macrophages and inducing apoptosis in cancer cells.

To investigate the molecular mechanisms underlying the biological effects of apigenin, we developed of a new genome-wide approach to identify direct targets of this flavonoid. From these studies, I identified 160 candidate targets of apigenin that revealed unexpected mechanisms on how this dietary phytochemical modulates cellular functions such as apoptosis, immune and DNA damage response signaling pathways. In addition, I observed that apigenin interacts with RNA binding proteins including the heterogeneous nuclear RiboNucleoProtein A2 (hnRNPA2) and affects splicing genome-wide, providing a novel mechanism on how this flavone regulates apoptotic cell fate through modulation of mRNA processing. Altogether, this investigation offers a fresh view on how flavonoids influence human health, by impacting multiple cellular targets with moderate affinity. Thus, in contrast to pharmaceutical drugs selected to have high affinity and specificity for main hubs of biological pathways, the effect of flavonoids would be distributed across the entire cellular network with consequent benefits on human health.

In addition, these results support the use of functional foods rich in flavonoids as an alternative for the treatment and prevention of inflammatory diseases and cancer.

iii

Dedication

To my parents

For planting the seed of curiosity in me

A mis padres

Por sembrar la semilla de la curiosidad en mi

Acknowledgements

I am deeply thankful with my advisor Dr. Andrea Doseff for her guidance, scientific education and advices throughout these years and to my co-advisor Dr. Erich Grotewold for his collaboration and scientific contribution to my education. My sincere appreciation to Dr. Dennis Guttridge, Dr. Dawn Chandler and Dr. Tsonwin Hai for accepting being in my committee and their scientific contributions to my education. I am especially grateful with Dr. Timothy Eubank for his help and collaboration with animal models of cancer.

Especial gratitude to Dr. Kengo Morohashi for teaching me his expertise on PD-seq and for his incredible contribution to the identification of apigenin targets. I would like to thank Dr. Alper Yilmaz, Dr. Xiaokui Mo, Mrs. Katherine Mejia-Guerra, Mr. Erich

Mukundi and Mr. Francisco Padilla-Obregon for their help with bioinformatical analyses.

My sincere acknowledgment to the former and current members of Doseff and Grotewold laboratories, especially Dr. Arti Parihar, Dr. Horacio Cardenas, Dr. Greg Hostetler, Dr.

Oliver Voss, Dr. Yadira Malavez, Dr. Antje Feller, Dr. M. Elba Gonzalez-Mejia, Dr. Wei

Li, Dr. Marcelo Pereira, Dr. Isabel Casas, Ms. Silvia Duarte, Ms. Catalina Pineda, Mr.

Luis D. Prada and Mr. Roberto Alers. It was a pleasure to work with each of them.

Especial thanks to Mr. Bledi Brahimaj for his help on cloning all FRET constructs and

Ms. Joanna Li for their help with protein purifications and pull downs. I would like to thank the visiting scholars Ms. Mayra Diosa-Toro and Ms. Laura

Rojas for their contributions on the miRNA manuscript and Mrs. Giovanna Merchand for her help with macrophage phenotyping in the PyMT model. I would like also to thank

Drs. Tom Schmittgen and Jinmai Jiang for their help with microRNAs. Thanks to Drs.

Jessica Cooperstone and Ken Riedl for their help with the preparations of celery-based apigenin diets. Especial gratitude to Dr. Kouji Kuramochi for providing the apigenin beads. I am very grateful with Drs. Wolf Frommer, Lexie Friend, Adrian R. Krainer, Ann

C. Williams and Philip B. Wedegaertner for constructs. Finally and I would like to acknowledge my funding sources, Pelotonia, the Food Innovation Center and the Public

Health Preparedness for Infectious Diseases Program.

Vita

July, 28th 1984………………………………..…Medellin, Colombia.

2001 – 2006…………………………………….B.S., Biology. University of Antioquia. Medellin, Colombia.

2008 – 2015……………………………………Graduate Research Associate. Working towards Ph.D., Molecular Cellular and Developmental Biology. The Ohio State University, Columbus, OH.

Publications

Arango D, Diosa-Toro M, Rojas-Hernandez LS, Cooperstone JL, Schwartz SJ, Mo X, Jiang J, Schmittgen TD, Doseff AI. Dietary apigenin reduces LPS-induced expression of mir-155 restoring immune balance during inflammation. 2015. Mol Nutr Food Res 59: 763-772.

Arango D*, Morohashi K*, Yilmaz A, Kuramochi K, Parihar A, Brahimaj B, Grotewold E, Doseff AI. Molecular bases for the action of a dietary flavonoid revealed by the comprehensive identification of apigenin human targets. 2013. Proc Natl Acad Sci 110: E2153-E2162.

Duarte S*, Arango D*, Parihar A, Hamel P, Yasmeen R, Doseff AI. Apigenin protects endothelial cells from lipopolysaccharide (LPS)-induced inflammation by decreasing caspase-3 activation and modulating mitochondrial function. 2013. Int J Mol Sci 14: 17664-17679.

Arango D, Parihar A, Villamena FA, Wang L, Freitas MA, Grotewold E, Doseff AI. Apigenin induces DNA damage through the PKCδ-dependent activation of ATM and H2AX causing down-regulation of genes involved in cell cycle control and DNA repair. 2012. Biochem Pharmacol 84: 1571-1580.

Hostetler G, Riedl K, Cardenas H, Diosa-Toro M, Arango D, Schwartz SJ, Doseff AI. Flavone deglycosylation increases their anti-inflammatory activity and absorption. 2012. Mol Nutr Food Res 56: 558-569.

*Shared first authorship. vii

Fields of Study

Major Field: Molecular, Cellular and Developmental Biology

viii

Table of Contents

Abstract ……………………………………………………………………..…………… ii

Dedication………………………………………………………………………………...iv

Acknowledgements ……………………………………...………………………………. v

Vita………………………………………………………………………….……..…….vii

Table of contents…………...……………………………………………….……..……...ix

List of Figures…………………………………………………………………………..xvii

List of Tables…………………………………………………………………………...xxii

Chapter 1. Introduction…………………..…………..…………..…………..…………... 1

1.1 Origin of monocytes and macrophages……………………………………….. 2

1.2 Role of monocytes and macrophages in inflammation……………………….. 3

1.3 Malignant transformation of monocytic cells……...…………………...... 5

1.4 Regulation of apoptosis…………………...……...…………………...... 6

1.5 Mechanisms of breast carcinogenesis…..……...……………………...... …... 8

1.5.1 Role of monocytes and macrophages in breast cancer………………….. 10

1.5.2 Role of NF-κB in breast cancer…………………….………………….... 12

1.5.3 Transcriptome alterations in breast cancer……………………………… 13

1.5.3.1 Alternative splicing of RNA…………………..………………...... 14

1.6 Flavonoids ……………..……………..……………..………………………...... 15

1.6.1 Anti-carcinogenic activity of apigenin……………..……..………..….... 16

1.6.2 Apigenin and inflammation ………………………..……..…………...... 18

1.6.3 Molecular targets of apigenin………………………..……..…………… 18

Chapter 2. Material and Methods……………..…….…………………….…………….. 26

2.1 Chemicals and Reagents……………..…………………………..………...... 26

2.2 Cell Lines and Culture……………..…………………………..…………….. 28

2.3 Analysis of cell cycle and proliferation ……………..……..………………... 29

2.4 Caspase-3 activity and apoptosis ……………..……..………………………. 30

2.5 Intracellular measurement of ROS……………..……..……………………… 31

2.6 Alkaline comet assay……………..……..………………………………...... 31

2.7 Identification of histone phosphorylation by LC–MS ……………..………... 32

2.8 Western blots……………..……..………………………………………...…. 33

2.9 Immunoprecipitations and in vitro kinase assays……………..……………... 34

2.10 Immunofluorescence of γH2AX……………..……..………………………... 35

2.11 siRNA silencing……………..……..………………………………………… 35

2.12 RNA isolation and Reverse Transcriptase-PCR (RT-PCR) …………………. 36

2.13 Quantitative RT-PCR (qRT-PCR) analysis……………..……..…………….. 36

2.14 Microarray analysis……………..……..……………………………………... 37

2.15 High-throughput screening for miRs (HTS-miRs) ……………...……….….. 38

2.16 Validation of HTS-miRs……………..……..………………………………... 38

2.17 Analyses of gene expression and alternative splicing genome-wide……….. 39

2.17.1 RNA-seq libraries preparation……………..……..……………………. 39

2.17.2 Bioinformatical analyses of RNA-seq data………..……..…………….. 40

2.17.3 Analysis of public available RNA-seq data………………..…………....41

2.18 Analyses of alternative splicing by RT-PCR ………..……..……………..... 41

2.19 Preparation of apigenin-immobilized PEGA Beads ……..……..………….. 42

2.20 Phage display Screening ……………..……..………………………………. 43

2.21 Phage display coupled with Illumina® GAII next-generation

sequencing (PD-seq) ……………..……..………………………………...... 44

2.22 Analysis of PD-Seq data……………..……..……………………………….. 45

2.23 Cloning of plasmids ……………..……..………………………………...... 46

2.24 Recombinant protein expression and production………………..………….. 47

2.25 Pull-down Assays ……………..……..……………………………………... 49

2.26 Spectrophotometric Analyses ……………..……..…………………………. 50

2.27 Fluorescence Resonance Energy Transfer (FRET) ……..……..…………… 51

2.28 Amplified luminescent proximity homogeneous assay (ALPHA) ………… 52

2.29 Enzymatic assays……………..……..………………………………………. 52

2.30 Preparation of celery-based apigenin-rich extracts and diets…..…………… 54

2.31 Animal models ……………..……..………………………………………....55

2.31.1 Mouse models of inflammation……………..……..…………………....55

2.31.2 Measurement of NF-κB in vivo……………..……..…………………... 56

2.31.3 IKKβ knock-out (KO) mice……………..……..……………………..... 57

2.31.4 Mouse models of breast cancer development……..……..…………….. 57

2.31.5 Mouse xenografts……………..……..…………………………………. 58

2.32 Histology and immunohistochemistry (IHC) ………..……..………………. 58

2.33 TUNEL……………..……..………………………………………………… 60

2.34 Analyses of metastasis……………..……..…………………………………. 60

2.35 Isolation of mouse leukocytes…………………………………...…..……….60

2.36 Flow cytometry……………..……..………………………………………… 61

2.37 Co-cultures…………………………………………………………..……….62

2.38 Immunodetection of cytokines……………..……..………………………….62

2.39 Statistical analysis……………..……..………………………...…………….63

Chapter 3. Apigenin Induces DNA Damage in a PKCδ-Dependent Pathway Leading to Down-Regulation of Genes Involved in Cell Cycle Control and DNA Repair..……. 71

3.1 Abstract……………..……..………………………………………...……….71

3.2 Introduction……………..……..…………………………………….……… 72

3.3 Results……………..……..…………………………………………….…… 73

3.3.1 Apigenin induces DNA damage………..……..……………..………… 73

3.3.2 Apigenin induces H2AX phosphorylation ………..………..…..……… 74

3.3.3 Apigenin-induced DNA damage is ROS and caspase 3-independent…. 75

3.3.4 PKCδ and p38 are required for apigenin-induced DNA damage…...…. 75

3.3.5 Apigenin affects cell cycle progression and gene expression………….. 77

3.4 Discussion………..……..………………………………………...………….78

xii

Chapter 4. Dietary Apigenin Reduces LPS-Induced Expression of MiR-155 Restoring

Immune-Balance During Inflammation………..…..……………………….…………..90

4.1 Abstract ………..……..…………..……..………..……..………....……..… 90

4.2 Introduction………..……..……..………..……..………..……...…..……… 91

4.3 Results………..………..……..………..……..………..…………..……...… 93

4.3.1 Apigenin regulates inflammatory miR expression in

LPS-stimulated macrophages ……..……..…………..……..…..……… 93

4.3.2 Apigenin reduces the LPS-induced expression of miR-155

primary transcript…………..……..………..……..……..……...……… 95

4.3.3 Celery-based apigenin-rich diets reduce miR-155 in

LPS-induced inflammation………..……..………..……..………..……95

4.3.4 Celery-based apigenin-rich diets reduce LPS-induced miR-155

expression modulating inflammatory regulators…………....…..…...…97

4.3.5 Celery-based apigenin-rich diets decrease LPS-induced expression of

miR-155 and TNFα in vivo during inflammation……..……..…………98

4.3.6 Apigenin decreases NF-κB activity in lungs………..………………….99

4.3.7 Apigenin blocks LPS-induced lethality………..…….……..…………..100

4.3.8 Apigenin decreases miR-155 and TNFα in an IKKβ/NF-κB

mediated pathway ………..………..……..………..……....…………...101

4.4 Discussion………..………..……..………..……..………..……....………...103

Chapter 5: Dietary Apigenin Delays Breast Cancer Development and Metastasis by

Immune-Modulating Macrophages and Inducing Apoptosis in Malignant Cells……...117

xiii

5.1 Abstract …………..……..………..……..………..……..…..……..………..117

5.2 Introduction………..…..……..………..……..………..……..……..……….118

5.3 Results………..……..………..……..………..……..………..……..……….120

5.3.1 Celery-based apigenin-rich diets decrease breast tumor growth……….120

5.3.2 Apigenin delays breast cancer progression …………..……..…………. 121

5.3.3 Dietary apigenin decreases proliferation and induces apoptosis

in breast tumors………..………..……..………..……..……..………… 123

5.3.4 Dietary apigenin reduces macrophage infiltration in breast tumors…… 125

5.3.5 Dietary apigenin decreases the expression of macrophage chemo-

attractants …………..……..………..……..……..……..……………… 127

5.3.6 Dietary apigenin induces apoptosis in blood monocytes

re-establishing normal numbers of macrophage precursors …………... 128

5.3.7 Dietary apigenin blocks the cancer cell/macrophage

cross-communication decreasing chemokine expression

and triggering apoptosis in cancer cells and macrophages…………….. 134

5.3.8 Dietary apigenin decreases metastasis………..……..…………………. 139

5.4 Discussion………..…….…..……..………..……..………………………… 140

Chapter 6: Molecular Basis for the Action of a Dietary Flavonoid Revealed by the

Comprehensive Identification of Apigenin Human Targets………..…..……………… 164

6.1 Abstract ………..……..……..………..……..………..…………………..… 164

6.2 Introduction…………..……..………..……..……..……..…………………. 165

6.3 Results………..……..……..………..……..………..……..………..………. 166

xiv

6.3.1 PD-Seq identifies candidate cellular targets for apigenin……………… 166

6.3.2 Apigenin targets are enriched in three main categories ……………..… 169

6.3.3 Validation of putative apigenin targets………..……..………………… 170

6.3.4 Apigenin binds the glycin rich domain of hnRNPA2…………..……… 172

6.3.5 Structural relationships of specific and high affinity interaction

of apigenin and other flavonoids with hnRNPA2 ……….....…..……… 174

6.3.6 Apigenin inhibits hnRNPA2 oligomerization………..……..……..…… 177

6.3.7 Apigenin modulates alternative splicing in breast cancer cells……...… 178

6.3.8 Apigenin alters alternative splicing genome-wide………..………….… 179

6.3.9 Apigenin regulates the splicing of RNA-binding proteins………..…… 182

6.3.10 Apigenin regulates survival and apoptotic signaling pathways

through alternative splicing……..……..………..……..…………..…… 184

6.3.11 Apigenin modulates splicing in known substrates of the

direct targets of this flavone……..……..………..……..……….....…… 187

6.3.12 Apigenin affects the splicing of genes dysregulated

in breast cancer…..……..………..……..………..……..………..…….. 188

6.3.13 Effect of apigenin in gene expression in breast cancer………..…..…… 191

6.3.14 Apigenin decreases proliferation and induces apoptosis

in breast cancer cells without affecting non-carcinogenic

breast epithelial cells………..……..………..……..…………..…..…… 194

6.3.15 Apigenin decreases the expression of cell cycle progression

genes and modulates splicing of apoptotic molecules in vivo.…………195

6.4 Discussion………..…..……..………..……..………..……..………….…… 197

Chapter 7: Conclusion and Future Directions……….…..……..……..….…..….……... 247

References…………..……..………..……..………..……..………..……....……..…… 257

Appendix A: HnRNPA2 regulates NF-κB phosphorylation…………..…..…..…..…… 299

Appendix B: List of Abbreviations…………..…..…..…..……………………………..303

xvi

List of Figures

Figure 1.1. Hematopoiesis..……………………………………………….…………….. 20

Figure 1.2. Simplified representation of the apoptosis and

NF-κB signaling pathways.……………………………………………….... 21

Figure 1.3. Human and mouse models of breast cancer progression. ………………..... 22

Figure 1.4. General model of cancer initiation.………………………………………… 23

Figure 1.5. Different classes of alternative RNA processing events.……………….….. 24

Figure 1.6. Structure and classification of selected flavonoids………………………… 25

Figure 3.1. Apigenin induces DNA damage in leukemia cells…………………………. 84

Figure 3.2. Apigenin induces H2AX phosphorylation..………………………………... 85

Figure 3.3. Apigenin-induced DNA damage is mediated by PKCδ and p38.…………... 86

Figure 3.4. Apigenin induces ATM and γH2AX phosphorylation in a

PKCδ and p38-dependent pathway..………………………………………... 87

Figure 3.5. Apigenin affects cell cycle progression of THP-1 cells

by down-regulating cell cycle and DNA repair genes..…………………….. 88

Figure 3.6. Working model of apigenin-induced DNA damage..………………………. 89

Figure 4.1. Identification of apigenin-regulated miRs in LPS-induced

inflammation..……………………………………………………………… 108

xvii

Figure 4.2. Apigenin reduces the expression of LPS-induced primary miR-155

transcript..………………………………………………………………….. 109

Figure 4.3. Celery-based apigenin rich diets reduce LPS-induced

miR-155 expression in macrophages..…………………………………….. 110

Figure 4.4. Celery-based apigenin rich foods and pure apigenin

restored the expression of miR-155 targets..………………………………. 111

Figure 4.5. Celery-based apigenin rich foods reduce LPS-induced

expression of miR-155 and TNFα in vivo. …………………………………112

Figure 4.6. Apigenin inhibits NF-κB activity in vivo. ………………………………… 113

Figure 4.7. Apigenin decreases LPS-induced lethality..……………………………….. 114

Figure 4.8. Apigenin regulates miR-155 and TNFα in an IKKβ/NF-κB mediated

pathway.……………………………………………………………………. 115

Figure 4.9. Working model of the immune-regulatory activity of apigenin.…………... 116

Figure 5.1. Celery-based apigenin-rich diets decrease breast tumor growth.………….. 147

Figure 5.2. Apigenin delays breast cancer development………………………………. 148

Figure 5.3. Dietary apigenin reduces proliferation and increases

apoptosis in breast tumors.…………………………………………………….. 149

Figure 5.4. Apigenin decreases proliferation and induces apoptosis

in PyMT cells in vitro.……………………………………………………... 150

Figure 5.5. Apigenin reduces macrophage infiltration.………………………………... 151

Figure 5.6. Apigenin decreases M2-like macrophages.………………………………... 152

Figure 5.7. Apigenin reduces the expression of NF-κB-dependent

chemokines in breast tumors.……………………………………………… 153 xviii

Figure 5.8. Apigenin does not affect macrophage progenitors in bone marrow……….. 154

Figure 5.9. Strategy to study leukocyte populations in blood and spleens…………….. 155

Figure 5.10. Apigenin decreases the numbers of blood monocytes.…………………... 156

Figure 5.11. Apigenin induces apoptosis in peripheral monocytes.…………………… 157

Figure 5.12. Apigenin decreases blood leukocytes without affecting

splenic leukocytes.………………………………………………………. 158

Figure 5.13. Apigenin blocks macrophage:PyMT crosstalk

inhibiting CCL2 expression………………………………………………159

Figure 5.14. Apigenin blocks the macrophage:cancer cell paracrine loop reducing

NF-κB phosphorylation and CCL2 expression in both cancer cells

and macrophages.………………………………………………………… 160

Figure 5.15. Apigenin blocks the macrophage-induced cancer cell survival Inducing

apoptosis in both macrophages and PyMT cells…………………………. 161

Figure 5.16. Dietary apigenin decreases pulmonary metastasis.………………………. 162

Figure 5.17. Model of the immune-regulatory activity of apigenin in breast cancer….. 163

Figure 6.1. Synthesis of apigenin-beads and PD-Seq strategy outline.………………... 209

Figure 6.2. Analysis of the MKET clone.………………………………………………210

Figure 6.3. Calculation of normalized In-frame-aligned Counts

Per Gene-model (nICPG).…………………………………………………. 211

Figure 6.4. Hierarchical clustering analysis of the PD-Seq results.……………………. 212

Figure 6.5. Conventional phage display identifies clones that bind

to the A-beads and are highly enriched in the PD-Seq.…………………… 213

Figure 6.6. Apigenin targets are enriched in three main categories.…………………… 214

xix

Figure 6.7. Validation of apigenin targets.…………………………………………… 215

Figure 6.8. hnRNPA2 directly binds apigenin through the glycine-rich domain

(GRD).……………………………………………………………………... 216

Figure 6.9. Mapping of the apigenin-binding site to hnRNPA2.……………………… 217

Figure 6.10. Mapping the hnRNPA2-apigenin structural binding signatures.………… 218

Figure 6.11. Binding affinity of the interaction of hnRNPA2 with apigenin

determined by UV/Vis spectroscopy.…………………………………… 219

Figure 6.12. Development of a flavonoid nanosensor.…………………………………220

Figure 6.13. Binding affinity of the interaction of hnRNPA2 with apigenin

determined by the FRET nanosensor.…………………………………… 222

Figure 6.14. Apigenin quenches CFP fluorescence……………………………………. 223

Figure 6.15. Flavonoid structural relationship provided by the FRET-based flavonoid

nanosensor.………………………………………………………………. 224

Figure 6.16. Apigenin affects hnRNPA2 dimerization…………………………………225

Figure 6.17. Expression of hnRNPA2 in breast epithelial cells.………………………. 226

Figure 6.18. Apigenin regulates alternative splicing of hnRNPA2 substrates in breast

cancer cells.………………………………………………………………. 227

Figure 6.19. Apigenin regulates splicing genome-wide.………………………………. 228

Figure 6.20. Apigenin regulates the splicing of RNA-binding proteins.………………. 229

Figure 6.21. Apigenin regulates survival and apoptotic signaling pathways

through alternative splicing.………………………………………………230

Figure 6.22. Apigenin affects splicing in genes regulating survival and apoptotic

pathways.………………………………………………………………… 231

Figure 6.23. Modulation of splicing by apigenin targets.……………………………… 232

Figure 6.24. Dysregulated mRNA processing in TNBC.……………………………… 233

Figure 6.25. Apigenin affects the splicing of genes dysregulated in TNBC.………….. 234

Figure 6.26. Effect of apigenin in gene expression in breast cancer..…………………. 235

Figure 6.27. Apigenin modulates the expression of genes dysregulated in TNBC.….... 236

Figure 6.28. Apigenin induces apoptosis and decreases proliferation

in triple negative breast cancer cells, but had no effect on

non-carcinogenic cells..………………………………………………….. 237

Figure 6.29. Apigenin modulates gene expression and splicing in vivo……………….. 238

Figure 7.1. Immune-modulatory and anti-carcinogenic mechanisms of apigenin.…….. 256

Figure A1. HnRNPA2 regulates NF-κB-p65 phosphorylation..………………………. 302

xxi

List of Tables

Table 2.1. List of antibodies…………………….…………………….………………… 64

Table 2.2. List of primers used for qRT-PCR…………………….…………………….. 66

Table 2.3. List of primers used for splicing analysis…………………….……………… 68

Table 2.4. List of clones generated in this study…………………….…………………. 69

Table 2.5. List of primers for cloning and site-directed mutagenesis used

in this study…………………….……………………………………….…... 70

Table 6.1. Summary of reads obtained by PD-seq………………….………………….. 240

Table 6.2. Summary of MCS and MKET reads…..………………….………………… 241

Table 6.3. Identified Apigenin Targets…………………….…………………………... 242

Table 6.4. Biding affinities of different flavonoids to the FRET nanosensor………….. 244

Table 6.5. Summary of reads obtained by RNA-seq…..………………….…………… 245

Table 6.6. List of genes in which apigenin restored splicing to profiles

observed in NBT……………………….………………………………….. 246

xxii

Chapter 1

Introduction

Diet provides health benefits against several cancers and inflammatory diseases [1, 2].

Flavonoids are the most abundant dietary phytochemicals emerging as key anti- carcinogenic and anti-inflammatory molecules. Yet, the mechanisms underlying their anti-cancer and anti-inflammatory activities are poorly defined. The goal of this project was to study the immune-modulatory and anti-carcinogenic mechanisms of the flavonoid apigenin. My contributions included the development of a new genome-wide approach to identify direct targets of small molecules that applied to apigenin, revealed unexpected novel mechanisms of action. Importantly, the nature of my findings provide a paradigm shifting in the field as I showed that apigenin interacts with numerous proteins, unlike chemotherapeutic drugs developed to act on fewer targets. From these studies, I demonstrated that apigenin interacts with RNA-binding proteins and affects splicing.

These observations have paramount implications, as abnormal splicing is central to diseases development. In addition, I implemented the use of a newly formulated celery- based apigenin-rich diet in mouse models of inflammation and breast cancer demonstrating that this diet, as well as apigenin, have anti-inflammatory and anti-

1 carcinogenic activities by modulating monocyte/macrophage biology in an NF-κB- dependent pathway. Hence, I established that apigenin restores homeostasis upon harmful threats by acting on malignant and immune cells. The results presented here support the use of functional foods rich in flavonoids as an alternative for the treatment and prevention of inflammatory disorders including bacterial infection and breast cancer.

1.1 Origin of monocytes and macrophages

The cells of the immune system originate in the bone marrow (BM) from hematopoietic stem cells (HSC, Fig. 1.1) [3]. HSC produce common lymphoid and myeloid progenitors. The lymphoid progenitors differentiate into natural killers (NK cells), B lymphocytes (B cells) and T lymphocytes (T cells) [3]. The myeloid progenitors give rise to megakaryocytes, erythrocytes and granulocyte/monocyte progenitors (GMP), which further differentiate into basophils, eosinophils, neutrophils and monocytes (Fig.

1.1) [4]. In addition, macrophages derive from circulating monocytes and localize into tissues [5]. Monocytes patrol the blood stream for 24-48 h before undergoing apoptosis

[6]. Inflammatory stimuli such as lipopolysaccharide (LPS), a component of gram- negative bacteria, malignant transformation into monocytic leukemia, tumor-derived factors like chemokine C-C motif ligand 2 (CCL2) or differentiation into macrophages promote monocyte survival [6-9].

1.2 Role of monocytes and macrophages in inflammation

Monocytes and macrophages are responsible for the innate defense against external and internal threats, e.g. bacteria, viruses and malignant cells, by initiating, sustaining and helping resolve inflammation [5]. Pattern recognition receptors, including Toll like receptors (TLRs), expressed by monocytes and macrophages, are responsible for the detection of pathogens [5, 10]. Thirteen TLRs that recognize diverse pathogen associated molecular patterns (PAMPs) have been identified in mammals [5, 10]. Binding of ligands to TLRs triggers an inflammatory response [5, 10]. For instance, binding of LPS to TLR4 induces the activation of IRAK1/4 (IL-1 receptor-associated kinase-1 and 4), the recruitment of the MyD88 adaptor and the ubiquitin ligase TRAF6 (TNF receptor- associated factor 6) to form the MyD88/IRAK1/4/TRAF6 complex (Fig. 1.2) [11, 12].

The formation of this complex is important for the activation of a downstream kinase cascade mediated principally by MAPKs (mitogen activated protein kinases) or TAK1

[TGF-β activated kinase 1, (Fig. 1.2)] [11, 12]. Subsequently, these kinases activate the

IKK (Inhibitor Of Kappa Light Polypeptide Gene Enhancer In B-Cells) kinase complex, which is composed by two kinases (IKKα, IKKβ) and IKKγ, a regulatory subunit [13].

The activated IKK complex phosphorylates the inhibitor of NF-κB (Nuclear Factor

Kappa-light-chain-enhancer of activated B cells), Iκ-Bα, stimulating Iκ-Bα degradation, thereby freeing NF-κB [13]. The NF-κB axis entails two sub-pathways, the canonical and the non-canonical [14]. The canonical pathway, which is comprised of the p65 and p50 subunits is regulated by IKKβ [14], which directly phosphorylates and activates NF-κB-

65 subunit on residue Ser536 [14]. The non-canonical cascade, consisting of the subunits

3 p52 and RelB is activated by IKKα [15]. Once activated, NF-κB moves to the nucleus and activates the expression of pro-inflammatory cytokines, chemokines, anti-apoptotic molecules and inflammatory microRNAs (miRs, Fig. 1.2) [13, 16-18]. The NF-κB- dependent expression of pro-inflammatory molecules such as tumor necrosis factor α

(TNFα) and miR-155 regulate the intensity and duration of the immune response [19-22].

Dysregulation of the mechanisms that control the immune response leads to inflammatory disorders [23]. For example, sepsis is characterized by an exacerbated immune response to components of pathogens such as LPS leading to organ failure and death [24]. Sepsis affects ~1 million people per year in the USA, reaching ~30-50% mortality [25]. High morbidity and mortality are associated with sepsis and current successful therapies are limited [25]. Several non-steroidal anti-inflammatory drugs

(NSAIDs) are currently used to ameliorate inflammatory diseases, but their long-term consumption is often accompanied by adverse effects such as cardiovascular complications [26], prompting the need for alternative therapeutic approaches.

The IKKβ/NF-κB axis is a hallmark of sepsis [23]. The importance of IKKβ/NF-κB in sepsis is exemplified by studies showing that deletion of IKKβ in monocyte/macrophages or chemical inhibition of NF-κB reduces LPS-induced mortality in murine models of inflammation [27, 28]. Thus, identifying approaches that target the

IKKβ/NF-κB axis to restore proper immune-regulation will furnish new paths for the treatment of sepsis. In chapter 4, we showed that apigenin exerts effective anti- inflammatory activity reducing LPS-induced expression of TNFα and miR-155 in an NF-

κβ/IKKβ dependent pathway in macrophages, thereby restoring immune balance.

Besides sepsis, other disorders involving the immune system include cancer, which is characterized by an insufficient immune response followed by the re-education of monocyte/macrophages towards a cancer-prone phenotype [29]. Moreover, malignant transformation of cells from the immune cells results in leukemia [30]. Together, inflammatory and proliferative disorders are the leading cause of deaths worldwide [31], inflicting an immense pressure to identify immune-modulatory and anti-proliferative approaches to re-establish homeostasis upon harmful stimuli.

1.3 Malignant transformation of monocytic cells

Monocytic leukemia arises by the malignant transformation of monocyte/granulocyte precursors, characterized by the evasion of apoptosis, uncontrolled proliferation of immature monocytes and impaired hematopoiesis in the bone marrow [30]. Acute myeloid leukemia (AML) is the most common type of leukemia in adults, with an estimated of ~18,000 new cases and ~10,000 deaths every year in the USA [32].

Spontaneous apoptosis during normal monocytes life span is executed by the cysteine-aspartic protease 3 (caspase-3), a key component of the apoptotic pathway, in a protein kinase Cδ (PKCδ)-dependent pathway [33]. PKCδ, a serine threonine kinase, is activated during spontaneous monocyte apoptosis and phosphorylates caspase-3 promoting its activation [33]. Additionally, the chaperone heat shock protein 27 (Hsp27) associates with caspase-3 inhibiting its activation (Fig. 1.2) [34]. Transformation into

AML is achieved by pro-survival signals that lead to the inactivation of caspase-3 [35] followed by constitutively overexpression of cell cycle progression genes [36]. Hence,

5 caspase-3 is a central player in the execution of apoptosis and monocyte cell fate. Re- activation of monocyte apoptotic fate by modulating the regulators of caspase-3 activity encompasses a promising avenue for AML treatment. In chapter 3, we showed that apigenin induces DNA-damage in a PKCδ-dependent pathway leading to the activation of caspase-3 and induction of apoptosis in the acute monocytic leukemia cell line THP-1.

1.4 Regulation of apoptosis

Resistance to apoptosis is a hallmark of malignant cells [37]. Apoptosis is regulated through two main signaling cascades, the intrinsic and the extrinsic pathways (Fig. 1.2)

[38]. Intrinsic apoptosis is activated by DNA damage or reactive oxygen species (ROS), induced by chemotherapeutic drugs or radiotherapy [38]. This process is characterized by permeabilization of the mitochondria membrane, loss of mitochondrial trans-membrane potential and the release of cytochrome c (CYC) and DIABLO from the mitochondria to the cytoplasm (Fig. 1.2) [39, 40]. In the cytoplasm, CYC binds the apoptotic peptidase activating factor 1 (APAF1) and the inactive caspase-9 (procaspase-9), stimulating the activation of caspase-9 [41, 42], which subsequently activates caspase-3 (Fig. 1.2) [43].

The intrinsic pathway is regulated by inhibitors of apoptosis (IAPs) including BIRC2

(Baculoviral IAP repeat containing 2), BIRC3, BIRC5, BIRC7, XIAP (X-linked IAP), and NAIP (NLR-family apoptosis inhibitory protein, also known as BIRC1), that block apoptosis by binding to caspases (Fig. 1.2) [44]. During apoptosis, DIABLO is released from the mitochondria into the cytoplasm and associates to IAPs, halting their inhibitory effect [40]. In addition, the intrinsic pathway is controlled by members of the B-cell

6 chronic lymphoid leukemia/Lymphoma 2 (Bcl-2) family of cell death regulators, involved in the formation of the outer mitochondria membrane pore [45]. The anti- apoptotic members of the BCL2 family including BCL2, BCL2L1 (BCL2-like protein 1),

BCL2L2 and MCL1 (Myeloid Cell Leukemia 1) prevent mitochondria permeabilization and the release of CYC and DIABLO from the mitochondria [46]. The pro-apoptotic members, BAX (BCL2 associated X protein), BAK1 (BCL2 antagonistic/killer 1), BID

(BH3 interacting domain death agonist), BAD (BCL-associated agonist of cell death),

BIK (BCL2-interacting killer), BCL2L11 and BBC3 (BCL2 binding component 3) promote mitochondrial membrane permeabilization (Fig. 1.2) [46].

The extrinsic pathway is triggered by the interaction of death ligands such as FASL

[Apoptosis stimulating fragment (FAS) ligand], TNFα or TRAIL (TNF-related apoptosis inducing ligand) to the death receptors FAS, TNFRS1A (tumor necrosis factor receptor superfamily, member 1A) and TNFRS10A/B, respectively [47]. Binding of death ligands causes receptor oligomerization, followed by the recruitment of the adapter protein Fas- associated Death Domain (FADD) and initiator caspases, including caspase-8 and caspase-10, forming the death inducing signaling complex (DISC) [48]. DISC formation activates caspase-3 [49], a process that is inhibited by CFLAR (CASP8 and FADD-Like

Apoptosis Regulator), also known as cFLIP [50]. DISC can also cleave BID, connecting the extrinsic with the intrinsic apoptotic pathways (Fig. 1.2) [51]. In this study, we showed that apigenin promotes apoptosis in leukemia cells (chapter 3), monocyte/macrophages (chapter 5) and breast cancer cells (chapters 5 and 6) by modulating the activity or splicing of key regulators of the apoptotic pathway.

1.5 Mechanisms of breast carcinogenesis

Breast cancer (BC) is the most common invasive malignancy among women and the second leading cause of cancer-related deaths [52]. More than 250,000 cases and ~40,000 deceases are associated with BC in the USA per year [52]. BC is a progressive disease initiated by a proliferative lesion in the epithelium such as atypical hyperplasia (Fig.

1.3A) [53], which progresses to late stages including carcinoma in situ and invasive carcinoma (Fig. 1.3A) [53]. Breast cancer is traditionally classified based on the expression of cellular receptors, i.e. progesterone (PR), estrogen (ER) and human epidermal growth factor receptor 2 (Her2, also known as neu or ERBB2) [54]. Triple negative breast cancer (TNBC), comprising ~15-20% of all BC cases, refers to a metastatic subtype of mammary malignancies characterized by the lack of expression of the three receptors (PR-, ER- and Her2-) [55]. Thus far, no cure is available for TNBCs

[55]. Another subtype of mammary tumors, accounting for ~10-15% of all BC cases, is the metastatic Her2+ breast cancer, characterized by overexpression of Her2 and low expression of ER and PR [56]. A common treatment for this type of BC is based on administration of a monoclonal antibody against Her2, also named Trastuzumab. Yet, nearly 70% of the patients with metastatic Her2+ tumors are resistant to Trastuzumab

[56]. The lack of treatments for metastatic breast cancers has increased the interest on finding therapies effective at early stages that may help delay or eliminate tumor progression. However, the genetic heterogeneity of breast cancers, the difficulties in early detection methods and the lack of clinical studies has limited our understanding of the molecular mechanisms underlying the progression from proliferative breast lesions into

8 advanced carcinomas.

To reproduce the pathological stages characteristics of human breast cancer progression, mouse models have been generated. Murine models mimicking human

TNBC development are not yet available [57, 58]. However, animals bearing mutations in the BRCA1 gene [58-60], found in ~10% of BC patients [61], show features of TNBC, but only in ~50% of the tumors [58]. Thus, these models are not considered 100% penetrant TNBC. In addition, BRCA1-deficient models presented low incidence of metastasis (~15%), suggesting that tumor progression to metastatic carcinoma was incomplete [57-60]. Transgenic mice overexpressing oncogenes such as Her2 and c-myc in the mammary epithelium developed Her2+ breast carcinomas [62, 63]. Yet, tumor formation required pregnancies, had long latency (>24 weeks old) and low incidence of metastasis [62-64]. FBV-MMTV-PyMT (Mammary Tumor Virus-Polyoma Virus Middle

T antigen, referred as PyMT+) transgenic mice that express the oncogene PyMT in the mammary gland were generated as a model that exhibits 100% tumor penetrance and more than 90% incidence of metastases, short latency (12 weeks old) and is pregnancy- independent [64, 65]. Importantly, PyMT+ mice resemble the stages of tumor progression characteristic of human BC, including early proliferative lesions such as hyperplasia (H) and adenoma (A), as precursors of early and late invasive carcinomas (EC and LC, respectively, Fig. 5.3B), and reproduce the molecular changes observed in human BCs such as over-expression of Her2 and low expression of PR and ER [66]. Hence, the

PyMT model constitutes a well-accepted pre-clinical animal model to study the mechanisms of breast cancer development. Using the PyMT model, we showed, in

9 chapter 5, that apigenin delays the progression of breast tumors to invasive and metastatic carcinoma and studied the mechanisms for the anti-carcinogenic activity of apigenin in vivo.

1.5.1 Role of monocytes and macrophages in breast cancer

The immune system is considered an extrinsic tumor suppressor machinery that can recognize and eliminate malignant cells [67]. However, the tumor microenvironment re- educates immune cells into pro-carcinogenic [67]. Macrophages constitute the highest population of tumor infiltrating leukocytes, playing a key role in the regulation of cancer progression [67]. Macrophages are typically classified as M1 and M2 [68]. Macrophages are polarized into an M1 phenotype, also known as classically activated macrophages, by interferon gamma (IFNγ), TNFα or TLRs ligands such as LPS [69]. Once activated, M1 macrophages release pro-inflammatory cytokines including interferons and interleukins and produce nitric oxide (NO) and ROS that help killing malignant cells [70]. M1 macrophages have antigen-presenting capabilities, hence activating cytotoxic cells to fight tumors [68]. Thus, M1 macrophages work as soldiers to fight against cancer cells.

In contrast, M2 macrophages, also known as alternatively activated macrophages, are polarized by IL-10, IL-4, tumor growth factor-beta (TGF-β), among others [68]. Cancer and stroma cells (cells of the tumor microenvironment) produce growth factors/chemokines such as colony stimulating factor 1 (CSF1) and CCL2 that stimulate monocyte survival and proliferation, in an NF-κB-mediated pathway, leading to amplification of monocyte numbers in bone marrow and peripheral blood [9, 71-73].

Macrophages derive from monocytes, which are attracted to the tumors by the action of chemokines including CCL2, CXC chemokine ligand 12 (CXCL12), vascular endothelial growth factor A (VEGFA) and CSF1 [72-76]. Subsequently, tumor associated macrophages (TAMs) are polarized to a tumor-prone M2 phenotype with different functions within the tumor microenvironment: 1) Release of growth factors and chemokines that induce tumor cell survival and proliferation [77, 78]; 2) Release of angiogenic factors, including VEGFA, that stimulate tumor vascularization and metastasis [75]; 3) Inhibition of cytotoxic cells, thereby promoting tumor evasion of the immune system [79].

Experiments in the PyMT model demonstrated that macrophages infiltrate breast tumors in high numbers and induce angiogenesis triggering metastasis [80, 81]. Depletion of macrophages using pharmacological inhibitors of the CSF1 receptor (CSF1R), anti-

CCL2 monoclonal antibodies or genetically deleting CSF1 or CCL2, showed decreased breast cancer growth and metastasis [71, 79, 82-86]. Thus, macrophages are a key component of the tumor microenvironment mediating breast cancer progression by promoting cell survival, angiogenesis, invasion, and metastasis. Targeting macrophage infiltration is a promising alternative for the treatment of breast cancer.

Macrophages originate from monocytes (Fig. 1.1), which are found in peripheral blood, spleens and bone marrow [87]. Monocytes expand during carcinogenesis resulting in higher macrophage differentiation and infiltration into tumors [76, 81, 87]. Monocytes are classified in non-classical and pro-inflammatory or classically activated monocytes

(Fig. 1.1) [4, 88, 89]. In addition, immature myeloid cells, referred as myeloid derived

11 suppressor cells (MDSC), are also a source of infiltrating macrophages (Fig. 1.1). MDSC constitute a heterogeneous population from two lineages, granulocytes (G-MDSC) and monocytes (Mo-MDSC) [89, 90]. In tissues, TAMs are recruited from pro-inflammatory monocytes and Mo-MDSC (Fig. 1.1) [89, 90]. Similar to TAMs, increasing numbers of circulating MDSC are correlated with poor BC prognosis [91, 92]. Moreover, MDSC inhibit immune cytotoxic cells, thereby promoting tumor evasion of the immune system

[92]. In chapter 5, we showed that apigenin delays breast cancer progression by immune- modulating monocytes and macrophages.

1.5.2 Role of NF-κB in breast cancer

NF-κB is constitutively activated in breast cancer [93-96]. NF-κB activation is dependent on its phosphorylation by IKKβ and Casein Kinase 2 (CK2). Accordingly, dominant negative versions of these kinases decreased NF-κB activity [97, 98]. Inhibition of NF-κB decreases cell proliferation and induces apoptosis in breast cancer cell lines

[93]. In vivo, inducible inhibition of NF-κB, by overexpressing a dominant negative version of the NF-κB inhibitor, IκBα, in the epithelium, induced cancer cell apoptosis, impaired proliferation and decreased macrophage infiltration resulting in reduced tumor burden, angiogenesis and metastasis in the PyMT model [99, 100]. Demonstrating that

NF-κB is also an important regulator of immune cells survival [101], pharmacological inhibition of NF-κB induced apoptosis in monocytes and macrophages [101-103]. In addition, deletion of IKKβ, a direct activator of NF-κB, in macrophages, inhibited M2 polarization and promoted macrophage tumoricidal activity causing tumor regression in a

12 murine model of ovarian cancer [104]. Hence, NF-κB is a key player during breast carcinogenesis by regulating epithelial cell proliferation and macrophage functions.

1.5.3 Transcriptome alterations in breast cancer

A driver mutation refers to a genomic alteration that confers selective growth advantage, promoting cancer [105]. It is estimated that for cancer initiation, 5 to 10 driving mutations are necessary [106]. These mutations generally occur in oncogenes or tumor suppressors, resulting in genome-wide alterations of transcriptome diversity, e.g gene expression and mRNA processing, giving the transformed cell a proliferative advantage, resistance to apoptosis, evasion of the immune system and/or metastatic potential (Fig. 1.4) [107]. Microarray analyses in 65 tumor human samples permitted classify BC, based on gene expression patterns, into five subtypes including: Luminal A,

Luminal B, Basal-like (also referred as TNBC), Her2+ and normal-like subtypes [108,

109]. A recent meta-analysis study of global gene expression in TNBC identified 206 genes that are recurrently dysregulated in TNBC compared to non-TNBC and normal breast tissues [110]. These 206 genes are proposed as aggressiveness markers in TNBC, comprising among others genes that regulate cell cycle and chromosome instability including cyclins such as CCNB1, and the cyclin-dependent kinase inhibitor 1A

(CDKN1A, also known as p21) [110]. Hence, there is a remarkable contribution of dysregulated gene expression to breast carcinogenesis (Fig. 1.4). In chapter 6, we studied the effect of apigenin on gene expression in the triple negative human breast cancer cell line MDA-MB-231 using genome-wide approaches, and determined the overlapping

13 between the genes dysregulated in TNBC with those affected by apigenin.

1.5.3.1 Alternative processing of mRNA

In addition to driving mutations affecting changes in gene expression, alternative processing of mRNA is key in the pathophysiology of human cancers [111]. Abnormal isoform expression of molecules involved in apoptosis, metabolism, migration and epithelial-mesenchymal transition (EMT) contribute to tumorigenesis [112-119]. More than 90% of all human genes undergo alternative mRNA processing, underscoring the importance of these mechanisms for transcriptome diversity [120]. Alternative transcript isoforms result from the differential inclusion of subsets of exons and introns through mechanisms that include alternative splicing (AS), alternative transcription start site

(ATSS) or alternative polyadenylation (APA) [120]. Eight main classes of mRNA processing events (also referred as splicing events [120]) are recognized (Fig. 1.5): skipped exons (SE), retained introns (RI), mutually exclusive exons (MXE), alternative 5′ splice site (A5SS) and alternative 3′ splice site (A3SS) events are generated through AS.

ATSS results in transcripts with an alternative 5’ first exon (AFE), while APA produces alternative 3’ last exons (ALE) or isoforms harboring shorter or longer 3′ UTRs (TUTR,

Fig. 1.5) [120]. These processes are regulated by RNA binding proteins, e.g. heterogeneous nuclear RiboNucleoProteins (hnRNPs), that either weaken or strengthen exon inclusion [121]. Interestingly, hnRNPs, for example hnRNPA2, are generally overexpressed in several malignancies including breast cancer [122-125], contributing to the abnormal expression of alternative transcript isoforms during tumorigenesis [125].

However, the specific RNA-binding proteins regulating each class of splicing event have not been determined yet. In chapter 6, we showed that apigenin modulates mRNA processing in triple negative breast cancer cells genome-wide.

Alternative splicing isoforms have been described in apoptotic genes [126]. For example, the splice isoform caspase-9a encodes the functional apoptotic caspase-9, which is responsible for inducing cell death. In contrast, the splice isoform caspase-9b, lacking exons 3–6, encodes a caspase-9 protein that exhibits a dominant-negative activity and inhibits apoptosis [127]. In the case of the cFLIP protein, also known as CFLAR, a transcript with an alternate last exon (exon 7 instead of 14), encodes a shorter isoform, cFLIPS, which prevents the activation of specific death receptors [128]. Notably, the anti- apoptotic isoforms of these genes are generally overexpressed in malignant cells conferring resistance to apoptosis [129, 130]. Furnishing approaches that restore normal gene expression and/or alternative mRNA processing will be advantageous, helping to reduce the typical resistance to apoptosis found in cancer cells, thereby contributing to the treatment and prevention of breast cancer.

1.6 Flavonoids

Flavonoids are ubiquitously found plant molecules characterized by a C6-C3-C6 core structure (Fig. 1.6A). According to modifications of this core structure, they are classified in several subgroups including flavanones, flavones, flavonols, isoflavones and anthocyanins (Fig. 1.6B-F) [131]. Moreover, flavonoids are usually found in plants linked to sugars (glycosylated, Fig. 1.6G) increasing their diversity [131]. More than

8,000 flavonoids have been described, constituting the most abundant class of dietary phytochemicals in our diet [131]. Flavonoids are associated to several biological activities including anti-allergic, anti-microbial, anti-proliferative, anti-viral, and anti- inflammatory functions [131-135]. However, their mechanisms of action remain largely unknown.

1.6.1 Anti-carcinogenic activity of apigenin

Apigenin [4’,5,7-trihydroxyflavone, (Fig. 1.6B)], is a flavone abundant in celery and parsley, two main components of the Mediterranean diet, and chamomile tea [136, 137].

Epidemiological studies reported apigenin as the only flavonoid, among six tested, which lowers the incidence of ovarian cancer, when ingested as part of the Mediterranean diet

[138]. We previously showed that apigenin decreases proliferation and induces apoptosis in leukemia and solid tumor cell lines, but had no effect on non-malignant immortalized epithelial cells [139]. Luteolin, a flavone structurally related to apigenin, differing by an –

OH group in ring B (Fig. 1.6B), the flavonol quercetin (Fig. 1.6C), and the isoflavone genestein (Fig. 1.6E) were also shown to have anti-carcinogenic activity in cancer cell lines [140-142]. In contrast, naringenin, a flavanone structurally different from apigenin by a double bond in ring C (Fig. 1.6D), had not effect on cell proliferation [139]. These findings suggest a structural relationship between flavonoids and their anti-cancer activity.

Treatment with apigenin of leukemia cells resulted in the PKCδ and p38-dependent phosphorylation of Hsp27, a protein that inhibits caspase-3, thereby allowing caspase-3-

16 dependent execution of apoptosis [139, 143]. Apigenin increased the production of ROS in leukemia cell lines, but this effect was not required for apigenin-induced apoptosis

[139]. Additionally, apigenin induced the release of CYC into the cytoplasm and decreased expression of the anti-apoptotic proteins BCL2 and MCL1 promoting apoptosis in leukemia cells [144].

Apigenin promoted apoptosis through the intrinsic pathway in cancer cell lines by inducing the expression of BAX, while decreasing XIAP and BCL2L1, resulting in caspase-3 activation [145-149]. Moreover, apigenin activated the extrinsic apoptotic pathway by increasing the expression of the death receptor TNFRSF10B (also known as death receptor 5 or DR5) and reducing the expression of its inhibitor cFLIP (CFLAR)

[150, 151]. As a consequence, apigenin sensitized tumor cells to TRAIL-induced apoptosis [150-152].

Apigenin halted proliferation of several cancer cell lines by disrupting cell cycle progression leading to an arrest in G1 or G2/M, depending on the cell type, independently of p53 [144, 145, 153-155]. Moreover, apigenin decreased proliferation by inhibiting survival signaling pathways, including NF-κB and the PI3K/AKT/mTOR cascades [144,

156-159]. In mouse xenograft models, apigenin decreased lung, breast, colon and prostate tumor growth by increasing apoptosis and reducing proliferation of cancer cells [144,

160-163]. In a spontaneous model of prostate cancer, apigenin was found to delay tumor progression by inhibiting the PI3K/AKT signaling pathway [164], demonstrating that apigenin is an anti-carcinogenic flavonoid in vitro and in vivo.

1.6.2 Apigenin and inflammation

Apigenin decreased the expression of inflammatory cytokines such as TNFα, IL-8,

IL-1β and COX-2 in LPS-treated peripheral blood mononuclear cells (PBMC) [165], in phorbol 12-myristate 13-acetate (PMA)-stimulated mast cells [166], in TNFα stimulated endothelial cells [167] or in Helicobacter pylori-infected gastric cells [168]. We previously showed that apigenin decreases LPS-induced mortality in vivo and reduces the expression of pro-inflammatory cytokines in monocytes and macrophages by abrogating the transcriptional activity of NF-κB [169], through the inhibition of IKKβ [169].

Apigenin decreased LPS-induced acute lung injury by inhibiting TNFα and COX-2 expression and decreasing leukocyte infiltration in brancheoalveolar lavage fluids [170].

We also previously reported that apigenin inhibits IL-8-induced neutrophil migration, suggesting its ability to modulate chemotaxis [171]. Epidemiological studies have correlated the consumption of apigenin with a lower incidence of cardiovascular diseases

[172]. Thus apigenin has potent immune-modulatory and anti-carcinogenic activities in vitro and in vivo. Yet, the underlying molecular mechanisms are poorly understood. The goal of this project was to study the immune-modulatory and anti-carcinogenic mechanisms of the flavonoid apigenin using molecular, biochemical, cellular and in vivo approaches.

1.6.3 Molecular targets of apigenin

Apigenin modulates a broad range of signaling cascades including NF-κB

PI3K/AKT/mTOR and apoptosis [139, 143, 169, 173], demonstrating that this flavone

18 offers health benefits by affecting multiple cellular functions. However, the precise mechanisms of how apigenin regulates these pathways are yet to be uncovered. The cellular targets of apigenin remain largely unknown, inflicting significant challenges for the understanding on how this flavonoid provides health benefits. Few direct targets of apigenin have been described. Apigenin inhibited the activity of purified CK2 kinase in vitro by competing with its ATP binding site (IC50 of ~1 µM) [174]. Recent studies using pull downs with apigenin-linked beads and recombinant purified proteins showed that apigenin interacts with ribosomal protein S9 (RPS9) [175]. Screening of a chemical library found that apigenin interacts with purified MUC1, a protein that associates with

IKKβ and NF-κB promoting their activities in breast cancer cell lines [176]. In addition, apigenin inhibits MUC1 dimerization in vitro, as demonstrated using recombinant purified proteins (IC50 ~75 µM) [177]. Fluoresce spectroscopy approaches found that apigenin interacts at the interface of α and β tubulin heterodimers disrupting microtubule formation [178]. Yet, despite these findings, it is unclear whether apigenin exerts its beneficial effects either by significantly affecting the activity of just a few molecules or through additive gains from modest effects on a large number of cellular proteins. Hence, the pleotropic activities of apigenin, the indirect regulation of signaling pathways and the lack of knowledge on apigenin targets urges a comprehensive identification of its direct cellular targets, as an essential step to understand the underlying immune-modulatory and anti-carcinogenic mechanisms of this dietary phytochemical. The contribution of my work to the field included the comprehensive identification of the direct targets of apigenin.

Figure 1.1. Hematopoiesis. HSC: hematopoietic stem cells. CMP: common myeloid progenitors. CLP: common lymphoid progenitors. GMP: granulocyte-monocyte progenitors. GP: granulocyte progenitors. G-MDSC: myeloid-derived suppressor cells from granulocyte lineage. Mo-MDSC: myeloid-derived suppressor cells from monocyte lineage. Adapted from Schouppe et al 2012 [89], and Ginhoux et al 2014 [179].

20 Figure 1.2. Simplified representation of the apoptosis and NF-!B signaling pathways.

Figure 1.3. Human and mouse models of breast cancer progression. A. Schematic representation of human breast cancer development. The classic model of human breast cancer progression proposes that neoplastic evolution initiates in epithelial cells, progresses to atypical hyperplasia, evolves to early in situ and culminates as invasive carcinoma. Adapted from Bombonati et al 2011 [53]. B. Hematoxylin and eosin (H&E) stainings of mammary gland sections from PyMT+ mice, representing the different stages of tumorigenesis. The murine PyMT+ model of breast cancer progression states that normal epithelium progresses to hyperplasia, advances to adenoma, evolves to early carcinoma and culminates into late invasive carcinoma [66]. Pictures were taken by Daniel Arango Tamayo.

22 Figure 1.4. General model of cancer initiation.

23 Figure 1.5. Different classes of alternative mRNA processing events. Skipped exon (SE), retained intron (RI), mutually exclusive exons (MXE), alternative 5’ splice site (A5SS) and alternative 3’ splice site (A3SS) are generated by alternative splicing. Alternative 5’ first exon (AFE) results from different transcription start sites. Alternative 3’ last exons (ALE) or isoforms harboring shorter or longer 3& UTRs (Tandem 3’ UTR, TUTR) are produced by alternative polyadenylation and cleavage of mRNA.

24 Figure 1.6. Structure and classification of selected flavonoids. A. Characteristic C6- C3-C6 core structure. Examples of flavonoids representing the different subclasses. B. Flavanones. C. Flavonols. D. Flavanones. E. Isoflavones. F. Anthocyanins G. In plants, flavonoids are commonly found linked to sugars. For example, apigenin can be glycosylated at positions 7 (apigenin 7-O-glucoside) or 6 (apigenin 6-C-glucoside).

25 Chapter 2

Material and Methods

2.1. Chemicals and reagents

DAPI (4’,6-diamidino-2-phenylindole), DHE (dihydroethidium), DCFDA (2,7-

dichlorofluorescein diacetate), protein A agarose beads, RPMI-1640 (Roswell Park

Memorial Institute medium-1640), DMEM (Dulbecco's Modified Eagle Medium),

DMEM/F12 medium, IMDM (Iscove's Modified Dulbecco's Medium), P/S

(penicillin/streptomycin), pENTR-D-TOPO vector, Gateway LR Clonase®, TRIzol®,

DNAse-I, TaqMan® Universal PCR Master Mix, SYBR® Green PCR Master Mix,

ThermoScriptTM RT-PCR system, TaqMan® miR Reverse Transcription kit, DNA oligonucleotides and the miR assay kits for murine miR-155 (miR-155-5p strand, assay

ID 002571), miR-lethal-7a (miR-let-7a, assay ID 000377), small nucleolar RNA-202

(snoRNA-202, assay ID 001232) and lipofectamine-2000 were purchased from Life

Technologies (Carlsbad, CA).

Apigenin, luteolin, chrysoeriol, naringenin, eriodictyol, quercetin, kaempferol, flavopiridol, genistein, apigenin 7-O-glucoside, apigenin 6-C-glucoside, DMSO

(dimethyl sulfoxide), LPS (Lipopolysaccharide, E. coli serotypes

O111:B4 and 0128:B8), proteinase K, H2O2, Kolliphor EL, NMPA (normal melting

point agarose), LMPA (low melting point agarose), FBS (fetal bovine serum),

hydrocortisone, cholera toxin, insulin, Na-N-lauryl sarcosinate, Triton X-100, PMSF

(Phenyl-methane-sulfonyl-fluoride), Na-glycerophosphate, Na-pyrophosphate, NaF,

orthovanadate, chymostatin, leupeptin, antipain, pepstatin, imidazole, UDP-glucose

(Uridine 5'-diphosphoglucose), sodium glycine, NAD+ (Nicotinamide Adenine

Dinucleotide), EGTA (Ethylene Glycol Tetra-acetic Acid), sodium isocitrate, reduced

glutathione, BSA (bovine serum albumin), DTT (Dithiothreitol), bromophenol blue,

propidium iodide (PI) and ethidium bromide were obtained from Sigma (St. Louis, MO).

KCl, Na2HPO4, KH2PO4, NaOH, MnCl2, MgCl2, PIPES [Piperazine-N,N′-bis(2-

ethanesulfonic acid], Hepes [4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid],

H3PO4, sulfuric acid, acetone, ACN (acetonitrile), TFA (trifluoroacetic acid), Xylene, formalin, methanol, ethanol, isopropanol, chloroform, paraformaldehyde and the

Enhanced chemo-luminescence (ECL) Western Blotting Substrate were purchased from

Thermo Fisher Scientific (Waltham, MA).

NaCl, EDTA (Ethylenediaminetetraacetic Acid), Tris [Tris-(hydroxymethyl)- aminomethane], glycine, SDS (sodium dodecyl sulfate) and agarose were from Research

Products International (RPI, Prospect, IL). Caspase-3 inhibitor DEVD-FMK and the caspase-3 substrate DEVD-AFC were from Enzyme System Product (Livermore, CA).

The ROS (Reactive Oxygen Species) inhibitor EUK-134 was from Cayman Chemical

(Ann Arbor, MI). Epidermal Growth Factor (EGF) was purchased from Prepotech

(Rocky Hill, NY). Nonidet-NP40, Tween-20 and glycerol were purchased from Amresco

(Framingham, MA). Acrylamide was purchased from BIORAD (Hercules, CA). The p38

inhibitor SB203580 and the PKCδ inhibitor rottlerin were from Calbiochem (San Diego,

CA). Polyethyleneglycol-polyacrylamide copolymer (PEGA) beads, T7 Select® Human

Breast Tumor cDNA phage library and nitrocellulose membranes were from EMD

Millipore (Cambridge, MA). [γ-32P]-ATP was from Perkin Elmer (Waltham, MA),

phosphatidyl-serine and diacylglycerol were from Avanti Polar Lipids (Alabaster, AL).

Purified histone-2B, D-luciferin, Collagenase D and RNAse A were from Roche

(Indianapolis, IN). Powdered AIN-93G control diet was purchased from Harlan

(Indianapolis, IN). Ultrapure T4 ligase was obtained from Enzymatics (Beverly, MA).

Phusion Hot Polymerase and the restriction enzymes KpnI and SpeI were purchased from

New England Biolabs (Ipswich, MA). Antibodies used in this work are described in

Table. 2.1.

2.2 Cell lines and culture

THP-1 human monocytic leukemia cells and RAW 264.7 murine macrophages were

cultured in endotoxin-free RPMI-1640 supplemented with 5% FBS and 1% P/S. MDA-

MB-231 human TNBC cells, HeLa human cervical carcinoma cells and A549 human

lung adenocarcinoma cells were cultured in DMEM supplemented with 10% FBS and 1%

P/S. MCF-10A human immortalized breast epithelial cells were grown in DMEM/F12

medium supplemented with 10% FBS, 1% P/S, 20 ng/ml EGF, 0.5 mg/ml

hydrocortisone, 100 ng/ml cholera toxin and 10 µg/ml insulin. L-929 human fibroblasts

were grown in IMDM medium supplemented with 10% FBS and 1% P/S to obtain L-cell

conditioned media, rich in CSF1. All cell lines were obtained from ATCC (Manassas,

VA).

Bone Marrow Derived Macrophages (BMDMs) were differentiated by incubating the

flushing of mice femurs and tibia in RPMI-1640 supplemented with 10% FBS, 30% L-

cell media and 1% P/S in non-treated tissue culture plates for 7 days, followed by two

washes in PBS (Phosphate Buffer Solution) before harvesting for further experiments.

The PyMT murine breast cancer cell line was obtained from Dr. Timothy Eubank (The

Ohio State University) and cultured in DMEM supplemented 10% FBS and 1% P/S as

previously described [180]. All cells were cultured at 37°C in 5% CO2 environment.

2.3 Analysis of cell cycle and proliferation

For cell cycle analysis, cells were collected after treatment, washed twice with PBS prior to fixation in 70% ethanol, washed again with PBS twice and stained with 50 µg/ml propidium iodide (PI) containing 0.2 mg/ml DNAse-free RNAse A for 30 min at room temperature and immediately analyzed by flow cytometry.

Cell proliferation was determined by the MTS assay using the CellTiter 96 Aqueous

One Solution according to manufacturer’s protocol (Promega, Madison, WI). For this

purpose, 6,000 cell/100 µl of culture medium were treated with increasing concentrations

of apigenin or diluent DMSO control for 24 or 48 h. The reduction of tetrazolium salts to

formazan was measured by changes in absorbance at 490 nm using the EnSpire

multimode plate reader (PerkinElmer, Waltham, MA).

2.4 Caspase-3 activity and apoptosis

The percentage of apoptotic cells was determined by staining with calcein A/M. Cells

(1 x 106/ml) were collected after treatment, washed once with PBS and incubated with 1

µg/ml calcein A/M for 30 min followed by addition of 50 µg/ml propidium iodide for 5 min. Cells were rinsed twice with PBS and 200 cells were scored using a fluorescence microscope (Olympus BX40 equipped with the Optronics DEI 750E CE Digital Camera), as previously described [139]. Cells stained green (calcein A/M+) and red (PI+ positive)

or red alone were considered apoptotic, while green cells in the absence of red were

considered alive. The percentage of apoptotic cells was expressed as the number of PI+

positive over total number of cells and multiplied by 100.

Caspase-3 activity was used as an enzymatic indicator of apoptosis. Cells (1 x 106) were lysed in 50 µl of NP-40 lysis buffer containing 0.5% NP-40, 50 mM Tris, pH 7.4,

150 mM NaCl, 10 mM Na-glycerophosphate, 5 mM Na-pyrophosphate, 50 mM NaF, 1 mM orthovanadate, 1 mM DTT, 0.1 mM PMSF, 2 µg/ml of the protease inhibitor cocktail CLAP (chymostatin, leupeptin, antipain, and pepstatin) for 20 min at 4°C. Cell lysates (10 µg total protein) were mixed with 20 µM DEVD-AFC in 100 µl of cytobuffer containing 100 mM PIPES, 20% glycerol, 2 mM EDTA and 1 mM EDTA and caspase-3 activity was determined as nM of AFC released per time per mg of protein, as previously described [139]. AFC was determined using a Cytofluor 400 fluorometer (Filters: excitation 400 nm, emission 508 nm; Perspective Co., Framingham, MA).

2.5 Intracellular measurement of ROS

Cellular production of reactive oxygen species (ROS) was measured by fluorescence

microscopy. Briefly, THP-1 cells (1x106/ml) were pre-treated with 20 µM EUK-134 for 1 h, prior to the addition of 50 µM apigenin and incubated by an additional 1 h. Cells were rinsed with PBS twice and incubated in RPMI-1640 (without phenol red) in the presence of 10 µM DHE, a specific superoxide anion (O2-) probe, or 20 µM DCFDA, a general

ROS probe, for 30 min at 37°C, rinsed with PBS twice and visualized using the fluorescence microscope (Olympus BX40 equipped with the Optronics DEI 750E CE

Digital Camera). Fluorescence intensities were measured from 50 randomly selected cells using the ImageJ software (National Institute of Health, Bathesda, MD).

2.6 Alkaline comet assay

DNA damage was assessed by the alkaline comet assay (single cell gel electrophoresis) as described [181]. Briefly, cell suspension (20 µl; 106 cells/ml) were

mixed with 80 µl of 0.6% LMPA and dipped onto microscope slides, pre-coated with 1%

NMPA, and covered with coverslips at 4°C for 5 min. Coverslips were removed and

slides were covered with 0.6% LMPA and incubated at 4°C for 5 min with coverslips.

Coverslips were removed and slides were immersed overnight in lysis buffer (2.5 M

NaCl, 100 mM EDTA, 10 mM Tris, 1% Na-N-lauryl sarcosinate, 10% DMSO, 1% Triton

X-100, pH 10). After lysis, slides were washed with PBS and DNA denatured in alkaline

buffer (300 mM NaOH, 1 mM EDTA, pH 13) for 30 min followed by electrophoresis at

28 V and 300 mA for 30 min. Slides were then rinsed 3 times with 400 mM Tris, pH 7.5

and stained with 0.2 µg/ml ethidium bromide for 1 min. Comets were analyzed using the

Optronics DEI 750E CE Digital Output (Optronics, Goleta, CA) mounted on an Olympus

BX40 fluorescence microscope. Fluorescence intensities were measured from 100 randomly selected cells for each biological sample using the Comet Assay Software

Project (CASP, [182]) and the extent of DNA damage reported as % Tail DNA.

2.7 Identification of histone phosphorylation by LC–MS

Liquid chromatography–mass spectrometry (LC–MS) was performed by Dr. Liwen

Wang from Dr. Michael Freitas’ lab at The Ohio State University as described [183].

Briefly, 107 cells were washed with 10 mM Tris–HCl pH 7.5 and resuspended in lysis buffer (0.1% NP-40, 10 mM Na-glycerophosphate, 5 mM Na-pyrophosphate, 50 mM

NaF, 1 mM sodium orthovanadate, 1 mM DTT, 0.1 mM PMSF, 2 µg/ml CLAP. After centrifugation, the pellets were collected and washed with Tris–HCl buffer and extracted with 0.4 N sulfuric acid followed by overnight precipitation in 80% acetone. The resulting precipitate was collected, dried, and dissolved in 100 µl of 20% acetonitrile

(ACN) and 0.05% trifluoroacetic acid (TFA) solution. Histone modifications were characterized by LC–MS using reversed phase high performance liquid chromatography

(HPLC, Waters model 2690, Milford, MA) coupled to a MicroMass Q-TOF mass spectrometer (MicroMass, Wythenshawe, Manchester, UK). Histone mixtures were separated on a 1.0 mm x 150 mm C18 column (Discovery Bio wide pore C18 column, 5

µM, 300 Å, Supelco, Bellefonte, PA). Mobile phase A contained ACN with 0.05% TFA.

Mobile phase B contained 0.05% TFA in HPLC grade water. Starting with 20% B, the

gradient increased linearly to 30% B in 2 min, from 30% B to 35% B in 8 min, from 35%

B to 50% B in 20 min, from 50% B to 60% B in 5 min, from 60% B to 95% B in 1 min

and stayed in 95% for 4 min at the flow rate of 25 µl/min. LC–MS data was deconvoluted

using the Masslynx 4.0 software package (MicroMass, Wythenshawe, Manchester, UK).

Phosphorylation of histones is indicated by the presence of species with a mass shift of

+80 Da compared to the original mass.

2.8 Western blots

Protein expression or phosphorylation in cellular lysates was determined by western

blot. Cell lysates to detect phosphorylated γH2AX were prepared from 1x106 cells in 50

µl of 20 mM Hepes pH 7.4, 150 mM NaCl, 1 mM EDTA, 1.5 mM MgCl2, 0.2% Tween-

20, 10 mM Na-glycerophosphate, 5 mM Na-pyrophosphate, 50 mM NaF, 1 mM orthovanadate, 1 mM DTT, 0.1 mM PMSF, 2 µg/ml CLAP for 30 min on ice followed by sonication using a Branson Sonifier 450 (output control 3, duty cycle 30, pulses 3, Brason

Ultrasonics Corporation, Danbury, CT) and centrifuged at 14,000 x g for 10 min at 4°C.

Cell lysates for immunodetection of all other proteins were prepared from 1x106 cells in

50 µl of NP40 lysis buffer for 30 min on ice followed by centrifugation at 14,000 x g for

10 min at 4°C. Tissue lysates were prepared from 50 mg tissue homogenized in 400 µl

NP40 lysis buffer using a Brinkmann Polytron Homogenizer (Brinkmann Instruments

Inc, Westbury, NY) followed by centrifugation at 14,000 x g for 10 min at 4°C. Equal amounts of protein were separated by SDS-PAGE using SDS-Tris-Glycine buffer (1%

SDS, 250 mM Tris, 1.92 M Glycine) transferred onto nitrocellulose membranes using

Tris-Glycine buffer (250 mM Tris, 1.92 M Glycine, 20% methanol) and immunoblotted

with primary antibodies followed by horseradish peroxidase (HRP)-conjugated secondary

antibodies and visualized by enhanced chemo-luminescence using the ECL Western

Blotting Substrate. Antibodies used for western blot are described in Table 2.1.

2.9 Immunoprecipitations and in vitro kinase assays

PKCδ activity was evaluated by in vitro kinase assays. For this purpose, cell lysates

were prepared from 10x106 cells in 500 µl of NP-40 lysis buffer by incubating 30 min on ice followed by centrifugation at 14,000 x g. Soluble protein fraction (250 µg) was immunoprecipitated overnight at 4°C with 200 ng anti-PKCδ antibodies or isogenic IgG as control, followed by 1 h incubation with protein A-agarose beads. Immunoprecipitates were rinsed three times with NP-40 lysis buffer and twice with kinase buffer (25 mM

Hepes pH 7.4, 10 mM MnCl2, 1 mM MgCl2, 1 mM DTT, 0.1 mM PMSF) and subjected

to in vitro kinase assays for 1 h at 37°C in the presence of 20 ml kinase buffer containing

2.5 µCi of [γ-32P] ATP, 0.5 mM ATP, 200 µg/ml phosphatidyl-serine, 20 µg/ml

diacylglycerol and 2.5 µg of H2B as exogenous substrate. Reactions were stopped by the

addition of 10 ml 5X Laemmli buffer (2% SDS, 10% glycerol, 0.02% bromophenol blue,

60 mM Tris-HCL pH 6.8), boiled, resolved by SDS-PAGE and subsequently transferred

to membranes. Phosphorylated H2B was visualized by autoradiography and the same

membranes were re-blotted with anti-PKCδ antibodies.

2.10 Immunofluorescence of γH2AX

To evaluate γH2AX levels by immunofluorescence, cells (1x106) were fixed in 100 µl of 2% paraformaldehyde for 10 min at room temperature, rinsed twice with PBS, collected in slides by cytospinning centrifugation and subsequently permeabilized in

0.2% Triton X-100 at 4°C for 15 min and blocked with PBS containing 1% FBS and 100

µg/ml human total IgG for 30 min at room temperature. Slides were incubated with 350

µg/ml anti-γH2AX antibodies for 1 h at RT, rinsed with PBS twice and incubated with

350 µg/ml anti-mouse antibodies Alexa Fluor 488 conjugated for 1 h at room

temperature. Slides were then rinsed twice with PBS and stained 5 min with 0.05 µg/ml

DAPI at room temperature, washed twice with PBS and visualized using the Optronics

DEI 750E CE Digital output mounted on Olympus BX40 fluorescence microscope.

2.11 siRNA silencing

To knock-down the expression of PKCδ and p38, 10 x 106 THP-1 cells were transfected with 100 nM siRNAp38 (Cell Signaling, Cat: 6386), siRNA-PKCδ (Qiagen,

Cat:1027283) or siRNA scramble control (Qiagen, Valencia, CA; Cat: 1027284) in 100

µl of nucleofector solution using the Amaxa nucleofector program V-001 (Lonza, Basel,

Switzerland) as previously described [143]. For gene silencing of hnRNPA2, 5 x 105

cells/ml were transfected with siRNAs against hnRNPA2 (5’-GGAACAGUUC

CGUAAGCUC-3’) or scramble control using lipofectamine-2000. Forty-eight after

transfections cells were collected and efficiency of silencing was determined by western-

blot.

2.12 RNA isolation and reverse transcriptase-PCR (RT-PCR)

Total RNA was isolated from 1 x 106 cells or ~50 mg tissue using 1 ml TRIzol. After

lyses, RNA was extracted with 200 µl chloroform, precipitated with 100 µl isopropanol

and washed with 75% ethanol twice. RNA pellets were air dried and resuspended in 50 µl

RNAse-free water. RNA (1 µg) was treated with 1 U of DNAse I for 10 min at room

temperature. DNAse-treated RNA was reverse transcribed to cDNA by incubating 1 µg

RNA, 1 µM oligo(dT) primers, 1mM dNTP, 0.5 µM DTT at 65°C for 5 min followed by

incubation with 2 U RNAse-out, 0.75 U Thermoscript-RT and 1X cDNA synthesis buffer

at 50°C for 60 min. Reactions were stopped by incubating at 85°C for 5 min.

2.13 Quantitative RT-PCR (qRT-PCR) analysis

RNA expression was measured by qRT-PCR. For this purpose, a 20 µl mixture

containing 2 µl of template cDNA (~20 ng), 0.25 µM of each primer and 10 µl SYBR

Green Master Mix was run in an ABI 7900HT RT-PCR system using the Sequence

Detection System (SDS) 2.4 software (Applied Biosystems) and the following

conditions: 95°C for 10 min, 40 cycles of 95°C for 1 min, 60°C for 1 min, and 72°C for 1

min. Fold change in expression was calculated as: 2ΔCt(treatment)/Average[2ΔCt(controls)], where ΔCt = (Ct gene of interest - Ct Internal control). All selected genes were normalized to the expression of at least one internal control. Internal controls are described in the figure legends. All primers used for qRT-PCR are described in Table 2.2.

2.14 Microarray analysis

Genome-wide mRNA expression in THP-1 cells was analyzed by microarrays. Total

RNA was isolated from 1 x 106 THP-1 cells treated for 3 h with diluent DMSO or 50 µM apigenin. RNA integrity was analyzed using 5 ng RNA in the Agilent 2100 Bioanalyzer

(Agilent Technology, Santa Clara, CA). RNA integrity number (RIN) higher than 9.5 was

found acceptable for further experiments. cDNA was synthesized and hybridized to the

Human Gene 1.0 ST Array Data Set chip (Affymetrix, Santa Clara, CA) at the Nucleic

Acid Shared Resource (The Ohio State University). The array contained 28,869

annotated human transcripts with 764,885 distinct probes. All experiments were

performed in biological triplicates. Dr. Xiaokui Mo from the Center for Biostatistics, The

Ohio State University, provided assistance on the microarray analysis. Signal intensities,

background correction, and normalization were performed using the Affymetrix

Expression Consol software and applying the RMA method (Affymetrix). One-way

ANOVA was used to analyze the normalized data using R 2.9.0. The variance shrinkage

method was employed to improve the estimates of variability and statistical tests of

differentially expressed genes [184]. Genes significantly changing (p < 0.01) between

groups with a fold change greater than 1.2 were further analyzed using Ingenuity

Pathways Analysis (IPA) software (Ingenuity Systems, Redwood City, CA). IPA

identified the biological functions, based on the Gene Ontology (GO), that were most

significant to the data set.

2.15 High-throughput screening of miRs (HTS-miRs)

To evaluate the expression of miRs in mouse macrophages, total RNA was isolated

from 1 x 106 RAW 264.7. DNAse-treated RNA (1 µg) was converted to cDNA by priming with a 10 µM mix of 450 looped primers specific to mature mouse miRs (Mega

Plex kit, Life Technologies) using the High-Capacity cDNA Reverse Transcription Kit

(Life Technologies), following manufacture’s instructions and as previously reported

[185]. qRT-PCR was performed using the TaqMan® Universal PCR Master Mix and 312

TaqMan® miR probes in an Applied Biosystems 7900HT qRT-PCR equipped with a 384

well reaction plate and the Sequence Detection System (SDS) 2.4 software (Life

Technologies). Liquid-handling and the Zymak Twister robots were used to increase

throughput and reduce error, as previously reported [185]. MiRs expression was

normalized against 5S rRNA [∆Ct (miR – 5S rRNA)], a common standard when different

conditions are compared. Fold change was calculated as: 2-∆Ct(treatment)/2-∆Ct(control). The

statistical analyses of HTS-miRs were performed by Dr. Xiaokui Mo. MiRs expression

levels were fit to a general linear model and statistical significance between treatments

was determined by pair-wise comparisons [186]. Results were display by Volcano plots,

p-value vs. Log2[Fold Change]), using p-value less than 0.05 and fold change greater than

2 as threshold [187]. Data were analyzed using SAS 9.3 (SAS, Inc; Cary, NC).

2.16 Validation of HTS-miRs

For miR validation, 200 ng DNAse-treated RNA was reversed transcribed using

TaqMan® miR assay kits for mouse miR-155 and miR-let-7a following manufacturer’s

38 conditions and expression determined by qRT-PCR in an Applied Biosystems 7900HT qRT-PCR system using a 96 well reaction plate and following conditions: 40 cycles of

95°C for 1 min, 60°C for 1 min, and 72°C for 1 min. Validated miR-155 and miR-let-7a expression levels were normalized to the expression of the internal control snoRNA-202.

Fold change was calculated as: 2-∆Ct(treatment)/2-∆Ct(control), where ∆Ct = CtmiR - CtsnoRNA202.

2.17 Analyses of gene expression and alternative splicing genome-wide

To determine the effect of apigenin on gene expression and alternative splicing genome-wide, we performed RNA-seq in MDA-MB-231 cells treated with 50 µM apigenin or diluent DMSO for 48 h.

2.17.1 RNA-seq libraries preparation

RNA-seq libraries were generated in collaboration with Dr. Kengo Morohashi from

Dr. Erich Grotewold’s lab. Briefly, RNA was isolated by the Trizol method, as described in section 2.12. RNA quality was evaluated with the Agilent Bioanalyzer 2100 and was considered acceptable when RIN was higher than 9.5. Barcoded Illumina libraries were generated from two independent biological replicates by the TrueSeq RNA kit, following manufacture’s instructions (Illumina, San Diego, CA) obtaining a total of four libraries (2

DMSO and 2 Apigenin). The four libraries were mixed and sequenced in one lane using single-end 50-mer reads in the Illumina HiSeq 2000 sequencer (Illumina) at the Nucleic

Acid Shared Resource at The Ohio State University. A total of 150 million reads were obtained for all four RNA-seq libraries.

2.17.2 Bioinformatics analyses of RNA-seq data

RNA-seq data analyses were performed by Mrs. Maria Katherine Mejia-Guerra, Mr.

Francisco Padilla-Obregon and Eric Mukundi from Dr. Erich Grotewold’s lab. Raw

RNA-seq data was demultiplexed using Illumina CASAVA pipeline version 1.8.2. Per

base quality and duplication levels of each library were evaluated using FASTQC,

followed by alignment to human genome assembly 19 (hg19) using TopHat v1.4.1 by

default parameters [188]. The annotations for all human gene models were obtained from

RefSeq hosted at University of California-Santa Cruz and used for a semi-guide junctions

alignment. Uniquely and properly aligned reads were kept for further analysis. To

evaluate replicate reproducibility, raw read counts per exon/gene were loaded into R and

normalized for library sizes using DESeq 1.10.1 (R version 2.15.2), variance stabilized

and compared by hierarchical clustering (complete linkage method) using the Euclidean

distances between the samples and by principal component analysis using DESeq 1.10.1.

Differentially expressed genes were obtained using a negative binomial test with a false

discovery rate (FDR) < 0.05 [189].

To determine differences in alternative splicing, normalized reads were analyzed using

the software mixture of isoforms (MISO) with default parameters using two different

approaches [120]. First, the exon-centric analysis estimates the expression of exons and

computes differences in the major classes of mRNA processing events (also referred as

splicing events, see section 1.5.3.1 and Fig. 1.5), i.e skipped exon (SE), retained intron

(RI), alternative last exon (ALE), alternative first exon (AFE), alternative 3’ splice site

(A3SS), alternative 5’ splice site (A5SS), mutually exclusive exons (MXE) and tandem

3’UTR (TUTR). Second, the isoform-centric analysis estimates the expression of whole transcript-isoforms. To determine the inclusion/exclusion ratio of an exon/isoform we determined the percent-spliced-index or PSI [Ψ = #_inclusion_reads/(#_inclusion_reads

+ #_exclusion_reads)]. The difference between apigenin and DMSO-treated MDA-MB-

231 cells was expressed as delta PSI (ΔΨ = Ψapigenin - ΨDMSO) and statistical significance was achieved when Bayes Factor (BF) was higher than 20 (BF > 20) for each biological replicate, as previously described [190]. AS events significantly changed in both biological replicates with an absolute number of PSI !ΔΨ! > 0.05 were kept for further analysis. Enrichment of AS events affected by apigenin in the mRNA processing categories was determined using a hypergeometric distribution analysis.

2.17.3 Analysis of public available RNA-seq data

RNA-seq data sets from normal breast tissue (NBT, n=3) and triple negative breast cancer (TNBC, n=6) were downloaded from GEO (Gene Expression Ominbus, series

GSE52194, [115]). To determine changes in genes expression and mRNA processing between TNBC and NBT, RNA-seq data sets were analyzed as described in section

2.17.2.

2.18 Analyses of alternative splicing by RT-PCR

To validate the changes in alternative splicing identified by RNA-seq, exon-specific primers were designed using the software Primer3 [191]. Gene sequences were obtained from UCSC (University of California Santa Cruz) genome browser [192]. A 20 µl

mixture containing 2 µl of cDNA (20 ng) template, 0.25 µM of each primer, 0.2 mM

dNTPs and 1U Taq Polymerase (Life Technologies) was run using the following

conditions: 95°C for 5 min, 32 cycles of 95°C for 30 sec, 60°C for 30 sec, and 72°C for 2

min, followed by 72°C for 5 min. Primers used to amplify splice forms are listed in Table

2.3. Splice variants were resolved in 1-2% agarose gels and resolved by electrophoresis at

120V for 40 min. To determine the inclusion/exclusion ratio of an exon/isoform we

determined the percent-spliced-index by densitometry as Ψ=(density of isoform

X)/Σ(density of all isoforms). Statistical significance between treatments was determined by two-

tailed t-test.

2.19 Preparation of apigenin-immobilized PEGA beads

Apigenin-immobilized PEGA (amino polyethyleneglycol-polyacrylamide copolymer)

beads were kindly provided by Dr. Kouji Kuramochi from the Kyoto Prefectural

University, Kyoto, Japan). Briefly, 0.40 mmol/mg PEGA beads were washed three times

with pyridine and subsequently mixed with 3.3 mole equivalents (to the amino group

loaded on the beads) of 4-nitrophenyl bromoacetate. Beads were stirred at room

temperature for 3 h, filtered, and washed three times with approximately 10 ml of each,

dichloromethane (CH2Cl2), methanol, and N,N’-dimethylformamide (DMF) and subsequently mixed with 2.2 mol-equivalents of apigenin and 1.8 mole equivalents of

K2CO3 to the bromoacetyl group loaded on the beads. The resulting suspension was stirred at room temperature for 3 days, filtered, and washed three times with approximately 10 ml of each CH2Cl2, methanol, and H2O. The apigenin-immobilized

PEGA beads were vacuum-dried. Filtrates were neutralized by the addition of 1 M

aqueous HCl. After layer separation, the aqueous layer was extracted seven times with 50

ml of EtOAc. The combined organic layer was washed with brine, dried over anhydrous sodium sulfate (Na2SO4) and concentrated in vacuo. The unloaded apigenin was recovered from the supernatant by silica gel chromatography (Chloroform/MeOH =

20:1). The amount of apigenin immobilized on 1 mg of dried PEGA beads was estimated to be ~0.14 mmol as determined by subtracting the amount of unloaded apigenin from the amount of apigenin used in the reaction. These beads are referred throughout the text as

A-beads. Acetylated PEGA beads were used as controls and were referred through the

text as C-beads. C-beads were loaded with acetic anhydride (Ac2O) according to the

procedure reported previously [193].

2.20 Phage display screening

The phage display screening was performed using a T7 Select® Human Breast Tumor cDNA phage library (EMD Biosciences, Cat. No. 70644-3) in collaboration with Dr.

Kengo Morohashi from Dr. Erich Grotewold’s lab. The original library containing 1010

plaque formation units (pfu) was amplified by plate lysate amplification, according to the

manufacture’s protocol (EMD Biosciences). Pre-clearing of the amplified library was

done by incubating 2 ml of T7 phage (109 pfu/ml) with 200 µl (10 mg/ml) of C-beads at

4°C overnight. The pre-cleared phage suspension (1 ml) was incubated with 100 µl of 10

mg/ml A- or C-beads at 4°C overnight and washed 10 times with 1 ml of TBS-T buffer

(20 mM Tris pH8.0, 150 mM NaCl, 0.05% Tween 20), followed by elution with 100 µl

of 1% SDS for 10 min at room temperature. Five microliters of each eluted fractions was

inoculated into 3 ml of Escherichia coli Rosetta-gami 5615 (EMD Bioscience) as bacteria

host, in Luria Broth (LB) medium containing 50 µg/ml carbenicillin and incubated for 3 h

at 37°C, constant shaking. Phage-infected bacteria were centrifuged at 1,500 x g for 5 min and supernatants containing phage particles were used for next round of incubation with C- or A-beads (also referred as bio-panning). Phage titers for each bio-panning step

were evaluated by counting plaque formation units (pfu/ml, see Fig. 6.4A), according to

manufacture’s protocol (EMD Biosciences).

2.21 Phage display coupled with Illumina® GAII next-generation sequencing (PD- seq)

Phage DNA was isolated from the input and the elution fractions obtained in the first and second round of bio-panning using either A- or C-beads (referred as A-E or C-E respectively) by phenol/chloroform extraction, and amplified by PCR using T7 flanking insert primers. The primers include three consecutive random nucleotides at the 5’ regions to help cluster recognition in Illumina® GAII. Amplified PCR fragments were used for preparation of Illumina® libraries according to manufacture’s protocol (Illumina,

San Diego, CA) with some modifications. Briefly, 1.5 pmol of PAGE-purified grade

indexed adapter oligos were ligated with 100 ng of amplified cDNA with Ultrapure T4

ligase for 15 min at room temperature. Ligation products were purified using a PCR

purification kit (Qiagen, Valencia, CA), followed by amplification using Phusion Hot

Polymerase and sequenced by Illumina® GAII.

2.22 Analysis of PD-Seq data

PD-seq sequencing data was analyzed in collaboration with Dr. Alper Yilmaz from Dr.

Erich Grotewold’s lab. Sequences corresponding to empty clones, named multi-cloning site clones (MCS) and contaminant clone referred as MKET, were subtracted from the total number of sequences. Remaining reads were aligned to human coding sequences using SeqMap [194], allowing two mismatches. Normalized in-frame-aligned counts per gene-model (nICPGs) were calculated as follows: The number of aligned sequences per gene model were counted, and in cases when a read aligned to multiple coding sequences, the number of reads was divided by the number of aligned coding sequences. For instance, a read that aligns to two coding sequences, A and B, results in 0.5 as a weighted count (see Fig. 6.3 for detailed description). If a second read matches to A but not B, then

the total ICPG for A will be 1.5 and for B will be 0.5. Thus, by using weighted counts,

the total count number is identical to total number of reads. To determine the normalized

ICPG for a given gene A, ICPGA was divided by the sum of all ICPGs, and multiplied by

106 for better handling. All genes appearing in at least one library were used for further

analysis. A binomial test was used to predict significant changes between two samples.

Log10(nICPG) was used for clustering, and log10(nICPG) was arbitrary converted to zero if nICPG was zero for a particular gene. Clustering and heat-map representations were calculated by applying the hierarchical clustering using MeV software [195] with average linkage clustering. Enrichment of functional categories was analyzed using DAVID

(Database for Annotation, Visualization, and Integrated Discovery) [196]. The total number of genes appearing in the heat map (15,568 genes) was used as a background data

set to obtain enrichment of functional categories.

2.23 Cloning of plasmids

Clones used in this study are described in Table 2.4. pEGFP-hnRNPA2 and pET9c-

hnRNPA2 were kindly provided respectively by Drs. Lexie Friend [197] and Adrian R.

Krainer [198]. p-EGFP-BAG1L, and pEGFP-ARHGEF1 were obtained from Drs. Ann C.

Williams [199] and Philip B. Wedegaertner [200], respectively. pGEX-2T-hnRNPA1 was

a gift by Dr. Ralph Nichols [201]. The different hnRNPA2 cDNA fragments were

amplified using the pET9c-hnRNPA2 clone as template and cloned into a pENTR-D-

TOPO vector. The following primers were used to generate by PCR amplification

different fragments of hnRNPA2 (primers used for cloning are described in Table 2.5):

hnRNPA21-341 full length, PAO-351 and PAO-338; hnRNP-A21-263, PAO-351 and PAO-

337; hnRNP-A21-255, PAO-351 and PAO 1232; hnRNP-A21-248 PAO-351 and PAO 1233; hnRNP-A21-189, PAO-351 and PAO-372; hnRNP-A2190-341, PAO-374 and PAO-338; hnRNP-A2264-341, PAO-352 and PAO-338. Point mutations in hnRNPA2 were created

utilizing the Quick Change Site-directed Mutagenesis Kit (Stratagene, Cedar Creek, TX).

The following primers were used for site-directed mutagenesis (Table 2.5): hnRNP-A21-

248/P239R, PAO-289 and PAO-290; hnRNP-A21-255/R254G, PAO-291 and PAO-292; hnRNP-

A21-255/R254N, PAO-293 and PAO-294; hnRNP-A21-255/Y250G, PAO-295 and PAO-296. All

mutations were confirmed by sequencing. To generate the GST-tagged and 6xHis-tagged

fragments the pENTR-D-TOPO containing hnRNPA2 fragments were cloned into

pDEST15 or pDEST17 vectors, respectively, by recombination using the Gateway LR

Clonase® enzyme mix and the conditions suggested by the manufacturer.

To generate the fluorescent indicator protein plasmids (pFLIP, see Table 2.4 for clone description), hnRNP-A2264-341 cDNA was amplified by PCR from pET9c-hnRNPA2

vector using the PAO-379 and PAO-378 primers pair (Table 2.5). Fragments were cloned

into the pENTR-D-TOPO vector and transferred to the 6xHis-tagged pFLIP vectors by

recombination using the Gateway LR Clonase® enzyme mix. pFLIP vectors, previously

used to generate biosensors, were generously provided by Dr. Wolf Frommer [202]. The

hnRNPA2264-341 fragment was cloned into pFLIP1 or pFLIP2 vectors containing the N-

terminal-CFP and C-terminal-YFP tags or into pFLIP4 vector containing the N-terminal-

GFP and the C-terminal-mKO (monomeric Kusabira-Orange). pFLIP2 and pFLIP4

vectors hold the Gateway® recombination linkers flanked by the KpnI and SpeI restriction

sites. Different versions of FLIP2-hnRNPA2264-341 (referred as pFLIP2-1, 2-2 and 2-3-

hnRNPA2264-341) and FLIP4-hnRNPA2264-341 (referred as pFLIP4-1 and 4-2-hnRNPA2264-

341) were generated by digestion and self-ligation using KpnI and/or SpeI sites, to improve

FRET emission.

2.24 Recombinant protein expression and production

To obtain GST-tagged hnRNPA2 proteins, BLR(DE)LysS cells transformed with

GST-hnRNPA21-341, GST-hnRNP-A21-263, GST-hnRNP-A21-255, GST-hnRNP-A21-248,

GST-hnRNP-A21-189, GST-hnRNP-A2190-341, GST-hnRNP-A2264-341, GST-hnRNP-A21-

248/P239R, GST-hnRNP-A21-255/R254G, GST-hnRNP-A21-255/R254N, GST-hnRNP-A21-255/Y250G

or GST vector alone were grown in LB containing 100 µg/ml ampicillin until reaching an

OD600 = 0.5. Next, protein expression was induced with 1 mM IPTG for 2 h at 30°C.

Cells were harvested by centrifugation at 14,000 x g and resuspended in lysis buffer (1

mg bacteria /ml, PBS pH 7.4, 1 mM DTT, 0.1 mM PMSF, 2 µg/ml CLAP, 1% Tween 20

and 10 mg/ml lysozyme), and sonicated (Output Control: 8; Duty Cycle 80%; 5 cycles,

10 pulses each cycle). Proteins were purified by glutathione-binding affinity

chromatography, following manufacturer’s suggestions (EMD Biosciences). Briefly, 150

µl pre-equilibrated beads (50% slurry) were added to 10 ml bacteria lysates and

incubated at 4°C for 2 h. Next, lysates were transferred to a column, flow-trough

collected and beads washed three times with 1 ml PBS pH 7.4. Flow-trough was collected

after each wash. Beads were eluted with 1 ml of elution buffer (10 mM reduced

glutathione, 50 mM Tris, pH 8.0). GST-tagged proteins (1 ml) were dialyzed in 500 ml

TBS buffer (1:500 dilution, 20 mM Tris, pH 7.6, NaCl 150 mM, 1 mM DTT, 0.1 mM

PMSF) for 3 h at 4°C twice, aliquoted and stored at -80°C until further used.

To obtain 6xHis-hnRNPA2 and 6xHis-FLIP-hnRNPA2 proteins, BLR(DE)LysS cells

were grown in LB containing 100 µg/ml ampicillin until reaching an OD600 = 0.5. Protein expression was induced with 1 mM IPTG for 2 h at 30°C. After induction, cells were collected and lysed (1 mg bacteria/ml buffer), by sonication with buffer containing 20

mM Tris, pH 8.0, 1 mM DTT, 0.1 mM PMSF, 2 µg/ml of CLAP using the above-

mentioned conditions. His-tagged proteins were purified by binding affinity to nickel-

charged beads (EMD Biosciences). Bacteria lysates (1g bacteria/10 ml) were incubated

with 125 µl pre-equilibrated beads (50% slurry) at 4°C for 2 h, washed three times with 1

ml aliquots of 20 mM Tris-HCl, pH 8.0, three times eluted with 20 mM imidazole and

eluted with 1 ml of 50 mM imidazol. Elution fractions (1 ml) containing the 6xHis-

hnRNPA2 and 6xHis-FLIP-hnRNPA2 proteins were dialyzed in 500 µl (1:500 dilution)

20 mM Tris buffer, pH 8.0 for 3 h at 4°C, twice, aliquoted and stored at -80°C until

further used.

2.25 Pull-down assays

To validate the interaction of apigenin candidate targets with apigenin, we performed

pull-down assays. Briefly, GST-pull-downs were carried out by incubating 100 nM of

purified recombinant GST-hnRNPA2 proteins or GST alone with 150 µg of pre-

equilibrated A- or C-beads in 100 µl of TBS buffer pH 7.6, containing 1mM DTT, 0.1

mM PMSF, 2 µg/ml of CLAP for 12 h at 4°C in a rotator. The beads were spun down at

1,500 x g for 30 sec and supernatants were recovered and kept at -70°C until further used.

Beads were then washed three times by adding 400 µl TBS buffer followed by

centrifugation at 1,500 x g for 30 sec each. After the last wash, the bound proteins were

eluted with 20 µl of 2% SDS. Both, bound and ¼ of supernatant fractions were resolved

in 12 % SDS-PAGE and immunoblotted with anti-GST antibodies. For pull-downs using

cellular lysates, 1 x 106 HeLa cells transiently transfected with hnRNPA2-GFP,

ARHGEF1-GFP, GFP-BAG1 or GFP vector alone (see Table 2.4 for vector description) or 2 x 106 MDA-MB-231 cells were lysed in TBS-T buffer for 20 min on ice, followed

by sonication using a Branson Sonifier 450 (output control 3, duty cycle 30, pulses 3) and

centrifuged at 14,000 x g for 10 min at 4°C. To test the binding of hnRNPA2-GFP,

ARHGEF1-GFP, GFP-BAG1 or GFP to apigenin, cell lysates (200 µg) were incubated

with 300 µg of A- or C-beads in 100 µl of TBS buffer, pH 7.6 containing 1mM DTT, 0.1 mM PMSF, 2 µg/ml of CLAP for 12 h at 4°C. To evaluate the binding of Hsp70 and

MSI2 to apigenin, cell lysates (1 mg) were incubated with 600 µg of A- or C-beads in

100 µl of TBS buffer, as described above. The beads were spun down at 1,500 x g for 30 sec and supernatants were recovered and kept at -70°C until further used. The bound proteins were eluted with 20 µl of 2% SDS. Bound and supernatant (1/4 of the total supernatant) fractions were analyzed by 12% SDS-PAGE and immunoblotted anti-GFP for HeLa cell lysates and anti-MSI2 or anti-Hsp70 for MDA-MB-231 cell lysates.

2.26 Spectrophotometric analyses

The change in the UV-visible absorption of apigenin was determined by a spectrophotometric method. Free apigenin (10 µM) was incubated with increasing concentrations of GST-hnRNPA2 or GST alone (0.2, 0.5, 1, 2, 5, 10 and 20 µM) in 150

µl of TBS buffer, pH 8.0 containing protease inhibitor cocktail for 15 min at 37°C using a

96-well plate (SensoPlate 96W Sterile, Greiner, Monroe, NC). Absorption spectra ranging from 260 to 450 nm were measured using a spectrofluorometer plate reader

(FlexStation3, Molecular Devices, Sunnyvale, CA). The dissociation constant of the complex, KD, was calculated with the Benesi-Hilderbrand method [203, 204]:

l[CApigenin]/ΔA= (1/[ChnRNPA2])(1/εK) + (1/ε)

where CApigenin and ChnRNPA2 are concentrations of apigenin and GST-hnRNPA2

respectively, ΔA is the change in absorbance, ε is the extinction coefficient of the

complex, l is the path-length (0.4 cm for this system) and K is the association constant,

corresponding 1/KD. KD was therefore determined by plotting (lCApigenin/ΔA) vs.

(1/ChnRNPA2).

2.27 Fluorescence resonance energy transfer (FRET)

Emission spectra of purified FLIP-hnRNPA2C proteins were determined using a spectrofluorometer plate reader (FlexStation3, Molecular Devices, Sunnyvale, CA) by exciting Cyan Fluorescent Protein (CFP) at 405 nm and recording emission over the range of 460-600 nm. CFP shows a maximum peak at 480 nm while Yellow Fluorescent

Protein (YFP) shows a maximum peak at 530 nm. FRET was determined as the intensity of fluorescence at 530 nm divided by the intensity of fluorescence at 480 nm (YFP/CFP ratio). Binding of flavonoids to FLIP-hnRNPA2190-341 was assessed by incubating 1 µM of purified FLIP-hnRNPA2190-341 with increasing concentrations of apigenin, luteolin,

chrysoeriol, naringenin, eriodictyol, quercetin, kaempferol, flavopiridol, genistein,

apigenin 7-O-glucoside or apigenin 6-C-glucoside (ranging from 1 to 100 µM) or diluent

DMSO as control in 200 µl of 20 mM Tris, pH 8.0 at 37°C for 3 h.

The dissociation constants of the complexes, KD, were calculated by fitting the

YFP/CFP ratio curves to the equation for the binding of a ligand to a protein:

S = (r - Rmin)/(Rmax - Rmin) = [L]bound/[P]total = n[L]/(KD + [L])

where S, saturation [L], ligand concentration; [L]bound, concentration of bound ligand;

n, number of equal binding sites; [FLIP]total, total concentration of FLIP nanosensor; r,

ratio; Rmin, minimum ratio in the absence of ligand; and Rmax, maximum ratio at

saturation with ligand [202]. Saturation curves were obtained and KD determined by non- linear regression using the GraphPad Prism. Levels of statistical significance between means in FRET experiments were determined by one-way ANOVA.

2.28 Amplified luminescent proximity homogeneous assay (ALPHA)

Dimerization of hnRNPA2 was determined by ALPHA using GSH-acceptor and anti-

His-donor beads according to manufacture’s instructions (Perkin Elmer, Waltham, MA).

Briefly, bacterial expressed purified 6xHis-hnRNPA2 (125 nM) was incubated with 125 nM GST-hnRNPA2 for 1 h at room temperature in 10 µl TBS buffer, pH 7.6 containing 1 mM DTT, 0.1 mM PMSF, 2 µg/ml of CLAP. After 1 h, 20 µg/ml GSH-acceptor and 20

µg/ml anti-His-donor beads in 20 µl TBS buffer and incubated for additional 6 h at room temperature. Diluent DMSO, 100 µM apigenin or 100 µM naringenin were added to the

mix and incubated for 15 min at room temperature. Arbitrary luminescent units were determined using the EnSpire multimode plate reader with ALPHA technology (Perkin

Elmer), and expressed divided by 1,000. Statistical significance was determined by one- way ANOVA using the GraphPad Prism software.

2.29 Enzymatic assays

To determine UDP-glucose 6-dehydrogenase (UGDH) activity, 1 x 106/ml MDA-MB-

231 cells were treated with 50 µM apigenin, 50 µM naringenin or diluent DMSO for 3 h.

Ten million cells were homogenized by douncing (10 strokes) in lysis buffer (10 mM

Tris-HCl pH 8.7, 50 mM KCl, 1.5 mM MgCl2, 0.1 mM PMSF, 2 µg/ml CLAP), and

subsequently centrifuged at 20,000 x g for 15 min at 4°C. Cell lysates (250 µg of protein)

were incubated in 200 µl buffer containing 1 mM UDP-glucose, 100 mM sodium glycine

pH 8.7 and 1 mM NAD+ at room temperature. Activity was determined by assessing the

change in NAD+ absorbance at 340 nm for 30 min using the EnSpire multimode plate

spectrophotometer reader (Perkin Elmer).

Isocitrate dehydrogenase 3 (IDH3) activity was determined in mitochondria

preparations from MDA-MB-231 cells treated with 50 µM apigenin, 50 µM naringenin

or diluent DMSO for 3 h. Mitochondria were isolated from 2 x 107 cells by dounce homogenization (100 strokes) in 400 µl of mitochondria isolation (MI) buffer containing

20 mM Tris pH 7.2, 0.8 M sucrose, 40 mM KCl, 2 mM EGTA, 1 mg/ml BSA, 0.1 mM

PMSF, 2 µg/ml CLAP, and centrifuged and 1,500 g for 10 min at 4°C. Pellets were resuspended in 400 µl MI buffer and centrifuged at 17,000 g for 30 min at 4°C. Pellets containing mitochondrial fraction were resuspended in 200 µl MI buffer and lysed by three rounds of freeze and thaw. Purity of the isolated fractions was evaluated by Western blots using anti-cytochrome C antibodies, a specific mitochondrial marker, and GAPDH, a cytoplasmic marker. IDH3 activity was evaluated by incubating 250 µg of mitochondrial protein in 200 µl buffer containing 100 mM K2HPO4, 100 mM KH2PO4, 8 mM MgCl2, 500 µM NAD+, and 2 mM sodium isocitrate, pH 7.6. Activity was

determined by assessing the change in NAD+ absorbance at 340 nm for 30 min using the

EnSpire multimode plate reader.

Enzymatic units were calculated using the following formula: Enzymatic units =

(ΔA340*Vf*d.f.)/(ε*mg*l), where ΔA340 is the change in absorbance at 340 nm over time;

Vf is the final reaction volume; d.f., dilution factor; ε, extinction coefficient of NAD+ determined to be 6.22; mg, amount of protein; and l, light path estimated to be 0.68.

Levels of statistical significance between treatments were determined by one-way

ANOVA.

2.30 Preparation of celery-based apigenin-rich extracts and diets

Celery-based apigenin rich extracts with increased aglycone content were prepared as previously described [205]. Briefly, fresh celery leaves, naturally abundant in apigenin 7-

O-apiosylglucoside (apiin), were juiced using a common kitchen juicer. Celery juice was incubated at room temperature for 3 h and then lyophilized. To obtain enzyme-treated celery extract (ECE), lyophilized celery was heated in 1.5 N H3PO4 for 90 min at 100°C to convert apiin into apigenin 7-O-glucoside. After neutralization with 10% KOH to pH

5.0, the preparation was incubated with ground raw (unheated) almonds (20% dry wt, obtained from grocery store) for 2 h at 50°C to convert apigenin 7-O-glucoside into aglycone apigenin by β-glucosidases present in the almond powder. The resulting aglycone-rich mixture was lyophilized and powder extracted in 70% methanol. ECE was dried under N2 gas and reconstituted in 300 µL DMSO. Flavone concentration was

determined by HPLC at the Nutrient and Phytochemical Analytical Shared Resource

(NPASR) and samples were diluted in DMSO to 10 mM apigenin equivalents and stored

at -20°C for further use.

The AIN-93G control diet supplemented with 10% w/w ECE (referred hereafter as apigenin-diet) was prepared as previously described [205]. Briefly, 10% (w/w) lyophilized enzyme-treated celery was combined with powdered AIN-93G diet. Distilled

H2O was added to make a paste and pellets were formed with a Simple Tablet Maker

(Wholistic Research Co., Royston, Herts, UK). The same procedure was used to make pellets of the AIN-93G control diet. Pellets were dried at 40°C to a constant weight and stored at -20°C until used for feeding. In serum from mice consuming apigenin-diet for 7 days apigenin reached ~ 1 µM concentration, as demonstrated by HPLC analyses and previously reported [205].

2.31 Animal models

All procedures were approved by the Ohio State University Institutional Animal Care

Committee protocol A0208.

2.31.1 Mouse models of inflammation

Male C57BL/6J mice, 6-8 weeks of age (Jackson Laboratories, Bar Harbor, ME), were used after 7 days acclimation. Mice received 50 mg/kg apigenin or diluent DMSO intraperitoneally (i.p.) 3 h prior administration of 37.5 mg/kg LPS (E. coli 0128:B8) in

500 µl of PBS or 500 µl PBS as vehicle by i.p. for additional 3 h. Alternatively, mice were fed ad libitum for 7 days with either control or apigenin diet prior receiving 37.5 mg/kg LPS or PBS by i.p. for 3 h. Mice were euthanized with CO2, lungs, broncho- alveolar lavage fluids (BALFs) and serum were collected and stored at -80oC.

2.31.2 Measurement of NF-κB in vivo

BALB/C-Tg(NF- κB-RE-luc)-Xen male mice, 6-8 weeks of age, systemically expressing luciferase under the control of NF-κB were provided by Dr. Denis Guttridge from the The Ohio State University. Mice were injected i.p. with 100 µl DMSO as vehicle or apigenin (50 mg/kg of body weight, dissolved in DMSO), 3 h prior to administration of a lethal dose of LPS (37.5 mg/kg in 100 µl volume) for 6 h. After treatment, mice were injected i.p. with D-luciferin (150 mg/kg of body weight in 100 µl

PBS) and immediately anesthetized using 1% isoflurane in an anesthesia chamber. Mice were imaged using an ultrasensitive camera consisting of an image intensifier coupled to a charge-coupled device (CCD; Xenogen IVIS Imaging System; Xenogen, Alameda,

CA). Subsequently, mice were sacrificed and vital organs excised, placed in a Petri dish and imaged. The images were processed using the Living Image Software (Caliper Life

Sciences, Hopkinton, MA) and expressed as photons/sec.

To determine luciferase enzyme activity, ~50 mg of tissue were homogenized in 400

µl of luciferase lysis buffer (Promega, Cat. No. E153A) using a Brinkmann Polytron

PTA 7K1 homogenizer (Brinkmann instruments Inc., Westbury, NY) and 20 µg of protein homogenates were used to evaluate luciferase activity using the Luciferase Assay

Kit according to the manufacturer’s protocol (Promega) and normalized by protein concentration.

2.31.3 IKKβ knock-out (KO) mice

C57/BL6J-LysM-cre+/+IKKβF/F (IKKβKO) mice, harboring myeloid specific knock-out

of IKKβ, and C57/BL6J-LysM-cre-/-IKKβF/F (IKKβWT), which are wild-type for IKKβ,

were kindly provided by Dr. Michael Karin [206] and used to obtain bone marrow

derived macrophages (see section 4.3.8).

2.31.4 Mouse models of breast cancer development

FVB/N-Tg-Mouse Mammary Tumor Virus-Polyoma Virus T antigen (MMTV-PyVT)-

634Mul/J (referred hereafter as PyMT+) and wild-type FVB mice (referred hereafter as

PyMT-) were obtained from Jackson Laboratory (Ban Harbor, ME). PyMT+ males were mated with PyMT- females to generate PyMT+ females and offspring was genotyped according to Jackson Laboratory’s protocols, as previously described [180]. Immediately

after weaning, 3-week old PyMT+ female mice were randomly separated into two groups

and injected intraperitoneally (refer hereafter i.p.) daily with 100 µl vehicle (containing:

20% DMSO, 30% kolliphor EL, 10% ethanol and 40% PBS) or 25 mg/kg apigenin

dissolved in vehicle. PyMT- female mice, that do not develop tumors, were treated with

vehicle and used as age-matched controls. Mice were sacrificed when they were 4, 7, 9 or

12-weeks-old. For dietary interventions, three-weeks-old PyMT+ female mice were

randomly separated in two groups and fed ad libitum, starting immediately after winning,

with either control or apigenin diets (see section 2.30). Mice were euthanized when they

were 12-weeks-old. Mammary glands, blood and other organs were collected. Mammary

glands were weighed, fixed in 10% formalin or snap frozen and stored at -80oC for

further use.

2.31.5 Mouse xenografts

Severe combined immuno-deficient (SCID) female mice (6-10 weeks old) were

injected with 1 x 106 MDA-MB-231 breast cancer cells in 100 µl DMEM into the

mammary fat pad. Mice were injected by i.p. daily for 28 days with vehicle (20% DMSO,

10% ethanol, 30% kolliphor EL and 40% PBS) or 25 mg/kg apigenin dissolved in 100 µl

vehicle,. Tumor length (l) and width (w) were measured three times a week using a caliper and tumor volume calculated with the formula w * l2/2. After 28 day, tumors were procured, fixed in 10% formalin or snap frozen and stored at -80oC for further used.

2.32 Histology and immunohistochemistry (IHC)

Right inguinal mammary glands and the apical, cardiac and diaphragmatic lung lobes,

inflated with 1 ml PBS, were fixed in 10% buffered formalin, paraffin embedded,

longitudinally sectioned and mounted onto glass slides. Tissues sections were

deparaffinized by performing three 5 min incubations in xylene, followed by rehydration

in a series of 100%/95%/70% ethanol solutions before staining with hematoxylin and

eosin (H&E) using the Histo-Perfect H&E staining kit (BBC Biochemicals, Mount

Vernon, WA).

For all IHC analysis, antigen retrieval was conducted by incubating deparaffinized

slides for 20 min on 10 mM citrate buffer, pH 6, at 125oC in a pressure cooker. Slides

were washed twice with dH2O, incubated with 3% H2O2 in PBS for 15 min and blocked

58 with 2% isotype control serum, 0.1% Tween-20 and 1% BSA in PBS for 30 min at room temperature. Slides were stained with mouse anti-α-SMA (alpha-smooth muscle actin, clone 1A4, Sigma, 1:100), rabbit anti-Ki67 (SP6, 1:100) or rat anti-F4/80 (CI:A3-1, 1:20) antibodies (see Table 2.1 for antibody description) for 1 h at room temperature in a humidified chamber, rinsed 3 times in PBS, followed by incubation with 1:200 secondary biotin-conjugated antibodies (Vector Laboratories, Burlingame, CA) for 1 h at room temperature. Slides were developed using the Vectastain ABC kit according to manufacture’s instructions (Vector Laboratories) and counter-stained with hematoxylin for 30 seconds. Slides were visualized under the light microscope (Olympus BX40 equipped with the Optronics DEI 750E CE Digital Camera), and analyzed blindly in 20 random fields at 40X using the software ImageJ. The proliferation index was calculated as the number of cells stained positive for Ki67, divided by the total number of cells per field at 40X magnification. The number of macrophages infiltrated in the peri- and intra- tumoral regions was determined as cells stained positive with F4/80 antibodies divided by the area of each region, as determined using ImageJ. Data are represented as macrophages/mm2.

Dr. Priyadharsini Nagarajan, an expert breast cancer pathologist at The Ohio State

University, provided assistance scoring the four morphological stages characteristics of breast cancer development in the PyMT model (Chapter 5, Fig. 5.2). To score hyperplasia

(H), adenoma (A), early carcinoma (EC), and late carcinoma (LC), H&E-stained mammary sections were scanned with an Aperio Scanscope XT scanner (Leica,

Wetzlar, Germany) and the area of each stage was analyzed by morphometric

59 measurements using the ImageJ software, as previously described [207]. In addition, to score adenomas and early carcinomas, mammary gland sections were immune-stained with anti-α-SMA antibodies. Adenomas are positively stained with anti-α-SMA antibodies, while early carcinomas lack expression of α-SMA.

2.33 TUNEL

Apoptosis was assessed in deparaffinized mammary gland sections using the Terminal

Uridine Nick-End Labeling (TUNEL) kit and following manufacturer’s instructions

(EMD Millipore, Billerica, MA). Results are represented as TUNEL positive counts per field (cpf) in 20 random fields at 40X.

2.34 Analyses of metastasis

Metastatic burden was calculated in H&E-stained lung sections by measuring width

(w) and length (l) of the metastatic nodules with the ImageJ program, and applying the sphere volume formula l * w2/2. Metastatic index was calculated by counting the number of metastatic nodules in the apical, cardiac and diaphragmatic lung lobes and divided by the number of lobes.

2.35 Isolation of mouse leukocytes

To isolate bone marrow cells, femurs were flushed with PBS and leukocytes were centrifuged at 1,200 rpm for 5 min and then resuspended in 1 ml cold distilled water

(hypotonic solution) for 1 min to lyse erythrocytes. Splenic leukocytes were isolated by

mechanically disrupting 100 mg spleens, using a scalpel, in 2 ml of 2 mg/ml collagenase

D containing 0.1 mg/ml DNAse I in HBSS buffer, pH 7.4, followed by incubation at

37oC for 20 min in the same solution. Digested tissue was passed through a mesh, centrifuged at 1,200 rpm for 5 min and rinsed in 1 ml cold hypotonic solution for 1 min on ice. To isolate peripheral blood leukocytes, 500 µl of blood were rinsed in 10 ml cold hypotonic solution. After lysis, bone marrow, splenic and blood leukocytes were resuspended in PBS containing 1% FBS (fetal bovine serum), counted by trypan blue exclusion and immediately analyzed by flow cytometry.

2.36 Flow cytometry

Bone marrow cells, splenocytes and blood leukocytes (1 x 106 cells/ml) were stained

with 1:150 fluorochrome-conjugated primary antibodies in PBS containing 1% FBS on

ice for 30 min followed by two washes with PBS. For intracellular stainings, cells were

incubated with BD Cytofix/Cytoperm™ (BD Pharmagen, Franklin Lakes, NJ) for 30 min

followed by staining with fluorochrome-conjugated primary antibodies in

permeabilization buffer (Biolegend, San Diego, CA). All samples were acquired with a

LSRII flow cytometer (BD Pharmagen) and analyzed with FlowJo software (Ashland,

OR). Fluorochrome-conjugated antibodies are described in Table 2.1. All antibodies were

tested with their respective isotype controls.

2.37 Co-cultures

For syngenic co-cultures, BMDMs were stained with 2.5 µM eFluor670 and PyMT

cells were labeled with 5 µM eFluor450 cellular dyes (eBiosciences) dissolved in PBS, at

a cell density of 107 cells/ml for 20 min at 37°C. PyMT cells and BMDMs were cultured

alone or in combination, at 1:1 or 1:10 PyMT:BMDM ratios, in serum-free DMEM

media for 1 h before treatment with diluent DMSO, 10 µM apigenin, 10 µM Bay-11-

7082 or concomitantly treated with apigenin and Bay-11-7082 for different periods of

time. Supernatants were collected and used to measure cytokine expression. Apoptosis

was evaluated using the FITC-Annexin V/7’AAD kit as previously reported [34]. NF-κB

phosphorylation was evaluated by incubating with anti-phospho-NF-κB-p65S536

antibodies (93H1, 1:250, Table 2.1) or isogenic IgG for 30 min on ice followed by two

rinses in PBS and staining with anti-rabbit IgG-alexa 488 conjugated secondary

antibodies (1:250, Table 2.1) for 30 min on ice. CCL2 expression was determined by

incubating with anti-CCL2-PE antibodies (2H5, 1:150, Table 2.1). Each cell population

was analyzed by flow cytometry as described in the section 2.37.

2.38 Immune-detection of cytokines

Mouse CCL2 was measured using the BD OptEIA™ - Mouse MCP-1 ELISA kit (BD

Pharmagen) and Mouse TNFα was analyzed using the DuoSet TNFα ELISA kit (R&D

Systems, Minneapolis, MN) according to the manufacturer’s suggestions. Absorbance at

450 nm was measured in the EnSpire plate reader and data analyzed as previously

reported [169].

2.39 Statistical analyses

All data are expressed as mean ± SEM. Statistical tests used in this study include two- tailed student’s t-test, One- and Two-way ANOVA followed by Bonferroni’s post hoc.

Number of biological replicates, statistical tests employed, and significance are stated in the text and figure legends. Genome-wide and HTS statistical analyses are described in sections 2.14, 2.15 and 2.17.2.

Table 2.1. List of antibodies Catalog Antibody Clone Type Application Company No. anti-mouse IgG-biotin - Polyclonal PK-4002 IHC Vectastain anti-mouse IgG-HRP - Polyclonal NA931V WB GE Healthcare anti-mouse-AF-488 - Polyclonal A11029 IF Life Technologies anti-rabbit IgG-biotin - Polyclonal PK-4001 IHC Vectastain anti-rabbit IgG-HRP - Polyclonal NA934V WB GE Healthcare anti-rabbit-AF-488 Polyclonal A11034 FC Life Technologies

anti-rat IgG-AF-633 - Polyclonal A21094 IHC-IF Life Technologies anti-rat IgG-biotin - Polyclonal PK-4004 IHC Vectastain ATM Ab-3 Polyclonal PC116 WB EMD Millipore ATR - Polyclonal ab2905 WB Abcam B220-FITC RA3-6B2 Monoclonal 553088 FC BD Pharmagen CCL2-PE 2H5 Monoclonal 12-7096-41 FC eBioscience CCR2-PE 475301 Monoclonal FAB5538P FC R&D Systems CD117-APC 2B8 Monoclonal 553356 FC BD Pharmagen CD11b-APC M1/70 Monoclonal 561690 FC BD Pharmagen CD11b-FITC M1/70 Monoclonal 553310 FC BD Pharmagen CD16/32-PE-Cy7 93 Monoclonal 25-0161-81 FC eBioscience CD206 - Polyclonal ab64693 IHC-IF Abcam CD25-BV650 PC61 Monoclonal 102037 FC Biolegend CD3-FITC 17A2 Monoclonal 561798 FC BD Pharmagen CD3-PE 17A2 Monoclonal 100205 FC Biolegend CD4-PerCP/Cy5.5 GK1.5 Monoclonal 100433 FC Biolegend CD49b-PE-Cy7 DX5 Monoclonal 108921 FC Biolegend CD8-APC 53-6.7 Monoclonal 100711 FC Biolegend Cleaved Caspase-3 D175 Monoclonal 9661 IHC Cell Signaling Cytochrome-C - Polyclonal 4272 WB Cell Signaling F4/80 Cl:A3-1 Monoclonal MCA497 IHC-IF AbD Serotec Foxp4-Alexa-488 150D Monoclonal 320011 FC Biolegend GAPDH FL-335 Polyclonal sc-25778 WB SC Biotech GFP - Polyclonal ab290 WB Abcam Gr-1-FITC RB6-8C5 Monoclonal 553127 FC BD Pharmagen Gr-1-PerCP/Cy5.5 RB6-8C5 Monoclonal 108427 FC Biolegend GST-HRP 8-326 Monoclonal MA4-004 WB Thermo Fisher hnRNPA2/B1 DP3B3 Monoclonal sc-32316 WB SC Biotech Hsp70 C92F3A-5 Monoclonal SPA-810 WB Enzo IKKβ 2C8 Monoclonal 2370 WB Cell Signaling

Continued 64

Table 2.1 continued Catalog Antibody Clone Type Application Company No. Ki67 SP6 Monoclonal RM-9106 IHC Thermo Fisher Ly6C-PE-Cy7 HK1.4 Monoclonal 128017 FC Biolegend Ly6G-PerCP/Cy5.5 1A8 Monoclonal 127615 FC Biolegend MSI2 EP1305Y Monoclonal ab76148 WB Abcam NF-κB-p65 C-20 Polyclonal sc-372 WB SC Biotech NK1-1-PE PK136 Monoclonal 557391 FC BD Pharmagen p38 - Polyclonal 9212 WB Cell Signaling p-ATM S1981 EP1890Y Monoclonal ab81292 WB Abcam p-ATR S428 EPR2184 Monoclonal ab178407 WB Abcam p-H3 S10 - Polyclonal 9701 IHC Cell Signaling p-NF-κB-p65 S536 93H1 Monoclonal 3033 WB Cell Signaling p-p38 T180/Y182 - Polyclonal 9211 WB Cell Signaling p-γH2AX S139 JBW301 Monoclonal 05-636 WB EMD Millipore PKCδ C-20 Polyclonal sc-937 WB, IP SC Biotech PR C-19 Polyclonal sc-538 IHC SC Biotech Rabbit IgG - - sc-2027 IP SC Biotech sca-1-PE D7 Monoclonal 553108 FC BD Pharmagen αSMA 1A4 Monoclonal A2547 IHC Sigma β-tubulin AA2 Monoclonal sc-80011 WB SC Biotech

Table 2.2. List of primers used for qRT-PCR

Gene Name Sequence PAO number

Forward 5’-GTAACCCGTTGAACCCCATT-3’ PAO-296 18S Reverse 5’-CCATCCAATCGGTAGTAGCG-3’ PAO-297 Human Forward 5- GGACACTCCTTGCCAAATGCAG-3’ PAO-523 BACH1 Reverse 5’-TGACCTGGTTCTGGGCTCTCAC-3’ PAO-524 Human Forward 5'-GAAGACGAGACGGGTTGCA-3' PAO-633 CCNA2 Reverse 5'-AGGAGGAACGGTGACATGCT-3' PAO-634 Human Forward 5'-TCTGGATAATGGTGAATGGACA-3' PAO-631 CCNB1 Reverse 5'-CGATGTGGCATACTTGTTCTTG-3' PAO-632 Human Forward 5’-CTCCAGGAAGAGGAAGGCAA-3’ PAO-509 CCNE1 Reverse 5’-TCGATTTTGGCCATTTCTTCA-3’ PAO-510 Human Forward 5’-CTATTTGGCTATGCTGGAGG-3’ PAO-511 CCNE2 Reverse 5’-TCTTCGGTGGTGTCATAATG-3’ PAO-512 Human Forward 5’-ACCGTCACTATGGACCAGC-3’ PAO-517 CDC25A Reverse 5’-TTCAGAGCTGGACTACATCC-3’ PAO-518 Human Forward 5’-ATGGAGAACTTCCAAAAGGTGGA-3’ PAO-521 CDK2 Reverse 5’-CAGGCGGATTTTCTTAAGCG-3’ PAO-522 Human Forward 5'-GACTCTCAGGGTCGAAAACG-3' PAO-1121 CDKN1A Reverse 5'-GGATTAGGGCTTCCTCTTGG-3' PAO-1122 Human Forward 5’-GAGCTCACTCAGACCCCAAG-3’ PAO-513 E2F2 Reverse 5'-AACAGGCTGAAGCCAAAAGA-3’ PAO-514 Human Forward 5'-CTGTGGACCTCATCCAGAAGCA-3' PAO-618 FEN1 Reverse 5'-CCAGCACCTCAGGTTCCAAGA-3' PAO-619

Human Forward 5'-GAGAGCAGAAGACCGAAAGC-3' PAO-1123 GADD45A Reverse 5'-TGCAGAGCCACATCTCTGTC-3' PAO-1124 Human Forward 5’-ACTTTGGTATCGTGGAAGGACT-3’ PAO-230 GAPDH Reverse 5’-GTAGAGGCAGGGATGATGTTCT-3’ PAO-231 Human Forward 5'-GGCTACGGAGGTGGTTATGA-3' PAO-416 HNRNPA2B1 Reverse 5'-ACCCCCAAAGTTTCCACTCT-3' PAO-417 Human Forward 5’-TCAAGAGGCGAACACACAAC-3’ PAO-515 MYC Reverse 5’-GGCCTTTTCATTGTTTTCCA-3’ PAO-516 Human Forward 5'-GCTGGTTGTGAGCATTCGTG-3' PAO-622 POLH Reverse 5'-TGCTCAAGAAGCTGGTGATGTC-3' PAO-623 Human Forward 5'-AATGCAAAGTGTGTGCAAGC-3' PAO-624 RAD1 Reverse 5'-GTCCCTGGCATAGGACTTGA-3' PAO-625

Continued

Table 2.2 continued

Gene Name Sequence PAO number Human Forward 5'-TCACGGTTAGAGCAGTGTGG-3' PAO-1115 RAD51 Reverse 5'-GGCAACAGCCTCCACAGTAT-3' PAO-1116 Human Forward 5'-GGCGAGAGGAGCACAGATAC-3' PAO-218 RELA Reverse 5'-CGGCAGTCCTTTCCTACAAG-3' PAO-219 Human Forward 5'-TCGCCTGGTTCTTTTTGCA-3' PAO-620 XRCC2 Reverse 5'-TCTGATGAGCTCGAGGCTTTC-3' PAO-621 Mouse Forward 5'-GAAGGCGGTGATTTTAACGA-3' PAO-230 CAP Reverse 5'-TCCAGCGATTTCTGTCACTG-3' PAO-231 Mouse Forward 5'-CCACTCACCTGCTGCTACTCAT-3' PAO-310 CCL2 Reverse 5'-TGGTGATCCTCTTGTAGCTCTCC-3' PAO-311 Mouse Forward 5'-CAGAGCCAACGTCAAGCA-3' PAO-1291 CXCL12 Reverse 5'-AGGTACTCTTGGATCCAC-3' PAO-1292 Mouse Forward 5'-AACAGCCCCTATCCTCCTTC-3' PAO-1107 FOXO3a Reverse 5'-GGCATGGTCTGAGGAATCAT-3' PAO-1108 Mouse Forward 5'-TTCACCACCATGGAGAAGGC-3' PAO-292 GAPDH Reverse 5'-GGCATGGACTGTGGTCATGA-3' PAO-293 Mouse Forward 5'-GGTGGAGGTCGTACAAGCAT-3' PAO-1135 PR Reverse 5'-CTCATGGGTCACCTGGAGTT-3' PAO-1136 Mouse Forward 5'-CTGTTAATGCTAATTGTGATAGG-3' PAO-1143 pri-miR-155 Reverse 5'- GCTAACAGGTAGGAGTCAGTCAG-3 PAO-1145 Mouse Forward 5'-TGCATATCCCTTATCCTCTGG-3' PAO-1144 pre-miR-155 Reverse 5'- GCTAACAGGTAGGAGTCAGTCAG-3 PAO-1145 Mouse Forward 5'-CAAACTCGGAGAGGTTCTGC-3' PAO-1101 SMAD2 Reverse 5'-GCCAGCCGTATCTCTGGTTA-3' PAO-1102 Mouse Forward 5'-CCCCAAAGGGATGAGAAGTT-3' PAO-393 TNFα Reverse 5'-CACTTGGTGGTTTGCTACGA-3' PAO-394 Mouse Forward 5'-TTACTGCTGTACCTCCACC-3' PAO-1293 VEGFA Reverse 5'-ACAGGACGGCTTGAAGATG-3' PAO-1294

Table 2.3. List of primers used for splicing analysis

Gene Sequence Gene Location PAO number BAX Forward 5’-GAGAGGTCTTTTTCCGAGTGG-3’ Exon 4 PAO-758 Reverse 5’-GGAGGAAGTCCAATGTCCAG-3’ Exon 6 PAO-759 BCL2L11 Forward 5'-AGACCAAATGGCAAAGCAAC-3' Exon 2 PAO-1073 Reverse 5'-TCTTGGGCGATCCATATCTC-3' Exon 4 PAO-1074 BIRC5 Forward 5'-GGACCACCGCATCTCTACAT-3' Exon 2 PAO-673 Reverse 5'-TCTCCGCAGTTTCCTCAAAT-3' Exon 3 PAO-674 Caspase-9 Forward 5'-AGACCAGTGGACATTGGTTC-3' Exon 2 PAO-462 Reverse 5'-GGTCCCTCCAGGAAACAAA-3' Exon 7 PAO-463 CCNL2 Forward 5'-AGCCCGTGCCTCTACTACTG -3' Exon 4 PAO-1190 Reverse 5'- AAGAAACCAATGGGGACGAT-3' Exon 8 PAO-1191 CFLARL Forward 5'-CGAGGCAAGATAAGCAAGGA-3' Exon 12 PAO-545 Reverse 5'-GGCAGAAACTCTGCTGTTCC-3' Exon 14 PAO-546 CFLARS Forward 5'-CGAGGCAAGATAAGCAAGGA-3' Exon 6 PAO-547 Reverse 5'-CACATGGAACAATTTCCAAGAA-3' Exon 7 PAO-548 DIABLO Forward 5'-GGCTCTGAAGAGTTGGCTGT-3' Exon 2 PAO-962 Reverse 5'-GCTGCCATCTCTGAAAGACC-3' Exon 6 PAO-964 HIF1A Forward 5'-TGCATCTCCATCTCCTACCC-3' Exon 13 PAO-930 Reverse 5'-AGCTGTCTGTGATCCAGCATT-3' Exon 16 PAO-931 HNRNPH1 Forward 5'-GTACACATGCGGGGATTACC-3' Exon 8 PAO-1194 Reverse 5'-TCCCATCATTTGGCTACCAT-3' Exon 11 PAO-1195 HRAS Forward 5'-GTGGGGAACAAGTGTGACCT-3' Exon 3 PAO-965 Reverse 5'-ATCTCACGCACCAACGTGTA-3' Exon 5 PAO-966 NAIP Forward 5'-CCACAAGTGAAAGCAATCTTGA-3' Exon 11 PAO-1214 Reverse 5'-CTTTGCCTCTTGAAACCACTG-3' Exon 12 PAO-1215 RBM3 Forward 5'-TTGAACTGCCATGTCCTCTG-3' Exon 2 PAO-1207 Forward 5'-AGCAGACTTGCCTGCATGAT-3' Exon 5 PAO-1208 SRSF5 Forward 5'-TGCTCCACCTGTAAGAACAGAA-3' Exon 6 PAO-1188 Forward 5'-ATTTAGGTCGGTGTGCATCC-3' Exon 7 PAO-1189 TICAM2 Forward 5'-CCTTCGAGCTTCCTGACAAC-3' Exon 1 PAO-1143 Forward 5'-CATTAACCCCTGACTCACAGC-3' Exon 4 PAO-1144 Reverse 5'-TGTCCATGCAGACCCATTTA-3' Exon 5 PAO-1145 TSC1 Forward 5'-CGTCTCCTTTTTGCGTTCTC-3' Exon 7 PAO-956 Reverse 5'-AGAGCAAGACCCTGTCTCCA-3' Intron 12 PAO-957 Reverse 5'-CAGTCTGTCCAGCACTTCCA-3' Exon 15 PAO-958

Table 2.4. List of clones generated in this study.

Clone Tag Backbone Vector Selection Marker AB number Source hnRNPA2 (Full length) GFP pEGFP-N1 Kan AB598 Dr. Lexie Friend hnRNPA2 (Full length) - pET9c Kan AB577 Dr. Adrian Krainer hnRNPA2 (Full length) - pENTR Kan AB599 Daniel Arango hnRNPA2 (Full length) GST pDEST15 Amp AB624 Daniel Arango hnRNPA2 (Full length) His pDEST17 Amp AB728 Daniel Arango hnRNP-A21-263 - pENTR Kan AB621 Daniel Arango hnRNP-A21-255 - pENTR Kan AB1009 Daniel Arango hnRNP-A21-248 - pENTR Kan AB1010 Daniel Arango hnRNP-A21-189 - pENTR Kan AB681 Daniel Arango hnRNP-A2190-341 - pENTR Kan AB682 Daniel Arango hnRNP-A2264-341 - pENTR Kan AB617 Daniel Arango hnRNP-A21-263 GST pDEST15 Amp AB637 Daniel Arango 1-255 69 hnRNP-A2 GST pDEST15 Amp AB1004 Ms. Joanna Li hnRNP-A21-248 GST pDEST15 Amp AB1005 Ms. Joanna Li hnRNP-A21-189 GST pDEST15 Amp AB685 Daniel Arango hnRNP-A2190-341 GST pDEST15 Amp AB686 Daniel Arango hnRNP-A2264-341 GST pDEST15 Amp AB638 Daniel Arango hnRNP-A21-248/P239R GST pDEST15 Amp AB1014 Mr. Dasean Nardone hnRNP-A21-255/R254G GST pDEST15 Amp AB1015 Mr. Dasean Nardone hnRNP-A21-255/R254N GST pDEST15 Amp AB1016 Mr. Dasean Nardone hnRNP-A21-255/Y250G GST pDEST15 Amp AB1017 Mr. Dasean Nardone pFLIP1-hnRNP-A2264-341 His, CFP, YFP pFLIP1 Amp AB724 Mr. Bledi Brahimaj pFLIP2-hnRNP-A2264-341 His, CFP, YFP pFLIP2 Amp AB725 Mr. Bledi Brahimaj pFLIP2-1-hnRNP-A2264-341 His, CFP, YFP pFLIP2 Amp AB739 Mr. Bledi Brahimaj pFLIP2-2-hnRNP-A2264-341 His, CFP, YFP pFLIP2 Amp AB740 Mr. Bledi Brahimaj pFLIP2-3-hnRNP-A2264-341 His, CFP, YFP pFLIP2 Amp AB741 Mr. Bledi Brahimaj pFLIP4-hnRNP-A2264-341 His, CFP, YFP pFLIP4 Amp AB727 Mr. Bledi Brahimaj pFLIP4-1-hnRNP-A2264-341 His, CFP, YFP pFLIP4 Amp AB802 Mr. Bledi Brahimaj pFLIP4-2-hnRNP-A2264-341 His, CFP, YFP pFLIP4 Amp AB803 Mr. Bledi Brahimaj hnRNPA1 GST pGEX-2T Amp AB908 Dr. Ralph Nichols BAG1L (Full length) GFP pEGFP-C1 Kan AB906 Dr. Ann Williams ARHGEF1 (Full length) GFP pEGFP-N1 Kan AB805 Dr. Philip Wedegaertner Table 2.5. List of primers for cloning and site-directed mutagenesis used in this study

PAO Number Sequence PAO-351 Forward 5'-AAGGAAAAAAGCGGCCGCCATGGAGAGAGAAAAG-3' PAO-338 Reverse 5'-TTATAGGCGCGCCCGTATCGGCTCCTCCCA-3' PAO-337 Reverse 5'-AAGGCGCGCCCATATCCAGGTCCTCCACCA-3' PAO-372 Reverse 5'-AAGGCGCGCCCAGAACTCTGAACTTCCTGC-3' PAO-374 Forward 5'-AAGGAAAAAAGCGGCCGCCGGAAGAGGAGGCAAC-3' PAO-352 Forward 5'-AAGGAAAAAAGCGGCCGCCGGCAACCAGGGTG-3' PAO-377 Forward 5'-AAGGAAAAAAGCGGCCGCCGGTACCATGGAGAGAGAAAA-3' PAO-379 Reverse 5'-TTATAGGCGCGCCCACTAGTGTATCGGCTCCTC-3' PAO-378 Forward 5'-AAGGAAAAAAGCGGCCGCCGGTACCGGCAACCAGGGTG-3' PAO-1232 Reverse 5’-TTATAGGCGCGCCCTCCTCTTCCTCCTCC-3’ PAO-1233 Reverse 5’-TTATAGGCGCGCCCGGGGCTACCTCCAAA-3 PAO-1089 Forward 5'-AAATTGCCACCTCCACGTCCTCCTCCATACC-3' PAO-1090 Reverse 5'-GGTATGGAGGAGGACGTGGAGGTGGCAATTT-3' PAO-1091 Forward 5'-GCGCGCCCTCCTCCTCCTCCTCCATAA-3' PAO-1092 Reverse 5'-TTATGGAGGAGGAGGAGGAGGGCGCGC-3' PAO-1093 Forward 5'-CGCGCCCTCCATTTCCTCCTCCATAACCGG-3' PAO-1094 Reverse 5'-CCGGTTATGGAGGAGGAAATGGAGGGCGCG-3' PAO-1095 Forward 5'-CTCTTCCTCCTCCACCACCGGGGCTACCTCCAA-3' PAO-1096 Reverse 5'-TTGGAGGTAGCCCCGGT GGTGGAGGAGGAAGAG-3' T7 insert up Forward 5’-NNNATGCTCGGGGATCCGAATT-3’ T7 insert down Reverse 5’-NNNAACCCCTCAAGACCCGTTTAG-3’

Chapter 3

Apigenin Induces DNA Damage in a PKCδ-Dependent Pathway Leading to Down-

Regulation of Genes Involved in Cell Cycle Control and DNA Repaira

3.1 Abstract

Apigenin exhibits anti-proliferative activities through mechanisms that remain poorly

understood. This study demonstrated that apigenin induces DNA damage in monocytic leukemia cells in p38 and PKCδ dependent pathways, yet independent of ROS production or caspase-3 activity. Treatment of monocytic leukemia cells with apigenin induced the phosphorylation of the Ataxia-Telangiectasia Mutated (ATM) kinase and histone H2AX, two key regulators of the DNA damage response pathway, without involving the Ataxia-

Telangiectasia Mutated and Rad-3-related (ATR) kinase. Silencing and pharmacological inhibition of PKCδ abolished ATM and H2AX phosphorylation, whereas inhibition of p38 attenuated H2AX phosphorylation independently of ATM. Consistently, the activation of ATM delayed cell cycle progression at G1/S and led to down-regulation of genes involved in G1/S transition and DNA repair. Taken together, this study defined the signaling networks responsible for the induction of DNA damage by apigenin in

a Arango D, Parihar A, Villamena FA, Wang L, Freitas MA, Grotewold E, Doseff AI. 2012. Apigenin induces DNA damage through the PKCδ-dependent activation of ATM and H2AX causing down- regulation of genes involved in cell cycle control and DNA repair. Biochem Pharmacol 84:1571-1580.

leukemia resulting in genome-wide changes in gene expression that derive in cell cycle

arrest and apoptosis.

3.2 Introduction

Apigenin has been shown to regulate several signaling pathways in leukemia [139,

143, 208], yet the precise mechanisms by which apigenin exert its anti-carcinogenic

effects remain poorly understood. Apigenin induced apoptosis in human monocytic

leukemia cells in a p38 and PKCδ-dependent phosphorylation of Hsp27 leading to

activation of caspase-3 [139, 143]. In addition, apigenin induced the production of ROS,

although dispensable for the execution of cell death [139, 143, 209].

Commonly used chemotherapeutic drugs exert their anti-carcinogenic effects by

inducing DNA-damage [210-212]. Genotoxic damage, including DNA double strand breaks (DSBs), results in the activation of a signaling network characterized by the activation of the Ataxia Telangiectasia Mutated (ATM) and the Ataxia-Telangiectasia

Mutated and Rad-3-related (ATR) kinases, among others [213-217], leading to the phosphorylation of the histone H2AX at Ser-139 (γH2AX) [213, 217, 218] a hallmark of

DSBs [219]. Activation of ATR, ATM and H2AX have an important role in triggering a

DNA damage response including stimulation of DNA repair, activation of cell cycle checkpoints, and eventually induction of apoptosis [220]. PKCδ activation was observed in response to a variety of apoptotic and genotoxic stimuli [221-223] and is essential for etoposide-induced apoptosis [222], whereas the activation of the p38 pathway has been associated with stress-activated response and UV-radiation [224].

In spite of the growing significance of apigenin as an anti-cancer agent, its mechanisms of action remain elusive. In this chapter, we established a mechanism responsible for apigenin-induced apoptosis by inducing DNA damage in a p38 and PKCδ dependent pathways, revealing a complex signaling network responsible for the apoptotic effects of this dietary compound in leukemia.

3.3 Results

3.3.1 Apigenin induces DNA damage

To understand the early mechanisms associated with apigenin-induced apoptosis in monocytic leukemia, we examined whether apigenin promoted DNA damage using the comet assay in THP-1 monocytic leukemia cells. THP-1 cells treated with the diluent

DMSO showed negligible levels of Tail DNA percentage (% Tail DNA). Cells treated for

3 h with different concentrations of apigenin resulted in a dose-dependent increase in tail

DNA percentage. The % Tail DNA increased from 10% to 20% in cells treated with 25

µM and 50 µM apigenin respectively, reaching almost 30% in cells treated with 100 µM apigenin (Fig. 3.1A and B). Cells treated for 1h with 1 mM H2O2, a known inducer of

DNA-damage [225], showed approximately 50% Tail DNA, as previously reported (Fig.

3.1A and B) [225]. Next, we investigated the time-relationship of apigenin-induced DNA

damage and the activation of caspase-3, a key effector of apoptosis [139]. Cells treated

for different time points with 50 µM apigenin, a concentration previously reported as the

IC50 for THP-1 cells [139], showed a significant increase in % Tail DNA as early as 1 h

after apigenin treatment, an increase that continued even after 9 h (Fig. 3.1C). Caspase-3

activity was barely detectable at 3 h after apigenin treatment, but significantly increased after 6 h (Fig. 3.1D). Collectively, these results indicate that apigenin induces DNA damage prior to the activation of caspase-3.

3.3.2 Apigenin induces H2AX phosphorylation

To gain insights into the mechanisms regulating apigenin-induced DNA damage, histone modifications were analyzed by LC-MS in THP-1 cells treated with 50 µM apigenin for 12 h. An increase in a molecular weight peak corresponding to phosphorylated H2AX (Red arrow, 15133 Daltons; referred as γH2AX) was observed in apigenin-treated cells, compared with controls (Fig. 3.2A). No other histone modification changes were observed in apigenin-treated cells. The levels of non-phosphorylated

H2AX were similar in apigenin-treated cells and control cells, as indicated by the identical height of the 15053 Daltons peak (Black arrow, Fig. 3.2A). Analysis of γH2AX by western blot in THP-1 cells treated with 50 µM apigenin for various times showed a

~2.5 fold increase as early as 1 h after apigenin treatment reaching maximum levels at ~5 h (Fig. 3.2B). Formation of γH2AX nuclear foci, a hallmark of DSB, was detected by immunofluorescence in cells treated with 50 µM apigenin for 3h (Fig. 3.2C), a time when apoptotic bodies, characteristic of caspase-3-induced apoptosis, was not observed, as indicated by DAPI staining (Fig. 3.2C). Collectively, these results demonstrate that apigenin induces γH2AX, preceding the activation of the apoptotic program.

3.3.3 Apigenin-induced DNA damage is ROS and caspase 3-independent

Apigenin induced ROS production as well as the activation of caspase-3, p38 and

PKCδ [139, 143]. In agreement with previous results using electron paramagnetic

resonance [139], we observed that treatment with 50 µM apigenin significantly induced

ROS production at 1 h, as demonstrated by a ~4 fold increase in DCFDA fluorescence

- (Fig. 3.3A, black bars) and ~4 fold increase in superoxide anion (O2 ) levels using DHE

fluorescence (Fig. 3.3A, white bars). Pretreatment with 20 µM EUK, a ROS inhibitor,

- inhibited by 3-fold O2 (Fig. 3.3A). To investigate the molecular mechanisms involved in apigenin-induced DNA damage, we examined the % Tail DNA in THP-1 cells pretreated for 1 h with 20 µM EUK-134 (EUK), 20 µM DEVD-FMK (a caspase-3 inhibitor), 10 µM

SB203580 (a p38 inhibitor), or 15 µM rottlerin (a PKCδ inhibitor) followed by the addition of 50 µM apigenin for 3 h. Inhibition of ROS or caspase-3 had no significant effect in the % Tail DNA observed in apigenin-treated cells (Fig. 3.3B, white vs. black bars). In contrast, the % Tail DNA was reduced by almost 50% in cells pretreated with the p38 inhibitor SB203580, while inhibition of PKCδ decreased the % Tail DNA to levels found in controls (Fig. 3.3B). Taken together, these findings indicate that apigenin- induced DNA damage is independent of ROS production and precedes caspase-3 activation but dependent on p38 and PKCδ activities.

3.3.4. PKCδ and p38 are required for apigenin-induced DNA damage

To investigate the mechanisms involved in apigenin-induced DNA damage we evaluated the signaling network mediating γH2AX. Pretreatment with 10 µM SB203580

for 1 h prior the addition of apigenin significantly reduced γH2AX (Fig. 3.3C, compare

lanes 5-7 and lanes 2-4), whereas inhibition of PKCδ with 15 µM rottlerin completely

abrogated apigenin-induced γH2AX (Fig. 3.3C, compare lanes 8-10 and lanes 2-4). In

addition, inhibition of PKCδ resulted in a slight increase of p38 phosphorylation in

apigenin-treated cells (Fig. 3.3C, compare lanes 8-10 and lanes 2-4), while inhibition of

p38 reduced PKCδ activity to levels found in controls, as indicated by the reduced

phosphorylation of H2B in in vitro kinase assays (Fig. 3.3D, compare lanes 4 and 3).

Next, we evaluated the effect of apigenin on ATM and ATR phosphorylation, key

regulators of γH2AX. ATM phosphorylation increased 30 min after treatment with 50

µM apigenin compared with control cells, an increase that remained even after 6 h, (Fig.

3.4A). In contrast, no change in ATR phosphorylation was observed in cells treated with

apigenin (Fig. 3.4A). To investigate the role of PKCδ and p38 on apigenin-induced ATM

phosphorylation, cells were pretreated with 15 µM rottlerin or 10 µM SB203580 for 1 h

prior the addition of 50 µM apigenin. Inhibition of PKCδ resulted in a reduction of ATM

phosphorylation to levels observed in control cells (Fig. 3.4B, compare lanes 8-10 and

lanes 2-4). In contrast, inhibition of p38 had no effect on apigenin-induced ATM

phosphorylation (Fig. 3.4B, compare lanes 5-7 and lanes 2-4). To further define the

signaling network, PKCδ or p38 were silenced with siRNA-PKCδ or siRNA-p38

respectively, and phosphorylation of H2AX and ATM were evaluated in cells treated

with 50 µM apigenin or diluent control for 3 h. Silencing of PKCδ decreased apigenin- induced γH2AX and ATM phosphorylation to levels found in siRNA-control cells, but had no effect on apigenin-induced p38 phosphorylation (Fig 3.4C, compare lane 6 and

lane 2). However, silencing of p38 reduced apigenin-induced γH2AX, but had no effect on apigenin-induced ATM phosphorylation (Fig 3.4C, compare lane 4 and lane 2). Taken together, these results indicate a complex crosstalk between p38 and PKCδ, both capable of regulating γH2AX during apigenin-induced DNA damage and suggest a central role of

PKCδ in the phosphorylation of ATM.

3.3.5 Apigenin affects cell cycle progression and gene expression

To investigate the biological response to apigenin-induced DNA damage, we examined cell cycle in THP-1 cells treated with DMSO or with increasing concentrations of apigenin for 24 h. Nocodazol, a treatment that induces G2/M arrest in leukemia cells, was used as control [226]. We observed a significant decrease of cells with 4N content,

G2, from ~25% found in control cells to 10% in cells treated with 50 µM apigenin (Fig.

3.5A and B). This effect was accompanied by a significant accumulation of Sub-G1 cells, an indicator of apoptosis, from 5% in control cells to ~15 and 20% in cells treated with

25 and 50 µM apigenin, respectively (Fig. 3.5A and B) and an increase of cells with 2N content, G1, from 40% in control cells to 50% in cells treated with 25 or 50 µM apigenin, indicative of a G1 arrest (Figs. 3.5A and B). No significant differences were observed in cells treated with 10 µM apigenin. These results indicate that apigenin induces G1 arrest in THP-1 cells while increasing the Sub-G1 cell population.

To investigate the effects of apigenin on gene expression genome wide, we performed microarray analysis in THP-1 cells treated with 50 µM apigenin or diluent control for 3 h.

Apigenin changed significantly the expression of ~8.5% (2,390 genes) of all the genes

represented in the array, 81% (1936 genes) of which were down-regulated. Gene

Ontology (GO) analysis showed that the biological pathways most significantly affected by apigenin comprise gene expression, cell cycle, post-translation modification, DNA repair, and cell death (Fig. 3.5C). Consistent with the GO analysis, heat map representation of the cell cycle genes showed that apigenin induced a significant change in 259 genes involved in the cell cycle, corresponding to ~10% of all cell cycle genes represented in the microarray (Fig. 3.5D). In addition, apigenin significantly changed 140

DNA repair genes, corresponding to ~5.5% of all DNA repair genes represented in the array (Fig. 3.5E).

In agreement with the microarray data, qRT-PCR showed that the expression of cell cycle genes including CCNE1, CCNE2, E2F2, MYC, and CDC25A was significantly decreased in cells treated with 50 µM apigenin for 3 h, when compared to controls (Fig.

3.5F). Moreover, apigenin decreased the expression of DSBs repair genes such as

BACH1, FEN1, XRCC2, POLH and RAD1 (Fig. 3.5F). Consistent with the microarray data, CDK2 expression was not affected, suggesting a specific set of G1/S genes being modulated by apigenin (Fig. 3.5F). All together, our results indicate that apigenin induces

DNA damage leading to down-regulation of genes involved in cell cycle regulation and

DNA repair while inducing cell death.

4. Discussion

This study demonstrated that apigenin-induced activation of the apoptotic pathway is preceded by p38 and PKCδ-dependent induction of DNA damage, but independent of

ROS production. Apigenin induced activation of ATM and H2AX in a PKCδ-dependent, whereas p38 regulated H2AX phosphorylation but had no effect in ATM, suggesting a complex regulatory network (Fig. 3.6). Consistently, apigenin exhibited a down- regulation of DNA repair and cell cycle progression genes.

Commonly used chemotherapeutic drugs, such as etoposide, campothecin, and doxorubicin are topoisomerase poisons that trigger apoptosis by inducing DSBs leading to DNA damage [210-212]. Flavonoids such as genistein, luteolin and apigenin, induced topoisomerase dependent DNA damage in vitro [227, 228], however, whether they modulate topoisomerase in vivo has not been reported yet. Recently, apigenin was shown to bind DNA in vitro and localize to the nuclear matrix [229], suggesting that apigenin may promote DNA damage by intercalating with DNA and inhibiting the action of topoisomerases. Consistent with our results, others showed that apigenin induced DNA damage in glioma cell lines [230]. Previous studies showed that, in HL-60 leukemia cells, apigenin induced DNA fragmentation as a result of caspase-3 activation [231, 232]. We found that, while DNA damage increased over the experimental time, it was detected prior to the activation of caspase-3, observed only 3 h after apigenin treatment.

Consistently, the caspase-3 inhibitor DEVD-fmk failed to block apigenin-induced DNA damage. Thus, it is possible that one mechanism by which apigenin triggers apoptosis is by inducing DNA damage, which subsequently results in the activation of the apoptotic cascade.

We previously reported that apigenin induces a transient production of ROS [139], prompting us to investigate whether apigenin-induced DNA damage was ROS

related. Notably, we found that apigenin-induced DNA damage was ROS independent.

Although ROS induces oxidative DNA damage and single strand beaks, these events

might be rapidly repaired after removal of the free radicals. Moreover, we previously

reported that apigenin-induced caspase-3 activation was ROS independent [139], further

supporting the hypothesis that the generation of ROS is not involved directly in the

cytotoxic effects of apigenin. The flavone luteolin was shown to induce DNA damage

and apoptosis by affecting topoisomerase II activity in leukemia cells [233]. Consistent

with our results, luteolin-induced DNA damage was ROS independent [233]. In contrast,

the flavonol quercetin induced ROS-dependent DNA damage [233], exerting its DNA

damaging effects via metal-catalyzed oxidation with the subsequent generation of ROS

[234-236]. Thus, while flavonols induce DNA damage via production of ROS, flavones

such as apigenin may exert its clastogenic effect by inhibiting topoisomerases, further

aggravated by the activation of PKCδ.

DNA damage was accompanied by an increase of H2AX phosphorylation, an

important player in the DNA damage response pathway, which is modulated by ATM

and ATR kinases [217-220]. We found that apigenin induced robust ATM

phosphorylation at 30 min post-treatment, consistently with the kinetics of γH2AX, but

had not effect on ATR phosphorylation, suggesting that apigenin-induced γH2AX is

dependent on ATM. To further elucidate the mechanisms of apigenin-induced DNA

damage we assessed the role of PKCδ and p38, kinases activated in response to genotoxic agents hence regulating DNA-damage response pathways [222, 223, 237-239]. PKCδ and p38 activities arose with similar kinetics as apigenin-induced DNA damage. Herein, we

found that pharmacological inhibition and silencing of PKCδ completely abolished

apigenin-induced DNA damage, γH2AX and p-ATM, indicating that these events are

dependent on the activity of PKCδ. In contrast, inhibition of p38 only partially attenuated

apigenin-induced DNA damage and γH2AX and had not effect on ATM phosphorylation,

suggesting that PKCδ is the main player in promoting apigenin-induced DNA damage.

Pharmacological inhibition of p38 has been shown to decrease stress-induced γH2AX in leukemia, keratinocyte and myeloma cells [238-241]. Consistent with our results, it has been shown that p38 phosphorylates H2AX in vitro [238] and regulates DNA-damage

response downstream of ATM in response to chemotherapeutic drugs such as

doxorubicin and cisplatin [242]. In addition, pharmacological inhibition of PKCδ

abrogated etoposide-induced DNA fragmentation, γH2AX and ATM phosphorylation in

epithelial cells [222, 223], indicating the key role of PKCδ in DNA damage response

pathway. Although several models have been suggested, the mechanisms on how PKCδ

regulates the DNA damage response remain elusive [223, 237, 243, 244]. Based on our

findings, we propose that apigenin-induced PKCδ activity inhibits endogenous DNA

repair while promoting DNA strand breaks and apoptosis.

In response to DSBs, ATM phosphorylates many cell cycle checkpoint-related factors

such as γH2AX, p53, and CHK1 leading to cell cycle arrest, and eventually apoptosis

[245]. We found that apigenin decreased the G2/M population, followed by an increase

of cells in G1 and Sub-G1 in THP-1 cells, which are deficient on p53. Exposure of a wide

array of malignant cells to apigenin induced G2/M or G1 arrest regardless of their Rb and

p53 status, indicating that the modulation of cell cycle by apigenin is p53 independent

and cell type specific [155, 246-250]. Our results further support these findings due to the p53-deficient status of THP-1 cells [251]. The reasons for the diverse effect of apigenin on cell cycle regulation are yet to be uncovered but may reflect the existence of cell type specific signaling cascades.

To determine the mechanisms responsible for apigenin-induced cell cycle arrest and apoptosis we performed genome wide mRNA expression analysis by microarray.

Apigenin significantly down-regulated the expression of genes involved in cell cycle control and DNA repair. These observations were confirmed by qRT-PCR and were consistent with the induction of DNA damage and cell cycle arrest by apigenin (Fig. 3.5).

Microarray analysis was previously performed in prostate cancer cells treated with luteolin [252]. Consistent with our findings, among the biological pathways most significantly affected by that flavone were gene expression, cell cycle, cell death and

DNA repair [252]. Interestingly, the expression of ~80% of the genes affected by luteolin were also down-regulated [252]. We observed more than 500 genes that were commonly changed in both studies constituting potential transcriptional targets of flavones. CCNE1,

CCNE2, CDC25A, E2F2 and MYC, among others, were down-regulated in prostate as well as leukemia cells. E2F2 and MYC are transcription factors that regulate the expression of genes involved in cell cycle progression such as CCNE1 and CCNE2 [253].

CCNE1 and CCNE2 bind to cyclin dependent kinase 2 (CDK2) promoting its activity and the cell cycle transition from G1 to S [254]. Additionally, apigenin decreased the expression of genes involved in the response and repair of DSBs such as BACH1, FEN1,

XRCC2, DNA polymerase η (POLH) and RAD1, indicating that this flavones down- modulates DNA repair mechanisms increasing its cytotoxic activity. Besides the genes confirmed by qRT-PCR, apigenin decreased the expression of components of the basal transcription machinery including POLR2D (RNA polymerase II polypeptide D),

GTF2H4 (General transcription factor II H4), GTF2E1, GTF2B, and TBP (TATA box binding protein). The effect of apigenin on general transcriptional mechanisms may explain the overall down-regulation of genes affected by apigenin.

In conclusion, apigenin induced DNA damage activation and H2AX phosphorylation through ATM but independent of ATR. Activation of ATM and H2AX was PKCδ and p38 dependent leading to transcriptional down-regulation of genes involved in cell-cycle control and DNA repair hence triggering apoptosis.

Figure 3.1. Apigenin induces DNA damage in leukemia cells. A. Comet assays of THP-1 cells treated with indicated concentrations of apigenin or diluent DMSO for 3 h or with 1 mM H2O2 for 1 h. B. % Tail DNA determined by alkaline comet assay in cells treated as described in (A). C. % Tail DNA in cells treated with 50 #M apigenin for different lengths of time or diluent DMSO for 9 h. D. Caspase-3 activity was determined in cells treated as described in (C). Data represents Mean ± SEM, n = 3. *p < 0.05, **p<0.01, compared to DMSO control. One-way ANOVA. Adapted from Arango et al 2012, Biochem Pharmacol [255].

84 Figure 3.2. Apigenin induces H2AX phosphorylation. A. Epigenetic changes were analyzed by LC-MS in THP-1 cells treated with DMSO or 50 µM apigenin for 12 h (upper and lower panel, respectively). Black and red arrows indicate peaks corresponding to non-phosphorylated and phosphorylated H2AX, respectively. In collaboration with Drs. Liwen Wang and Michael A. Freitas. B. Lysates from THP-1 cells treated with 50 µM apigenin for different time periods or diluent DMSO for 6 h were immunoblotted with "H2AX antibodies, membrane were re-blotted with #-tubulin antibodies. C. THP-1 cells were immuno-stained with anti-"H2AX antibodies and counterstained with DAPI 3 h after treatment with 50 µM apigenin or DMSO. All results shown are representative of three independent experiments. Adapted from Arango et al 2012, Biochem Pharmacol [255].

85 Figure 3.3. Apigenin-induced DNA damage is mediated by PKC! and p38. A. THP-1 cells pre-treated with 20 µM EUK-134 for 1 h prior the addition of 50 µM apigenin or diluent DMSO for an additional hour were stained with DCFDA (black bars) or DHE (white bars) and visualized under the fluorescence microscope. Fluorescence intensity was determined using the ImageJ software. Data represent Mean ± SEM, n = 3. **p < 0.01 for EUK + Api compared to Api. One-way ANOVA. In collaboration with Dr. Arti Parihar. B. Comet assays of THP-1 cells pretreated for 1 h with 20 µM EUK-134, 20 µM DEVD-FMK, 10 µM SB203580 or 15 µM rottlerin prior to the addition of 50 µM apigenin or DMSO for 3 h (white and black bars respectively). Data represent the Mean ± SEM, n=4, *p < 0.05 and **p < 0.01 to apigenin treated cells. One-way ANOVA. C. Lysates from THP-1 cells pretreated for 1 h with 10 µM SB203580, 15 µM rottlerin or DMSO prior the addition of 50 µM apigenin or diluent DMSO for the times indicated were immunoblotted with "H2AX, p-p38, p38 and #-tubulin antibodies. D. Lysates from THP-1 cells treated with 50 µM apigenin in the presence or absence of 10 µM SB203580, 15 µM rottlerin or DMSO (indicated as -) were immunoprecipitated with anti-PKC! antibodies or isogenic IgG control and subsequently subjected to in vitro kinase assays, phosphorylated H2B was visualized by autoradiography. The same membrane was immunoblotted with anti-PKC! antibodies. Results are representative of three independent experiments. Adapted from Arango et al 2012, Biochem Pharmacol [255]. 86

Figure 3.4. Apigenin induces ATM and "H2AX phosphorylation in a PKC! and p38-dependent pathway. A. Lysates of THP-1 cells treated with 50 µM apigenin for the indicated times or diluent DMSO for 6 h were immunoblotted with anti-phospho-ATM (p-ATM), anti-ATM, anti-phospho-ATR (p-ATR) and anti-ATR antibodies. B. THP-1 cells were pretreated for 1 h with 10 µM SB203580 or 15 µM rottlerin prior to the addition of 50 #M apigenin for different time periods or diluent DMSO for 3 h. ATM phosphorylation was analyzed by western blot. C. Lysates from THP-1 cells transfected with siRNA-control, siRNA-p38 or siRNA-PKC! and subsequently treated with 50 #M apigenin or diluent DMSO for 3 h were analyzed by western blots with anti-"H2AX, p- ATM, PKC!, p-p38 or p38 antibodies. In all panels, the same membranes were re-blotted with anti-b-tubulin antibodies to ensure equal loading. Results are representative of three independent experiments. Adapted from Arango et al 2012, Biochem Pharmacol [255]. 87 Figure 3.5. Apigenin affects cell cycle progression of THP-1 cells by down-regulating cell cycle and DNA repair genes. A. Cell cycle distribution was analyzed in THP-1 cells treated for 24 h with various doses of apigenin, DMSO or 200 ng/ml nocodazol for 24 h. All results shown are representative of three independent experiments. B. Percentage of cells in different stages of the cell cycle as indicated in (A). Data represent the Mean ± SEM, n=3, *p < 0.05, Two-Way ANOVA. C. Gene expression analysis of THP-1 cells treated with 50 mM apigenin or diluent control for 3 h. Genes significantly changing between groups were analyzed based on Gene Ontology. Bars correspond to functional categories significantly enriched in the data sets. D. Heat map representation of cell cycle genes significantly modulated by apigenin (259 genes corresponding to 10.4% of total cell cycle genes). E. Heat map representation of DNA repair genes significantly modulated by apigenin (140 genes corresponding to 5.5% of total DNA repair genes). F. mRNA expression of selected genes was analyzed by qRT-PCR and normalized to the expression of GAPDH. Data represents the Mean ± SEM. n=4, *p < 0.05. Adapted from Arango et al 2012, Biochem Pharmacol [255].

88 Figure 3.6. Working model of apigenin-induced DNA damage. Apigenin induces DSBs, ATM and H2AX phosphorylation in a PKC!-dependent pathway, while p38 modulates apigenin-induced DNA damage independent of ATM. Apigenin-induced down-regulation of cell cycle control genes and ATM activation led to cell cycle arrest at the G1/S transition. Down-modulation of genes involved in DNA repair by apigenin indicates that cells may be unable to repair apigenin-induced DNA damage, hence triggering apoptosis. Adapted from Arango et al 2012, Biochem Pharmacol [255].

89 Chapter 4

Dietary Apigenin Reduces LPS-Induced Expression of MiR-155 Restoring Immune-

Balance During Inflammationb

4.1 Abstract

High incidence of inflammatory diseases afflicts the increasing aging-population

infringing a great health burden. Nutraceuticals, including apigenin, are emerging as

important anti-inflammatory approaches due to their health benefits, lack of adverse

effects and reduced costs. MicroRNAs (miRs) play a central role in inflammation by

regulating gene expression, yet how nutraceuticals affect miRs is poorly understood. The

aim of this study was to identify miRs involved in the anti-inflammatory activity of

apigenin and apigenin-rich diets and determine their immune regulatory mechanisms in

macrophages and in vivo. A high-throughput quantitative real-time PCR screen of 312

miRs in macrophages revealed that apigenin reduced LPS-induced miR-155 expression.

Analyses of miR-155 precursor and primary transcript indicated that apigenin regulated

miR-155 transcriptionally. Apigenin-reduced expression of miR-155, led to the increase

of anti-inflammatory regulators FOXO3a and SMAD2 in LPS-treated macrophages.

b Arango D, Diosa-Toro M, Rojas-Hernandez LS, Cooperstone JL, Schwartz SJ, Mo X, Jiang J, Schmittgen TD, Doseff AI. 2015. Dietary apigenin reduces LPS-induced expression of miR-155 restoring immune balance during inflammation. Mol Nutr Food Res 59:763-772. 90

Studies using IKKβ knock out bone marrow derived macrophages showed that the

effect of apigenin on miR-155 is regulated by the NF-κB/IKKβ axis. In vivo, apigenin inhibited NF-κB activity in the lungs. Moreover, apigenin or a celery-based apigenin rich diet reduced LPS-induced expression of miR-155 and decreased TNFα in lungs from

LPS-treated mice. These results demonstrate that apigenin and apigenin-rich diets exert effective anti-inflammatory activity reducing LPS-induced expression of miR-155 in an

NF-κB/IKKβ dependent pathway, thereby restoring immune balance.

4.2 Introduction

Inflammation is the first line of defense against pathogens and its proper regulation is essential for physiological homeostasis [256]. Dysregulated inflammation, a major contributor to the pathophysiology of sepsis, is characterized by exacerbated production of inflammatory mediators and uncontrolled immune function [257, 258]. Several nonsteroidal anti-inflammatory drugs (NSAIDs) are currently used to ameliorate inflammation, but their long-term consumption is often accompanied by adverse effects including cardiovascular diseases [26]. Hence, as the incidence of inflammatory diseases increases worldwide, there is great interest in identifying alternative approaches to restore proper immune balance. Nutraceuticals are emerging as potential immune modulators, due to their lack of adverse effects, low cost and easy administration [1].

Inflammatory stimuli, including microbial components such as lipopolysaccharide

(LPS), trigger a signaling cascade that activates NF-κB [259]. NF-κB increases the

expression of inflammatory regulators including pro-inflammatory cytokines such as

TNFα and miRs [260]. MiRs are single-stranded non-coding RNAs that modulate gene

expression by binding to the complementary regions of specific mRNA targets enabling

mRNA degradation or inhibiting translation [261]. MiRs are transcribed as long primary

transcripts or pri-miRs that are processed into ~60-100 nucleotide hairpins, named

precursors or pre-miRs. Pre-miRs are further processed by the endoribonuclease Dicer

into mature miRs, averaging 18-22 nucleotides in length [261].

MiRs can act as positive or negative regulators of inflammation [260]. MiR-155 is

induced by the NF-κB axis and plays a central role by regulating the duration and intensity of the immune response [18]. Several studies demonstrated that miR-155 regulates TNFα expression levels, a main immune-regulator, by increasing mRNA stability and translation [19-22]. In addition, other targets of miR-155, such as SMAD2

(smooth-muscle-actin and MAD-related 2) [262], a suppressor of the inflammatory molecules TNFα and iNOS (inducible nitric oxide synthase) [263], and FOXO3a

(Forkhead Box O3) [264], an inhibitor of NF-κB [265], are also important modulators of inflammation. Yet, despite the central role of miRs in inflammation, how anti- inflammatory dietary compounds affect miRs remains limited studied.

We showed that apigenin decreases LPS-induced lethality by reducing NF-κB activity and TNFα production [169]. In addition, apigenin reduced LPS-induced endothelial cell death by restoring normal metabolic function [266]. Highlighting the specificity of apigenin, naringenin lacked anti-inflammatory activities [139, 267]. Apigenin, similar to other flavonoids, is usually found in plants linked to sugars (glycosylated) [268]. We

reported that glycosylated flavonoids showed reduced absorption and anti-inflammatory activity [205]. To overcome these limitations, we developed celery-based apigenin-rich diets with increased aglycone (non-glycosylated) content [205]. Mice fed with these diets showed increased absorption of apigenin, reaching serum concentrations that effectively reduced inflammation in macrophages [205]. Yet, the immune-regulatory activity of celery-based apigenin-rich diets in vivo has not been studied.

Here, we used a high-throughput quantitative reverse transcription-PCR (qRT-PCR) screening to identify the miRs differentially regulated by apigenin during LPS-induced inflammation in macrophages. MiR-155 was the only miR affected by apigenin during

LPS-induced inflammation. We found that apigenin regulates pri-miR-155 at the transcriptional level. Importantly, we showed that consumption of a celery-based apigenin rich diet results in the reduction of LPS-induced miR-155 and inflammatory modulators in vivo at inflammatory organ sites. These studies demonstrate that celery- based diets rich in apigenin confer immune-regulatory activity in vivo, reaching levels of effectiveness similarly found with pure apigenin and suggest a mechanism by which dietary flavones, through miRs regulation, contribute to restore immune balance.

4.3 Results

4.3.1 Apigenin regulates inflammatory miR expression in LPS-stimulated macrophages

To identify miRs responsible for the immune-regulatory activity of apigenin, a high- throughput screening of miRs was conducted, as previously described [185]. Total RNA

was isolated from macrophages treated concurrently with 100 ng/ml LPS and 50 µM

apigenin (Api+LPS) or diluent DMSO (LPS), or treated with PBS and apigenin (Api) or

vehicle PBS and DMSO (DMSO) for 8 h. Changes in miR expression between LPS and

DMSO (Fig. 4.1A) or LPS and Api+LPS (Fig. 4.1B) were represented as Volcano plots

and considered significant when p values were less than 0.05 and fold changes were

greater than 2 fold (Fig. 4.1A and B, gray areas). Out of the 312 miRs included in the

array, 154 were expressed in macrophages. LPS significantly changed the expression of

five miRs as compared with macrophages treated with DMSO (Fig. 4.1A, black dots).

LPS increased subtly the expression of miR-715 (2.2 fold, p = 0.031), miR-677 (2.9 fold,

p = 0.048), miR-692 (5.4 fold, p = 0.045) and greatly increased miR-155 (82.3 fold, p =

0.00096), whereas miR-490 was moderately decreased by LPS (-4.6 fold, p = 0.033, Fig.

4.1A). Our results indicated that miR-155 was the only miR highly affected by LPS in macrophages, in agreement with previous reports [18, 20, 269], while very few other miRs were only slightly changed by LPS. To evaluate the effect of apigenin on miRs involved in inflammation, the high-throughput screening was used to identify miRs affected by apigenin in LPS-treated macrophages. Apigenin decreased the LPS-induced

expression of miR-155 by ~120.5 fold (p = 0.0049, Fig. 4.1B), whereas a subtle increase

of miR-let-7a by ~5.8 fold was observed (p = 0.035, Fig. 4.1B).

We next performed validation of miR-155 and miR-let-7a, the only two miRs affected

by apigenin during LPS-induced inflammation as identified by the high-throughput

screening. Consistent with the high-throughput screening results, apigenin reduced the

LPS-induced expression of miR-155 by ~45-fold reaching levels found in controls (Fig.

4.1C). In contrast, miR-let-7a expression was unaffected in all conditions tested (Fig.

4.1D), and therefore was not further studied. Altogether, these results identified miR-155 as the only miR regulated by apigenin during LPS-induced inflammation in macrophages.

4.3.2 Apigenin reduces the LPS-induced expression of miR-155 primary transcript

To investigate the molecular mechanism involved in the regulation of miR-155 by apigenin, the expression of pri-miR-155 and pre-miR-155 were evaluated in macrophages treated with 100 ng/ml LPS in the presence of 50 µM apigenin (Api+LPS) or diluent

DMSO (LPS) or treated with PBS in the presence of apigenin (Api) or DMSO for 8 h.

LPS increased pri-miR-155 expression by ~50 fold, as compared with DMSO or apigenin controls (Fig. 4.2A). Apigenin significantly decreased LPS-induced pri-miR-155 to levels found in macrophages treated with DMSO control (Fig. 4.2A, Api+LPS vs. LPS). The precursor pre-miR-155 was induced ~100 fold in the presence of LPS as compared with control (Fig. 4.2B, LPS vs. DMSO). Apigenin reduced pre-miR-155 in LPS-treated macrophages as compared with LPS-treated macrophages (Fig. 4.2B, Api+LPS vs. LPS).

These findings suggest that apigenin regulates miR expression at the level of transcription, thereby resulting in lower levels of mature miR-155 (Fig. 4.2C).

4.3.3 Celery-based apigenin-rich diets reduce miR-155 in LPS-induced inflammation

We previously described a celery-based aglycone apigenin-rich extracts (ECE) and showed their effectiveness in reducing LPS-induced NF-κB activity in macrophages

[205]. To study the effect of ECE in inflammation, we first performed a dose response

with pure apigenin added concurrently with 100 ng/ml LPS for 8 h in macrophages and

determined its effect on the expression of miR-155. Apigenin reduced LPS-induced miR-

155 expression in a dose dependent manner reaching statistical significance at 5 µM and

at 25 µM apigenin the level found in controls (Fig. 4.3A).

Next, macrophages were treated with 100 ng/ml LPS or diluent PBS control (Fig.

4.3B, black and white bars respectively) in the presence of ECE (25 µM apigenin-

equivalent, as determined by HPLC analyses [205]), 25 µM pure apigenin, diluent

DMSO or 50 µM naringenin, a structurally related flavonoid with no anti-inflammatory activity [267]. ECE, apigenin and naringenin had no effect on basal miR-155 expression as compared with non-stimulated macrophages (Fig. 4.3B, white bars). ECE significantly reduced LPS-induced miR-155 expression (Fig. 4.3B, ECE vs. DMSO, black bars) to levels observed in non-stimulated macrophages (Fig. 4.3B, DMSO, white bar). Similar results were obtained in cells treated with LPS in the presence of pure apigenin (Fig.

4.3B, Apigenin vs. DMSO, black bars). In contrast, naringenin had no significant effect on LPS-induced miR-155 expression (Fig. 4.3B, Naringenin vs. DMSO, black bars).

Together, these results showed that the celery-based apigenin-rich extracts reduce macrophage miR-155 expression in inflammation to levels achieved with pure apigenin.

4.3.4 Celery-based apigenin-rich diets reduce LPS-induced miR-155 expression

modulating inflammatory regulators

Inflammatory regulators act with distinct kinetics that is in part dependent on the

stimulus and the length of stimulation [270, 271]. Hence, we determined the kinetics of

miR-155 expression in macrophages treated with 100 ng/ml LPS. We found that miR-155

expression significantly increased at 8 h after LPS stimulation and continued similarly

high at 24 h, in agreement with previous studies (Fig. 4.4A and [269]). Next, we studied

the expression of SMAD2 and FOXO3a, targets of miR-155 that regulate inflammation

[262, 264]. We found that the expression of SMAD2 and FOXO3a was not significantly

decreased in macrophages stimulated with LPS for 8 h compared with non-stimulated

macrophages treated with PBS diluent control (Fig. 4.4B, gray vs. white bars). Yet, at 24

h, SMAD2 and FOXO3a expression were significantly decreased in LPS-treated

macrophages (Fig. 4.4B, black vs. white bars).

We next examined the effect of apigenin and celery-based apigenin-rich extracts in the

LPS-induced expression of SMAD2 and FOXO3a. Macrophages were stimulated with

100 ng/ml LPS (Fig. 4.4C-D, back bars) or diluent PBS (Fig. 4.4C-D, white bars) in the

presence of 25 µM apigenin, ECE (at 25 µM apigenin-equivalent) or diluent DMSO for

24 h, a time corresponding to a high level of miR-155 and low levels of both SMAD2 and

FOXO3. ECE or apigenin, when administered alone, had no effect on the basal

expression of SMAD2 or FOXO3a (Fig. 4.4C-D, white bars). The addition of ECE to

LPS-stimulated macrophages significantly increased the expression of SMAD2 and

FOXO3a (Fig. 4.4C-D, ECE vs. DMSO, black bars). Similar effects were observed with

apigenin (Fig. 4.4C-D, Apigenin vs. DMSO, black bars). These results indicate that both the celery-based apigenin-rich diet and pure apigenin effectively modulate miR-155 biological targets during inflammation.

4.3.5 Celery-based apigenin-rich diets decrease LPS-induced expression of miR-155 and TNFα in vivo during inflammation

We previously reported that apigenin reduces LPS-induced mortality and decreases the expression of TNFα in vivo [169]. In this model, LPS increased miR-155 expression in mouse lungs [272]. Previous data from our group showed that apigenin decreased the expression of pro-inflammatory chemokines and leukocyte infiltration in lungs from septic mice [273]. Hence, we examined the effects of apigenin in LPS-induced miR-155 expression in lungs procured from mice treated with 50 mg/kg apigenin or DMSO 3 h prior stimulation with 37.5 mg/kg LPS or diluent PBS for additional 3 h. We found that

LPS increased miR-155 expression in lungs by ~14 fold as compared with lungs from mice treated with vehicle DMSO or apigenin alone (Fig. 4.5A, LPS vs. DMSO or Api).

Administration of apigenin significantly decreased LPS-induced miR-155 expression to levels found in control mice (Fig. 4.5A, Api+LPS vs. LPS). In addition, qRT-PCR analyses showed that apigenin significantly reduced LPS-induced TNFα in the lungs

(Fig. 4.5B, Api+LPS vs. LPS).

We next evaluated the effect of apigenin-diet in LPS-induced miR-155 expression in vivo. Mice were fed ad libitum with control or apigenin diets for 7 days prior stimulation with 37.5 mg/kg LPS or diluent PBS for 3 h. Remarkably, apigenin diet significantly

decreased the LPS-induced miR-155 expression in the lungs by ~3.5 fold as compared

with LPS-treated mice fed with control diet (Fig. 4.5C, Apigenin-Diet+LPS vs. Control

Diet+LPS), reaching similar levels found in non-stimulated mice (Fig. 4.5C, Apigenin-

Diet+LPS vs. Apigenin-Diet+PBS). Consistently, we found that lungs from LPS-treated mice fed with apigenin-diet had a significant ~3.5 fold reduction of TNFα as compared with LPS-treated mice fed with control diet (Fig. 4.5D, Apigenin-Diet+LPS vs. Control

Diet+LPS). Underscoring the physiological activity of the celery-based apigenin rich diet, we observed that apigenin-diet decreased TNFα protein levels in broncho-alveolar lavage fluids (BALF) and serum of LPS-treated mice (Fig. 4.5E and F). Together, these results demonstrate the immune-regulatory effectiveness of apigenin-rich diets in vivo, highlighting the potential benefits of dietary interventions in the restoration of proper immune function.

4.3.6 Apigenin decreases NF-κB activity in lungs

NF-κB is a key regulator of the innate immune response and miR-155 expression

[274]. Apigenin and ECE reduce LPS-induced NF-κB activity in macrophages,

suggesting its ability to modulate a master regulator of inflammation [169, 205]. Based

on these findings, we evaluated the effect of apigenin on NF-κB activity in lungs from

septic mice. For this purpose, transgenic mice expressing luciferase under the control of

6x-NF-κB responsive elements were treated with PBS or 50 mg/kg apigenin 3 h prior

LPS treatment for 6 h, a time in which LPS induces high levels of luciferase activity, the

indicator of NF-κB activity [275]. Mice treated with LPS showed high luciferase

expression in lungs and spleens at 6 h after LPS treatment (Fig. 4.6A and B), as previously reported [276]. Apigenin administration decreased LPS-induced luciferase expression in lungs, but had no effect in splenic luciferase levels (Fig. 4.6A), in agreement with previous results from our group showing that apigenin does not affect

LPS-induced inflammation in spleens [273]. Consistent with these findings, luciferase activity assays in tissue homogenates showed that apigenin decreased the expression of luciferase in lungs of LPS mice by 3 fold, but had no effect on NF-κB activity in the spleens (Fig. 4.6B). These findings provide evidence that dietary apigenin modulates NF-

κB activity in an organ-specific manner in vivo.

4.3.7 Apigenin blocks LPS-induced lethality

We previously showed that apigenin induces survival in mice receiving a lethal dose of LPS for 72 h [169]. To investigate whether apigenin provided long-term survival during acute inflammation, mice were treated with 50 mg/kg apigenin or DMSO 3 h prior stimulation with 37.5 mg/kg LPS or diluent PBS. No mortality occurred in mice receiving either apigenin or vehicle (Fig 4.7), in accordance with our previous results

[169]. All mice pretreated with diluent DMSO and receiving LPS died within 7 days (Fig

4.7). Administration of apigenin prior LPS challenge promoted 70% survival during a 7 days period evaluation (Fig. 4.7, Api+LPS).

Next, to investigate if apigenin administered after LPS challenge promotes survival during acute inflammation, mice were treated with 50 mg/kg apigenin or DMSO 3 h post- stimulation with LPS. Apigenin also promoted 70% survival during a 7 days period

100

evaluation when administered 1 h after challenge with LPS (Fig. 4.7, LPS+Api). These

results demonstrated that apigenin supports survival in septic mice, suggesting its potential on clinical interventions.

4.3.8 Apigenin decreases miR-155 and TNFα in an IKKβ/NF-κB mediated pathway

Throughout immune response, NF-κB activity has a bimodal kinetics with a first peak of activation within 15 min of inflammatory challenge and the second peak occurring between 4-8 h post-stimulus [271, 277]. In agreement with these reports, we observed

that mouse macrophages treated with LPS display an increase of NF-κB phosphorylation

at 15 min and a second high peak of NF-κB activation at 8 h following inflammatory stimulus (Fig. 4.8A). We previously reported that apigenin inhibits the first peak of NF-

κB phosphorylation when added prior LPS challenge [169]. To evaluate whether apigenin affects NF-κB activation when administered after inflammatory stimulus, we treated mouse macrophages with DMSO or 25 µM apigenin 1, 3 and 6 h post-challenge with 100 ng/ml LPS for 8 h and tested the phosphorylation of NF-κB by immunoblot using anti- phosho- NF-κB-p65-Ser536 antibodies. Apigenin decreased NF-κB phosphorylation at 8 h

when added 1 and 3 h after LPS treatment but had no effect when administered 6 h post-

inflammatory stimulus (Fig. 4.8B).

To evaluate the role of NF-κB in the regulation of miR-155 and TNFα by apigenin, we

differentiated bone marrow macrophages (BMDMs) from LysM-Cre+/-IKKβF/F (IKKβKO) harboring myeloid-lineage specific knockout of IKKβ (see section 2.31.3), the upstream activator of NF-κB activity, upon cre expression [206]. Macrophages from IKKβF/F mice

101

lacking cre were used as wild-type (IKKβWT). After seven days differentiation, IKKβWT

and IKKβKO macrophages were treated with 25 µM apigenin or DMSO one hour after

challenge with 100 ng/ml LPS for 8 h. The expression of TNFα in LPS-treated IKKβKO

macrophages was reduced by ~3-fold when compared to wild-type (Fig. 4.8C, IKKβWT

grey bars vs. IKKβKO grey bars). Apigenin, added 1 h post-LPS, significantly decreased

by ~3 fold the expression of TNFα in wild-type to similar levels found in IKKβKO

macrophages treated with LPS and DMSO (Fig. 4.8C, IKKβWT grey vs. black bars compared to IKKβKO grey bars). The levels of TNFα were similar between IKKβWT and

IKKβKO macrophages treated with LPS and apigenin (Fig. 4.8C, IKKβKO black bars). In addition, apigenin had no effect on TNFα expression in IKKβKO macrophages (Fig. 4.8C,

IKKβKO grey vs. black bars).

Next, we determined the expression of miR-155 in IKKβWT and IKKβKO macrophages treated with 25 µM apigenin or DMSO one hour after challenge with 100 ng/ml LPS for

8 h. MiR-155 levels were reduced by ~3-fold in LPS-treated IKKβKO compared to wild- type macrophages (Fig. 4.8D, IKKβWT grey bars vs. IKKβKO grey bars). Apigenin, added

1 h post-LPS, significantly decreased by ~3 fold the expression of miR-155 in wild-type,

but not in IKKβKO macrophages, treated with LPS and DMSO (Fig. 4.8D, IKKβWT grey

vs. black bars compared to IKKβKO grey bars). MiR-155 expression was similar between

IKKβWT and IKKβKO macrophages treated with LPS and apigenin (Fig. 4.8D, IKKβKO

black bars). Altogether, these results indicate that apigenin has immune-modulatory activity by inhibiting the IKKβ/NF-κB axis even when added after LPS stimulation.

102

4.4 Discussion

Increased incidence of inflammatory diseases and health care costs has ignited the

interest in the use of nutraceuticals for the prevention and treatment of these diseases.

Yet, the underlying immune-regulatory mechanisms associated to dietary compounds

remain unclear. Here we showed that the dietary flavonoid apigenin regulates the

expression of miR-155 and TNFα during LPS-induced inflammation in macrophages and

in vivo, thereby helping to restore immune balance.

Using a high-throughput miRs screening in macrophages, we found miR-155

expression dramatically increased (~80 fold) by LPS, in agreement with other reports [18,

20, 269]. Previous studies showed a subtle transient increase of miR-132 (~1.5 fold) in

bone marrow macrophages stimulated with LPS [269] and a small decrease of miR-125b

(-1.3 fold) in RAW264.7 macrophages stimulated with LPS for 6 h [20]. The differences on the LPS-affected miRs identified may rely on the length of stimulation and the cell

type used. Two miRs, miR-155 and miR-let-7a, were modulated by apigenin in LPS-

induced macrophages. However, only miR-155 was validated and therefore further

studied. Differences between high-throughput screenings and qRT-PCR have been

reported [269, 278], and may reflect noise inherent to high-throughput platforms,

suggesting the need to use more stringent threshold for screening analyses.

Few reports have investigated the biological relationship of flavonoids and miRs

[279, 280]. Apigenin improved glucose tolerance by decreasing maturation of miR-103 in

epithelial cells [281], indicating that apigenin may modulate different miRs depending on

103

the cell type and stimulus used. In addition, quercetin decreased LPS-induced expression

of miR-155 in macrophages [282]. We showed that apigenin and apigenin-diet, but not

naringenin, reduced LPS-induced miR-155 expression in macrophages. Similarly,

apigenin, but not naringenin, induces apoptosis of various cancer cell lines [139] and

lacked anti-inflammatory activity [267]. Naringenin also failed to interact with the direct

targets of apigenin (Chapter 6), suggesting that despite structural similarities, naringenin

and apigenin elicit significantly different biological activities.

Regulation of miR levels by polyphenols has been associated to indirect effects on the

transcription of miRs or to direct interaction with mature miRs [282, 283]. Apigenin

binds yeast RNA in vitro [284], yet whether apigenin binds miRs has not been shown.

Nevertheless, treatment with apigenin reduced LPS-induced pri-miR-155, pre-miR-155

and matured miR-155 expression, indicating that apigenin decreases miR-155 at the

transcriptional level. The expression of miR-155 in LPS-treated IKKβKO macrophages

was reduced by ~3 fold when compared to wild-type macrophages, suggesting that miR-

155 expression is regulated in part by the IKKβ/NF-κB axis during inflammation, as

previously reported [18, 285-287]. Apigenin reduces NF-κB activity by inhibiting IKKβ

in mouse macrophages [169], indicating that apigenin regulates the transcription of pri-

miR-155 in an IKKβ/NF-κB mediated pathway. Consistently, we found, using IKKβKO

macrophages, that apigenin regulates miR-155 expression in an IKKβ/NF-κB dependent pathway even when administered after inflammatory challenge (Fig. 4.8D). In addition,

we observed, using IKKβKO and IKKβWT macrophages, that the regulation of TNFα

expression by apigenin is mediated by the IKKβ/NF-κB axis (Fig. 4.8C).

104

Post-transcriptional regulation of miR-155 by KHSRP (KH-type splicing regulatory protein) has been reported in LPS-treated macrophages [269], highlighting the role of

RNA binding proteins in the control of miR-155 maturation during LPS-stimulation.

Recent studies observed that apigenin inhibits maturation of miRs such as miR-103 by inhibiting the activation of TRBP (Tar RNA-binding protein) [281]. The lack of effect of apigenin on the LPS-induced miR-155 maturation compared to the effect of apigenin reported for miR-103 maturation might reflect the ability of apigenin to modulate different sets of RNA-binding proteins implicated in miR maturation. Indeed, we have identified several RNA-binding proteins as direct targets of apigenin (Chapter 6).

Together, these results suggest that in macrophages, apigenin reduces pri-miR-155 transcriptionally during inflammation, through the inhibition of the IKKβ/NF-κB axis, leading to the decrease of mature miR-155.

MiR-155 has been suggested as an immune modulatory checkpoint [285], by targeting several molecules involved in the regulation of the immune response including SMAD2 and FOXO3a [262, 264]. SMAD2 have anti-inflammatory activity by suppressing, among others, TNFα production, a main marker of inflammation [263], while FOXO3a is an inhibitor of NF-κB activity, a master regulator of TNFα expression [265].

Additionally, miR-155 increases TNFα levels by stabilizing its mRNA and potentiating

TNFα translation [19-22]. Hence, miR-155 is a main regulator of the key inflammatory molecule TNFα. In LPS-treated macrophages, SMAD2 and FOXO3a expression was significantly decreased at 24 h. Differences in the kinetics of expression between miR-

155 and some of its targets have been previously reported, showing delays of 8 to 12 h

105

[288, 289]. These differences have been attributed to a combination of the rate of

transcription, the rate of miR loading into the RISC complex and the rate of mRNA decay

[290]. Hence, our observation on the lag of time between miR-155 induction and

SMAD2/FOXO3a silencing are aligned with reports on other miR-155 targets. We

observed that apigenin and ECE restored the expression of SMAD2 and FOXO3a to

levels found in non-stimulated cells. Consistently, we previously reported that apigenin

and ECE decreased LPS-induced TNFα in macrophages [205]. Thus, apigenin relieves

the inhibition of SMAD2 and FOXO3a and decreases TNFα levels by reducing miR-155,

thereby restoring immune-balance (Fig. 4.9).

LPS induces dysregulated inflammation leading to sepsis, severe lung injury, organ

failure and death [266, 291]. We have previously shown that apigenin improves survival of mice during the first week after LPS administration, with 100% survival on the 3-day evaluation [292]. The present study shows that apigenin effectively reduced mortality associated with inflammation by ~70% for a period of 7 days, even when administered after LPS-challenge suggesting that apigenin can induce long lasting and possibly complete recovery from the lethal effects of acute endotoxemia. Highlighting the physiological activity of apigenin, we found that intraperitoneal apigenin reduced NF-κB activity in vivo, as supported by the reduced level of luciferase activity in lungs from transgenic mice that express luciferase under the control of NF-κB responsive elements.

However, these effects seem to be organ specific, as apigenin reduced NF-κB activity in lungs, but had no effect in spleens. Apigenin reduced miR-155 and TNFα levels in lungs from LPS-treated mice (Fig. 4.5A and B). Yet, while promising effects have been shown

106

with apigenin, limited solubility makes this route of administration unfeasible for clinical applications. Overcoming this common limitation of flavonoids, we demonstrated that a celery diet rich in apigenin reduced LPS-induced miR-155 and effectively restored TNFα expression in vivo. Effective concentrations of flavonoids normally range ~5-50 µM in cellular models [169, 205, 266, 282, 293]. Importantly, our studies showed that in vivo, concentrations of apigenin of ~ 1 µM [28], found in serum of mice fed with the celery- based apigenin rich diets effectively confer immune-regulatory activity by decreasing miR-155 and TNFα levels in LPS-treated mice. Future experiments are guarantee to evaluate the therapeutic as well as the preventive potential of this diet.

Together, these findings identify miR-155 as a central apigenin-regulated miR in inflammation and provide evidence of the underlying mechanism by which apigenin and diets rich in apigenin contribute to restore homeostasis, highlighting the benefits of dietary interventions as a strategy to restore proper immune function in vivo.

107

Figure 4.1. Identification of apigenin-regulated miRs in LPS-induced inflammation. Macrophages were treated with 100 ng/ml LPS in the presence of 50 µM apigenin (Api+LPS) or DMSO vehicle (LPS) or with PBS in the presence of apigenin (Api) or DMSO (DMSO) for 8 h. A. MiRs high-throughput qRT-PCR-based array from macrophages treated with LPS or DMSO. Expression values were converted to Log2[Fold change(LPS/DMSO)]. B. MiRs high-throughput qRT-PCR-based array from macrophages treated with LPS or Api+LPS. Expression levels were converted to Log2[Fold Change(Api+LPS/LPS]. For A and B, Log2(Fold change) was compared to p- values and represented as Volcano plots. Dots represent mean of three independent biological replicates. C. Individual qRT-PCR of miR-155 or, D. miR-let-7a. MiR expression is normalized to the levels of snoRNA202 (internal control) and represented as fold change relative to DMSO. Mean ± SEM, n=5, *p < 0.05. One-way ANOVA. In collaboration with Drs. Jinmai Jiang and Tom Schmittgen. Adapted from Arango et al 2015, Mol Nutr Food Res [294].

108 Figure 4.2. Apigenin reduces the expression of LPS-induced primary miR-155 transcript. A. Macrophages were treated with 100 ng/ml LPS in the presence of 50 µM apigenin (Api+LPS) or diluent DMSO (LPS), apigenin with PBS (Api) or both vehicles PBS and DMSO (DMSO) for 8 h. Pri-miR-155 expression was analyzed by qRT-PCR. B. qRT-PCR of pre-miR in the same samples described in (A). For B and C, expression is normalized to the levels of 18S and represented as fold change relative to DMSO. Mean ± SEM, n=4, *p < 0.05. One-way ANOVA. C. Schematic representation of miR processing. Primary transcript, pri-miR-155, is transcribed and cleaved into a precursor or pre-miR-155 and subsequently exported to the cytoplasm and further processed into mature miR-155 by a Dicer-containing complex. Adapted from Arango et al 2015, Mol Nutr Food Res [294].

109 Figure 4.3. Celery-based apigenin rich diets reduce LPS-induced miR-155 expression in macrophages. A. Macrophages were treated with 100 ng/ml LPS in the presence of different concentrations of apigenin or diluent DMSO (indicated as +/-) or with diluents PBS and DMSO (-/-) for 8 h. B. Macrophages were treated with 100 ng/ml LPS (black bars) or PBS (white bars) concurrently with ECE (25 µM apigenin equivalents), 25 µM pure apigenin, 50 µM naringenin or DMSO for 8 h. MiR-155 expression was determined by qRT-PCR and normalized to the levels of snoRNA202. Data represent fold change relative to cells treated with DMSO. Mean ± SEM, n=5, *p < 0.05. One-way ANOVA. Adapted from Arango et al 2015, Mol Nutr Food Res [294].

110 Figure 4.4. Celery-based apigenin rich foods and pure apigenin restored the expression of miR-155 targets. A. Mir-155 expression in macrophages treated with 100 ng/ml LPS for different periods of time as determined by qRT-PCR. B. SMAD2 and FOXO3a were determined by qRT-PCR in same samples described in (A). For A and B, data represent fold change relative to time 0 h. Mean ± SEM, n=3, *p < 0.05. C. Macrophages were treated with 100 ng/ml LPS (black bars) or PBS (white bars) in the presence of ECE (25 µM apigenin equivalents), 25 µM apigenin or DMSO for 24 h. SMAD2 expression was evaluated by qRT-PCR. D. FOXO3a expression was examined by qRT-PCR in same samples described in (C). For C and D, data represent fold change compared to cells treated with PBS and DMSO (DMSO, white bars). Mean ± SEM, n=5, *p < 0.05. One-way ANOVA. MiR-155 expression was normalized to the levels of snoRNA202, while SMAD2 and FOXO3a were normalized to the expression of GAPDH. Adapted from Arango et al 2015, Mol Nutr Food Res [294].

111 Figure 4.5. Celery-based apigenin rich foods reduce LPS-induced expression of miR-155 and TNF! in vivo. A. Lungs were procured from male C57/BL6J mice treated with 50 mg/kg apigenin or DMSO for 3 h prior administration of 37.5 mg/kg LPS or PBS for additional 3 h. MiR-155 was determined by qRT-PCR and normalized against snoRNA202 expression. B. TNF# was determined by qRT-PCR and normalized to GAPDH expression in same lung samples used in (A). For A and B, data represent fold change expression relative to DMSO. Mean ± SEM, n = 8. * p < 0.05. C. Lungs were obtained from mice fed with control or apigenin diets for seven days prior administration of 37.5 mg/kg LPS or PBS by i.p. for 3 h. MiR-155 expression was determined by qRT- PCR. D. TNF# was determined by qRT-PCR in same lung samples used in (C). E. TNF# was determined in broncho-alveolar lavage fluids (BALFs) by ELISA. F. TNF# was determined in serum by ELISA. For C-F, data represent fold change compared to control diet. Mean ± SEM, n = 6. * p < 0.05. For A-D, One-way ANOVA. For E-F, Student’s t- test. Adapted from Arango et al 2015, Mol Nutr Food Res [294].

112 Figure 4.6. Apigenin inhibits NF-"B activity in vivo. NF-"B-RE-luc transgenic mice were treated with apigenin (50 mg/kg of body weight) or vehicle 3 h prior the administration of 37.5 mg/kg LPS for 6 h. A. Lungs and spleens were excised, imaged and luciferase activity expressed as photons/sec/cm2/steradian. B. Luciferase activity was assayed in tissue homogenates and expressed as luciferase activity relative to LPS. Data represent mean ± SEM, n=5. * p < 0.05. Student’s t-test.

113 Figure 4.7. Apigenin decreases LPS-induced lethality. Kaplan-Maier survival curves were generated for C57/BL6J male mice receiving 50 mg/kg apigenin 3 h before injection of 37.5 mg/kg LPS (Api+LPS), apigenin 3 h after LPS (LPS + Api), DMSO 3 h prior LPS (LPS), DMSO and PBS (DMSO) or apigenin and PBS (Api). Mice were monitored every 3 h for the first 3 days and daily for up to 7 days. n = 8, * p < 0.05. In collaboration with Dr. Horacio Cardenas.

114 Figure 4.8. Apigenin regulates miR-155 and TNF! in an IKK#/NF-"B mediated pathway. A. Immunoblots of cell lysates from RAW 264.7 macrophages stimulated with 100 ng/ml LPS for the indicated periods of time. B. RAW 264.7 cells stimulated with 100 ng/ml LPS followed by treatment with 25 µM apigenin, added 1, 3 or 6 h after LPS challenge (LPS + Api, lanes 3-5). Cells were harvested at 8 h after LPS. Cells stimulated with 100 ng/ml LPS and DMSO (added 1 h after LPS, lane 2), treated with PBS and DMSO (DMSO, lane 1) or challenged with PBS and 25 µM apigenin (Api, lane 6) were used as controls. Cell lysates were resolved by SDS-PAGE and analyzed by immunoblot. For A and B, n=1. C-D. IKK!WT or IKK!KO bone marrow derived macrophages were stimulated with 100 ng/ml LPS for 1 h prior addition of 25 µM apigenin (LPS + Api) or diluent DMSO (LPS) for additional 7 h. Cells treated with DMSO and PBS (DMSO) were used as control. C. TNF! expression was evaluated by ELISA in the cell media. D. MiR-155 expression was determined in RNA isolated from macrophages by qRT-PCR and normalized against snoRNA202 expression. Data represent mean ± SEM. n=5. * p < 0.05. One-way ANOVA.

115 Figure 4.9. Working model of the immune-regulatory activity of apigenin. Adapted from Arango et al 2015, Mol Nutr Food Res [294].

116 Chapter 5

Dietary Apigenin Delays Breast Cancer Development and Metastasis by Immune-

Modulating Macrophages and Inducing Apoptosis of Breast Cancer Cellsc

5.1 Abstract

Breast cancer is the second leading cause of cancer-related deaths among women.

Despite improvements in early detection and availability of treatments, mortality associated to breast cancer remains significantly high, prompting the need to identify alternative approaches. Tumor associated macrophages (TAMs) promote cancer progression and metastasis. Hence, targeting the oncogenic functions of macrophages

constitute an alternative approach for cancer treatment. Using a murine pre-clinical model

of breast cancer development, we showed that both intraperitoneal administrations of

apigenin and dietary interventions with a celery-based apigenin-rich food delay the onset

of tumor progression and pulmonary metastasis. Our results show that apigenin induces

apoptosis and inhibits proliferation in breast tumors. In addition, we found that apigenin

halts macrophages infiltration in the tumor microenvironment by reducing the expression

of NF-κB-dependent chemokines and promoting apoptosis in circulating macrophage

c Arango D, Pereira MSF, Duarte S, Nagarajan P, Eubank TD, Doseff AI. 2015. Dietary Apigenin Delays Breast Cancer Development and Metastasis by Immune-Modulating Macrophages and Inducing Apoptosis of Breast Cancer Cells. In preparation. 117

precursors. Apigenin disrupts the macrophage/cancer cell paracrine loop resulting in reduced NF-κB activation and CCL2 expression, a macrophage chemotactic factor.

Moreover, apigenin blocks the macrophage-induced survival of malignant cells triggering apoptosis in cancer cells. These results reveal a new anti-carcinogenic mechanism of dietary flavonoids and support the use of apigenin-rich functional foods as a potential immune-regulatory strategy for breast cancer interventions.

5.2 Introduction

Breast cancer is the second leading cause of death among women world-wide [52].

Despite improvements in early detection methods and the availability of surgical techniques and chemotherapeutic methods, lethality associated to this neoplasia remains significantly high [52], prompting the need to identify new approaches for the prevention and treatment of breast cancer.

Breast cancer is initiated by driver mutations in epithelial cells that confer

proliferative advantage and resistance to apoptosis [106, 107, 295]. Uncontrolled

proliferation precedes the appearance of hyperplasia, the earliest stage of carcinogenesis

(chapter 1, section 1.5) [53]. Yet, progression to metastatic breast carcinoma requires

infiltration of adjuvant cells to the tumor microenvironment including macrophages,

lymphocytes and fibroblasts [80, 84, 296, 297]. Macrophages constitute the most

abundant population of tumor infiltrating leukocytes and they are key regulators of cancer

progression by releasing growth and angiogenic factors that induce tumor cell

proliferation, survival and dissemination [67, 75, 78].

118

Macrophages originate from monocytes (chapter 1, Fig. 1.1), which are found in peripheral blood, spleens and bone marrow [87]. Monocyte numbers expand during carcinogenesis resulting in higher macrophage differentiation and infiltration into tumors

[76, 81, 87]. Monocytes are classified in different subpopulations including non-classical monocytes and pro-inflammatory monocytes [4, 88, 89]. In addition, immature myeloid cells, referred as myeloid derived suppressor cells (MDSC), are also a source of infiltrating macrophages (chapter 1, Fig. 1.1). MDSC constitute a heterogeneous population from two lineages, granulocytes (G-MDSC) and monocytes (Mo-MDSC) [89,

90]. Pro-inflammatory monocytes and Mo-MDSC are recruited to tissues, where they differentiate into macrophages (chapter 1, Fig. 1.1) [89, 90], by the action of chemokines such as CCL2, CXCL12 and VEGFA [74-76]. The transcription factor NF-κB increases the expression of the chemokines CCL2 and VEGFA [17, 298], and stimulates macrophage infiltration in breast tumors [99]. In addition, NF-κB promotes cancer cell and monocyte/macrophage survival [93, 101, 299].

FBV-MMTV-PyMT (Mammary Tumor Virus-Polyoma Virus Middle T antigen, referred as PyMT+) transgenic mice express the oncogene PyMT in epithelial cells of the mammary gland and constitute a pre-clinical murine model of breast cancer development.

PyMT+ female mice exhibit 100% tumor penetrance and more than 90% incidence of metastases, short latency (12 weeks old) and pregnancy-independent (Chapter 1, section

1.5 and [64, 65]). Breast cancer development in the PyMT+ mice resembles the stages of tumor progression found in human mammary cancer, including hyperplasia (H), adenoma

(A), early carcinoma (EC) and late invasive carcinomas (LC), and reproduce the

119

molecular changes observed in human breast cancers such as over-expression of Her2

and low expression of PR and ER [66]. In addition, PyMT+ tumors are highly infiltrated

by macrophages, which are fundamental for progression to metastatic carcinoma in this

model [80, 81, 295].

In this chapter, we showed that apigenin and celery-based apigenin rich diets delay breast cancer progression and metastasis by a dual anti-proliferative and immune-

regulatory effect. Dietary apigenin decreased tumor associated macrophages by inhibiting

the expression of NF-κB-dependent chemokines and promoting apoptosis in circulating

macrophage precursors. In addition, apigenin disrupted the macrophage/cancer cells

cross-talk inducing apoptosis in malignant cells and macrophages. These results uncover

the mechanisms involved in the anti-carcinogenic activity of apigenin and support the use

of foods rich in apigenin as a immune-regulatory strategy for breast cancer interventions.

5.3 Results

5.3.1 Celery-based apigenin-rich diets decrease breast tumor growth

To evaluate the effect of apigenin on tumor growth in a mouse model of breast cancer

development, 3-weeks old PyMT+ female mice were administered vehicle or 25 mg/kg apigenin intraperitoneally (i.p) daily until reaching 12-weeks of age. PyMT- mice, which do not develop cancer, were treated with vehicle and used as age-matched controls. A 3- fold increase in the weight of mammary glands was observed in PyMT+ compared to

PyMT- 12-week old mice (Fig. 5.1A and B, PyMT- vs. PyMT+, both treated with vehicle).

Apigenin reduced tumor burden by ~2-fold compared to 12-week old PyMT+ mice

120

treated with vehicle (Fig. 5.1A and B, PyMT+/apigenin vs. PyMT+/vehicle), indicating that this flavone delays breast cancer growth. To test whether a celery-based diet rich in aglycone-flavones has anti-carcinogenic properties, PyMT+ female mice were fed ad libitum with control or apigenin diets, immediately after weaning (3-weeks old), until animals reached 12 weeks of age. Apigenin-rich diet reduced tumor growth by ~2-fold compared with mice fed with control diet (Fig. 5.1C and D, Apigenin Diet vs. Control

Diet). In addition, the reduction of tumor weight in mice fed with apigenin diet was

similar to the effect observed when we used intraperitoneal administration of apigenin

(Fig. 5.1A and C, red bars). These results indicate that functional foods rich in apigenin

confer effective anti-cancer activities comparable to pure apigenin.

5.3.2 Apigenin delays breast cancer progression

Breast cancer is a progressive disease, initiated as a proliferative lesion (hyperplasia,

H), which progresses to adenoma (A), advances to early carcinoma (EC) and further

evolves into late invasive carcinoma (LC) [66]. To evaluate the effect of apigenin on

breast cancer development, 3-week old PyMT+ female mice were administered daily 25 mg/kg apigenin or vehicle by i.p until mice reached 4, 7, 9 or 12 weeks of age. PyMT+

mice treated with vehicle showed a progressive increase in tumor burden compared to

PyMT- mice (Fig. 5.2A, blue vs. grey lines), in agreement with previous studies [84]. At

12 weeks, apigenin significantly reduced mammary gland mass compared to mice administered vehicle (Fig. 5.2A, red vs. blue lines). However, no effect on breasts weight was observed in 4, 7 or 9-week old animals receiving apigenin compared to mice treated

121

with vehicle (Fig. 5.2A, red vs. blue lines). Next, mammary gland tissue sections were

stained with H&E (Fig. 5.2B), and the area of each developmental stage was measured on

digital images. To accurately score adenoma and early carcinoma, which look

morphologically similar using H&E stainings, mammary tissue sections were stained

with anti-α-SMA (α-smooth muscle actin) antibodies [66]. Lobular regions stained

positive with anti-α-SMA antibodies were scored as adenoma, whereas regions stained

negative for anti-α-SMA antibodies are referred as early carcinoma (Fig. 5.2C). Our

results showed that apigenin delayed the appearance of adenoma in 4-week old mice (Fig.

5.2D, week 4, green bars) and the area of late carcinoma in 12-week old mice (Fig. 5.2D,

week 12, black bars). The distribution of breast cancer stages found at 7 and 9 weeks was

not significantly changed by apigenin (Fig. 5.2D, weeks 7 and 9), but less prominent

areas of adenoma (green) in the control group were observed (Fig. 5.2D, weeks 7 and 9).

These results suggest that apigenin has a dual effect on breast cancer development, at

week 4 during the transition from hyperplasia to adenoma, and at week 12 by delaying

progression from early to late carcinoma.

Progressive loss of progesterone receptor (PR) expression is found in ~30% of human

mammary tumors and is also observed in the PyMT model of breast cancer [52, 66].

Hence, we evaluated PR levels in breast tumors from 4- and 12-week old PyMT+ mice, times in which apigenin delayed tumor progression. Our results show that PR levels are decreased 6-fold in 4-week old PyMT+ animals receiving vehicle compared to age- matched PyMT- mice (Fig. 5.2E, blue vs. grey bars). Apigenin treatment in 4-week old

PyMT+ mice increased PR expression to levels found in normal mammary glands (Fig.

122

5.2E, red vs. blue and grey bars). At 12 weeks, PR levels were decreased ~12-fold in breasts from vehicle-treated PyMT+ compared to PyMT- mice (Fig. 5.2F, blue vs. grey bars). We found that apigenin increased ~4-fold PR expression compared to mice treated with vehicle (Fig. 5.2F, red vs. blue bars). However, PR expression in apigenin mice remained significantly low compared to normal levels found in PyMT- mice (Fig. 5.2F,

red vs. grey bars). Altogether, these results indicate that apigenin delays breast cancer

progression.

5.3.3 Dietary apigenin decreases proliferation and induces apoptosis in breast

tumors

Malignant breast transformation is characterized by high proliferation of epithelial

cells [295]. To evaluate the effect of apigenin on tumor proliferation at early stages of

breast carcinogenesis, mammary sections from 4-week old PyMT+ female mice receiving

apigenin or vehicle by i.p. were stained with anti-Ki67 antibodies, a proliferation marker,

by IHC. We found that apigenin administration decreased by 2-fold tumor proliferation in

mammary glands compared to PyMT+ mice treated with vehicle (Fig. 5.2G).

Subsequently, we evaluated the percentage of cells stained positive with anti-Ki67

antibodies (Ki67+) in 12-week old mice. Our results showed that apigenin reduced 20%

tumor proliferation compared to mice receiving vehicle (Fig. 5.3A), suggesting that this

flavone affects tumor growth at early and late stages of breast carcinogenesis. The minor

effect on proliferation detected in apigenin-treated mice at late stages, suggested another

mechanism responsible for the reduction in tumor mass. Thus, we assessed whether

123

apigenin promotes tumor apoptosis at week-12 by the TUNEL assay. We observed that

apoptosis was increased by ~1.5-fold in tumors from apigenin-treated mice compared to

animals receiving vehicle (Fig. 5.3B). Next, to determine the effect of diets rich in

aglycone apigenin on proliferation and apoptosis of breast tumors, Ki67 staining and

TUNEL assay were performed in mammary gland sections from 12-week old PyMT+

female mice fed with control or apigenin diets. Mice consuming apigenin-diet showed

~20% decreased proliferation and 2-fold increase in apoptosis compared to mice fed the

control diet (Fig. 5.3C and D). These results reveal that apigenin, administered

intraperitoneally or as part of a celery-based diet, decrease proliferation and promote

apoptosis in breast tumors.

To gain insights on the anti-proliferative and pro-apoptotic functions of apigenin on

breast cancer cells, proliferation, apoptosis and cell cycle distribution were evaluated in a

PyMT cell line treated in vitro with increasing concentrations of apigenin or diluent

DMSO for 24 h (Fig. 5.4). The flavonoid apigenin affected proliferation of PyMT cells in

a dose-dependent manner, as determined by the MTS assay (Fig. 5.4A). Ten-micromolar

apigenin showed a non-significant 15% decrease in cell proliferation compared to 30%

and 60% reduction observed in cells treated with 25 and 50 µM, respectively (Fig. 5.4A).

Staining with Calcein A/M and propidium iodide (PI) showed a 20% and 45% of

apoptosis in cells treated with 25 and 50 µM apigenin, respectively, compared with only a

basal 8% of apoptotic cells observed in cells treated with DMSO (Fig. 5.4B-C). Caspase-

3 activity was negligible in DMSO-treated PyMT cells, but increased 2.5 fold in cells treated with 25 µM apigenin and 7-fold in the presence of 50 µM apigenin (Fig. 5.4D).

124

Next, we examined the cell cycle profile of PyMT cells treated with 25 or 50 µM

apigenin or diluent DMSO for 24 h by propidium iodide staining and flow cytometry. A

significant increased of cells arrested in G2/M was observed in the presence of 50 µM

apigenin (55%) compared to ~38% in DMSO-treated cells. This effect was accompanied

by a 2-fold reduction of cells in G1 (Fig. 5.4E-F), indicative of G2/M arrest, as

previously shown for other breast cancer cell lines [300]. Altogether, these results show

that dietary apigenin decreases proliferation and promotes apoptosis of breast cancer cells

in vitro and in vivo.

5.3.4 Dietary apigenin reduces macrophage infiltration in breast tumors

The limited effect of apigenin on proliferation and apoptosis in late stages suggests

that additional mechanisms are responsible for delaying progression to invasive

carcinoma. Macrophages populate the tumor microenvironment and stimulate the

transition to late metastatic carcinoma [81, 84, 296]. We previously showed that apigenin

immune-modulates macrophage activity during LPS-induced inflammation (Chapter 4

and [169]). Hence, we evaluated the effect of this flavone on TAMs by IHC, staining mammary gland tissue sections with anti-F4/80 antibodies, a specific marker of macrophages. We found that the presence of tumoral macrophages increased between weeks 9 and 12, as previously described [79-81, 84, 296]. We observed that intra-tumoral macrophages stayed at baseline, whereas peri-tumoral numbers increased ~15-fold in 9- weeks old mice (Fig. 5.5A, week 9 vs. week 4). At 12 weeks, both intra-tumoral and peri- tumoral macrophages were significantly increased compared to weeks 4 and 9 (Fig.

125

5.5A). Yet, their peri-tumoral numbers doubled the amount observed in intra-tumoral

regions, as previously shown [296, 301-303]. To evaluate the effect of apigenin on macrophage infiltration, we stained mammary gland sections, using anti-F4/80 antibodies, from 9- and 12-week old PyMT+ female mice receiving vehicle or 25 mg/kg

apigenin. Flavone administration had no effect on the numbers of intra- or peri-tumoral

macrophages at week 9 (Fig. 5.5B, red vs. blue bars). However, apigenin significantly

decreased by ~2-fold the number of peri-tumoral macrophages in 12-week-old mice

compared to vehicle-treated mice (Fig. 5.5C, peri-tumoral, red vs. blue bars), reaching the

level found in the intra-tumoral region (Fig. 5.5C, red bars), without affecting the

numbers of macrophages in the intra-tumoral area (Fig. 5.5C, intra-tumoral, red vs. blue

bars). The amount of peri-tumoral macrophages in apigenin-treated mice at week 12

remained ~3-times higher (~48 macrophages/mm2) compared to the ~14 macrophages/mm2 found at week 9, (Fig. 5.5C vs. Fig. 5.5B, peri-tumoral, red bars).

Notably, mice consuming the apigenin diet, showed a 2-fold decrease in peri-tumoral macrophages at week 12 of age (Fig. 5.5D, peri-tumoral, red vs. blue bars), without reducing the numbers found in the intra-tumoral region (Fig. 5.5D, intra-tumoral, red vs. blue bars). These results indicate that apigenin and a celery-based rich diet effectively decrease macrophage infiltration at late stages of breast cancer progression.

Macrophages in breast tumor microenvironments are predominantly M2-like pro- tumoral cells [79]. Thus, to determine whether apigenin decreases M2-macrophages, we evaluated the number of cells stained positive for F4/80 and CD206 (also known as mannose receptor, a specific M2 marker) in the peri-tumoral region of breast tumor

126

sections of 12-week old mice. Our results show that ~80% of all the macrophages found

in the peri-tumoral region stained positive for CD206, suggesting they are M2

macrophages (Fig. 5.6), in agreement with previous studies [79]. Apigenin decreased by

2-fold the number of M2 cells (Fig. 5.6B, green bars), but had no effect in the number of

M1 macrophages (Fig. 5.6B, F4/80+CD206-, orange bars). These results indicate that

apigenin reduces the number of M2 macrophages in the peri-tumoral region.

5.3.5 Dietary apigenin decreases the expression of macrophage chemo-attractants

Macrophages are attracted to tumors by chemokines such as CCL2, VEGFA and

CXCL12, which are produced by cells of the tumor microenvironment such as cancer and

immune cells [74-76]. To study the effect of apigenin on chemokine expression, we

evaluated the levels of CCL2, VEGFA and CXCL12 by qRT-PCR in breast tumors from

12-week old PyMT+ female mice treated with 25 mg/kg apigenin or vehicle or fed with either control or apigenin-diet. We observed that apigenin administered by i.p. reduced by 3-fold the expression levels of CCL2 and by ~5-fold VEGFA compared with mice receiving vehicle, without affecting CXCL12 expression (Fig. 5.7A). Notably, consumption of apigenin diet diminished by 2.5-fold CCL2 and by 4-fold VEGFA

expression, but had no effect on CXCL12 expression compared with mice fed with

control diet (Fig. 5.7B).

The expression of CCL2 and VEGFA is regulated by the transcription factor NF-κB.

[17, 298]. We previously showed that apigenin blocks NF-κB activity in vitro and in vivo in mouse models of acute inflammation (Chapter 4 and [169, 205]). NF-κB is

127

constitutively activated in breast cancers [93, 94]. Hence, we studied the effect of apigenin on NF-κB activation in protein lysates obtained from breast tumors of 12-week

old PyMT+ female mice by immuno-blots using anti-phospho-p65-Ser536 antibodies, a

marker of NF-κB activation. Our results showed that administration of intraperitoneal

apigenin decreases by ~2-fold NF-κB-p65 phosphorylation in breast tumor lysates

compared to mice receiving vehicle (Fig. 5.7C). Similarly, PyMT+ consuming the

apigenin-rich diet showed a ~2-fold reduction in phospho-p65-Ser536 compared to mice

fed with control diet (Fig. 5.7D). Taken together these results indicate that apigenin and a

celery-based apigenin rich diet decreases NF-κB activation and the expression of the

macrophage chemotactic factors CCL2 and VEGFA at late stages of breast

carcinogenesis.

5.3.6 Dietary apigenin induces apoptosis in blood monocytes re-establishing normal

numbers of macrophage precursors

Macrophages are recruited to tissues from monocytes precursor [304], which are

produced in the bone marrow by differentiation from hematopoietic stem cells (HSC, chapter 1, section 1.1 and Fig. 1.1) [304]. Two populations of HSCs are recognized in mouse, named KSL [kit+(also known as CD117)/Sca-1+ lineage] and KL [kit+/Sca-1-

lineage]. KL cells differentiate into granulocyte/monocyte progenitors (GMPs), which

give rise to granulocytes and monocytes [4]. To determine the effect of apigenin on

monocyte progenitors, we isolated bone marrow cells from 12-week old PyMT+ female

mice treated with vehicle or 25 mg/kg apigenin or PyMT- mice treated with vehicle as

128

age-matched controls. Isolated bone marrow cells were stained with FITC-conjugated

lineage (Lin-FITC) markers (anti-CD3, anti-B220, anti-Gr-1, anti-CD11b, anti-NK1-1

and anti-Ter119 antibodies), anti-Sca-1-PE, anti-CD117-APC and anti-CD16/32-PE-Cy7 antibodies and determined the different cell populations by flow cytometry. Cells stained negative for lineage markers and positive for Sca-1 and CD117 (Lin-CD117+Sca-1+) correspond to the KSL population (Fig. 5.8A). Cells stained negative for lineage markers and Sca-1 but positive for CD117 (Lin-CD117+Sca-1-) are KL cells (Fig. 5.8A). Next, we gated the KL population and determined the percentage of GMPs, which are stained positive for CD16/32 (Lin-CD117+Sca-1-CD16/32+, Fig. 5.8A). Our results show that the

percentages of KSL and KL populations were similar between PyMT- and PyMT+ mice administered vehicle and apigenin-treated PyMT+ animals (Fig. 5.8B). In addition, the

number of GMPs was similar in PyMT- and PyMT+ mice treated with vehicle and in the

presence of apigenin (Fig. 5.8C). Thus, apigenin has no effect on HSC and GMP

populations in bone marrow.

To determine the effect of apigenin on differentiated myeloid cells, we stained bone

marrow cells with anti-CD11b antibodies, a myeloid marker, and anti-Gr-1 antibodies, a

granulocytic/monocytic marker (Fig. 5.8D). Cells stained positive for CD11b and Gr-1

(CD11b+Gr-1+) correspond to granulocytes and cells stained positive for CD11b but

negative for Gr-1 (CD11b+Gr-1-) are monocytes. We found a ~1.5 fold increase in the

number of granulocytes in PyMT+ mice, compared to PyMT- animals treated with vehicle

(Fig. 5.8E, PyMT+/Vehicle compared to PyMT-/Vehicle), as previously reported [305].

Apigenin had no effect on the number of granulocytes and monocytes compared to

129

PyMT+ receiving vehicle in bone marrow (Fig. 5.8E, Apigenin vs. Vehicle). These results indicate that apigenin does not alter the number of myeloid cells in the bone marrow.

Spleens also constitute an important reservoir of monocytes, the macrophage precursors [87]. To determine the effect of apigenin on the number of monocyte populations in spleens, splenic leukocytes were obtained from 12-week old PyMT+ female mice treated with vehicle or 25 mg/kg apigenin or age-matched PyMT- mice treated with vehicle as controls. Next, splenic leukocytes were stained with anti-CD11b, anti-Ly6G, anti-Ly6C and anti-CCR2 antibodies and analyzed by flow cytometry (Fig.

5.9). Pro-inflammatory monocytes were identified as CD11b+Ly6G-Ly6CIntCCR2+

(referred as Ly6CIntCCR2+), while Mo-MDSC were recognized as CD11b+Ly6G-

Ly6CHiCCR2+ (referred as Ly6CHiCCR2+, Fig. 5.9). We found a slight non-significant increase in the numbers of Ly6CIntCCR2+ and Ly6CHiCCR2+ populations in PyMT+ compared to PyMT- mice treated with vehicle (Fig. 5.10A, blue vs. gray). Apigenin treatment had no effect on Ly6CIntCCR2+ and Ly6CHiCCR2+ populations compared to

PyMT+ treated with vehicle (Fig. 5.10A, blue vs. gray). In parallel, we examined the number of lymphocytes and natural killer (NK) cells, immune cells with anti-tumor activity [306]. For this purpose, splenic leukocytes were stained with anti-CD3, anti-

CD4, anti-CD8 and anti-CD49b antibodies and analyzed by flow cytometry (Fig. 5.9).

+ + Thelper lymphocytes were identified as CD3 CD4 , cytotoxic lymphocytes were recognized as CD3+CD8+, whereas NK cells are positive for CD49b+ (Fig. 5.9). We observed that splenic CD3+CD4+ and CD3+CD8+ lymphocytes were significantly decreased by ~1.5–fold in vehicle-treated PyMT+ compared PyMT- mice receiving

130

vehicle (Fig. 5.10A, blue vs. grey). In contrast, NK cells remained similar between

PyMT+ and PyMT- mice both treated with vehicle (Fig. 5.10A, blue vs. grey).

Intraperitoneal apigenin had no effect on the number of splenic lymphocytes or NKs compared to PyMT+ animals receiving vehicle (Fig. 5.10A, red vs. blue bars). Next, we evaluated the number of splenic leukocytes in 12-week old PyMT+ consuming apigenin diet compared with PyMT+ mice fed with control diet (Fig. 5.10B, red vs. blue bars). We observed that apigenin diet had no effect on the number of splenic Ly6CIntCCR2+,

Ly6CHiCCR2+, CD3+CD4+, CD3+CD8+ or CD49+ leukocytes compared to mice fed with control diet (Fig. 5.10B, red vs. blue bars).

Next, we determined the effect of apigenin on circulating monocytes, another reservoir of macrophage precursors [81, 87]. For this purpose, blood leukocytes were stained and analyzed as described for splenic leukocytes (Fig. 5.9). We found that the numbers of circulating Ly6ChiCCR2+ and Ly6CIntCCR2+ cells were increased by ~4-fold in PyMT+ compared to PyMT- mice treated with vehicle (Fig. 5.10C, blue vs. grey bars), as previously reported [307, 308]. Apigenin treatment reduced by ~3.5-fold the number of blood Ly6ChiCCR2+ and Ly6CIntCCR2+ populations compared to mice receiving vehicle (Fig. 5.10C, red vs. blue bars), to similar amounts observed in normal mice (Fig.

5.10C, red vs. grey bars). In contrast, the numbers of CD3+CD4+, CD3+CD8+ or CD49+ populations remained similar in PyMT- treated with vehicle and PyMT+ receiving vehicle or apigenin (Fig. 5.10C). Subsequently, we determined the numbers of circulating monocytes and lymphocytes in 12-week old PyMT+ mice fed with control or apigenin diets. We observed that apigenin diet decreases by ~2-fold the numbers of Ly6ChiCCR2+

131

and Ly6CIntCCR2+ populations in blood compared to mice consuming control diet (Fig.

5.10D, red vs. blue bars), without affecting lymphocytes or NK cells (Fig. 5.10D, red vs.

blue bars). These results indicate that apigenin and a celery-based rich diet decrease the

numbers of circulating macrophage progenitors, without affecting blood lymphocyte populations.

To determine whether apigenin induces apoptosis of leukocyte, we determined the percentage of cells stained positive with annexin-V (annexin-V+) in both spleens and

blood from 12-week old PyMT- treated with vehicle, PyMT+ mice receiving vehicle or 25

mg/kg apigenin by i.p or PyMT+ mice fed with control or apigenin diets. We found that

intraperitoneal apigenin has no effect on the percentage of apoptotic cells in splenic

Ly6CIntCCR2+, Ly6CHiCCR2+, CD3+CD4+, CD3+CD8+ or CD49+ populations compared

to mice receiving vehicle (Fig. 5.11A, red vs. blue bars). In addition, apigenin diet did not change the percentage of splenic leukocyte populations undergoing apoptosis compared to PyMT+ mice fed with control diet (Fig. 5.11B, red vs. blue bars). Yet, intraperitoneal administration of apigenin increased by ~2-fold the percentage of circulating

Ly6CIntCCR2+ and Ly6CHiCCR2+ cells undergoing apoptosis compared to PyMT+ or

PyMT- mice treated with vehicle (Fig. 5.11C, red vs. blue and grey bars), without

inducing cell death in lymphocytes or NK cells (Fig. 5.11C, red vs. blue and grey bars).

Apigenin diet induced ~2-fold apoptosis in blood Ly6CIntCCR2+ and Ly6CHiCCR2+ cells

compared to mice fed with control diet (Fig. 5.11D, red vs. blue bars), but had no effect

on the percentage of apoptosis in CD3+CD4+, CD3+CD8+ or CD49+ populations (Fig.

5.11D, red vs. blue bars).

132

Next, we determined whether apigenin affects the total number of blood and splenic

leukocytes. We observed that the number of circulating leukocytes per ml of blood was

significantly increased in 12-week old PyMT+ compared to PyMT- mice treated with vehicle, as determined by trypan blue exclusion (Fig. 5.12A, blue vs. grey bars).

Treatment with 25 mg/kg apigenin decreased by ~2-fold the number of blood leukocytes compared to vehicle-treated PyMT+ mice (Fig. 5.12A, red vs. blue bars), reaching levels

found in normal mice (Fig. 5.12A, red vs. grey bars). Similarly, 12-week old PyMT+ mice fed with apigenin diet showed a 2-fold reduction in circulating leukocytes compared to

PyMT+ mice fed with control diet (Fig. 5.12B, red vs. blue bars). The number of splenic

leukocytes per gram of tissue was similar in both 12-week old PyMT+ and PyMT- mice treated with vehicle, as determined by trypan blue exclusion (Fig. 5.12C, blue vs. grey bars). Mice treated with apigenin had similar number of splenic leukocytes compared to

PyMT+ mice receiving vehicle (Fig. 5.12C, red vs. blue bars). Similarly, apigenin diet

had no effect on the number of splenic leukocytes compared to PyMT+ mice fed with control diet (Fig. 5.12D, red vs. blue bars). The weight of the whole spleen was increased by 1.5-fold in PyMT+ compared to PyMT- mice treated with vehicle (Fig. 5.12E, blue vs. grey bars). However, apigenin treatment had no effect on the weight of spleens compared to PyMT+ animals receiving vehicle (Fig. 5.12E, red vs. blue bars). In contrast, spleens weight was reduced by ~20% in PyMT+ mice fed with apigenin diet compared to animals

consuming control diet (Fig. 5.12E, red vs. blue bars). All together, these results indicate

that apigenin and a celery-based apigenin rich diet induces apoptosis in circulating

monocytes, the macrophage precursors, decreasing their numbers to the amounts found in

133

normal mice.

5.2.7 Dietary apigenin blocks the cancer cell/macrophage cross-communication

decreasing chemokine expression and triggering apoptosis in cancer cells and

macrophages

The cancer cells/macrophage paracrine loop induces the release of chemokines and

growth factors that promote cancer cell survival, recruitment of macrophages and metastasis [75, 78]. To investigate the anti-carcinogenic mechanisms of apigenin, we

studied the effect of this flavone on the macrophage/cancer cell cross-communication.

Syngenic PyMT cancer cells and bone marrow derived macrophages were cultured alone

[referred as PyMT alone (1:0), or macrophages alone (0:1)]) or both cell types were

cultured at equal amounts (1:1), or macrophages at ten fold excess (1:10) for 6, 24 and 48 h in serum free media. We found that the CCL2 amounts in cell media was increased by

~15 fold at 6 h, reaching a maximum of ~30 fold when tumor cells were co-cultured with macrophages (1:1 ratio) for 24 and 48 h compared to either PyMT cells or macrophages cultured alone, as determined by ELISA (Fig. 5.13A, black vs. white and dashed bars).

No differences in the levels of CCL2 produced were observed when equal concentration of cancer cells and macrophages were co-cultured (1:1 ratio) or when ten times more macrophages than PyMT cells were used (1:10 ratio, Fig. 5.13A, grey vs. black bars). To evaluate whether apigenin affects the levels of CCL2, PyMT cells and macrophages were cultured alone or together for 1 h, in serum free media, prior to the addition of 1, 5, 10 or

25 µM apigenin or diluent DMSO for additional 6 h. We found that apigenin decreased

134

CCL2 levels in a dose-dependent manner (Fig. 5.13B). One micromolar apigenin had no

effect on the amount of CCL2 produced compared to co-cultures treated with DMSO

(Fig. 5.13B). In contrast, a significant decreased in CCL2 levels was observed in the presence of 5 µM apigenin as compared to co-cultures treated with DMSO (Fig. 5.13B).

Higher concentrations of 10 and 25 µM apigenin reduced by ~6-fold the amount of CCL2

compared to co-cultures treated with DMSO, reaching basal levels found when

macrophages were cultured alone (Fig. 5.13B).

In order to determine whether the effect of apigenin on CCL2 levels is mediated by

NF-κB, a transcription factor that controls CCL2 expression [104, 309], PyMT cells were

co-cultured with macrophages at 1:1 ratio, or each cell type was cultured independently

for 1 h prior treatment with diluent DMSO, 10 µM apigenin (Api), 10 µM Bay-11-7082

(Bay), a NF-κB inhibitor, or concomitantly treated with Bay and apigenin (Bay + Api)

for additional 6 h in serum free media. We found that 10 µM apigenin decreased by ~5-

fold CCL2 amounts compared to DMSO-treated cells (Fig. 5.13C grey vs. blue bars, 1:1).

In addition, 10 µM Bay reduced by ~5-fold CCL2 to similar levels found in co-cultures

treated with just apigenin (Fig. 5.13C grey vs. red bars, 1:1). Co-treatment with a

combination of Bay + Api had no additive effect on CCL2 expression (Fig. 5.13C, khaki

vs. gray and red bars), suggesting that apigenin inhibits CCL2 expression in an NF-κB-

mediated pathway.

To evaluate whether the effect of apigenin is specific to PyMT cells or macrophages,

we pre-labeled PyMT cells with eFluor450 and macrophages with eFluor670. Next,

eFluor450-PyMT cells were combined with eFluor670-macrophages (1:1 ratio) or each

135

cell type was cultured alone for 1 h prior treatment with diluent DMSO, 10 µM apigenin,

10 µM Bay or Bay + Api for additional 6 h in serum free media. Cells were

permeabilized and stained with anti-CCL2-PE antibodies or anti-phospho-NF-κB-

p65Ser536 antibodies followed by FITC-conjugated secondary antibodies and analyzed by flow cytometry. To determine the levels of CCL2 and phospho-NF-κBp65Ser536 in each cell population, we gated based on the staining with eFluor450 and eFluor670, and determined the percentage of cells positive for phospho-NF-κB-p65Ser536 (p-p65+) or

CCL2 (CCL2+) within each population. We observed that ~20% PyMT cells and ~10%

macrophages expressed CCL2 when cultured alone (Fig. 5.14A blue bars in 1:0 and 0:1).

However, in co-cultures, 42% of PyMT cells and 30% of macrophages express CCL2

(Fig. 5.14A, blue bars, 1:1). Apigenin decreased by ~20% the number of PyMT cells and

macrophages expressing CCL2 in co-cultures (Fig. 5.14A, red vs. blue bars, 1:1), but had no effect when cells were cultured independently (Fig. 5.14A, red vs. blue bars, 1:0 and

0:1). The Bay inhibitor reduced by 30% expression of CCL2 in both PyMT cells and macrophages in co-cultures, reaching a similar level observed with apigenin alone (Fig.

5.14A, grey vs. blue bars, 1:1). However, Bay had no effect on the percentage of cells expressing CCL2 when cultured alone (Fig. 5.14A, grey vs. blue bars, 1:0 and 0:1).

Concomitant treatment with Bay and Api had no additive effect on the expression of

CCL2 (Fig. 5.14A, khaki vs. grey and red bars).

Next, we assessed the phosphorylation status of NF-κB-p65 subunit (p-p65+) by flow cytometry. Our results show that ~28% PyMT cells and ~10% macrophages have phosphorylated NF-κB-p65 in cells cultured independently (Fig. 5.14B blue bars in 1:0

136

and 0:1). In contrast, in co-cultures, there is a significant increase to ~50% of the PyMT

cells and ~38% of the macrophages presenting phosphorylated NF-κB-p65 (Fig. 5.14B, blue bars, 1:1). Apigenin decreased by ~2-fold the number of cells stained with anti-

phospho-NF-κB-p65 antibodies in both PyMT cells and macrophages in co-cultures (Fig.

5.14B, red vs. blue bars, 1:1), but had no effect on cells cultured independently (Fig.

5.14B, red vs. blue bars, 1:0 and 0:1). In contrast, Bay inhibitor significantly reduced the percentage of phospho-NF-κB-p65 positive PyMT cells incubated alone (Fig. 5.14B, grey vs. blue bars, 1:0). Yet, the effect of Bay on PyMT cells alone was not significant different from PyMT cells treated with just apigenin (Fig. 5.14B, grey vs. red bars, 1:0).

In addition, Bay had no affect in the phosphorylation of NF-κB-p65 in macrophages cultured alone (Fig. 5.14B, grey vs. blue bars, 0:1). We observed that Bay inhibited by

~2-fold the co-cultured-induced phosphorylation of NF-κB-p65 in both PyMT cells and

macrophages to similar levels found in co-cultures treated with apigenin (Fig. 5.14B,

grey vs. red bars, 1:1). Co-treatment with Bay and Api had no additive effect on the

percentage of cells with phosphorylated NF-κB-p65 compared to cells treated with each

compound independently (Fig. 14B, khaki vs. gray and red bars). Altogether, these

results indicate that apigenin decreases the cancer cell/macrophage paracrine loop

reducing NF-κB activation and CCL2 expression in both PyMT cells and macrophages.

Macrophages promote cancer cell survival [78]. To evaluate the effect of

macrophages on PyMT cell survival, we cultured eFluor450-labeled PyMT cells in

combination with Fluor670-labeled macrophages (1:1 ratio) or each cell type

independently in serum-free medium for 24, 48 and 72 h. Cells were collected and

137

stained with annexin-V and 7-AAD, markers of apoptosis, followed by flow cytometry

analysis. Next, we gated based on eFluor450 intensity for PyMT cells and eFluor670 for

macrophages, and determined the percentage of apoptosis within each population as the

sum of cells positive for just annexin-V and positive for 7’AAD and annexin-V. Our

results showed that ~12% of PyMT cells undergo apoptosis at 24 h, reaching 30%

apoptosis at 48 and 72 h when cultured alone in serum free media (Fig. 5.15A,

1:0/PyMT). In contrast, the percentage of macrophages undergoing apoptosis remained below 10% during the 72 h time course when cells were cultured independently (Fig.

5.15A, 0:1/macrophages). We found that when cells were co-cultured, only 5% the PyMT cells undergo apoptosis at 48 and 72 h, representing a 6-fold reduction compared with the percentage of apoptotic cells found when cultured in the absence of macrophages (Fig.

5.15A, 1:1/PyMT vs. 1:1/macrophages). In contrast, the percentage of macrophages undergoing apoptosis remained below ~10% when cultured in the presence of PyMT cells (Fig. 5.15A, 1:1/macrophages). These results indicate that the presence of macrophages increases PyMT cell survival.

Next, we determined the effect of apigenin on macrophage-induced PyMT survival.

For this purpose eFluor450-PyMT cells were combined with eFluor670-macrophages (at a 1:1 ratio) or each cell type was cultured independently for 1 h prior treatment with diluent DMSO, 10 µM apigenin, 10 µM Bay or Bay + Api for additional 72 h, a time of maximum PyMT cell apoptosis, in serum-free medium and the percentage of apoptosis was determined in each population using annexin-V/7’AAD staining by flow cytometry.

Apigenin has no significant effect on the percentage of PyMT cells undergoing apoptosis

138

when they were cultured alone (Fig. 5.15B, red vs. blue bars, 1:0). However, apigenin

increased by 6-fold the percentage of apoptotic PyMT cells when co-cultured with

macrophages (Fig. 5.15B, red vs. blue bars, 1:1/PyMT). In addition, apigenin induced by

~25% apoptosis in macrophages cultured alone or in combination with PyMT cells (Fig.

5.15B, red vs. blue bars, 0:1 and 1:1/macrophages). Bay increased apoptosis in PyMT

cells reaching 60% in cells cultured alone and ~35% when co-cultured with macrophages

(Fig. 5.15B, grey vs. blue bars, 1:0 and 1:1/PyMT). In addition, the NF-κB inhibitor induced ~30% apoptosis in macrophages cultured alone or in combination with PyMT cells (Fig. 5.15B, grey vs. blue bars, 0:1 and 1:1/macrophages). Combination of Bay and apigenin had no additive effect on apoptosis compared to treatment with each compound alone (Fig. 5.15B, khaki vs. red and grey bars). Altogether, these results indicate that apigenin blocks the macrophage-induced PyMT cell survival during serum starvation, promoting apoptosis in both PyMT cells and macrophages.

5.2.8 Dietary apigenin decreases metastasis

Metastases arise in ~30% of all breast cancer patients [310]. In the PyMT+ mice model, a 90% penetrance of pulmonary metastases has been reported [64, 65]. Since, macrophages promote progression to metastasis and based on the effect of apigenin and diets rich in this flavone on macrophage biology, we investigated the incidence of metastasis in lung tissues from 12-week old PyMT+ female mice administered vehicle or

25 mg/kg apigenin or fed the control or apigenin diet by H&E staining. We found that intraperitoneal apigenin decreased by ~4-fold the number and burden of metastatic

139

nodules compared to animals treated with vehicle (Fig. 5.16A-C). Notably, consumption

of apigenin diet reduced by ~12-fold the volume and by 4-fold the number of metastatic

nodules compared to mice fed with control diet (Fig. 5.16D-F). The results obtained here

reveal the potent anti-metastatic activity of a celery-based apigenin-rich diet and support

the use of flavonoid-rich functional foods as an alternative strategy for breast cancer

interventions.

5.3 Discussion

The results presented here demonstrate that apigenin and a celery-based apigenin rich

diet have anti-carcinogenic activity by regulating both cancer and immune cells (Fig.

5.17). Dietary apigenin decreases tumor-associated macrophages by promoting apoptosis

of peripheral monocytes, key macrophage progenitors, and reducing the expression of

NF-κB-dependent chemokines, thereby inhibiting infiltration (Fig. 5.17). Our studies showed that apigenin blocks the macrophage-induced survival of malignant cells

resulting in breast cancer cell apoptosis (Fig. 5.17). Hence, the anti-proliferative and

immune-modulatory activities of apigenin work in concert to delay tumor progression

and metastasis.

Previous studies reported that apigenin decreases lung, breast, colon and prostate

tumor growth in mouse xenograft models by increasing apoptosis and reducing

proliferation of cancer cells [144, 160-163]. Apigenin, administered by gavage, delayed

tumor burden in a spontaneous model of prostate cancer [164]. In addition,

intraperitoneal administration of apigenin decreased Helicobacter pylori-induced gastric

140

cancer formation in mice [311]. Yet, whether this flavone has anti-carcinogenic activity by immune-regulating the pro-tumoral functions of immune cells had not been

investigated. We showed that apigenin, and most significant, a celery-based aglycone- apigenin rich diet, hinder spontaneous breast cancer development in part by immune- modulating macrophages.

Our results revealed a dual effect on tumor development, in the transition from hyperplasia to adenoma at early stages and during the progression to late and metastatic carcinoma at a later stage. It has been proposed that at early stages, transformed cells acquire a proliferative advantage and resistance to apoptosis [107, 295], mediated in the

PyMT model by over-activation of the PI3K/AKT1 kinases [312]. The PyMT oncogene interacts with PI3K, inducing its activation and the phosphorylation of the downstream target AKT1 [313, 314]. In addition, over-activation of PI3K/AKT1 down-regulates PR expression [315, 316], a marker of breast cancer progression, which lack of expression is a characteristics of the PyMT model [66, 315, 316]. Inactivation of PI3K signaling using pharmacological inhibitors decreased proliferation and triggered apoptosis in PyMT- transformed cell lines [317], and AKT1 knock-out in PyMT+ mice reduced by ~50%

mammary tumor formation [318]. Inhibition of PI3K and AKT by apigenin has been

previously shown in various cancer epithelial cells [158, 173, 319, 320]. In addition,

apigenin inhibits PI3K and AKT activation leading to decreased tumor burden in a mouse

model of prostate cancer development [164]. Moreover, PyMT-induced breast tumors are

also characterized by over-expression of Her2 [66], which promotes proliferation by

activating RAS and MAPKs [295, 312, 321, 322]. It was previously shown that apigenin

141

reduces Her2 expression in breast cancer cells [323, 324]. Hence, the early effect of

apigenin on tumor proliferation may be mediated by its ability to inhibit the

PI3K/AKT/Her2 pathways.

At late stages, infiltration of macrophages, which occurs between 9 and 12 weeks of

age, promote progression to late carcinoma, induction of angiogenesis and metastasis

[79-81, 84, 296]. Previous studies have shown that more than 80% of infiltrating

macrophages are polarized to an M2 pro-cancer phenotype [79]. In agreement with these

findings, we observed that the majority of macrophages in PyMT+ mammary sections are

M2. Supporting the role of macrophages in metastases, genetic knock-out or inhibition of

CCL2 signaling using anti-CCL2 antibodies decreased by ~50% tumor growth and pulmonary metastasis in the PyMT model [71, 76]. In addition, depletion of macrophages by genetic ablation or pharmacological inhibition of CSF1 in PyMT mice delayed the onset of late-stage carcinomas and blocked pulmonary metastasis but had no effect on tumor initiation and progression to hyperplasia, adenoma or early carcinoma [79, 84]. In agreement with these observations, we showed that reduction of macrophages in the tumor microenvironment by dietary apigenin conveyed a delayed in progression to late carcinoma and a decrease in pulmonary metastasis. Hence, we propose a mechanism in which apigenin impacts cancer cells proliferation at early stages and halts macrophage infiltration cutting the progression to metastatic carcinoma at later stages.

Apigenin, as well as other flavonoids such as quercetin, were found to mitigate macrophage infiltration to inflammatory sites in mouse models of LPS-induced acute lung injury, Helicobacter pylori-induced gastric cancer, and adipose tissue of obese mice

142

[170, 311, 325-327]. Moreover, epigallocatechin gallate (ECGC), the main flavonoid in green tea, reduced macrophage infiltration in allograft murine mammary tumors [328].

These studies suggest that several flavonoids have immune-regulatory activity by modulating macrophage recruitment.

Circulating pro-inflammatory monocytes (Ly6CIntCCR2+) and Mo-MDSC

(Ly6CHighCCR2+) are critical progenitors of macrophages [304], and their numbers are

significantly increased during breast cancer development in human patients and mice

[307, 308, 329]. We found that, intraperitoneal apigenin as well as diets rich in this

flavone induce apoptosis in blood monocytes reducing their populations to levels found in normal mice, without affecting the numbers of splenic or bone marrow monocytes. In addition to these findings, we observed that apigenin decreases the expression of the macrophage chemo-attractants CCL2 and VEGFA in mammary tumors. These results indicate that apigenin halts macrophage infiltration in the tumor microenvironment by inducing apoptosis in peripheral monocyte progenitors and downregulating the expression of chemokines. Yet, apigenin does not completely deplete macrophages, which can be explained by the lack of effect on the expression of other chemotactic factors such as CXCL12. We observed that apigenin decreases the macrophage peri- tumoral numbers to levels found in the intra-tumoral area. The functions of peri-tumoral vs. intra-tumoral macrophages are not well understood. However, the presence of peri- tumoral, but not intra-tumoral macrophages was associated with increased cancer cell proliferation, as measured by Ki67 expression, larger tumors and advanced pathological stages in biopsies from breast cancer patients [330]. Hence, apigenin is decreasing peri-

143

tumoral macrophages with pro-tumoral functions.

NF-κB promotes monocyte/macrophage and cancer cell survival by regulating the expression of anti-apoptotic proteins [93-95, 101, 331]. NF-κB is constitutively activated in breast cancer cells and human primary breast tumors [93, 94]. Inhibition of NF-κB, using an antagonistic peptide of the IKK complex, decreased proliferation and induced

apoptosis in human breast cancer cell lines [93, 94]. In addition, specific inhibition of

NF-κB in mammary epithelial cells by overexpression of a dominant negative version of

IκBα, decreased breast cancer development and tumor growth, and reduced by ~50%

macrophage infiltration in breast tumors [99, 100]. Thus, NF-κB is a key factor in the

regulation of tumor cell survival and infiltration of macrophages, suggesting that NF-κB

is a therapeutic target for breast cancer interventions. We previously showed that

apigenin and celery-based apigenin rich extracts inhibit the transcriptional activity of the

NF-κB complex in LPS-stimulated mouse macrophages [169, 205]. In addition,

intraperitoneally apigenin inhibits NF-κB activity in lungs from LPS-treated mice

(Chapter 4, Fig. 4.6). In this study, we showed that apigenin decreases p65

phosphorylation, an indicator of NF-κB activity, in breast tumors as well as in

PyMT:macrophage co-culture experiments. Consistently, we observed that apigenin

decreases the co-cultured induced survival of cancer cells during serum starvations

thereby promoting apoptosis in PyMT cells and macrophages, an effect that was

reproduced by a pharmacological inhibitor of NF-κB. No additive effect was observed in

co-cultures treated in combination with apigenin and Bay, suggesting that apigenin

144

blocks the PyMT:macrophage paracrine loop in an NF-κB-mediated pathway.

The NF-κB axis entails two sub-pathways, the canonical and the non-canonical

(chapter 1, section 1.2) [14]. The canonical pathway, which is comprised of the p65 and

p50 subunits is regulated by IKKβ [14], while the non-canonical cascade, consisting of

the subunits p52 and RelB is activated by IKKα [15]. We previously showed that

apigenin regulates NF-κB by decreasing the kinase activity of IKKβ, but independent on

IKKα [169]. In addition, the results presented here indicate that apigenin inhibits the

canonical NF-κB-p65 subunit. CCL2 and VEGFA expression is regulated by canonical

NF-κB [17, 298]. Hence, we propose that apigenin downregulates CCL2 and VEGFA

expression by inhibiting the NF-κB canonical pathway. In contrast, the expression of

CXCL12, another macrophage chemotactic factor [332], is regulated by several transcription factors including the non-canonical NF-κB complex, Myb, SP1, CEBPβ, p53 and STAT3 [333-337]. Thus, the inability of apigenin to affect CXCL12 expression may reflect its incapability to regulate the non-canonical NF-κB pathway or other transcription factor involved in CXCL12 expression. In addition, HIF1A also induces the expression of VEGFA in hypoxic regions [338]. Previous studies showed that apigenin inhibits the HIF1A-dependent VEGFA expression by promoting HIF1A degradation in various cancer cell lines [320, 339-342]. Thus, inhibition of HIF1A by apigenin may contribute to the reduction of VEGFA observed in breast tumors.

VEGFA is a key chemokine in metastasis [85]. Transgenic expression of VEGFA in mammary glands from macrophage-depleted PyMT mice induces pulmonary metastasis

[85]. Our studies showed that both intraperitoneal apigenin and apigenin-diet decreases

145

VEGFA expression in breast tumors and the number and burden of pulmonary metastatic nodules. In addition, apigenin, and other flavonoids such as quercetin and genistein, were shown to reduce the incidence of metastatic nodules in xenograft mouse models of melanoma, prostate, colon, ovarian and breast cancer [164, 343-347]. Clinical studies have shown that patients with multiple and larger metastases have less survival times

(~4-12 months) compared to patients with solitary metastases [348, 349]. Thus, the results obtained here are of potential clinical relevance and support the use of flavonoid- rich functional foods as a potent immune-regulatory strategy for breast cancer interventions.

146

Figure 5.1. Celery-based apigenin-rich diet decreases breast tumor growth. PyMT+ mice (3-week old) were administered 25 mg/kg apigenin or vehicle intraperitoneally (i.p.) daily until mice reached 12-weeks olds. PyMT- mice receiving vehicle were used as age- matched controls. A. Mammary gland weight of 12-week old mice. Mean ± SEM, n=8. * p < 0.05. One-way ANOVA. B. Pictures are representative of two specimens (N1, N2 for non-cancer control; V1, V2 for vehicle-treated PyMT+ and A1, A2 for apigenin-treated PyMT+) used in data shown in (A). C. Mammary gland weight of 12-week old mice fed ad libitum with apigenin-rich or control diets. Mean ± SEM, n=8. * p < 0.05. Two-tailed t-test. D. Pictures are representatives of mice fed with apigenin diet (D1, D2) or with control diet (C1, C2) used in panel (C). Mean ± SEM, n=8. * p < 0.05. Two-tailed t-test.

147 Figure 5.2. Apigenin delays breast cancer development. PyMT+ mice (3-week old) were administered 25 mg/kg apigenin or vehicle by i.p. daily until mice reached 4, 7, 9 or 12-week old. PyMT- mice receiving vehicle were used as age-matched controls. A. Mammary gland weights at 4, 7, 9 and 12 weeks of age. Mean ± SEM, n=5. * p < 0.05, Two-Way ANOVA. B. H&E staining of mammary glands. Scale bar represents 5 mm. Inserts scale bar indicates 100 µm. H, hyperplasia; A, adenoma; EC, early carcinoma; LC, late carcinoma. C. #-SMA stainings of mammary gland sections at 9 and 12 weeks of age. Scale bar: 250 µm. D. Each stage of breast tumor development was evaluated in sections used in (B) and represented as percentage of the area. Mean ± SEM, n=5. * p < 0.05, Two-Way ANOVA. E-F. PR mRNA expression was evaluated by qRT-PCR in mammary tissue of 4-week old mice (E) and 12-week old mice (F). Mean ± SEM, n=5. * p < 0.05, One-way ANOVA. G. Mammary gland sections from 4-week old mice were stained with anti-Ki67 antibodies. Scale bar: 100 µm. Mean ± SEM, n=5. * p < 0.05, two-tailed t-test. In collaboration with Dr. Priyadharsini Nagarajan. 148 Figure 5.3. Dietary apigenin reduces proliferation and increases apoptosis in breast tumors. Mammary glands from 12-week old PyMT+ mice receiving 25 mg/kg apigenin or vehicle daily by i.p. were stained with anti-Ki67 antibodies to determine the proliferation index (A) or by TUNEL to assess apoptosis (B). Proliferation index was determined as the number of Ki67+ cells divided by total number of cells in 20 random fields at 40X. TUNEL positive cells were scored in 20 random fields at 40X and represented as positive counts per field (cpf). Mammary glands were procured from 12- week old PyMT+ mice fed ad libitum with apigenin or control diets and used to estimate Ki67 index (C) or TUNEL-positive cells (D). Bar graphs represent the mean ± SEM, n= 5 to 8, * p < 0.05, two-tailed t-test. Scale bars: 100 µm.

149 Figure 5.4. Apigenin decreases proliferation and induces apoptosis in PyMT cells in vitro. PyMT cells were treated with different concentrations of apigenin or diluent DMSO for 48 h. Proliferation and apoptosis was determined as described in supplementary M&M. A. Proliferation was evaluated by the MTT assay and represented as percentage of proliferation. B. Percentage of apoptotic cells was analyzed by calcein A/M-PI staining. Graph bars represent the mean ± SEM, n=3. * p < 0.05, One-way ANOVA. C. Pictures are representative of data used in (C). Scale bar: 10 µm. Cell cycle distribution was analyzed by PI-staining followed by flow cytometry. D. Caspase-3 enzymatic activity was determined by the release of AFC. E. Histograms are representative of three independent experiments. F. Percentages of cells in different stages of cells cycle as indicated in (E). Mean ± SEM, n=3. * p < 0.05, Two-way ANOVA.

150 Figure 5.5. Apigenin reduces macrophage infiltration. A. Mammary glands were procured from PyMT+ mice receiving vehicle daily by i.p. for the indicated periods of time. The number of macrophages in the mammary gland sections was determined by staining with anti-F4/80 antibodies in the intra-tumoral and peri-tumoral regions and expressed as the number of cells stained positive for F4/80+ per area (mm2) in 20 random fields at 40X. Mean ± SEM, n=5. * p < 0.05, One-Way ANOVA. Mammary glands from 9-week old (B) or 12-week old (C) PyMT+ mice receiving vehicle or apigenin daily by i.p. D. Mammary glands were procured from 12-week old PyMT+ mice fed ad libitum with control or apigenin diets. The number of macrophages was determined as described in (A). Dotted red lines designate the border between intra-tumoral and peri-tumoral regions. Black arrows indicate intra-tumoral while red arrows designate peri-tumoral macrophages. Mean ± SEM, n=5. * p < 0.05, two-tailed t-test. Scale bars: 100 µm.

151 Figure 5.6. Apigenin decreases M2 macrophages. A. Immunofluorescence of mammary gland tissue sections of 12-week old PyMT+ mice receiving vehicle or 25 mg/kg apigenin daily by i.p. Immunofluorescence was performed using anti-F4/80 (Red) and anti-CD206 (Green) antibodies and DAPI (blue). White arrows indicate macrophages in the peri-tumoral region. “T” indicates tumor. Scale bar: 100 µm. B. Number of M2 macrophages stained positive for anti-F4/80 and anti-CD206 antibodies (F4/80+/CD206+) and M1 macrophages stained positive for F4/80 but negative for anti-CD206 antibodies (F4/80+/CD206-) were counted in the peri-tumoral region and expressed as the number of cells per area (mm2). Data represent the mean ± SEM, n=5. *p < 0.05, two-tailed t-test.

152 Figure 5.7. Apigenin reduces the expression of NF-"B-dependent chemokines in breast tumors. PyMT+ (3-week old) mice were administered 25 mg/kg apigenin or vehicle by i.p. daily (A-C) or fed with control or apigenin-diet (B-D) until mice reached 12 weeks of age. A-B. Expression of CCL2, VEGFA, and CXCL12 was determined by qRT-PCR in RNA isolated from mammary glands. Mean ± SEM, n=8. * p < 0.05, two- tailed t-test. C-D. Phosphorylation of p65Ser539 NF-"B was evaluated by immunoblot using mammary tissue protein lysates. Each lane represents tissue lysates from different mice. V1-V4 corresponds to vehicle-treated PyMT+, A1-A4 to apigenin-treated PyMT+, C1-C4 to PyMT+ mice fed with control diet and D1-D4 to PyMT+ mice fed with apigenin diet. Immunoblots were quantified by densitometry and normalized using total p65 expression. Bar graphs on the right represent the mean ± SEM, n=7. * p < 0.05 using two-tailed t-test.

153 Figure 5.8. Apigenin does not affect macrophage progenitors in bone marrow. A. Gating strategy used to analyze macrophage progenitors in the bone marrow by flow cytometry. Bone marrow cells were stained with FITC-conjugated lineage (Lin) markers (anti-CD3, anti-B220, anti-Gr-1, anti-CD11b, anti-NK1-1 and anti-Ter119 antibodies), anti-Sca-1, anti-CD117 (also known as kit) and anti-CD16/32 antibodies. Cells stained negative for lineage markers and positive for Sca-1 and CD117 (Lin-CD117+Sca-1+) correspond to the KSL population. Cells stained negative for lineage markers and Sca-1 but positive for CD117 (Lin-CD117+Sca-1-) are KL cells. Next, we gated the KL population and determined the percentage of granulocytes and macrophage progenitors (GMPs), which are stained positive for CD16/32 (Lin-CD117+Sca-1-CD16/32+) [350]. B. KL and KSL percentages in bone marrows from 12-week old PyMT+ administered 25 mg/kg apigenin or vehicle daily or PyMT- treated with vehicle. Mean ± SEM, n=4. C. Percentage of GMPs were determined in populations analyzed in (B). Mean ± SEM, n=4. D. Analysis of myeloid bone marrow cells by flow cytometry. Bone marrow cells were stained with anti-Gr-1 and anti-CD11b antibodies. Cell populations are identified as CD11b+Gr1+ (granulocytes) and CD11b+Gr1- (Monocytes). E. Percentage of different cell populations obtained from data shown in (D). Mean ± SEM, n=4.

154 Figure 5.9. Strategy to study leukocyte populations in blood and spleens. To examine monocytes, blood or splenic leukocytes were stained with anti-CD11b, anti-Ly6G, anti- Ly6C and anti-CCR2 antibodies and annexin-V, a marker of apoptosis. Cell populations were analyzed by flow cytometry. Pro-inflammatory monocytes were identified as CD11b+Ly6G-Ly6CIntCCR2+ and Mo-MDSC were recognized as CD11b+Ly6G- Ly6CHiCCR2+. To study lymphocytes, blood or splenic leukocytes were stained with anti-CD3, anti-CD4, anti-CD8 and anti-CD49b antibodies and annexin-V. Thelper lymphocytes were identified as CD3+CD4+, cytotoxic lymphocytes were recognized as CD3+CD8+, whereas NK cells are positive for CD49b+. Apoptosis was evaluated in each cell population by following the percentage of annexin-V-positive cells.

155 Figure 5.10. Apigenin decreases the numbers of blood monocytes. Splenic and blood leukocytes were obtained from 12-week old mice administered 25 mg/kg apigenin or vehicle daily by i.p. and age-matched control PyMT- mice receiving vehicle or from mice fed with control or apigenin diet. The numbers of different leukocyte populations were evaluated by staining with specific antibodies followed by flow cytometry analyses, as described in Fig. 5.9. A-B. Number of splenic leukocytes per g of spleen. C-D. Number of blood leukocytes per ml of blood. Data represent mean ± SEM, n=5. * p < 0.05. One- Way ANONA for A and C. Two-tailed t-test for B and D.

156 Figure 5.11. Apigenin induces apoptosis in peripheral monocytes. Splenic and blood leukocytes were obtained from 12-week old mice administered 25 mg/kg apigenin or vehicle daily by i.p. and age-matched control PyMT- mice receiving vehicle or from mice fed with control or apigenin diet. The percentage of apoptotic cells was determined by staining with annexin-V in the populations from Fig. 5.10. A-B. Percentage of apoptotic splenic leukocytes. C-D. Percentage of apoptotic blood leukocytes. Data represent mean ± SEM, n=5. * p < 0.05. One-Way ANONA for A and C. Two-tailed t-test for B and D.

157 Figure 5.12. Apigenin decreases blood leukocytes without affecting splenic leukocytes. PyMT+ 12-week old mice were treated with 25 mg/kg apigenin or vehicle daily by i.p. or fed with control or apigenin diets. PyMT- mice receiving vehicle were used as age-matched controls. A-B. Leukocytes were isolated from blood, counted by trypan blue exclusion and expressed as the number of leukocytes per ml of blood. C-D. Leukocytes were isolated from spleens and number of leukocytes per g of spleen determined by trypan blue exclusion. Weight of whole spleen was measured in 12-week old PyMT- mice receiving vehicle or PyMT+ mice treated with 25 mg/kg apigenin or vehicle (E) or fed with control or apigenin diets (F). Data represent the mean ± SEM, n=5. * p < 0.05. One-way ANOVA for A, C and E. Two-tailed t-test for B, D and F.

158 Figure 5.13. Apigenin blocks the PyMT:macrophage crosstalk inhibiting CCL2 expression. A-C. PyMT cancer cells and macrophages were cultured alone (referred as the ratio of PyMT to macrophages 1:0 and 0:1, respectively) or in combination (1:1 ratio) for the indicated periods of time (A); for 1 h prior addition of increasing concentration of apigenin for additional 6 h (B); or for 1 h before receiving diluent DMSO, 10 µM apigenin (Api), 10 µM BAY 11-7082 (Bay), or co-treated with 10 µM of Api and Bay (Bay + Api) for 6 h (C), in serum-free media. CCL2 expression was measured in the supernatants by ELISA. Data represent mean ± SEM, n=4. * p < 0.05, One-way ANOVA. In collaboration with Dr. Marcelo Pereira. 159 Figure 5.14. Apigenin blocks the macrophage:cancer cell paracrine loop reducing NF-"B phosphorylation and CCL2 expression in both cancer cells and macrophages. Efluor450-labeled PyMT cancer cells and efluor670-stained macrophages (M$) were cultured alone (referred as 1:0 and 0:1, respectively) or in combination (PyMT:M$, 1:1) for 1 h prior to the addition of diluent DMSO, 10 µM apigenin (Api), 10 µM BAY 11-7082 (Bay), or co-treated with Api and Bay (Bay + Api) for 6 h in serum-free media. Cells were permeabilized and stained with anti-CCL2-PE antibodies or anti-phospho-NF-"B-p65Ser536 antibodies followed by FITC-conjugated secondary antibodies and analyzed by flow cytometry. Efluor450+ cells were recognized as PyMT cells and efluor670+ were selected as macrophages. A. Percentage of cells positive for CCL2 (CCL2+) was determined by flow cytometry. B. Percentage of cells showing phosphorylation of NF-"B-p65Ser536 (p-p65+) as assessed by flow cytometry. Data represent mean ± SEM, n=5. * p < 0.05, One-way ANOVA. In collaboration with Dr. Marcelo Pereira.

160 Figure 5.15. Apigenin blocks the macrophage-induced cancer cell survival inducing apoptosis in both macrophages and PyMT cells. A-B. Efluor450-labeled PyMT cancer cells and efluor670-stained macrophages (M$) were cultured alone (1:0 and 0:1, respectively) or in combination (PyMT:M$, 1:1) for the indicated periods of time (A) or for 1 h before receiving diluent DMSO, 10 µM apigenin (Api), 10 µM BAY 11-7082 (Bay), or co-treated with 10 µM Api and Bay (Bay + Api) for additional 72 h (B), in serum free media. Cells were stained with Annexin V/7’AAD and the percentage of apoptosis for efluor450+PyMT and efluor670+macrophages cells was evaluated by flow cytometry. Data represent mean ± SEM, n=4. * p < 0.05, One-way ANOVA.

161 Figure 5.16. Dietary apigenin decreases pulmonary metastasis. Lungs from 12-week old PyMT+ mice receiving 25 mg/kg apigenin or vehicle daily by i.p. were stained with H&E. A. Pictures are representative of two specimens (V1, V2 for vehicle treated PyMT+ and A1, A2 for apigenin-treated PyMT+). B. Metastatic burden was calculated by measuring width (w) and length (l) of metastatic nodules and applying the sphere volume formula l * w2/2. C. Metastatic index was calculated by counting the metastatic nodules in the apical, cardiac and diaphragmatic lung lobes and divided by the number of lobes. Lungs from 12-week old PyMT+ mice fed ad libitum with control or apigenin diet were stained with H&E. D. Pictures are representative of two specimens (C1, C2 for PyMT+ fed with control diet and D1, D2 for apigenin-fed PyMT+ mice). E. Metastatic tumor burden. F. Metastasis index. Data represent mean ± SEM, n=7. * p < 0.05, two-tailed t- test. Scale bars: 100 µm. 162 Figure 5.17. Model of the immune-regulatory activity of apigenin in breast cancer. Apigenin inhibits NF-"B activation in breast cancer cells and macrophages and decreases the expression of NF-"B-dependent chemokines such as CCL2 and VEGFA. This effect, along with the induction of apoptosis in peripheral macrophage precursors (monocytes), results in reduced macrophage infiltration into the tumor microenvironment. Moreover, apigenin blocks the macrophage-induced survival of cancer cells. Hence, the immune- regulatory activity of apigenin delays tumor progression and metastasis.

163 Chapter 6

Molecular Basis for the Action of a Dietary Flavonoid Revealed by the Comprehensive Identification of Apigenin Human Targetsd

6.1 Abstract

Cellular targets for dietary phytochemicals remain largely unknown, posing significant challenges for the understanding of how nutraceuticals provide health value.

Here, we describe the comprehensive identification of apigenin-protein interactions using an innovative high-throughput approach that combines phage display with second- generation sequencing (PD-Seq). The 160 identified high-confidence candidate apigenin targets are significantly enriched in three main functional categories: GTPase activation, membrane transport and mRNA metabolism/alternative splicing. This last category includes splicing factors such as the Heterogeneous Nuclear RiboNucleoProtein A2/B1

(hnRNPA2/B1) and musashi-2 (MSI2). Using hnRNPA2 as a bona fide apigenin target, we determined the affinity of the hnRNPA2-apigenin interactions and found that apigenin binds to the C-terminal Glycine Rich Domain (GRD) of hnRNPA2, preventing hnRNPA2 homodimer formation, and perturbing the alternative splicing of several

d Arango D, Morohashi K, Yilmaz A, Kuramochi K, Parihar A, Brahimaj B, Grotewold E, Doseff AI. 2013. Molecular bases for the action of a dietary flavonoid revealed by the comprehensive identification of apigenin human targets. Proc Natl Acad Sci 110:E2153-E2162.

164

human hnRNPA2 substrates. Moreover, apigenin modulates splicing of a large number of

genes involved in cell death and survival and changes dysregulated breast cancer splice

variants to normal mRNA signatures, providing a novel mechanism on how this flavone

regulates apoptotic cell fate through modulation of splicing. Hence, in contrast to

pharmacological drugs designed for specific molecules, dietary phytochemicals affect

multiple cellular proteins, which in turn, cascade into quantitative effects on many more

downstream genes that combinatorially, result in the recognized health benefits of apigenin.

6.2 Introduction

Cellular targets of apigenin remain unknown, inflicting significant challenges for the understanding on how this flavonoid provides health benefits. A few direct targets have been described, including CK2 (casein kinase 2), MUC1 (mucin 1) and tubulin [174, 177,

178]. However, it is unclear whether this dietary compound exerts its beneficial effects

either by significantly affecting the activity of just a few molecules or through additive

gains from modest effects on a large number of cellular proteins. Apigenin modulates a

broad range of signaling cascades including the DNA-damage response and apoptotic

pathways, and the MAPK, AKT/PI3K/mTOR and NF-κB cascades (Chapters 3-5 and

[139, 143, 169, 173]), demonstrating that apigenin offers health benefits by affecting

multiple cellular functions. However, the precise mechanisms on how apigenin regulates

the signals of these pathways remain unknown. The pleotropic activities of apigenin, the

indirect regulation of signaling pathways and the lack of knowledge on its targets

165

prompted us to comprehensively identify the proteins that directly interact with this flavone.

In this chapter, I describe the identification of 160 human cellular targets for apigenin, using a newly developed method that combines phage display with second generation sequencing (PD-Seq), and permits the high-throughput discovery of small molecule-protein interactions. The identified protein targets are significantly enriched in three main functional categories corresponding to GTPase activation, membrane transport and mRNA metabolism/alternative splicing. Pull-down assays validated the binding of apigenin to proteins involved in mRNA metabolism including hnRNPA2/B1 (referred hereafter as hnRNPA2) and MSI2, important factors in the progression of tumorigenesis by the regulation of splicing, mRNA stability, and mRNA transport [122, 351]. As a consequence of these interactions, apigenin modulates splicing of a large number of downstream genes involved in cell death and survival. Our results provide a comprehensive example of how a dietary flavonoid interacts with multiple targets, and how these associations cascade into quantitative effects on splicing for many more genes, helping explain the broad effect of dietary phytochemicals in cellular homeostasis.

6.3 Results

6.3.1 PD-Seq identifies candidate cellular targets for apigenin

To comprehensively identify the direct cellular targets of apigenin, we coupled apigenin to PEGA beads after activation with 4-nitrophenyl bromoacetate (see section

2.19 and Fig. 6.1A). This method resulted in apigenin being coupled to the beads through

166

either one of the two –OH groups in the A-ring, or through the single –OH group in the

C-ring (Fig. 6.1A and Fig. 1.6B, apigenin), exposing different faces of this flavone to

proteins. PEGA beads were linked to an acetyl group as control (Fig. 6.1A). The

apigenin-loaded beads (referred hereafter as A-beads), or acetyl-loaded control beads (C-

beads), were used to screen, in parallel, a commercially available human breast cancer phage-display cDNA library (Fig. 6.1B). Three rounds of selection with C- and A-beads,

referred as bio-pannings, were performed and phage DNA from each of the fractions in

two rounds of bio-panning was collected and named Or-lib (original library), Input1, C-

E1 (control beads elution 1), A-E1 (apigenin beads elution 1), Input2, C-E2 and A-E2

(Fig. 6.1C). Indexed libraries for sequencing by Illumina® GAII were generated by

amplifying inserts with primers in the phage arms (Table 2.4, T7 insert up and T7 insert

down) and attaching Illumina® indexed adapters to allow multiplexed sequencing (See section 2.21 and Fig. 6.1D).

After obtaining a total of ~46 million indexed 35-base paired-end reads corresponding to the seven different libraries obtained (Or-lib, Input1, C-E1, A-E1,

Input2, C-E2 and A-E2; Fig. 6.1C and Table. 6.1), we eliminated reads corresponding to the phage multicloning site (MCS site) or a small identical fragment harboring what we have named the “MKET clone” (Fig. 6.2A). The MKET clone corresponded to 17 aminoacids (Fig. 6.2A), highly enriched in all seven fractions (Table 6.2), and aligned to the backbone of several commercial synthetic vectors (Fig. 6.2B), yet possibly an artifact

and contaminant during the phage display library preparation. From ~8.3 million filtered

reads obtained from the seven different fractions, 64% sequences aligned to open reading

167

frames in the human genome (Table 6.1, Filtered vs. Aligned reads). Next, to discover

putative apigenin targets, the following criteria to the identified sequences were applied:

a) inserts had to be in frame with the phage capsid (Table 6.1, in-frame reads), b) reads

had to match one or multiple protein-coding sequences in the human genome. In those

cases when reads aligned to multiple coding sequences, weighted counts were used to

obtain nICPGs (normalized In-frame-aligned Counts Per Gene-model; see section 2.22

and Fig. 6.3). c) sequences had to be significantly enriched in A-beads compared to C- beads after hierarchical clustering (see section 2.22 and Fig. 6.4A and B). Genes meeting

these three criteria were enriched in cluster I and were selected as apigenin candidate

targets (Fig. 6.4C). From a minimum of 15,568 genes represented in the original phage

display cDNA library, as established from combining sequence information from the

original and selected fractions, 160 genes were considered the candidate apigenin targets

(Table 6.3).

Simultaneously to PD-Seq, we characterized clones recovered after the third round of

bio-panning by the traditional phage display approach, by performing bacterial infections

and individual plaque analysis (Fig. 6.5A and B). This approach yielded a very small

number of clones, as previously reported [352, 353]. Out of six clones analyzed, three

corresponded to the MKET fragment (Fig 6.5C). The three remaining clones

corresponded to two different fragments of the C-terminal glycin-rich domain (GRD) of

the Heterogeneous Nuclear RiboNucleoProtein A2 (hnRNPA2263-341, Fig. 6.5C), also

identified as the top candidate by PD-Seq (Table 6.3). HnRNPA2, highly overexpressed

in several cancers [122-125], play fundamental roles in the progression of tumorigenesis

168

by regulating alternative splicing, mRNA stability, and mRNA transport [354].

6.3.2 Apigenin targets are enriched in three main categories

A comparison of the aminoacidic sequence of the 160 putative apigenin targets using

Pratt (Pattern Matching) and ScanProsite [355, 356], failed to reveal any obvious common protein domains or stretches of conserved amino acids. Analyses based on known functional annotations and gene ontology (GO) showed three main categories significantly (p < 0.01) over-represented among the 160 apigenin proteins including

GTPase activation, membrane transport and mRNA metabolism/alternative splicing identified in cluster I (Fig. 6.6A). In contrast, proteins in clusters II and IV, that portraits non-significant enrichments in the AE1 and AE2 fractions, respectively (Fig. 6.6B and

D), and molecules in cluster III, that includes enrichments in control beads (Fig. 6.6C), did not show over-represented biological pathways, providing mechanistic specificity for the 160 candidate apigenin targets. The “GTPase activation” functional category contains genes such as ARHGEF1 (Rho Guanine Nucleotide Exchange Factor1) involved in the activation of Rho GTPases [357], a family of proteins regulating cell polarity and cell migration, processes modulated by apigenin [319, 344]. The “membrane transport” category includes sodium, zinc, and calcium ion transporters, as well as mitochondrial membrane transport proteins, consistent with the mitochondrial envelope being a major apigenin accumulation site in sensitive cells [358]. The third category corresponds to factors involved in “alternative splicing and mRNA metabolism”, including hnRNPA2

(Table 6.3). Together, these three GO functional classes comprise 121 of the 160

169

candidate apigenin targets identified (Table 6.3).

6.3.3 Validation of putative apigenin targets

The identification of just hnRNPA2 by conventional phage display screening

conceded us the actual phage clone containing 78 aminoacids in the C-terminal region of hnRNPA2 (hnRNPA2263-341). As a first approach to validate apigenin candidate targets,

we incubated the hnRNPA2263-341-phage (φ-hnRNPA2263-341) with A-beads in the presence of 20 µM free apigenin or diluent DMSO as control, followed by bacterial infections and counting of colony forming units (cfu). We observed that free apigenin significantly decreased 2-fold the association of φ-hnRNPA2263-341 to A-beads compared to DMSO (Fig. 6.7A). To show specificity of hNRNPA2, φ-hnRNPA2263-341 was incubated with A-beads in the presence of 20 µM free naringenin, a flavanone structurally related to apigenin (Fig. 1.6), that lacks anti-proliferative and anti- inflammatory activity (Fig. 4.3 and [139]). Naringenin failed to compete for φ- hnRNPA2263-341 (Fig. 6.7A), demonstrating that apigenin specifically interacts with φ- hnRNPA2263-341.

As a second strategy of validation, we evaluated the interaction of apigenin with the

identified candidate targets in cellular lysates. For this purpose, HeLA cells were

transiently transfected with hnRNPA2-GFP followed by pull down of cell lysates with

either A- or C-beads, and immunoblotted with anti-GFP antibodies. We observed that

hnRNPA2-GFP specifically binds the A-, but not the C-beads (Fig. 6.7B, line 2 compared

to line 3). Showing specificity for hnRNPA2, GFP failed to interact with A-beads (Fig.

170

6.7B, line 1). Similar results were obtained for ARHGEF1-GFP and GFP-BAG1 (BCL2-

Associated Athanogene), an anti-apoptotic protein that enhances the activity of BCL2

(Fig. 6.7C and D, line 2) [359]. Next, we evaluated the association of apigenin with the

RNA-binding protein MSI2 [351], and Hsp70 (Heat shock protein 70), a chaperone-like

protein with anti-apoptotic function [360]. Cell lysates from the human breast cancer cell

line MDA-MB-231 were pull down with A- or C-beads and immunoblotted with anti-

MSI2 and anti-Hsp70 antibodies, respectively. We observed that MSI2 and Hsp70 bind to

A-beads but not C-beads (Fig. 6.7E and F).

The third validation approach consisted of testing the putative apigenin targets by

assessing their enzymatic activity. IDH3G (isocitrate dehydrogenase 3 gamma) is a

subunit of the human isocitrate dehydrogenases which catalyzes the NAD+-dependent oxidative decarboxylation of isocitrate to α-ketoglutarate [361], whereas UGDH (UDP- glucose-6-dehydrogenase) catalyzes the NAD+-dependent oxidation of UDP-glucose to

UDP-glucoronate [362]. To test the effect of apigenin on the activity of IDH3, enzymatic assays were conducted in isolated mitochondria from MDA-MB-231 human breast cancer cells treated with 50 µM apigenin, 50 µM naringenin or diluent DMSO used as

control. IDH3 activity was inhibited by more than 60% by apigenin compared to DMSO,

while naringenin had no effect (Fig. 6.7G). In addition, 50 µM apigenin, but not 50 µM naringenin, significantly decreased by 30% the UGDH enzymatic activity compared to diluent DMSO in breast cancer cell extracts (Fig. 6.7H). All together, these results demonstrate that PD-seq is a reliable new method to identify direct cellular targets of small compounds.

171

6.3.4 Apigenin binds to the glycin rich domain of hnRNPA2

The identification of hnRNPA2 as an apigenin target by both PD-Seq and

conventional phage-display and the importance of this protein in cancer, make the

apigenin-hnRNPA2 interaction an ideal candidate for further characterization. To

determine which region(s) in hnRNPA2 binds to apigenin, we generated clones harboring

fragments that represent the full-length protein (hnRNPA21-341), the last 78 amino acids

identified by phage display (Fig, 6.5C, hnRNPA2264-341), a peptide lacking these last 78 amino acids (hnRNPA21-263), the region missing the glycin-rich domain (hnRNPA21-189)

and the GRD (hnRNPA2190-341) fused to GST (Fig. 6.8). Each of these clones was expressed in bacteria and the purified proteins were tested for their ability to associate with A- or C-beads in pull-down experiments, which were analyzed by Western blot using antibodies against GST. We observed that the full-length hnRNPA21-341 (Fig. 6.8A, lane 2), the GRD (hnRNPA2190-341, Fig. 6.8B, lane 2) and the last 78 aminoacids

(hnRNPA2263-341, Fig. 6.8C, lane 2) interacted with A-beads. However, the hnRNPA21-263

protein, which lacks the last 78 amino acids, also bound A-beads (Fig. 6.8D, lane 2),

suggesting that the binding of hnRNPA2 to apigenin likely involves multiple sites.

Demonstrating that the interaction with apigenin requires the GRD, the deletion of this

domain (hnRNPA21-189) abolished A-bead binding (Fig. 6.8E, lane 2). The lack of

association of purified GST to A-beads validates its specificity for hnRNPA2 (Fig. 6.8F, lane 2). These results confirm the GRD as the hnRNPA2 domain with flavone-binding capacity.

The GRD domain, indicated in yellow in Fig. 6.9A, contains six YGGG amino acid

172

repeats (Fig. 6.9A, red letters), which are absent in the RNA-binding domain (RBD, Fig.

6.9A, indicated in blue). To evaluate whether the YGGG repeats are involved in the

recognition of apigenin by hnRNPA2, we generated versions of hnRNPA2 containing

two or just one YGGG, hnRNPA21-255 (i and ii) and hnRNPA21-248 (i, Fig. 6.9B), fused to

GST, and used these purified proteins in pull-down assays to evaluate the binding to C-

or A-beads. We found that hnRNPA21-255 associates with A-beads (Fig. 6.9C, lane 2).

However, hnRNPA21-248 failed to interact with the A-beads (Fig. 6.9D, lane 2). The difference between hnRNPA21-248 and hnRNPA21-255 proteins is the peptide GYGGGRG

(repeat ii) (Fig. 6.9B). The sequence corresponding to the repeat “i”, GYGGGP, present in hnRNPA21-248, lacks the arginine (R) in the last position and has a proline (P) instead.

Together, these results suggest that an R after the YGGG repeat might be key for the

recognition of apigenin or that the binding requires two YGGG repeats.

Arginine, and other amino acids harboring also an NH2 group in the side chain, such

as asparagine (N), can form hydrogen bonds with the -OH groups of apigenin, as

previously shown by docking analysis, where R and N in the docking site of tubulin

formed hydrogen bonds with apigenin [178]. This is consistent with the association of

apigenin to the last 78 amino acids of the GRD, a fragment that contains two YGGGNs

(Fig. 6.8C and Fig. 6.9A, repeats v and vi). To test the contribution of the amino acids

tyrosine (Y), R and N in the binding with apigenin, we generated different mutant

proteins with specific amino acid substitutions in the region of hnRNPA2 that associates

with apigenin (Fig 6.10A, green letters). We observed that clones in which Y at position

250 or R at residue 254 were substituted by glycine (G, hnRNPA21-255/Y250G and

173

hnRNPA21-255/R254G, respectively), failed to bind to the A-beads (Fig 6.10B and C, lane

2). However, the substitution in which R at position 254 was replaced by N, hnRNPA21-

255/R254N, allows hnRNPA2 interaction with A-beads (Fig 6.10D, lane 2), suggesting that

the residues Y, R or N in an YGGG(R/N) peptide are necessary for the association with

apigenin. To test whether YGGGR is sufficient for the hnRNPA2-apigenin interaction,

we substituted the proline P239 for an R, to obtain hnRNPA21-248/P239R, which harbors just

one YGGGR. This clone failed to interact A-beads (Fig. 6.10E, lane 2), suggesting that

just one YGGGR peptide is not sufficient for apigenin binding. Demonstrating the

importance of the YGGG(R/N) repeats on the interaction of hnRNPA2 with apigenin,

hnRNPA1, a closely related protein that only contains one YGGG followed by a G in the

last position (Fig. 6.10F, red letters), failed to interact with apigenin (Fig. 6.10G, lane 2).

These results suggest that the presence of two YGGGs, in which one has an N or an R in

the last position, may be necessary for the hnRNPA2-apigenin interaction. Further

experiments are currently ongoing to test additional structural constraints.

6.3.5 Structural relationships of specific and high affinity interaction of apigenin

and other flavonoids with hnRNPA2

To determine the affinity and specificity of the interaction between flavonoids and

hnRNPA2, two strategies were pursued. First, we took advantage of the ability of

apigenin and other flavones to absorb light at 310 and 370 nm (Fig. 6.11A). We

incubated 10 µM apigenin with increasing concentrations of purified full length GST-

hnRNPA2 or GST for 15 min at 37°C and measured the changes in apigenin absorbance

174

in a spectrophotometer. Absorbance of apigenin increases when incubated with GST-

hnRNPA2, but not with GST alone (Fig. 6.11A). Similar changes in absorption of

flavonoids have been reported as a consequence of their association with proteins [204,

363, 364]. The increased absorption at 370 nm was used to determine the dissociation

constant (KD, see section 2.26) of GST-hnRNPA2 for apigenin, which was estimated at

2.66 ± 1.09 µM (Fig. 6.11B and C).

The second strategy consisted on developing a genetically-encoded flavonoid nanosensor based on fluorescence resonance energy transfer (FRET). Towards this objective, we cloned hnRNPA2264-341 in a collection of vectors for expression as

translational fusions between fluorescent proteins with distinct excitation/emission

spectra and different linker lengths between the fluorescent proteins (Fig. 6.12A and Fig.

6.13B). Out of eight constructs tested, FRET, defined as the presence of two fluorescent

peaks, was observed in five of them (Fig. 6.12B and Fig. 6.13B, fluorescent peaks at 480

and 530 nm), for instance the constructs containing cyan fluorescent protein (CFP) and

yellow fluorescent protein (YFP). To test the ability of apigenin to affect FRET (Fig.

613A), the five CFP-hnRNPA2264-341-YFP constructs were incubated with increasing

concentrations of apigenin or diluent DMSO for 3 h at 37°C. Apigenin decreased

fluorescence at 480 and 530 nm (Fig. 6.12B and Fig. 6.13B), but in only one case,

pFLIP2-3-hnRNPA2264-341 (pFluorescent Indicator Protein2-hnRNPA2264-341, Fig. 6.13B), was the YFP/CFP ratio increased by apigenin in a concentration-dependent manner (Fig.

6.13C compared to Fig. 6.12C) indicating that apigenin increases energy transfer between

CFP and YFP upon binding to hnRNPA2264-341. This nanosensor was therefore used to

175

estimate the apparent KD of apigenin for this fragment, which was determined to be 22.99

± 7.70 µM (Fig. 6.13D).

The absorption spectra of apigenin partially overlap with CFP but not YFP (Fig.

6.14A). As a result, apigenin quenches CFP, with no effect in YFP fluorescence (6.14B and C). In contrast, the absorption spectrum of naringenin does not overlap CFP or YFP spectra (Fig. 6.14A), and therefore had no effect on CFP or YFP fluorescence (6.14B and

C). Hence, the absence of an isosbestic point in the nanosensor spectra is likely a consequence of flavonoids absorbing light at a wavelength that partially overlaps with the

CFP excitation spectrum (Fig. 6.14).

One advantage of the FRET-based method over the spectrophotometric approach is that it permits us to investigate the specificity and relative affinity of the interaction between hnRNPA2 and other flavonoids. We tested the effect of different flavonoids on the energy transfer of pFLIP2-3-hnRNPA2264-341. The flavone luteolin differs from

apigenin by an additional –OH in ring B, and also binds with hnRNPA2 although with

lower affinity (Table 6.4, Fig. 6.15A). However, a methyl group in the additional B-ring

–OH, as present in chrysoeriol, a common flavone in many medicinal plants, results in the complete inhibition of the binding (Table 6.4, Fig. 6.15A). Apigenin is often found in the diet as C- or O-glucosides [365]. Revealing the important role of the 7-O group in the association with hnRNPA2, apigenin 7-O-glucoside shows no binding to the FRET nanosensor (Table 6.4, Fig. 6.15B). In contrast, apigenin 6-C-glucoside (isovitexin)

significantly binds to hnRNPA2 with lower affinity than aglycone apigenin (Table 6.4,

Fig. 6.15B). Neither naringenin nor the related flavanone eriodictyol affected

176

FRET of the nanosensor (Table 6.4, Fig. 6.15D). Suggesting that the binding to

hnRNPA2 is not limited to flavones, and consistent with many biological activities of

flavones being shared by flavonols [135], quercetin and kaempferol also show a

significant association with hnRNPA2, yet with lower affinity than apigenin (Table 6.4,

Fig. 6.15E). No binding of hnRNPA2 to flavopiridol or to the isoflavone genistein was

detected, indicating that these biologically active compounds function through

mechanisms distinct from apigenin (Table 6.4, Fig. 6.15C and F).

6.3.6 Apigenin inhibits hnRNPA2 oligomerization

HnRNPA2 forms oligomers through its GRD region, required for the recognition of

mRNA by the RBD domain [366-368]. To determine if the binding of apigenin to the

GRD region affects hnRNPA2 dimerization, the Amplified Luminescent Proximity

Homogenous Assay (ALPHA) was used (Fig. 6.16A, see section 2.28). Recombinant

purified 6xHis-hnRNPA2 was incubated with GST-hnRNPA2 or GST alone for 1 h

before adding glutathione (GSH)-linked donor beads and anti-His acceptor beads for

additional 6 h (Fig. 6.16A). Demonstrating the ability of hnRNPA2 to form dimers,

6xHis-hnRNPA2 incubated with GST-hnRNPA2 showed ~20 RLUs (relative light units,

Fig. 6.16B, blue bar), compared to only ~4 RLUs observed when 6xHis-hnRNPA2 was

incubated with GST alone (Fig. 6.16B, grey bar) or the proteins were denatured by heat

(Fig. 6.16B, black bar). The addition of 100 µM apigenin to the GST-hnRNPA2/6xHis- hnRNPA2 heterodimers resulted in the reduction by ~2-fold RLUs compared to DMSO

(Fig. 6.16B, red vs. blue bars), whereas no decrease in RLU was observed in the presence

177

of 100 µM naringenin (Fig. 6.16B, blue vs. magenta bars). These results demonstrate that, in vitro, binding of apigenin to the GRD region inhibits hnRNPA2 dimerization.

6.3.7 Apigenin modulates splicing in breast cancer cells

Over-expression of hnRNPA2 has been reported in several human cancers, including breast [122]. To determine whether apigenin affects hnRNPA2 expression in cancer cells, we treated human triple negative breast cancer (TNBC) cells, MDA-MB-231, or the immortalized non-carcinogenic breast epithelial cell line MCF-10A with 50 µM apigenin, or diluent DMSO for 48 h. We observed that MDA-MB-231 cells express high levels of hnRNPA2 mRNA and protein (Fig. 6.17A and B), whereas MCF-10A cells have low expression of hnRNPA2 (Fig. 6.17A and B). In contrast, hnRNPA1 protein levels, a non- apigenin target that shares 25% of hnRNPA2 substrates [369], were similar in both

MDA-MB-231 and MCF-10A cells (Fig. 6.17A). Apigenin treatment had no effect on hnRNPA2 or hnRNPA1 levels in MDA-MB-231 or MCF10-A cells (Fig. 6.17A and B).

To determine the biological consequences of the association of apigenin with hnRNPA2, we evaluated the splicing of the known hnRNPA2 substrates caspase-9 and cFLIP (also known as CFLAR) [125]. The splicing isoform caspase-9a encodes a functional apoptotic protein caspase-9, which is responsible for inducing cell death. In contrast, the splice isoform caspase-9b, lacking exons 3–6, encodes a caspase-9 protein that exhibits a dominant-negative activity and inhibits apoptosis [127]. cFLIP has two splice isoforms, long (cFLIPL) and short (cFLIPS). An alternate last exon (exon 7 instead of 14), encodes cFLIPS that prevents the activation of specific death receptors [128]. To assess the effect

178

of apigenin on the levels of cFLIP and caspase-9 isoforms, MDA-MB-231 and MCF-10A

cells were treated with 50 µM apigenin, 50 µM luteolin, 50 µM naringenin or diluent

DMSO for 48 h. RNA was isolated and splicing determined by RT-PCR using exon

specific primers (Table 2.3). The different isoforms were quantified by densitometry and

expressed as percent-splice-index (PSI, Ψ = density splicing isoform X / ∑density of all

splicing isoforms). In MDA-MB-231, but not in MCF-10A cells, apigenin reduced Ψ of cFLIPS (Fig. 6.18A and D) and caspase-9b (Fig. 6.18B and E) without affecting the splicing of BIRC5 (Fig. 6.18C and F), an alternatively-spliced non-hnRNPA2 substrate

[369]. In agreement with its ability to also bind hnRNPA2 (Table 6.4 and Fig. 6.15A), the

flavone luteolin had a similar effect on splicing (Fig. 6.18). In contrast, naringenin, a non-

hnRNPA2-interacting flavonoid, showed no effect on splicing (Fig. 6.18). Taken together, these results demonstrate that apigenin, and likely other flavones such as luteolin, interact with hnRNPA2, inhibiting its dimerization and altering the alternative splicing patterns of hnRNPA2 substrates.

6.3.8 Apigenin alters alternative mRNA processing genome-wide

Based on the findings that apigenin regulates splicing, we sought to study its effect on

mRNA processing genome-wide. For this purpose, MDA-MB-231 cells were treated with

50 µM apigenin or diluent DMSO for 48 h, a time and concentration shown to affect

splicing (Fig. 6.18). mRNA was isolated and used to generate bar-coded cDNA Illumina

libraries from two independent biological replicates and sequenced by Illumina. We

obtained ~35 million 50-base single-end reads per library (~148 million total reads, Table

179

6.5). Next, the reads were aligned to the human genome assembly 19 (hg19), using

TopHat [188], obtaining ~93% alignment (Table 6.5). To analyze alternative mRNA

processing, two approaches were applied using the software Mixture of Isoforms

(MISO), the “exon-centric” and the “isoform-centric” analyses (chapter 2, section 2.17.2

and [190]). The “exon-centric analysis” estimates the expression of exons and computes differences in the major classes of mRNA processing events (also refereed as splicing events [120]) including skipped exon (SE), retained intron (RI), mutually exclusive exons

(MXE), alternative 5’ splice site (A5SS), alternative 3’ splice site (A3SS), alternative first exon (AFE), alternative last exon (ALE) and tandem 3’UTR (TUTR, chapter 1, section 1.5.3.1 and Fig. 1.5). Since the exon-centric method looks at individual splicing events, it does not detect the complexity of mRNA processing within one transcript- isoform. To overcome this limitation, we also used the “isoform-centric” analysis, which estimates the expression of the transcript-isoforms. However, the isoform-centric does not specify the types of mRNA processing events occurring within each isoform and is usually limited by incomplete annotation of the mRNA isoforms. Hence, the two approaches are complementary to analyze mRNA processing genome-wide [190]. To determine the inclusion/exclusion ratio of an exon/isoform, we determined the percent- spliced-index or PSI [Ψ = #_inclusion_reads / (#_inclusion_reads + #_exclusion_reads)]

[120, 190]. The differences in splicing between apigenin and DMSO-treated MDA-MB-

231 cells was expressed as the delta PSI between the two treatments (ΔΨ = Ψapigenin -

ΨDMSO) and statistical significance was achieved when Bayes Factor (BF) was higher

than 20 (BF > 20, chapter 2, section 2.17.2 and [120, 190]).

180

We observed 80,053 mRNA processing events in MDA-MB-231 cells treated with

DMSO and 82,267 in cells treated with apigenin, using the “exon-centric” approach.

From those, 1,437 mRNA processing events, corresponding to 801 genes, were significantly altered in apigenin-treated cells when compared with cells treated with

DMSO. In addition, we found 1,480 transcript-isoforms, corresponding to 1,142 genes, significantly changed in cells receiving apigenin compared to DMSO-treated cells, using the “isoform-centric” approach. Together, the isoform-centric and exon-centric analyses identified 1,600 genes differently spliced by apigenin in breast cancer MDA-MB-231 cells (Fig. 6.19A, the sum of all genes in the Venn’s diagram). Yet, only 343 genes overlapped between the two methods (Fig. 6.19A, yellow), demonstrating the need of both approaches to analyze splicing changes genome-wide.

To validate the computational analysis pipeline used to analyze mRNA processing genome-wide, we used RT-PCR with exon specific primers that recognize specifically different splicing isoforms, followed by separation of DNA fragments using agarose electrophoresis, as shown in Fig. 6.18. From the 1,600 genes affected by apigenin, 45 genes were randomly selected and studied by RT-PCR. Of those, 40 showed changes on alternative mRNA processing, as predicted by our computational analyses, demonstrating a ~90% validation of the computational pipeline. Comparisons of the delta PSI values

(⎟ΔΨ⎟=⎟Ψapigenin - Ψcontrol)⎟, the difference between apigenin and DMSO-treated MDA-

MB-231 cells) obtained by RNA-seq and RT-PCR showed high correlation corresponding to R2 = 0.89 (Fig. 6.19B), demonstrating high reliability of the genome- wide computational analysis.

181

An advantage of the exon-centric analysis is that allows studying the specific types of

mRNA processing events (chapter 1, Fig. 1.5). We found that the 1,437 splicing events

affected by apigenin fell into all eight splicing categories (Fig. 6.19C). The majority of

mRNA processing events, ~56%, were represented by SE and AFE. However, only RI

(~14%, 196 events) and ALE (~15%, 223 events) were significantly enriched in cells

treated with apigenin (p < 0.05, hypergeometric distribution, Fig. 6.19C, white numbers), when compared to the total ~80,053 splicing events identified in DMSO-treated MDA-

MB-231 cells (Fig. 6.19C). Together, A5SS, A3SS, MXE and TUTR represent only 9%

of all the events affected by apigenin, suggesting a minor effect of this flavone on those

types of mRNA processing categories. Overall, these results indicate that apigenin

modulates mRNA processing genome-wide, with a significant preference for two types of

events, RI and ALE.

6.3.9 Apigenin regulates the splicing of RNA-binding proteins

To gain insights into the biological role of the genes differentially spliced by apigenin

(Fig. 6.19A), we performed GO analyses based on the biological functions of the 1,600

genes modulated by apigenin (Fig. 6.20A), using the software Ingenuity Pathway

Analysis (IPA). We observed a high enrichment of molecules in the “RNA processing”

GO category (Fig. 6.20A), underscoring the importance of apigenin on the mechanisms

that regulate mRNA metabolism, and in agreement with the categories observed for the

direct targets of apigenin (Fig. 6.6). The 106 molecules in “RNA processing” included

three members of the core spliceosome (snRNPC, snRNPF and snRNPG), eight

182

molecules of the hnRNP family (hnRNPA1, hnRNPA2, hnRNPH1, hnRNPD, hnRNPF, hnRNPH1, hnRNPM and hnRNPU), five proteins of the serine arginine family of splicing factors (SRSF2, SRSF3, SRSF5, SRSF6, SRSF9) and five members of the RNA binding motif (RBM) family of splicing regulators (RBM3, RBM4, RBM14, RBM17 and

RBM39). Together, these families regulate exon/intron inclusion/exclusion that result in

SE and RI, two splicing categories responsible for 36% and 14%, respectively, of mRNA processing events differently regulated in apigenin-treated cells (Fig. 2.19C). In addition, the “RNA processing” category contained four molecules, CPSF3 (cleavage and polyadenylation specific factor 3), CPSF4, CSTF2 (cleavage stimulation factor 2) and

CSTF3, involved in alterative mRNA polyadenylation and cleavage, a mechanism involved in the generation of alternative 3’ last exons (ALE), a type of mRNA processing significantly enriched in apigenin-treated cells (Fig. 2.19C) [370].

Next, we looked at the effect of apigenin on hnRNPH1, SRSF5 and RBM3, splicing factors previously reported as substrates of hnRNPA2 [369], using RT-PCR followed by agarose electrophoresis. We observed that apigenin promoted the exclusion of intron 6 in

SRSF5 and intron 10 in hnRNPH1 as well as inclusion of exon 4 in RBM3 in MDA-MB-

231 cells compared to cells treated with DMSO (Fig. 6.20B). These results confirmed the changes predicted by our computational analysis of RNA-seq data and demonstrate that apigenin affects the splicing of RNA-binding proteins, hence altering the mRNA processing regulatory network.

183

6.3.10 Apigenin regulates survival and apoptotic signaling pathways through

splicing

In addition, other biological functions enriched in the genes differentially spliced in

apigenin-treated cells included gene expression, cell death and protein synthesis,

comprising 626 molecules of the 1,600 genes affected by apigenin (Fig. 6.20A).

Moreover, GO analysis based on biological pathways, revealed that the most significantly

enriched signaling cascades included the mTOR, p53, glucocorticoids, PI3K/AKT and

apoptosis (Fig. 6.21A). These cascades regulate gene expression, consistent with the

finding of “gene expression” as a biological function affected by apigenin (Fig. 6.21A

compared to Fig. 6.20A). The mTOR, PI3K/AKT, p53 and apoptosis signaling cascades

are mechanisms of cell death and survival, in agreement with the category “cell death” as

a biological function significantly enriched by apigenin treatment (Fig. 6.21A and

6.20A).

Next, we studied whether the 1,600 genes were enriched in annotated tumor

suppressors (TS), oncogenes (OG) or cancer driver genes (CDG) [105], but no

significantly enrichments were found in these groups (p > 0.05, Fig. 6.21B, grey bars

below threshold). These results are consistent with the nature of the carcinogenic role of

these genes relying on genetic mutations rather than in changes of splicing or expression

[105, 371]. All together, these results indicate that apigenin has a high significant impact

on the mRNA processing network itself and affects splicing of genes that regulate gene expression, cell death and protein synthesis.

Previous reports show that apigenin inhibits several proteins in the

184

mTOR/p53/PI3K/apoptosis pathways [139, 173, 372], the signaling pathways most significantly enriched among the genes differentially spliced in apigenin-treated cells

(Fig. 6.21A). Interestingly, apigenin did not modulate the splicing of the main hubs of these cascades, e.g. caspase-3, NF-κB, mTOR and p53 but rather altered the splicing of their regulators (Fig. 6.21C, yellow balloons). For example, the isoform variants of

cFLIP, TMED7-TICAM2 (Transmembrane Emp24 Protein Transport Domain

Containing 7-Tir-containing adaptor molecule-2) and TSC1 (tuberous sclerosis 1), which

contain alternative last exons. cFLIP splicing was previously described (Fig. 6.18A and

Fig. 6.22A). TMED7 and TICAM2 are two contiguous genes. TMED2 is a trans-

membrane protein and TICAM2 is a toll-like receptor-interacting protein [373]. An

alternative last exon in TMED7 creates a read-through into TICAM2 resulting in a

chimera protein, TMED-TICAM2 [also known as TAG (TRAM adaptor with GOLD

domain)], which interacts with toll like receptors and inhibits down-stream NF-κB

signaling [374, 375]. We found that apigenin increases the levels of the splicing isoform

TMED7-TICAM2 (Fig. 6.22B). TSC1 is a regulator of mTOR and an alternative last exon containing a portion of intron 11 results in a non-coding transcript [376]. Apigenin increases the levels of the TSC1 isoform with the alternate last exon (Fig. 6.22C).

The pro-apoptotic protein BCL2L11 (also known as BIM) is comprised of several isoforms including BCL2L11EL (Extra Long) and BCL2L11L (Long) [377]. BCL2L11EL

is the result of an in-frame retention of intron 2. This intron codes for an extra-domain,

which is phosphorylated by survival kinases such as AKT1 and MAPKs, resulting in the

inhibition of the pro-apoptotic function of BCL2L11 [378-381]. In contrast, BCL2L11L

185

lacks intron 2 and promotes apoptosis [379-381]. Our results show that apigenin

decreases BCL2L11EL while increasing the levels of BCL2L11L (Fig. 6.22D). NAIP, also known as BIRC1, is an inhibitor of apoptosis [382]. Retention of intron 11 in this gene generates a non-coding transcript (NAIP-RI, retained intron) [376]. Hence, increase in

NAIP due to RI in cells treated with apigenin may reduce NAIP protein expression allowing execution of apoptosis (Fig. 6.22E) [383].

In addition, our results indicate that some of the SE changes induced by apigenin corresponded to several regulators of cell death and survival including DIABLO, BAX,

HRAS and HIF1A. Exclusion of exon 4 in DIABLO produces a short isoform known as

SMAC3 [384]. SMAC3, but not DIABLO, promotes apoptosis by inducing XIAP ubiquitination and degradation [384]. We observed that apigenin increases the levels of the SMAC3 isoform (Fig. 6.22F). In addition, inclusion of exon 5 in BAX shifts its

coding frame producing BAXε, a short version of BAX that lacks the BH2 and trans-

membrane domains required for the binding to the anti-apoptotic protein BCL2 [385].

Our results showed an increase of BAXε in apigenin-treated cells (Fig. 6.21G). HIF1A encompasses the splice variants HIF1A826 and HIF1A735 [386]. HIF1A735, the isoform

increased in apigenin-treated cells (Fig. 6.22H), lacks the transactivation domain

impairing its transcriptional activity [387]. Moreover, HRAS comprises two isoforms, p21 and p19 [388]. When translated, HRASp19, generated by inclusion of exon 4 works as a negative regulator of HRASp21 and blocks its survival signal resulting in reduced proliferation and induction of apoptosis [388, 389]. We found that apigenin promotes inclusion of exon four in HRAS transcripts (Fig. 6.22I). In all the cases described here,

186

apigenin increased the anti-proliferative or pro-apoptotic splice versions of the genes

(Fig. 6.22 and Fig. 6.21C). Hence, it is plausible that a mechanism in which apigenin exerts its anti-carcinogenic activity is by switching transcript variants of cancer cells from survival to pro-apoptotic versions thus eliminating their intrinsic resistance to apoptosis.

6.3.11 Apigenin modulates splicing of hnRNPA2 and MSI2 substrates

Among the apigenin targets identified by PD-seq, six proteins were directly

implicated in RNA metabolism (Table 6.3), including hnRNPA2, MSI2, CELF1

(CUGPB, ELav-Like Family Member 1), CELF4, UPF3B (Up-Frameshift Suppressor

3B), and SRRT (Serrate RNA Effector Molecule). We searched the literature to find

substrates of these RNA-binding proteins. The genes directly targeted by hnRNPA2,

MSI2 and CELF1 were previously identified in HEK293 human embryonic kidney cells,

K562 human leukemia cells, and C2C12 mouse muscle cells, respectively, by CLIP

(Crosslinking-immunoprecipitation)-seq [369, 390, 391]. Next, we intersected the genes

modulated by apigenin with the hnRNPA2, MSI2 and CELF1 substrates. We observed

that out the 1,600 genes differentially spliced by apigenin, 480 (~30%) corresponded to

targets of hnRNPA2, MSI2 or CELF1. From those, 163 are substrates of MSI2, 253 of

hnRNPA2 and 97 of CELF1 (Fig. 6.23A). Enrichment analyses found that the group of

genes in which splicing was modulated by apigenin, were enriched in the substrates of

MSI2 and hnRNPA2 (p < 0.05, hypergeometric distribution). The lack of enrichment of

CELF1 substrates in our analysis may be due to its main role on heart development [392],

but not in cancer, and that CELF1 substrates where identified using in mouse muscle cell

187

lines. A network representation of the MSI2 and hnRNPA2 substrates in which splicing

was altered by apigenin showed that MSI2 and hnRNPA2 regulate splicing of 126 genes

involved in cell death and survival, identified by GO analysis using IPA (Fig. 6.23B).

Among these genes, key regulators include BCL2L11, BAX and cFLIP (Fig. 6.23B,

orange). These results indicate that apigenin regulates the splicing of substrates of MSI2

and hnRNPA2. Further silencing experiments will be performed in order to evaluate the

role of these RNA-binding proteins on the modulation of mRNA processing by apigenin.

6.3.12 Apigenin affects the splicing of genes dysregulated in breast cancer

To evaluate whether apigenin modulates the splicing of genes dysregulated in breast

cancer, we analyzed public available RNA-Seq data sets from six human biopsies of

triple negative breast cancer (TNBC) and three samples from healthy donors (normal

breast tissue, NBT), downloaded from GEO (Gene Expression Ominbus, series

GSE52194, [115]). Applying the same computational pipeline described in section 6.3.8

to these data, we determined the genes which splicing is perturbed in breast cancer tissue

compared with normal breast tissue. To determine the inclusion/exclusion ratio of an

exon/isoform, we determined the PSI (Ψ). Statistical significance between NBT and

TNBC was considered when the Bayes factor (BF) was higher than 20 (BF > 20, see

section 6.3.8 and [120, 190]). From 86,378 mRNA processing events occurring in NBT and 88,582 in TNBC, 1097 events, corresponding to 630 genes were found dysregulated

in TNBC (Fig. 6.24A), using so far only the exon-centric approache. These aberrant

events correspond to all the eight mRNA processing categories. From those, ~31% (344

e The isoform-centric approach will be performed in order to complement the results obtained by the exon-centric method. However, the analysis is still ongoing. 188

events from all the 1097) are ALE, the only category significantly enriched in TNBC

(Fig. 6.24A). Changes in splicing corresponding to the AFE, A3SS, SE and RI categories

accounted for 58.3% of the events, while A5SS, MXE and TUTR contributed to the

remaining 10% of the dysregulated events found in TNBC (Fig. 6.24A). Next, the 630

genes with aberrant splicing in TNBC were subjected to GO analysis. We found that

most of the genes ~53 % (338 molecules) corresponded to cell death, proliferation and

cell migration (Fig. 6.24B). In addition, 46 genes are involved in RNA processing (Fig.

6.24B). When the analysis was done taking into consideration the signaling pathways, we found that the mTOR/EIF2, integrin-like kinase (ILK) and the cell-cell junction pathways

were significantly enriched in TNBC (Fig. 6.24C). These results uncover similitudes and

differences with the biological functions and signaling pathways affected by apigenin in

MDA-MB-231 cells. First, underscoring the importance of the mTOR/EIF2 pathway and

apoptosis in the anti-carcinogenic activity of apigenin, we found that this signaling

pathway is highly significantly enriched in TNBC compared to NBT and in apigenin-

treated breast cancer cells (Fig. 6.24C compared to Fig. 6.21A). Second, genes with

aberrant splicing in TNBC are enriched in 46 RNA-binding proteins including members

of the hnRNP family such as hnRNPA2/B1, hnRNPA1 and hnRNPU and SR splicing

factors including SRSF3, SRSF34, SRSF35 and SRSF37, demonstrating the

dysregulation of splicing factors in cancer, as previously shown [369]. ALE was the most

common type of abnormal events in TNBC (Fig. 6.24A). In addition, the percentage of

RI in aberrant splicing was slightly increased, from 5% in NBT to 8% among the events

dysregulated in TNBC, although not significantly, in contrast to the percentage of RI

189

observed in apigenin-treated cells (Fig. 6.24A compared to 6.19C). Moreover, SE events, representing 13% of all cancer-dysregulated events seem to have less contribution on the genes dysregulated in TNBC, despite the fact that is the most common category observed in NBT and TNBC (Fig. 6.24A).

A comparison of the genes with abnormal splicing in TNBC to those modulated by apigenin identified 110 genes in common, corresponding to 17.5% of the 630 genes

dysregulated in TNBC (Fig. 6.25A). From those, 33 undergo changes in splicing, when

exposed to apigenin, that restore the splicing profiles found in NBT (Fig. 6.25A). Hence,

we called these cases “restored” genes. There are 77 genes dysregulated in TNBC, in

which splicing was changed by apigenin, but the profile was different than what was

observed in NBT (Fig. 6.25A). The 110 genes in common included 84 splicing events affected in apigenin-treated cells and dysregulated in TNBC (Fig. 6.25B). From those,

57% events corresponded to RI and ALE (Fig. 6.25B), the two categories significantly enriched in apigenin-treated cells (Fig. 6.25B compared to Fig. 6.19C).

Next, we looked at the biological function of the 33 restored genes using GO analyses. We found that 11 were involved in cell death and proliferation, 10 in RNA processing and 6 in genes expression (Table 6.6). Subsequently, we performed isoform specific RT-PCR from mRNA of MDA-MB-231 and MCF-10A cells treated with 25 µM apigenin or DMSO for 48 h in two of the restored genes involved in cell death and mRNA processing, i.e. NAIP and CCNL2. The latter a protein involved in the regulation

of splicing during cell cycle progression [393, 394]. CCNL2 is a cyclin that contains a

serine-arginine domain, characteristics of the SR family [393, 394]. Overexpression of

190

CCNL2 in cancer cells induces apoptosis [394-396]. A splicing isoform of CCNL2 retaining intron 6 (Fig. 6.25C, CCNL2-RI) is predicted to produce a non-coding transcript [376], which may halt CCNL2 protein expression. MDA-MB-231 cells have higher levels of the CCNL2-RI isoform, compared to MCF-10A cells (Fig. 6.25C, lane 3 vs. 1). Apigenin decreased ~6-fold CCNL2-RI in MDA-MB-231 cells (Fig. 6.25C, lane 4 vs. 3), to levels found in MCF-10A cells treated with DMSO (Fig. 6.25C, lane 4 vs. 1). In addition, apigenin promoted ~4-fold intron 11 inclusion of the anti-apoptotic protein

NAIP in MDA-MB-231 cells compared to cells treated with vehicle (Fig. 6.25D, lane 4 vs. 3), to similar levels found in MCF-10A cells treated with DMSO (Fig. 6.25D, lane 4 vs. 1). In contrast, apigenin had no effect on the splicing of NAIP and CCNL2 in MCF-

10A cells (Fig. 6.25C and D, lane 2 vs. 1). These results suggest that apigenin restores aberrant-splicing profiles in a subset of TNBC-dysregulated genes to normal mRNA- signatures contributing to the anti-carcinogenic activity of this dietary phytochemical.

6.3.13 Effect of apigenin on gene expression in breast cancer

To evaluate whether the effects of apigenin on mRNA processing are independent of its ability to change gene expression, we determined the steady-state expression of genes,

(see section 2.17.2, chapter 2) in MDA-MB-231 cells treated with 25 µM apigenin or

DMSO for 48 h, using the DESeq software. Statistical significance between apigenin and

DMSO was considered when False Discovery Rate was less than 0.05 (FDR < 0.05).

Apigenin significantly changed the expression of 2,122 genes, out of the ~12,000 genes

expressed in MDA-MB-231 cells. Of those, the levels of 47.5%, corresponding to 1,008

191

genes, were down-regulated, whereas the expression of 52.5%, corresponding to 1,114 genes, were up-regulated. From these 2,122 differentially expressed genes, only 253, corresponding to ~12%, showed changes in both gene expression and mRNA processing

(Fig. 6.26A, brown). In addition, apigenin regulates splicing of 1,347 genes (Fig. 6.26, orange equals 1,600 genes affected in splicing minus 253 genes commonly affected in gene expression), without changing their overall expression. These results suggest that apigenin has the ability to regulate mRNA processing in addition of gene expression.

Next, we performed GO ontology analysis of the 2,122 genes differently expressed in apigenin-treated cells (Fig. 6.26B and C). The top four biological functions comprising

848 genes, out of the 2,122, corresponded to genes involved in cell cycle, DNA repair, cell death and cell proliferation (Fig. 6.26B). In addition, GO analysis of signaling pathways revealed the DNA damage response and cell cycle progression cascades such as

ATM, p53, BRCA1 and G2/M checkpoint, were the most affected by apigenin (Fig.

6.26C). GO analysis of the 253 genes commonly affected by splicing and gene expression in apigenin-treated cells, were also enriched in genes involved in cell cycle progression (Fig. 6.26D and E). Compared to the biological role of genes affected by splicing in apigenin-treated cells (Fig. 6.20A compared to Fig. 6.26B and Fig. 6.21A compared Fig. 6.26C), these results indicate that the changes in mRNA processing and genes expression observed in apigenin-treated cells impact different functions in the systems’ biology. Changes in genes expression mostly affect the mechanisms that regulate cell cycle progression and DNA damage response, while alterations in mRNA processing modulates the splicing regulatory network and the mechanisms that regulate

192

cell death and survival.

To evaluate the effect of apigenin on genes in which expression is aberrant in breast cancer, we applied DESeq analysis to the publically available data obtained from 6

TNBC and 3 NBT independent patient samples. We found 3,816 genes which expression is dysregulated in TNBC compared to NBT. An intersection of the 3,816 TNBC-altered

genes (Fig. 6.27A, addition of yellow and orange), with the 2,122 affected by apigenin in

expression (Fig. 6.27A, addition of brown and orange), found that 530 molecules were

common (Fig. 6.27A, orange). A further study from those 530 genes indicated that

apigenin restored the levels found in normal tissues in 346 genes (Fig. 6.27A, purple),

while the remaining 194 corresponded to genes, which were affected by apigenin, but in a

manner not found in NBT (Fig. 6.27A, light blue). Gene ontology analysis of the 346

genes demonstrated a highly significant enrichment in the mechanisms involved in cell

cycle progression such as the ATM and G2/M checkpoints and p53 signaling pathways

(Fig. 6.27B, light blue). These results resemble the enrichment in obtained when GO was

performed in 2,122 genes affected by apigenin (Fig. 6.26C). The main categories found

enriched in the other 194 genes that did not resemble the expression level found in NBT

included granzyme A signaling and alanine metabolism (Fig. 6.27C). Next, we evaluated

the steady state mRNA levels of some of the molecules involved in the ATM, G2/M

checkpoints and p53 cascades by qRT-PCR in MDA-MB-231 or MCF-10A cells treated

with 25 µM apigenin (+) or DMSO (-) for 48 h (Fig. 6.27D). We found that the

expression of cyclins CCNB1 and CCNA2, two well-known positive regulators of cell

cycle, and RAD51, a DNA-repair molecule, was increased in MDA-MB-231 cells when

193

compared to MCF-10A (Fig. 6.27D), and apigenin reduced their expression to levels

found in DMSO-treated MCF-10A cells (Fig. 6.27D). The addition of apigenin to MCF-

10A cells had not effect on the expression of these genes (Fig. 6.27D). In addition, the expression of the negative regulators of cell cycle such as GADD45A (growth arrest and inducible DNA damage 45-alpha) and CDKN1A (cyclin-dependent kinase inhibitor 1A)

is lower in MDA-MB-231 cells compared to the levels found in MCF-10A cells (Fig.

6.27D). The addition of apigenin increased the expression of GADD45A and CDKN1A1 in MDA-MB-231, but not in MCF-10A, to levels found in MCF-10A cells treated with

DMSO (Fig. 6.27D). To evaluate the reliability of the RNA-seq computational analysis, we determined the mRNA steady state levels of NF-κB-p65 subunit, a molecule that was not affected by apigenin treatments neither was dysregulated in TNBC compared to NBT, as determined by DEseq analysis. NF-κB-p65 expression was similar in MDA-MB-231

compared to MCF-10A cells treated with DMSO (Fig. 6.27D). Apigenin treatment had

no effect on the levels of NF-κB-p65 in MDA-MB-231 or MCF-10A cells (Fig. 6.27D).

These results indicate that apigenin modulates abnormal-expression of genes

dysregulated in breast cancer cells to levels found in non-carcinogenic breast cells.

6.3.14 Apigenin decreases proliferation and induces apoptosis in breast cancer cells

Apigenin modulates the splicing of apoptotic genes increasing the levels of their pro- apoptotic isoforms and restoring the expression of genes involved in cell cycle progression in breast cancer cells. To evaluate whether this genetic changes result in

differential induction of apoptosis in cancer cells compared to non-carcinogenic cells, we

194

evaluated apoptosis and proliferation in MDA-MB-231 and MCF-10A cells treated with

diluent DMSO or 50 µM apigenin for 48 h. Apigenin decreased ~40% proliferation of

MDA-MB-231 cells, but had no effect in MCF-10A cells (Fig. 6.28A). Consistent with the changes observed in known regulators of G2/M transition such as CCNB1, CCNA2,

GADD45A and CDKN1A (Fig. 6.27D), we found that apigenin arrested breast cancer cells in the G2/M phase of cell cycle (Fig. 6.28B). Moreover, addition of apigenin resulted in a 3-fold increase of apoptosis in MDA-MB-231 cells, but had no effect in

MCF-10A cells, as illustrated by calcein A/M and propidium iodide staining (Fig. 6.28D and E). In addition, we found that apigenin induced caspase-3 activity in MDA-MB-231 breast cancer cells (Fig. 6.28F). These results are consistent with the increase of pro- apoptotic splicing isoforms of genes involved in apoptosis (Fig. 6.21 and 6.22).

Altogether, these findings indicate that apigenin induces apoptosis and decreases proliferation in breast cancer cells without affecting non-carcinogenic epithelial breast cells.

6.3.15 Apigenin decreases the expression of cell cycle progression genes and modulates splicing of apoptotic molecules in vivo

To evaluate whether the changes in splicing and gene expression induced by apigenin in human breast cancer cell lines were occurring in vivo, we used a xenograft mouse model. For this purpose, human MDA-MB-231 breast cancer cells were injected into the mouse mammary fat pad and 24 h later 25 mg/kg apigenin or vehicle were administered daily for 28 days via i.p. We observed that up to day 16, apigenin has no significant effect

195

on tumor size when compared with mice receiving vehicle. Yet, starting at day 18, a

significant decrease in tumor volume was observed, reaching a ~2-fold reduction at 28

days (Fig. 2.29A and B).

To evaluate the effect of apigenin on tumor proliferation, tumor sections were stained

with anti-Ki67 antibodies, a proliferation marker, by IHC. We found that apigenin

administration decreased tumor proliferation by 2-fold compared with mice treated with

vehicle (Fig. 6.29C). Subsequently, we assessed whether apigenin promotes tumor

apoptosis. We observed that apoptosis was increased by ~2-fold in tumors from apigenin-

treated mice compared to animals receiving vehicle, as shown by TUNEL assays (Fig.

5.29D). Our results are in agreement with previous studies of apigenin in MDA-MB-231

xenografts [161].

To determine whether the genes affected by apigenin in cellular models were also

perturbed in in vivo, we evaluated the mRNA expression of molecules involved in cell cycle progression and the splicing of genes involved in regulation of apoptosis. Total

RNA isolated from tumors was used to evaluate the expression of cyclins CCNB1 and

CCNA2, positive regulators of cell cycle progression (Fig. 6.27), by qRT-PCR using human specific primers. We found that apigenin decreases by ~3-fold CCNB1 and by 2-

fold CCNA2 expression in MDA-MB-231 xenografts from apigenin-treated mice compared to animals receiving vehicle (Fig. 6.29E). In addition, we evaluated the

splicing of BCL2L11 and cFLIP, two apoptotic genes in which apigenin increased their

pro-apoptotic isoforms (Fig. 6.18A and 6.22), by RT-PCR using human exon specific

primers. Apigenin reduced the level of BCL2L11EL and cFLIPS by ~2-fold in tumors

196

compared to animals treated with vehicle (Fig. 6.29F). These results suggest that apigenin

increases pro-apoptotic spliced variants and decreases the expression of cell cycle

progression genes in vivo, thereby contributing to decrease tumor growth.

6.4 Discussion

The results shown here describe the development and implementation of a new innovative high-throughput strategy, PD-Seq, to comprehensively identify the human cellular targets of apigenin. Importantly, this method can be used for the target identification of any small molecule. Using PD-seq, we identified 160 candidate apigenin targets, significantly enriched in three main functional categories corresponding to

GTPase activation, membrane transport and mRNA metabolism/alternative splicing. This last category included splicing factors such as hnRNPA2 and MSI2. These results revealed a new mechanism of action of flavonoids by modulating alternative mRNA processing genome-wide, helping explain the biological effects of this dietary phytochemical.

The identification of just hnRNPA2 by conventional phage display underscores one of the main shortcomings of this method, the recovery of very few candidates likely due to the limited amount of clones that can be tested and because many putative targets fail to be properly amplified or selected through multiple bio-panning rounds [397]. Our results demonstrate that PD-Seq overcomes this constraint. Highlighting the usefulness of

PD-Seq for the identification of small molecule binding proteins, we validated the association of apigenin with seven proteins by different methods that included phage

197

competition, pull-downs and enzymatic assays. In addition, PD-seq found proteins that were previously described as direct targets of apigenin such as Mucin 1 (MUC1), identified by screening a chemical library for MUC1 inhibitors [177], suggesting the ability of PD-Seq to identify targets of apigenin. MUC1 is a membrane glycoprotein aberrantly expressed in breast cancer [398]. MUC1 C-terminal domain is cleaved and form homo-dimers that localize to the cytoplasm and nucleus where it interacts with transcriptions factors such as NF-κB [399, 400]. Apigenin was shown to directly bind the

C-terminal domain of MUC1 inhibiting its dimerization and disrupting MUC1 nuclear localization (IC50 ~75 µM) [177].

PD-seq also identified targets of apigenin shown to interact with other flavonoids. For example, hnRNPA2 interacts with proanthocyanidins, a group of oligomeric flavonoids, and seems to be responsible for the anti-viral activity of blueberry leaf extracts against

Hepatitis C Virus (HCV) [401]. Consistently, we found, using the FRET nanosensor, that hnRNPA2 interacts with several classes of flavonoids including flavones and flavonols.

Two other flavonoid targets identified in our studies include lactase (LCT), an enzyme that hydrolyzes a range of flavonol and isoflavone glycosides [402], and the UDP-

Glucose Dehydrogenase (UGDH), which activity is inhibited by the flavonol quercetin, resulting in decreased proliferation of breast cancer cells [403]. Yet, PD-seq failed to capture the previously identified apigenin targets tubulin, RPS9 (Ribosomal protein S9) and CK2 [174, 175, 178]. The interaction of apigenin with RPS9 was identified by pull- down of cell lysates and validated by pull-down of recombinant purified RPS9 protein with apigenin-beads [175]. However, the absence of RPS9 might be due to its

198

undetectable presence in the original library, and while its presence was detected in the

fraction AE2 in Cluster IV, this cluster lacked statistical significance of enrichment

(Cluster IV, Fig. 6.4), suggesting that potential apigenin targets with low read coverage

were missed in our screening. Yet, this type of shortcoming could be overcome by

increasing sequencing depth in future implementations of PD-Seq. Apigenin binds at the

interface between α-tubulin/β-tubulin, as determined by spectrofluorometry and in silico

docking analyses, inhibiting microtubule polymerization and inducing arrest of cells in

G2/M [178]. These observations uncover a limitation of phage display to identify small-

molecule interactions that require the quaternary structure of hetero-complexes such as

the interface of α-tubulin/β-tubulin. In addition, apigenin is generally believed to be a

CK2 inhibitor. Apigenin inhibits CK2 activity in purified preparations from rat livers and human cell lysates [174, 404]. However, the direct interaction of apigenin with CK2 has not been demonstrated in previous studies; neither was CK2 identified in our screening.

Hence, it is possible that apigenin inhibits CK2 indirectly via another peptide, yet to be identified. Supporting this idea, hnRNPA2 has been shown to interact with CK2 and promote its activation [405]. Thus, apigenin by binding to hnRNPA2 may reduce CK2 activity.

The finding that most of the apigenin targets correspond to one of three main categories (GTPase activation, membrane transport and splicing) has several interesting implications. First, polyphenols are often assumed to bind non-specifically to a large number of proteins [406, 407]. It is clear from our results that this is not the case, since there is a subset of preferentially bound targets. Moreover, the FRET nanosensor

199

demonstrated the specificity of hnRNPA2 for flavones and flavonols but not flavanones

or isoflavones, suggesting that, at least one of apigenin targets, has selective specificity

for flavonoid subgroups. Second, kinases and other proteins with ATP-binding proteins

have been generally believed to be main targets for flavonoids [408, 409]. Interestingly,

our screening identified only five kinases (ADCK1, EPHA5, MAST1, PFKP and ULK4;

Table 6.3) as candidate apigenin targets, despite more than 560 kinases being represented

in the library.

HnRNPA2, and its splice variant hnRNPB1, play fundamental roles in the

progression of tumorigenesis by regulating splicing, mRNA stability, and mRNA

transport [354]. Over-expression of hnRNPA2 has been reported in several human

cancers, including breast [122], and hnRNPA2 expression is recognized as a marker of

glioblastoma and lung cancer [123-125]. Consistently, we found that hnRNPA2 is

overexpressed in breast cancer cells compared to non-carcinogenic breast epithelial cells.

Apigenin binds to the GRD of hnRNPA2, at least in two different sites (Fig. 6.8), with a

dissociation constant that ranges 2-20 µM. The lower affinity (higher apparent KD) determined by FRET reflects most likely the differences between the two KD

determination methods, and the absence of additional apigenin recognition sites in

hnRNPA2264-341. However, one advantage of the FRET-based method over the

spectrophotometric approach is that it permits us to investigate the specificity and relative

affinity of the interaction between hnRNPA2 and other flavonoids. Our results show that

the –OH group in ring A is necessary for interaction with hnRNPA2, as demonstrated by

the lack of binding of 7-O-glucoside to the nanosensor. Highlighting the importance of

200

the 7-O group itself in the interaction with hnRNPA2 rather than a steric hindrance by the glucosyl group, apigenin 6-C-glucoside significantly binds to hnRNPA2 but with lower affinity than apigenin. Notably, and consistent with flavanones lacking the biological activities observed for apigenin, neither naringenin nor the related flavanone eriodictyol affected the FRET of the nanosensor. In addition, quercetin and kaempferol also show a significant interaction with hnRNPA2 but with lower affinity than apigenin, consistent with the fact that many biological activities of flavones are shared by flavonols [135]

The GRD domain is implicated in hnRNPA2 homodimerization, important for the participation of this protein on RNA binding. Indeed, a recent study reported that apigenin impairs the ability of hnRNPA2 to bind viral RNA preventing enterovirus-71 infection of mammalian cells [410]. The amino acids Y, R or N in the GRD-YGGG(R/N) repeats seem to be necessary for apigenin binding. Yet, only one YGGGR is not sufficient to support the interaction with A-beads, suggesting that either two YGGGs or other residues may mediate the association of hnRNPA2 with apigenin. To evaluate these possibilities, we are generating hnRNPA21-255 clones with substitutions in residues

Y235G, P239G, N243G as well as deleting amino acids between YGGG repeats i and ii

(Fig. 6.9B). The structural mechanisms that allow binding of flavonoids to their targets are not well understood. The general consensus indicates that flavonoids bind through hydrophobic interactions and hydrogen bonds between -OH groups in the flavonoid and side chains of amino acids [411, 412]. Recent structural analysis in silico suggested that apigenin interacts with tubulin through hydrogen bonds between –OH group in rings A and B and the aminoacids R, N and Q, present in the tubulin docking site [178]. Our

201

preliminary results on hnRNRPA2 also indicate that R, Y and N are key for apigenin binding, but the structural characteristics of the apigenin binding site need further

investigation. Once we accurately define the apigenin-binding site, further 3D docking

analyses will be performed in order to identify an apigenin interacting structural motif

common to apigenin targets.

We demonstrated that apigenin affects mRNA processing genome-wide. To our

knowledge, this is the first study uncovering the regulation of splicing by dietary

phytochemicals. RNA-seq analysis provides a reliable method for the identification of splicing genome-wide, judging by the high percentage of validation and the correlation between RT-PCR and RNA-seq (Fig. 6.19B), as previously reported [190]. However, the effect of apigenin on caspase-9 splicing observed by RT-PCR in MDA-MB-231, was not detected by the computational approach employed in this study. These results are likely to reflect the inability of MISO to recognize multi-exon skipping events, as present in

caspase-9b (skipping exons 6-9), and indicate that the number of mRNA processing

events affected by apigenin is higher than anticipated. Despites this limitation, we

identified 1,600 genes alternatively spliced after apigenin treatment falling into eight

types of mRNA processing events. Yet, only RI and ALE events were significantly

enriched by apigenin, as found in our analyses. The mechanisms that regulate the presence of different types of mRNA processing events are not well understood, but may involve several RNA binding proteins. In addition, it is becoming well accepted that the steady state levels of transcripts with alternative last exons are regulated through either mRNA stability, because these transcripts have alternative 3’UTRs with differential

202

recruitment of miRs or mRNA-destabilizing proteins, or by alternative cleavage and

polyadenylation of newly synthesized transcripts [370]. Among the apigenin targets, hnRNPA2 is a major regulator of mRNA stability by binding to AU rich sequences in the

3’UTR of its transcript substrates [413]. AGO1 (Argonaut 1), another apigenin target that we identified, is directly implicated in the recruitment of miRs to their complementary mRNA sequences, thus regulating the levels of transcripts with alternative 3’ last exons

[414, 415]. In addition, SRRT (Serrate RNA effector molecule, also known as ARS2) regulates 3’-end RNA cleavage and polyadenylation [416]. Hence, binding of apigenin to hnRNPA2, AGO1 or SRRT may result in differential abundance of ALE isoforms.

However, the effect of silencing or inhibition of AGO1, SRRT or hnRNPA2 on ALE has not been investigated so far.

The steady state levels of transcripts with RIs are regulated by non-sense mediated mRNA decay [417], a mechanism that degrades RI-bearing transcripts, or by altering alternative splicing of newly synthesized transcripts [417]. Changes in RI and SE events

were enriched in breast cancer cells upon silencing of MSI2, suggesting the role of MSI2

in these mechanisms, so far poorly understood [351]. UPF3B, another target of apigenin,

is a main regulator of non-sense mediated mRNA decay [418]. Thus, binding of apigenin

to MSI2 or UPF3B may contribute to differential accumulation of transcripts with

retained introns. In addition, RI is the most common type of AS event in plants [419],

which makes it tempting to speculate a conserved involvement of flavonoids in the

regulation of mRNA metabolism. Indeed, the bi-flavone isoginkgetin is used as a general

inhibitor of the spliceosome [420]. However, whether apigenin, isoginkgetin or other

203

flavonoids regulate mRNA processing in plants has not been studied.

We observed that apigenin affects the splicing of RNA-binding proteins, and RNA

processing was the function most significantly enriched among genes differentially

spliced in breast cancer cells treated with this flavone, as found in our GO analysis. These

results are consistent with previous studies showing that hnRNPs, including hnRNPA2,

cross-regulate the splicing of RNA-binding proteins, contributing to build a compensatory loop involved in the regulation of these proteins [369]. The molecules involved in mRNA processing include members of the core spliceosome, the hnRNP family, the serine arginine family and the RNA binding motif (RBM) family of splicing regulators. These families regulate exon/intron inclusion/exclusion that result in SE and

RI, two types of splicing encompassing 50% of the events differently spliced in apigenin- treated cells. In addition, molecules involved in alterative mRNA polyadenylation and cleavage were also differentially spliced in apigenin-treated cells. Thus, apigenin directly binds to splicing factors such as hnRNPA2 and MSI2 that can alter the mRNA processing regulatory network by modulating the splicing of more RNA-binding proteins.

A recent study showed that treatment of human breast cancer MCF7 cells with doxorubicin, a topoisomerase inhibitor, induced changes in alternative splicing of 248 genes [421]. From this number, 84 (~34%) corresponded to changes in ALE, demonstrating a high enrichment of ALE events during DNA damage [421]. Apigenin and other flavonoids have been shown to be topoisomerase inhibitors and promote DNA damage (Chapter 1). Further analysis of our data, found that 24, out of the 85 ALE events affected in doxorubicin-treated cells, were also regulated by apigenin, indicating that

204

although few of the changes induced by apigenin may be a consequence of its ability to

induce DNA damage, the vast majority are independent, and reflect the ability of

apigenin to regulate the mRNA processing network.

The comparison of genes in which apigenin affected mRNA processing with known

substrates of apigenin targets revealed that hnRNPA2, MSI and CELF1 shared 253, 163

and 97 splicing-modulated genes with those affected by apigenin. Lack of enrichment

with CELF1 substrates may reflect the fact that CELF1 is a prime regulator of heart

development [392], and that their substrates where identified in mouse cell lines, while

hnRNPA2 and MSI play fundamental roles in breast carcinogenesis by regulating cell

proliferation and migration [118, 351, 422, 423]. However, it is important to note that the

interaction of an RNA-binding protein with mRNA does not necessary imply changes in

transcript processing or abundance. In addition, substrates of each apigenin target may

vary according to the cell line or tissue employed. Hence, the contribution of each target

to the splicing events affected by apigenin needs to be addressed by genome-wide analysis of mRNA-processing events upon silencing of apigenin targets in breast cancer cells. Further studies are also necessary to determine whether the genomic changes observed in breast cancer can be extended to other malignancies.

Cancer cells are often resistant to apoptosis [38], an effect ascribed in part to the

inefficient activation of caspases due to the dysregulated expression of anti-apoptotic molecules and over-activation of survival signaling pathways [38]. It is becoming evident that molecules involved in the regulation of apoptosis generally have different splice variants with opposite roles: anti-apoptotic or pro-apoptotic functions [126], and cancer

205

cells frequently express the anti-apoptotic isoforms [119]. In this study, we showed that

apigenin modulates alternative splicing of genes involved in the mTOR, PI3K/AKT and

apoptotic pathways switching splice variants from the anti-apoptotic and pro-survival

isoforms to the more pro-apoptotic and anti-proliferative versions of molecules such as

cFLIP, caspase-9, BCL2L11, CCNL2, DIABLO, BAX, NAIP, TMED7-TICAM2, HRAS

and HIF1A, among others. Previous studies showed that apigenin induces apoptosis in

cancer cells by increasing the protein levels of BCL2L11L, and BAX [424, 425], decreasing the expression of cFLIPS [150], inhibiting the mTOR and AKT signaling pathways [173], or impairing the transcriptional activities of NF-κB and HIF1A [156,

339]. However, these effects have been shown to be indirect. This study uncovers a novel

mechanism on how apigenin regulates apoptotic and survival signaling pathways through

modulation of splicing.

We found little overlap (~12%) between genes affected by splicing with gene affected

in their expression, suggesting that the effect of apigenin on mRNA processing is

independent to gene expression. In addition, we observed that the genes in which

expression is affected by apigenin are mostly related to the mechanisms involved in cell

cycle progression and DNA damage response. In contrast, apigenin-induced alterations in

splicing seem to involve molecules in in cell death and survival. Thus, it is plausible to

propose a mechanism in which apigenin regulates the splicing regulatory network, by

directly associating with RNA-binding proteins and changing the splicing of mRNA

binding factors. This effect results in alterations of molecules involved in cell death and

survival that together with changes on mRNA steady state levels of genes involved in cell

206

cycle progression decrease proliferation and induce apoptosis in breast cancer cells.

From the 3,816 genes abnormally expressed in TNBC, 346 (9%) were restored to levels found in normal mammary tissues. In addition, from 630 genes differentially spliced in TNBC, 110 were also modulated by apigenin, and 33 (5%) were restored by apigenin to splicing profiles observed in NBT, indicating that between 5-10% of cancer- specific transcriptome signatures are returned to normal levels in apigenin treated cells.

This “restoration” has several implications. First, apigenin can delay cancer progression by affecting transcript variants that promote tumor development. However, the stage- specific alterations of splicing events still need to be determined. Hence, the PyMT model of breast cancer development is an important tool to determine splicing changes that occur during progression to metastatic carcinoma, and the effect of apigenin on such events in vivo. Second, by inhibiting the resistance to apoptosis, apigenin can sensitize cancer cells to chemotherapeutic drugs. Indeed, apigenin increases the efficacy of doxorubicin, fluorouracil, BCL2 inhibitors and TRAIL in various cancer cells lines [150,

158, 424, 426]. Preliminary results from our group demonstrated that apigenin sensitizes lung cancer cells to TRAIL-induced apoptosis, in part by increasing the expression of a splice variant that yields higher protein levels of TNFRSF10B (also known as DR5), the

TRAIL receptor (not shown). The findings that some of the apigenin-dependent changes in transcriptome diversity are observed in vivo furnishes a powerful alternative for the treatment of breast cancer either with apigenin alone or in combination with other chemotherapeutic drugs.

In summary, this study offers a novel view on how dietary phytochemicals influence

207

the systems network, by impacting multiple cellular targets with moderate affinity. Thus,

in contrast to pharmaceutical drugs selected to have high affinity and specificity for main

hubs of biological pathways, the effect of flavonoids would be distributed across the

entire network (Chapter 7), resulting in a fine-tuning effect, with consequent benefits on human health.

208

Figure 6.1. Synthesis of apigenin-beads and PD-Seq strategy outline. A. Scheme for the chemical synthesis of apigenin-immobilized (A-beads shown in orange) and acetylated control PEGA beads (C-beads shown in blue). The coupling of apigenin to the beads occurred at the end of a polyethylene glycol linker (PEGA beads). Depending on the apigenin -OH group participating in the coupling to the phenyl bromoacetate group, A-beads consist of a combination of three products. B. Schematic representation of the bio-panning steps in the screening of a phage display cDNA library generated from human breast cancer mRNA. Three rounds of bio- panning (3X), each including binding to the beads, wash, elution, and amplification were performed in parallel using A- or C- beads. C. Schematic representation of the fractions used to make the libraries for Illumina GAII sequencing. The pre-clearing and washing steps were skipped from the figure for simplicity. The original library was an aliquot of single amplified library purchased from Novagen. Original library (Ori-lib), input and elution fractions (referred as E) obtained from the first and second rounds of bio-panning using A- and C-beads were used to generate libraries for sequencing and were named A-E1, A-E2 and C-E1, C-E2, respectively. D. Schematic representation of Illumina GAII libraries preparation. PCR primers (indicated by arrows) at the cDNA insert and vector boundaries were used to amplify the cDNA-containing region and subsequently ligated to Illumina adapters (grey areas) and indexed sequences (red area). In collaboration with Dr. Kengo Morohashi. Adapted from Arango et al 2013, PNAS [427].

209 Figure 6.2. Analysis of the MKET clone. A. Nucleotide sequence of the phage MCS without insert or containing the MKET clone. B. Alignment of the MKET DNA and peptide sequences indicating its probable origin. Adapted from Arango et al 2013, PNAS [427].

210 Figure 6.3. Calculation of normalized In-frame-aligned Counts Per Gene-model (nICPG). Sequences were filtered and aligned to human coding sequences (cds). The number of sequences aligned in-frame to a single gene were considered as in-frame aligned counts (blue bar), whereas alignment of a single sequence to multiple genes were considered as weighted counts (purple bar). To calculate nICPGs, the number of aligned sequences per gene model was counted (blue bars), and in cases when a read aligned to multiple coding sequences (purple bar), the number of reads was divided by the number of aligned coding sequences. For instance, a read that aligns to two coding sequences, A and B, results in 0.5 as a weighted count. If a second read matches to A but not B, then the total ICPG for A will be 1.5 and for B will be 0.5. Thus, by using weighted counts, the total count number is identical to total number of reads. To determine the normalized ICPG for a given gene A, ICPGA was divided by the sum of all ICPGs, and multiplied by 106 for better handling. In collaboration with Dr. Alper Yilmaz. Adapted from Arango et al 2013, PNAS [427].

211 Figure 6.4. Hierarchical clustering analysis of the PD-Seq results. A. Heat map

representation of log10(nICPG) for the minimal 15,568 genes present in the phage-display library analyzed by hierarchical clustering using the average linkage method. The columns in the heat map correspond to those described in Fig. 6.1C. Selected clusters are

indicated by Roman numbers. B. Line graph indicating distribution of log10(nICPG) within clusters. The red line indicates log10(nICPG) of hnRNPA2. C. An enlarged view of Cluster I consisting of 160 genes significantly enriched in the A-E2 fraction. In collaboration with Dr. Alper Yilmaz. Adapted from Arango et al 2013, PNAS [427].

212

Figure 6.5. Conventional phage display identifies clones that bind to the A-beads and are highly enriched in the PD-Seq. A. Phage (() enrichment was determined by counting plaque formation units (pfu/ml) in elutions from A- (white bars) and C-beads (black bars) after each round of bio- panning (1st, 2nd and 3rd). B. Enrichment of selected clones was determined by PCR using T7 primers (Table 2.4) flanking the insert after every round of bio-panning. C. Peptide sequences of apigenin-binding peptides isolated by conventional phage display. Sequences in red correspond to the hnRNPA2/B1 peptide fragment and in black to the MKET non-coding fragment. In collaboration with Dr. Kengo Morohashi. Adapted from Arango et al 2013, PNAS [427].

213 Figure 6.6. Apigenin targets are enriched in three main categories. The bar graphs displays functional categories (FunCat) enriched in each clusters, and the x-axis (log- scale) indicate the p-value. Filled black bars correspond to FunCats enriched significantly (p < 0.01). A. Cluster I. B. Cluster II. C. Cluster III. D. Cluster IV. In collaboration with Dr. Kengo Morohashi. Adapted from Arango et al 2013, PNAS [427].

214 Figure 6.7. Validation of apigenin targets. A. A-beads were co-incubated with selected phages in the presence of 20 µM apigenin, naringenin or DMSO. Relative binding percentage to A-beads was determined by counting plaque formation units (pfu/ml) of

()hnRNPA2264-341 relative to DMSO control. Data represent the mean ± SEM, n = 3; * p < 0.05. B-D. Lysates from HeLa cells transiently expressing full-length hnRNPA2-GFP or GFP alone (B), ARHGEF1-GFP (C) or GFP-BAG1 (D) were pulled-down with A- or C-beads. Pull-downs were resolved by SDS-PAGE and analyzed by Western blot using anti- GFP antibodies. E-F. Pull-downs from MDA-MB-231 cell lysates were immunoblotted with anti-MSI2 (E) or anti-HSP70 (F) antibodies. All Western blots are representative of biological triplicates. G-H. MDA-MB-231 cells were treated with 50 µM apigenin, 50 µM naringenin or DMSO for 3 h. IDH3 (G) and UGDH (H) activities were measured in mitochondria and whole cells lysates, respectively, as described in section 2.29. Iodoacetic acid (1 M), an enzymatic inhibitor, was added as control. Data represent mean ± SEM, n = 4; * p < 0.05 determined by two-way ANOVA. Adapted from Arango et al 2013, PNAS [427].

215

Figure 6.8. hnRNPA2 directly binds apigenin through the glycine-rich domain (GRD). Different versions of recombinant affinity-purified GST-hnRNPA2 proteins were pulled-down with A- or C-beads (indicated as A or C). Pull-downs (bound) and supernatants fractions were resolved by SDS-PAGE and analyzed by Western blot using anti-GST antibodies. Arrows indicate the correct sized products; smaller bands present in some of the lanes correspond to degradation products. GRD corresponds to the glycine- rich domain, and RBD to the RNA-binding domain of hnRNPA2. A. Full length GST- hnRNPA21-341. B. GST-hnRNPA2190-341 containing the GRD. C. GST-hnRNPA2264-341 corresponds to the C-terminal 78 amino acid fragment present in (-hnRNPA2264-341, identified by conventional phage display screening. D. GST-hnRNPA21-263 lacks the C- terminal 78 amino acid fragment. E. GST-hnRNPA21-189 is the clone in which the GRD domain was deleted. F. GST alone. Gels are representative of three biological repeats. Adapted from Arango et al 2013, PNAS [427].

216 Figure 6.9. Mapping of the apigenin-binding site to hnRNPA2. A. Peptide sequence of hnRNPA2 obtained from Uniprot (Entry name: ROA2_HUMAN). The six YGGG repeats in the GRD (yellow) are shown in red letters. Blue area refers to the RBD. B. The GST-hnRNPA2 clones used in pull down assays. Vector sequence (grey) corresponds to the polylinker of the pDEST15 vector. Affinity-purified GST-hnRNPA21-255 (C) and GST-hnRNPA21-248 (D) proteins were pulled-down with either apigenin or control beads (indicated as A or C, respectively). Pull-downs (bound) and supernatants fractions were resolved by SDS-PAGE and analyzed by Western blot using anti-hnRNPA2 antibodies. Cartoons on the right represent the hnRNPA2 fragment used in pull down assays. The westerns are representative of two biological repeats.

217 Figure 6.10. Mapping the hnRNPA2-apigenin structural binding signatures. A. The hnRNPA2 constructs used in pull down assays showing amino acid substitutions in green boxes. Affinity-purified GST-hnRNPA21-255/Y250G (B), GST-hnRNPA21-255/R254G (C), GST-hnRNPA21-255/R254N (D), and GST-hnRNPA21-248/P239R (E), were pulled-down with A- or C-beads. Bound and supernatants fractions were resolved by SDS-PAGE and analyzed by Western blot using anti-hnRNPA2 antibodies. Cartoons on the right represent the hnRNPA2 peptides used in each pull down. F. Peptide sequence of hnRNPA1 obtained from Uniprot (Entry name: ROA1_HUMAN). The only YGGG repeat in the GRD (yellow) is highlighted in red letters. Blue region indicates the RBD of hnRNPA1. G. Affinity-purified GST-hnRNPA1 (full length) was pulled-down with A- or C-beads and analyzed by Western blot using anti-hnRNPA1 antibodies. Smaller bands present in some of the lanes correspond to degradation products. Gels are representative of two biological repeats.

218 Figure 6.11. Binding affinity of the interaction of hnRNPA2 with apigenin determined by UV/Vis spectroscopy. A. Apigenin (10 µM) was titrated with increasing concentrations of purified GST-hnRNPA2 (0, 0.2, 0.5, 1, 2, 5, 10, 20 µM) or GST (insert). Changes in absorption across the UV-visible spectrum were determined over the 250 to 450 nm range. B. Changes in absorbance of apigenin at 370 nm as determined in (A). C. Dissociation constant (KD) of the apigenin-hnRNPA2 complex calculated using the Benesi-Hilderbrand method as described in section 2.26. Data represent the mean ± SEM, n = 3. Adapted from Arango et al 2013, PNAS [427].

219

Figure 6.12. Development of a flavonoid nanosensor. A. hnRNPA2264-341 was cloned into different fluorescent indicator protein (FLIP) vectors (see section 2.27). Bacteria lysates expressing different versions of FLIP-hnRNPA2264-341 proteins were incubated with increasing concentrations of apigenin (0, 0.01, 0.1, 1, 5, 10, and 25 µM) for 3 h at 37°C. Continued 220 Figure 6.12 continued

B. Relative fluorescence units (RFU) were determined by spectrofluorometry (λext: 405 nm; λemi: 460-600 nm) and represented as emission spectra. C. The calculated YFP/CFP fluorescent ratios (530 nm/480 nm) are also represented over the 0-25 µM concentration range of apigenin. Mean ± SEM, n = 3. In collaboration with Mr. Bledi Brahimaj. Adapted from Arango et al 2013, PNAS [427].

221

Figure 6.13. Binding affinity of the interaction of hnRNPA2 with apigenin determined by the FRET nanosensor. A. Schematic representation of the flavonoid nanosensor. B. hnRNPA2264-341 (Fig. 6.8C) was cloned into the pFLIP2 vector in frame between the regions coding for the N-terminal-CFP and C-terminal-YFP fluorescent proteins. The affinity-purified FLIP2-3-hnRNPA2264-341 protein was incubated with increasing concentrations of apigenin (0, 1, 5, 10, 25, 50, and 100 µM) for 3 h at 37°C. Relative fluorescence units (RFU) were determined by spectrofluorometry (lext: 405 nm; lemi: 460-600 nm) and represented as emission spectra. C. The calculated YFP/CFP fluorescent ratios (530 nm/480 nm) are represented over the 0-100 µM apigenin concentration range. (F) Apigenin-dependent changes in YFP/CFP ratios were transformed into saturation curves as described in section 2.27. KD value was determined by using non-linear regression. Data represent the mean ± SEM, n = 3. In collaboration with Mr. Bledi Brahimaj. Adapted from Arango et al 2013, PNAS [427].

222 Figure 6.14. Apigenin quenches CFP fluorescence. A. Apigenin absorption spectra partially overlap with the CFP excitation spectrum. Absorption spectra of 100 µM apigenin and 100 µM naringenin were measured using a spectrofluorometer plate reader over the 285-475 nm range. The CFP and YFP excitation wavelengths were obtained from the literature [428]. B-C. Purified CFP (B) or YFP (C) were incubated with increasing concentrations (0, 1, 5, 10, 25, 50, and 100 µM) of apigenin or naringenin. Relative fluorescence units (RFU) were determined by spectrofluorometry and represented as emission spectra. CFP *ext: 405 nm, CFP *emi: 460-600 nm; YFP *ext: 475 nm, YFP *emi: 510-600 nm. Adapted from Arango et al 2013, PNAS [427].

223 Figure 6.15. Flavonoid structural relationship provided by the FRET -based flavonoid nanosensor. A-F. The affinity-purified FLIP2-3-hnRNPA2264-341 protein was incubated with increasing concentrations of the indicated flavonoids (0, 1, 5, 10, 25, 50, and 100 µM) for 3 h at 37°C. RFU were determined by spectrofluorometry (*ext: 405 nm; *emi: 460-600 nm) and represented as emission spectra. The calculated YFP/CFP fluorescent ratios (530 nm/480 nm) are also represented over the 0-100 µM concentration range for each flavonoid. The chemical structures of the corresponding flavonoids are shown on the right. Data represent the mean ± SEM, n = 3. A. Flavones: luteolin and chrysoeriol. B. Apigenin glucosides: apigenin 7-O glucoside and apigenin 6-C glucoside. C. Flavopiridol. D. Flavanones: naringenin and eriodictyol. E. Flavonols: quercetin and kaempferol. F. Genistein. Statistical significance of the variation of the observed YFP/CFP ratios over the tested flavonoid concentration range was conducted by one-way ANOVA. Black curves represent p < 0.05, and broken gray curves represent p + 0.05. In collaboration with Mr. Bledi Brahimaj. Adapted from Arango et al 2013, PNAS [427].

224 Figure 6.16. Apigenin affects hnRNPA2 dimerization. A. Schematic representation of ALPHA as described in section 2.28. B. Purified 6xHis-hnRNPA2 (125 nM) protein was incubated with either 125 nM native GST-hnRNPA2, 125 nM boiled GST-hnRNPA2 or 125 nM GST for 1 h at RT, followed by addition of GSH-donor and anti-His-acceptor beads for 6 h. Apigenin (100 µM), naringenin (100 µM) or diluent control (DMSO) were added for 15 min at RT. Data represent the mean ± SEM. n = 4. *p < 0.05. Adapted from Arango et al 2013, PNAS [427].

225 Figure 6.17. Expression of hnRNPA2 in breast epithelial cells. MDA-MB-231 breast cancer epithelial cells and MCF-10A non-carcinogenic epithelial breast cells were treated with 50 µM apigenin or diluent DMSO control for 48 h. A. Cell lysates were resolved by SDS-PAGE and immunoblotted with anti-hnRNPA2, anti-hnRNPA1 and GAPDH antibodies. A. hnRNPA2 expression was evaluated by qRT-PCR and normalized to the expression of GAPDH. Data represents mean ± SEM. n = 4. Adapted from Arango et al 2013, PNAS [427].

226 Figure 6.18. Apigenin regulates alternative splicing of hnRNPA2 substrates in breast cancer cells. A-C. MDA-MB-231 breast cancer cells were treated with 50 µM apigenin (Api), 50 µM luteolin (Lut), 50 µM naringenin (Nar) or diluent DMSO for 48 h. Total RNA was isolated and alternative splicing was analyzed by RT-PCR using exon specific primers for cFLIP (A), caspase-9 (B), and BIRC5 (C). GAPDH expression was used as loading control. Reactions were resolved in 2 % agarose gels. Splicing variants (boxes) are represented schematically on the right. (D-F) Graphs represent the percent- splice-index (PSI, %) of the indicated splice isoform. Data represent the mean ± SEM. n = 3. * p < 0.05. One-way ANOVA. Adapted from Arango et al 2013, PNAS [427].

227 Figure 6.19. Apigenin regulates splicing genome-wide. A. RNA-seq was performed in MDA-MB-231 cells treated with 50 µM apigenin or diluent DMSO for 48 h. Changes in mRNA processing were analyzed using MISO taking into consideration “isoform- centric” (green) and “exon-centric” (red) analyses. The numbers represent the genes in which splicing was altered by apigenin. B. Lineal regression of the &% determined by RT-PCR and MISO analyses of RNA-Seq. Each dot indicates a different gene. Data represent mean of n=3. C. The broken pies indicate the percentage of each mRNA processing category identified in MDA-MB-231 cells treated with DMSO (80,053 events) or apigenin (82,267 events), or affected by apigenin treatment (Api vs. DMSO, 1,437 events). Enrichment of apigenin-affected events in each category was determined by the hypergeometric distribution and are represented by white numbers = p < 0.05). In collaboration with Mrs. Katherine Mejia-Guerra and Mr. Francisco Padilla-Obregon.

228 Figure 6.20. Apigenin regulates the splicing of RNA-binding proteins. Genes in which apigenin modulated mRNA processing (1,600 genes) were analyzed based on biological function categories using Ingenuity Pathway Analysis (IPA). The x-axis indicate the –log10(p-value) and red line indicates a threshold of p = 0.05 [- log10(0.05)=1.3]. The 626 genes on the right represent unique molecules corresponding to the “gene expression”, “cell death” and “protein synthesis” biological functions. B. MDA-MB-231 breast cancer cells were treated with 50 µM apigenin (Api) or diluent DMSO for 48 h. Total RNA was isolated and alternative splicing analyzed by RT-PCR using exon specific primers (Table 2.3). PCR products were resolved by electrophoresis in 1 % agarose gels. Splicing events are schematically represented on the top. Blue lines indicate the predominant events in cells treated with DMSO, while red lines indicates the predominant event in apigenin-treated cells. Graphs represent the PSI (%) of the variants indicated by the red arrow. Data represent the mean ± SEM. n = 3. * p < 0.05. Two-tailed t-test.

229 Figure 6.21. Apigenin regulates survival and apoptotic signaling pathways through alternative splicing. Genes in which apigenin modulated AS (1,600 genes) were analyzed based on gene ontology using IPA. The x-axis indicate the –log10(p-value) and red line indicates a threshold of p = 0.05 [-log10(0.05)=1.3]. B. The bar graphs displays GO analysis compared to signaling pathways. A. Enrichment compared to annotated cancer driver genes, tumor suppressors or oncogenes. C. Simplified schematic representation of the mTOR, PI3K/AKT, p53 and apoptosis signaling pathways obtained form IPA analysis. Shown in yellow are genes differentially spliced in apigenin-treated cells.

230 Figure 6.22. Apigenin affects splicing in genes regulating survival and apoptotic pathways. MDA-MB-231 breast cancer cells were treated with 50 µM apigenin (Api) or diluent DMSO for 48 h. Total RNA was isolated and alternative splicing analyzed by RT- PCR using exon specific primers (Table 2.3). PCR products were resolved by electrophoresis in 1-2 % agarose gels. Splicing events are schematically represented on the top. Blue lines indicate the predominant events in cells treated with DMSO, while red lines indicates the predominant event in apigenin-treated cells. A. cFLIP. B. TMED7- TICAM2. C. TSC1. D. BCL2L11. E. NAIP. F. DIABLO. G. BAX. H. HIF1A. I. HRAS. Graphs represent the PSI (%) of the variants indicated by the red arrow. Data represent the mean ± SEM. n = 3. * p < 0.05. Two-tailed t-test.

231 Figure 6.23. Modulation of splicing by apigenin targets. A. The 1,600 genes in which apigenin modulated mRNA processing were compared to the known substrates of hnRNPA2, MSI2 and CELF1 [369, 390, 391]. Gene enrichment was determined by the hypergeometric distribution. The x-axis indicates the –log10(p-value) and the red line indicates a threshold of p = 0.05 [-log10(0.05)=1.3]. The 480 genes on the right indicate the number of unique genes on the three categories. B. Network analysis generated using the software Cytoscape of the apigenin targets (shown in red) and apigenin-affected genes involved in cell death (yellow), as determined by IPA analysis. Orange indicates validated splicing changes using RT-PCR (Fig. 6.22). 232

Figure 6.24. Dysregulated mRNA processing in TNBC. A. Public available RNA-Seq data sets from six human biopsies from breast tissue of TNBC patients and three samples from breast tissue of healthy donors (normal breast tissue, NBT) were analyzed using MISO, as shown in Fig. 6.19C. The broken pie graphs indicate the percentage of each mRNA processing category identified in NBT (86,378 events) or TNBC (88,582 events), or dysregulated in TNBC compared to NBT (TNBC vs. NBT, 1,097 events). Enrichment of apigenin-affected events in each category was determined by the hypergeometric distribution (right, white numbers = p < 0.05). The 630 genes with aberrant splicing in TNBC compared to NBT were analyzed based on biological function categories (B) or signaling pathways (C) using IPA. The x-axis indicate the –log10(p-value) and red line indicates a threshold of p = 0.05 [-log10(0.05)=1.3]. In collaboration with Mr. Eric Mukundi.

233 Figure 6.25. Apigenin affects the splicing of genes dysregulated in TNBC. A. The 630 genes showing aberrant splicing in TNBC (right) were compared to the 801 genes in which splicing was affected by apigenin (left) in MDA-MB-231 cells, as determined using MISO exon-centric analysis. From the common genes (orange), those changed by apigenin to splicing profiles found in NBT are indicated in purple, while the events affected by apigenin that did not resemble normal profiles are shown in light blue. B. The broken pie indicates the percentage of each mRNA processing category observed in 84 events altered in TNBC compared to NBT that were also affected in apigenin treated MDA-MB-231 cells. C-D. Splicing was evaluated by RT-PCR using exon specific primers in MDA-MB-231 or MCF-10A cells treated with 50 µM apigenin or diluent DMSO for 48 h. PCR products were resolved by electrophoresis in 1% agarose gels. Splicing events are schematically represented on the top. Blue lines indicate the predominant events in cells treated with DMSO, while red lines indicates the predominant event in apigenin-treated cells. Graphs represent the PSI (%) of the variants indicated by the red arrow. Data represent the mean ± SEM. n = 3. * p < 0.05. One-way ANOVA.

234 Figure 6.26. Effect of apigenin in gene expression in breast cancer. A. Changes in gene expression were evaluated using the DESeq software in MDA-MB-231 cells treated with 25 µM apigenin or DMSO for 48 h. Venn’s diagram representing the 2,122 differentially expressed genes in apigenin-treated cells (green) compared to the 1,600 genes in which apigenin affected mRNA processing (orange). The 2,122 differentially expressed genes in apigenin treated cells were analyzed using IPA based on their biological functional categories (B) or signaling pathways (C). The 253 genes modulated by apigenin commonly in gene expression and splicing (brown) were analyzed based on biological function categories (D) and signaling pathways (E). In collaboration with Mrs. Katherine Mejia-Guerra.

235 Figure 6.27. Apigenin modulates the expression of genes dysregulated in TNBC. A. The 3,816 genes differently expressed in TNBC were compared with the 2,122 affected in apigenin-treated cells. From the common genes, those changed by apigenin to levels found in NBT are represented in purple, while the genes affected by apigenin but did not resemble normal levels are shown in light blue. C. Signaling pathway analysis in the 346 genes restored by apigenin. D. Signaling pathway analysis of the 194 genes affected by apigenin that did not resemble normal levels. E. Gene expression was evaluated by qRT- PCR in MDA-MB-231 or MCF-10A cells treated with 50 µM apigenin (+) or diluent DMSO (-) control for 48 h. Mean ± SEM, n=4. * p < 0.05. Two-tailed t-test.

236 Figure 6.28. Apigenin induces apoptosis and decreases proliferation in triple negative breast cancer cells. MDA-MB-231 and MCF-10A cells were treated with 50 µM apigenin (+) or diluent DMSO control (-) for 48 h. A. Percentage of cell proliferation was evaluated by the MTT assay. B. MDA-MB-231 cells were stained with propidium iodide (PI) and cell cycle was analyzed by flow cytometry. Graphs are representative of n=3. C. Percentage of cells in each phase of cell cycle was quantified from data in (B). D. Percentage of apoptotic cells was assessed by calcein A/M and propidium iodide staining and visualized under the fluorescence microscope. E. Pictures are representative of (D) for MDA-MB-231 cells. Scale bar: 100 µm. F. Caspase-3 activity was evaluated in MDA-MB-231 cells by the DEVD-AFC assay. Data represent mean ± SEM, n = 3, * p < 0.05. Two-tailed t-test.

237 Figure 6.29. Apigenin modulates gene expression and splicing in vivo. SCID female mice were injected with 1 x 106 human epithelial breast cancer MDA-MB-231 cells into the mammary fat pad and 24 h later administered vehicle or 25 mg/kg apigenin daily by i.p. for 28 days. A. Tumor length (l) and width (w) were measured three times a week using a caliper and tumor volume was calculated with the formula w x l2/2. Mean ± SEM, n = 12. * p < 0.05. B. Pictures representative of four tumors from mice receiving vehicle (V1, V2, V3 and V4) or apigenin (A1, A2, A3, and A4) collected at day 28. IHC analysis of tumor sections stained with anti-Ki67 antibodies (C) or by TUNEL (D) in mice receiving apigenin or apigenin for 28 days. Mean ± SEM, n = 10. * p < 0.05. Scale bar: 100 µm. Continued

238 Figure 6.29 continued

E. Gene expression was evaluated by qRT-PCR using primers for human CCNB1 and human CCNA2 (Table 2.2), in RNA from tumors of mice receiving vehicle or apigenin for 28 days. F. Splicing was assessed by RT-PCR using human specific primers for cFLIP and BCL2L11 (Table 2.3) in same sample from (E). PCR products were resolved agarose gels and quantified by densitometry. Graphs represent the PSI (Ψ) of the variants indicated by the red arrow. For E-F, bar graphs represent mean ± SEM, n = 7. * p < 0.05.

239

Table 6.1. Summary of reads obtained by PD-seq

In-frame Library Total Reads Filtered Reads Aligned Reads Reads Original library 4,859,548 873,845 661,351 545,038 Input1 7,817,706 1,738,992 1,332,165 1,102,007 C-E1 7,318,918 1,704,578 1,135,646 761,372 A-E1 5,755,257 1,721,367 1,149,144 751,676 Input2 3,863,523 277,984 208,054 172,963 C-E2 6,250,021 610,713 320,373 176,654 A-E2 10,694,740 1,401,628 508,012 313,289 Total 46,559,713 8,329,107 5,314,745 3,822,999

*The Filtered Reads column corresponds to the number of reads minus the MKET clone contaminant and reads with no insert (MCS reads, Fig. 6.2).

240

Table 6.2. Summary of MCS and MKET reads

Library Total Reads MCS Reads* MKET Reads Original library 4,859,548 1,014,536 2,971,167 Input1 7,817,706 1,438,079 4,640,635 C-E1 7,318,918 1,637,884 3,976,456 A-E1 5,755,257 1,231,065 2,802,825 Input2 3,863,523 769,109 2,816,430 C-E2 6,250,021 1,485,813 4,153,495 A-E2 10,694,740 1,871,830 7,407,376 Total 46,559,713 9,448,316 28,768,384

* Multicloning site (No insert).

241

Table 6.3. Identified Apigenin Targets

ENSG Number Gene Name p-value* GO† ENSG Number Gene Name p-value* GO† ENSG00000122566 HNRNPA2B1 <1E-300 c ENSG00000056736 IL17RB 7.11E-15 b,c ENSG00000125351 UPF3B <1E-300 c, d ENSG00000187800 PEAR1 5.68E-14 b,d ENSG00000121236 TRIM34 <1E-300 c ENSG00000212882 AL139010.29 5.68E-14 b,d ENSG00000109906 ZBTB16 <1E-300 c ENSG00000178996 SNX18 1.14E-13 b,c,d ENSG00000143476 DTL <1E-300 b,c,d ENSG00000085433 WDR47 2.12E-13 c ENSG00000109132 PHOX2B 1.51E-273 d ENSG00000151303 AGAP11 2.27E-13 a ENSG00000197768 KRT17 3.81E-271 d ENSG00000184935 AC090510.4 9.09E-13 d ENSG00000168038 ULK4 3.89E-261 d ENSG00000116176 TPSG1 9.09E-13 b ENSG00000116544 DLGAP3 1.83E-245 b,d ENSG00000176009 ASCL3 9.09E-13 b ENSG00000105643 ARRDC2 1.88E-183 c ENSG00000142319 SLC6A3 9.09E-13 b,d ENSG00000137766 UNC13C 7.04E-133 b,d ENSG00000101489 CELF4 1.82E-12 c ENSG00000214773 AC112512.6 1.51E-123 d ENSG00000008311 AASS 1.82E-12 b ENSG00000213689 TREX1 6.34E-117 c,d ENSG00000185499 MUC1 3.64E-12 b,c,d ENSG00000187546 TMEM195 2.30E-110 b,d ENSG00000020256 ZFP64 1.64E-11 c,d ENSG00000076928 ARHGEF1 2.92E-107 a,b,c,d ENSG00000107262 BAG1 5.82E-11 c ENSG00000170832 USP32 2.15E-92 b,d ENSG00000166948 TGM6 1.16E-10 c ENSG00000134668 SPOCD1 3.14E-89 c ENSG00000063761 ADCK1 1.16E-10 c,d ENSG00000140279 DUOX2 3.14E-89 b ENSG00000011451 WIZ 4.66E-10 c,d ENSG00000137857 DUOX1 3.14E-89 b ENSG00000100441 KHNYN 4.66E-10 d ENSG00000169126 ARMC4 3.98E-89 d ENSG00000114859 CLCN2 4.66E-10 b,c,d ENSG00000103313 MEFV 6.28E-89 c ENSG00000110057 UNC93B1 9.31E-10 b,d ENSG00000182175 RGMA 8.43E-81 b,c ENSG00000110975 SYT10 2.84E-09 b ENSG00000008197 TFAP2D 2.16E-78 d ENSG00000123178 C13orf1 3.73E-09 c ENSG00000090920 FCGBP 2.76E-76 b ENSG00000138442 WDR12 3.73E-09 d ENSG00000159200 DSCR1 1.77E-74 c ENSG00000182077 PTCHD3 3.73E-09 b ENSG00000203989 RHOXF2B 7.37E-70 d ENSG00000157193 LRP8 7.45E-09 b,c,d ENSG00000131721 RHOXF2 7.37E-70 d ENSG00000175471 MCTP1 1.49E-08 b,c ENSG00000204620 AF196972.1 1.16E-69 d ENSG00000105245 NUMBL 7.45E-09 d ENSG00000058668 ATP2B4 1.11E-55 b,c,d ENSG00000112273 HDGFL1 4.88E-04 d ENSG00000080031 PTPRH 1.42E-38 b,c,d ENSG00000147454 SLC25A37 1.49E-08 b,c,d ENSG00000186517 ARHGAP30 4.46E-38 a,c,d ENSG00000213380 COG8 1.49E-08 b,d ENSG00000179542 SLITRK4 3.76E-37 b ENSG00000146950 SHROOM2 2.98E-08 b,d ENSG00000107897 ACBD5 4.93E-32 b,c,d ENSG00000122034 GTF3A 2.98E-08 c,d ENSG00000117385 LEPRE1 2.64E-30 c ENSG00000106484 MEST 1.19E-07 b,c ENSG00000035928 RFC1 2.52E-29 c,d ENSG00000119812 FAM98A 1.19E-07 d ENSG00000181295 AL031289.1 8.08E-28 d ENSG00000112701 SENP6 2.38E-07 c,d ENSG00000119929 CUTC 2.58E-26 d ENSG00000102452 VGCNL1 4.63E-07 b,c ENSG00000154914 USP43 5.17E-26 c,d ENSG00000145934 ODZ2 4.77E-07 b,d ENSG00000116991 SIPA1L2 8.27E-25 a,c,d ENSG00000144354 CDCA7 4.77E-07 c ENSG00000160007 GRLF1 6.62E-24 a,c ENSG00000165185 KIAA1958 4.77E-07 c,d ENSG00000103197 TSC2 1.32E-23 a,b,c,d ENSG00000159784 FAM131B 4.77E-07 d ENSG00000196440 ARMCX4 1.32E-23 d ENSG00000203943 SAMD13 3.81E-06 c ENSG00000074621 SLC24A1 3.43E-23 b,c,d ENSG00000166845 C18orf54 3.81E-06 c ENSG00000139223 ANP32D 1.06E-22 d ENSG00000127951 FGL2 3.81E-06 d ENSG00000107937 GTPBP4 4.02E-21 d ENSG00000215811 BTNL10 3.81E-06 b ENSG00000086200 IPO11 2.50E-20 d ENSG00000155984 TMEM185A 7.63E-06 b ENSG00000118096 IFT46 3.09E-20 d ENSG00000150938 CRIM1 7.63E-06 b,d ENSG00000033327 GAB2 9.43E-20 b,c,d ENSG00000026036 TNFRSF6B 1.53E-05 c,d ENSG00000092820 VIL2 1.20E-19 b ENSG00000138735 PDE5A 3.05E-05 c,d ENSG00000145861 C1QTNF2 4.34E-19 b ENSG00000087087 SRRT 3.05E-05 d ENSG00000149187 CELF1 4.34E-19 c ENSG00000168495 POLR3D 3.05E-05 d ENSG00000139842 CUL4A 2.17E-18 c,d ENSG00000134490 TMEM241 3.05E-05 b,c ENSG00000141424 SLC39A6 3.77E-18 b,c,d ENSG00000165125 TRPV6 3.05E-05 b,c,d ENSG00000181381 DDX60L 1.39E-17 c ENSG00000160293 VAV2 3.05E-05 a,c,d ENSG00000166341 DCHS1 2.78E-17 b ENSG00000101096 NFATC2 3.05E-05 c,d ENSG00000109814 UGDH 1.11E-16 d ENSG00000166471 TMEM41B 3.05E-05 b,c,d ENSG00000153902 LGI4 2.22E-16 c ENSG00000182261 NLRP10 4.08E-05 d ENSG00000115850 LCT 8.88E-16 b ENSG00000101220 C20orf27 6.10E-05 d ENSG00000107821 KAZALD1 3.55E-15 c ENSG00000126768 TIMM17B 6.10E-05 b,d

continued

242

Table 6.3 continued

ENSG Number Gene Name p-value* GO† ENSG Number Gene Name p-value* GO† ENSG00000010704 HFE 2.44E-04 b,c,d ENSG00000102858 MGRN1 1.95E-03 c,d ENSG00000113448 PDE4D 2.44E-04 b,c,d ENSG00000067057 PFKP 1.95E-03 d ENSG00000115290 GRB14 2.44E-04 b,d ENSG00000178952 TUFM 1.95E-03 d ENSG00000126803 HSPA2 2.44E-04 b ENSG00000181036 FCRL6 1.95E-03 b,c,d ENSG00000184650 ODF4 2.44E-04 b ENSG00000177034 MTX3 1.95E-03 b,c,d ENSG00000115275 MOGS 2.44E-04 b ENSG00000171160 MORN4 1.95E-03 c ENSG00000196792 STRN3 2.44E-04 b,c,d ENSG00000092847 AGO1 1.95E-03 c ENSG00000183783 KCTD8 2.44E-04 d ENSG00000162981 FAM84A 1.95E-03 a,c ENSG00000196381 ZNF781 2.44E-04 c ENSG00000185686 PRAME 1.95E-03 b ENSG00000187621 TCL6 2.44E-04 d ENSG00000152894 PTPRK 1.95E-03 b,c,d ENSG00000170734 POLH 2.44E-04 c,d ENSG00000124279 FASTKD3 1.95E-03 d ENSG00000105613 MAST1 2.44E-04 b,d ENSG00000112112 COL11A2 3.91E-03 c ENSG00000153944 MSI2 2.44E-04 c,d ENSG00000067829 IDH3G 7.39E-03 d ENSG00000124260 MAGEA10 4.88E-04 d ENSG00000151989 C2orf21 7.81E-03 b,c ENSG00000198216 CACNA1E 9.77E-04 b,c,d ENSG00000157551 KCNJ15 7.81E-03 b,d ENSG00000205334 AC074091.13 9.77E-04 d ENSG00000174307 PHLDA3 7.81E-03 b ENSG00000002016 RAD52 9.77E-04 e,d ENSG00000117984 CTSD 7.81E-03 d ENSG00000088756 ARHGAP28 9.77E-04 a,c,d ENSG00000157093 LYZL4 7.81E-03 d ENSG00000072182 ACCN4 9.77E-04 b,c,d ENSG00000176700 SCAND2 7.81E-03 c ENSG00000151693 DDEF2 1.95E-03 a,b,c ENSG00000145242 EPHA5 7.81E-03 b,c,d ENSG00000120251 GRIA2 1.95E-03 b,c,d ENSG00000155975 VPS37A 1.56E-02 b,c,d

*p-value of Log10(nICPGInput2) compared to Log10(nICPGE2-A) †Gene Ontology: a. GTPase activation, b. Membrane, c. Alternative Splicing, d. Others. Grey: Validated Apigenin Targets

243

Table 6.4. Biding affinities of different flavonoids to the FRET nanosensor

p-valuea 2 Flavonoid (one-way ANOVA) R KD (µM) Apigenin 0.00012 0.873 22.99 ± 7.70 Luteolin 0.00031 0.880 131 ± 78.89 Quercetin 0.00002 0.916 126.6 ± 60.18 Kaempferol 0.01873 0.815 27.13 ± 12.24 6-C-Apigenin 0.91532 0.897 60.88 ± 24.14 7-O-Apigenin 0.73034 0.468 N.B.* Chrysoeriol 0.10702 0.213 N.B. Naringenin 0.10737 0.120 N.B. Eriodyctiol 0.96653 0.233 N.B. Genistein 0.03077 0.037 N.B. Flavopiridol 0.99484 0.503 N.B. aStatistical significance of the YFP/CFP ratios over the tested flavonoid concentration range. R2: Coefficient of determination of saturation curves fitted by non-linear regression (see Materials and Methods). Kd: Dissociation constant. Data Represents Mean ± SEM, n = 3

244

Table 6.5. Summary of reads obtained by RNA-seq

Library Total Reads Aligned Reads % Alignment Apigenin-1 36,914,908 33,828,822 91,64 Apigenin-2 31,697,261 29,228,044 92.21 DMSO-1 45,603,321 43,122,500 94.56 DMSO-2 34,350,468 32,512,718 94.65 Total 148,565,958 138,692,084 93.35

245

Table 6.6. List of genes in which apigenin restored splicing to profiles observed in NBT

Ensembl Gene ID Gene Symbol Gene Name GO* ENSG00000221978 CCNL2 cyclin L2 a, b ENSG00000249437 NAIP NLR Family, Apoptosis Inhibitory Protein a ENSG00000140464 PML promyelocytic leukemia a, b ENSG00000151914 DST dystonin a ENSG00000196924 FLNA filamin A, alpha (actin binding protein 280) a ENSG00000105401 CDC37 cell division cycle 37 homolog b ENSG00000161692 DBF4B DBF4 homolog B (S. cerevisiae) b ENSG00000263001 GTF2I General Transcription Factor Iii b, d ENSG00000110066 SUV420H1 suppressor of variegation 4-20 homolog 1 b ENSG00000197323 TRIM33 tripartite motif-containing 33 b ENSG00000092439 TRPM7 transient receptor potential, subfamily M, member 7 b ENSG00000122566 HNRNPA2B1 heterogeneous nuclear ribonucleoprotein A2/B1 c ENSG00000239306 RBM14 RNA binding motif 14 c ENSG00000122406 RPL5 ribosomal protein L5 c ENSG00000138326 RPS24 ribosomal protein S24 c ENSG00000013441 CLK1 CDC-like kinase 1 c ENSG00000100201 DDX17 DEAD (Asp-Glu-Ala-Asp) box polypeptide 17 c ENSG00000123136 DDX39A DEAD (Asp-Glu-Ala-Asp) box polypeptide 39A c ENSG00000198563 DDX39B DEAD (Asp-Glu-Ala-Asp) box polypeptide 39B c ENSG00000198492 YTHDF2 YTH domain family, member 2 c ENSG00000114416 FXR1 fragile X mental retardation c ENSG00000185658 BRWD1 bromodomain and WD repeat domain containing 1 d ENSG00000109118 PHF12 PHD finger protein 12 d ENSG00000085415 SEH1L SEH1-like d ENSG00000196693 ZNF33B zinc finger protein 33B d ENSG00000215421 ZNF407 zinc finger protein 407 d ENSG00000188227 ZNF793 zinc finger protein 793 d ENSG00000166130 IKBIP IKK interacting protein e ENSG00000132680 KIAA0907 KIAA0907 e ENSG00000126214 KLC1 kinesin light chain 1 e ENSG00000124786 SLC35B3 solute carrier family 35, member B3 e ENSG00000119801 YPEL5 yippee-like 5 (Drosophila) e ENSG00000176731 C8orf59 C8orf59 e

* GO: a. Cell death; b. Proliferation; c. RNA processing; d. Gene expression. e. other.

246

Chapter 7

Conclusions and Further Directions

This research aimed to study the underlying molecular mechanism for the immune-

modulatory and anti-carcinogenic activities of the nutraceutical apigenin. Towards this

aim, we developed a new genome-wide approach that couples phage display with next

generation sequencing (PD-seq) to identify the direct cellular targets of apigenin,

revealing unexpected novel mechanisms of action. Importantly, the nature of these

findings provided a paradigm shifting in the field, as we showed that dietary

phytochemicals influence the systems network by impacting multiple (hundreds) cellular

targets with moderate affinity, unlike pharmaceutical drugs selected to have high affinity

and specificity. Thus, the effect of a dietary phytochemical would be distributed across

the entire network (Fig. 7.1A), resulting in a fine-tuning effect, with a consequent impact

on human health. Physiologically, we found that apigenin has anti-inflammatory and anti-

carcinogenic activity by immune-modulating monocytes and macrophage activation

during inflammation and cancer. Moreover, apigenin interfered the cancer cell/macrophage cross communication inducing apoptosis in malignant and immune cells, which results in decreased tumor growth and metastasis (Fig. 7.1B).

247

Our results showed that apigenin induces DNA strand breaks in leukemia, promoting a DNA damage response mediated by H2AX and ATM in a PKCδ-dependent pathway but independent on ROS production (Chapter 3). These findings contrast former publications where the activity of flavonoids was attributed mainly to their ability to modulate ROS [234-236, 429]. In this study, we demonstrated that apigenin exerts its biological activity by interacting with multiple proteins. The ability of apigenin to induce

DNA damage is not limited to leukemia but was also observed in glioblastoma and lung cancer cells [230, 430]. Flavonoids inhibited topoisomerase activity in vitro [227, 228].

Yet, whether apigenin promotes DNA damage in cells by inhibiting topoisomerases is not

known. We found that apigenin-induced DNA damage is dependent on PKCδ (Chapter

3). Previous reports showed that PKCδ phosphorylates and activates topoisomerase

activity in response to DNA damage [237]. In addition, pharmacological inhibition of

PKCδ abrogated DNA damage induced by etoposide [222], a topoisomerase inhibitor.

Hence, PKCδ is a key player in the induction of DNA damage by topoisomerase

inhibitors. Therefore, a mechanism can be proposed in which apigenin-induced PKCδ

activity affects topoisomerase activity promoting DNA strand breaks and apoptosis.

The identification of apigenin targets suggests that DNA damage might result from a

direct interaction with DNA repair proteins. Among the 160 targets identified, six are implicated in DNA repair i.e. DTL (Denticleless E3 ubiquitin ligase), CUL4 (Culin 4A),

RAD52 (DNA repair protein 52), POLH (DNA polymerase eta), RFC1 (Replication factor C1) and TREX1 (Three prime repair exonuclease 1). DTL and CUL4 are part of an ubiquitin ligase complex that regulates genome instability [431]. RAD52 is a

248

recombinase essential for homologous recombination repair [432]. POLH replicates DNA

when pyrimidine dimers are present, thereby bypassing the lesion [433]. RFC1 is

responsible for the formation of replication forks during DNA replication and repair, and

its silencing in HeLa epithelial cells stalls the fork ultimately leading to DNA stand

breaks [434]. Moreover, TREX1 expression is increased upon DNA damage and its silencing in human and mouse fibroblast induces DNA stand breaks [435, 436]. Thus,

inhibition of DNA repair proteins by apigenin can promote DNA damage (Fig. 7.1A, red,

DNA repair), which potentiated by the activation of PKCδ and a DNA damage response

pathway mediated by ATM, induces apoptosis (Chapter 3, Fig. 3.6). Further experiments

are necessary to validate the interactions of apigenin with its DNA-repair putative targets

and determine the contribution of each of these molecules to apigenin-induced DNA

damage and apoptosis.

Previous work from our group showed that apigenin decreases the transcriptional

activity of NF-κB by inhibiting IKKβ [169], an upstream activator responsible for the phosphorylation of p65-NF-κB subunit and Iκ-Bα [13]. Yet, inhibition of IKKβ by apigenin was indirect [169], suggesting that other upstream players of the pathway are responsible for the effect of apigenin on the IKKβ/NF-κB axis. Among the apigenin targets, ZFP64 (Zinc finger protein 64) and MUC1 were shown to promote the IKKβ/NF-

κB axis activation [176, 400, 437]. ZFP64 directly binds with NF-κB-p65 subunits potentiating p65 recruitment to the promoters of pro-inflammatory cytokines such as

TNFα [437]. However, since ZFP64 works downstream of IKKβ, it seems an unlikely functional target in our system. MUC1 dimers directly interact with both IKKβ and NF-

249

κB promoting the phosphorylation of NF-κB-p65 by IKKβ [176, 400]. Silencing of

MUC1 in breast cancer cells resulted in reduced NF-κB activity [176, 400]. Yet, the interaction of MUC1 with IKKβ/NF-κB during inflammation has not been studied.

Apigenin association with MUC1 inhibited its dimerization [177]. Thus, apigenin may inhibit the binding of MUC1 with IKKβ resulting in NF-κB inhibition (Fig. 7.1A, red,

MUC1). Further experiments are necessary to test this hypothesis.

Unexpectedly, our results revealed that apigenin associated with several RNA binding proteins. Among them, hnRNPA2 a main auto-antigen in autoimmune diseases such as arthritis rheumatoid, induces the expression of pro-inflammatory cytokines in lymph node T cells [438]. While the role of hnRNPA2 in inflammation has yet to be uncovered, our preliminary results showed that silencing of hnRNPA2 decreases NF-κB-p65 phosphorylation in breast cancer cells and LPS-treated macrophages reducing the expression of TNFα during inflammation (Appendix A). These results indicate that hnRNPA2 might regulate the IKKβ/NF-κB axis, but the mechanisms on how this RNA- binding protein promotes NF-κB activity remain unknown. Additional mechanisms can be proposed to explain the role of hnRNPA2 on the immune-modulatory activity of apigenin. First, hnRNPA2 associated with CK2 in HeLa epithelial cell lysates promoting its activity [405]. Since CK2 directly phosphorylates and activates IKKβ and NF-κB

[439-444], it is plausible to speculate that by binding to hnRNPA2, apigenin disrupts the interaction with CK2 affecting IKKβ/NF-κB activation. However, all my attempts to reproduce the binding of hnRNPA2 with CK2 in macrophages have been unsuccessful

(not shown). Thus, whether this is a functional interaction during inflammation is yet to

250

be investigated. Second, hnRNPA2 interacts with the cRel:p50 subunits of NF-κB,

working as a transcriptional co-activator of the NF-κB complex during mitochondrial

stress [445, 446]. Yet, whether this interaction happens during inflammation has not been

shown. In addition, the association of hnRNPA2 with NF-κB components occurs

downstream of IKKβ and therefore seems an unlikely mechanism for the regulation of

the IKKβ/NF-κB axis during inflammation. Finally, apigenin may be affecting the

splicing or mRNA stability of upstream regulators of the IKKβ/NF-κB axis (Fig. 7.1,

yellow). Splicing is an important mechanism for the resolution of inflammation by

increasing anti-inflammatory transcript isoforms that results in decreased inflammation

[447]. Our computational analysis of RNA-seq data found that apigenin affects

alternative transcript isoforms of TICAM2, IKKγ and CK2 (Fig. 7.1, yellow).

Interestingly, specific IKKγ splice isoforms define the utilization of the NF-κB or

interferon response factor (IRFs) pathways during viral infection [448]. Apigenin

increased an IKKγ transcript isoform that is predicted to lack the IKKβ interacting domains and therefore may block NF-κB activation. Moreover, apigenin increased the read-through transcript TMED7-TICAM2, a negative regulator of TLR signaling [374].

Hence, apigenin has the ability to regulate the IKKβ/NF-κB from different flanks.

However, further experiments are necessary to evaluate the effects from interacting with

MUC1 and hnRNPA2 and modulating mRNA processing or stability of upstream regulators of NF-κB that may explain the inhibition of NF-κΒ by apigenin.

Apigenin induces apoptosis in a variety of cancer cell lines [145-149]. We showed

that apigenin has anti-carcinogenic activity by triggering apoptosis in monocytic

251

leukemia, breast cancer cells and macrophages and in circulating monocytes from tumor-

bearing hosts (Chapters 3, 5 and 6). Yet, the underlying mechanisms for the apoptotic

activity of apigenin are still unknown. Among apigenin targets, HSP70, BAG1 and

TNFRSF6B are anti-apoptotic proteins. HSP70 (Heat shock protein 70) is a chaperon protein that interacts with the apoptotic DISC complex inhibiting the activation of

Caspase-8 [449, 450]. BAG1 interacts with BCL2 and enhances its anti-apoptotic activity

[451]. TNFRSF6B (also known as decoy receptor 3) competes for the binding of death ligands inhibiting the extrinsic apoptotic pathway [452, 453]. Overexpression of HSP70,

BAG1 and TNFRSF6B are common in cancer contributing to the apoptotic resistance observed in cancer cells [453-455]. Hence, inhibition of HSP70, BAG1 and TNFRSF6B by apigenin may eliminate cancer cell resistance to apoptosis triggering cell death (Fig.

7.1A, red, HSP70, BAG1, TNFRSF6B). In addition, we also observed that apigenin increased the levels of pro-apoptotic transcript variants in key regulators of the apoptotic and survival pathways including cFLIP, caspase-9, BCL2L11, CCNL2, DIABLO, BAX,

NAIP, TMED7-TICAM2, HRAS and HIF1A, among others (Fig. 7.1A, yellow).

Altogether, these observations imply that by inhibiting the resistance to apoptosis, apigenin can sensitize cancer cells to chemotherapeutic drugs. Indeed, preliminary data from our group indicates that apigenin sensitizes lung cancer cells to TRAIL-induced apoptosis by disrupting the interaction of HSP70 with the DISC complex and increasing the expression of a splice variant that yields higher protein levels of DR5, the TRAIL receptor (not shown). Thus, the use of apigenin in combination with chemotherapeutic drugs can increase their efficacy, even at lower doses, reducing the adverse effect

252

normally observed during cancer treatment [456].

Our results demonstrated that apigenin regulates mRNA processing genome-wide

(Chapter 6). These observations have paramount implications, as abnormal splicing is

central to diseases development. We found that the GO biological function “RNA

processing” was the most significantly enriched category among genes differentially

spliced in breast cancer cells treated with apigenin, consistently with previous studies

showing that hnRNPA2 cross-regulate the splicing of RNA-binding proteins [369]. Two

mRNA processing categories, RI and ALE, were enriched in the splicing events affected

by apigenin. Alternative polyadenylation and cleavage of mRNA generates transcripts

with different 3’ last exons (ALE), a mechanism regulated by SRRT, a candidate target of

apigenin. In addition, mRNA stability modulates the abundance of transcripts with

alternative last exons [370]. AGO1, an apigenin putative target, regulates mRNA stability

by recruiting miRs to their complementary mRNA sequences [414, 415], and hnRNPA2

is a major regulator of mRNA stability genome-wide by binding to AU rich sequences in

the 3’UTR of its transcript substrates [413]. Moreover, the steady state levels of transcripts with RIs are regulated by non-sense mediated mRNA decay or by altering

alternative splicing of newly synthesized transcripts [417]. UPF3B, another candidate

target of apigenin, is a main regulator of non-sense mediated mRNA decay [418], and

silencing of MSI2 induced alterations in RI and SE events in breast cancer cells, however

the underlying mechanism responsible for this observation is not known [351]. Hence,

binding of apigenin to hnRNPA2, AGO1, SRRT, UPF3B or MSI2 may result in

differential abundance of ALE and RI isoforms. Thus, we can propose a mechanism in

253

which apigenin modulates the splicing regulatory network, by directly associating with

RNA-binding proteins and changing the splicing of RNA processing factors (Fig. 7.1A,

red, RNA-binding proteins). This effect translates into changes of splicing in molecules involved in cell death and survival (Fig. 7.1A, yellow).

Apigenin inhibited monocyte/macrophage activation during inflammation and cancer and interfered the cancer cell/macrophage paracrine loop inducing apoptosis in cancer cells and macrophages, which results in decreased tumor growth and metastasis (Fig.

7.1B). The findings that apigenin induced apoptosis in macrophages and in circulating monocytes from cancer-bearing hosts indicates that some of the apoptotic mechanisms observed in cancer cells can be extrapolated to monocyte/macrophages. Here, it is

important to determine whether changes in the transcriptome observed in breast cancer

cells can be extended to other systems. Preliminary data from our lab showed that

apigenin modulates splicing in monocytic leukemia and lung cancer cells (not shown). In

addition, a comparison of the genome-wide analyses of gene expression performed in

breast cancer cells (chapter 6, Fig. 6.26) and monocytic leukemia (Chapter 3, Fig. 3.5)

found that 259 genes, out of 2,122 genes affected by apigenin in breast cancer cells, are

also modulated in monocytic leukemia, suggesting that although some of the apigenin

effects are common to several model systems, others will be tissue/cell specific. For

example, apigenin inhibited NF-κB in lungs but not in spleens (chapter 4), induced cell

death in blood monocytes but not splenic monocytes (chapter 5), and triggered apoptosis

in breast cancer cells (chapter 5 and 6), monocytic leukemia (chapter 3) and blood

monocytes (chapter 6) but not in non-carcinogenic epithelial cells (Chapter 6) or

254

lymphocytes (chapter 5). Hence, I anticipate that the effects of this flavonoid will depend on the intracellular bioavailability of apigenin and the concentration of its targets in different cell types/tissues during special stimulus (cancer or inflammation).

Another significant contribution from these studies was the use of a newly formulated celery-based apigenin-rich diet in mouse models of inflammation and breast cancer demonstrating that this diet, as well as apigenin, have anti-inflammatory and anti- carcinogenic activities by modulating monocyte/macrophage biology and reducing tumor growth and metastasis. These results support the use of functional foods rich in flavonoids as an alternative for the treatment and prevention of inflammatory disorders including sepsis and breast cancer. Clinical trials will be conducted in the near future to evaluate the effectiveness of celery-based apigenin-rich diets in human health.

In summary, apigenin is a multifunctional dietary phytochemical that exerts its biological activity by interacting with multiple proteins of the systems networks restoring homeostasis upon harmful threats such as inflammation and cancer.

255

Figure 7.1. Immune-modulatory and anti-carcinogenic mechanisms of apigenin. A. Apigenin interacts with multiple proteins (red) of the network including splicing factors apoptotic molecules, DNA repair proteins and GTPase exchange factors. In addition, apigenin affects mRNA processing of molecules involved the regulation of cell death and survival signaling pathways (yellow). B. Physiological model of apigenin. Apigenin inhibits monocyte/macrophage activation during inflammation and cancer and interferes the cancer cell/macrophage cross communication inducing apoptosis in cancer cells and macrophages, which results in decreased tumor growth and metastasis 256

References

1. Galland L. 2010. Diet and inflammation. Nutr Clin Pract 25:634-640.

2. Key TJ, Schatzkin A, Willett WC, Allen NE, Spencer EA, Travis RC. 2004. Diet, nutrition and the prevention of cancer. Public Health Nutr 7:187-200.

3. Doulatov S, Notta F, Laurenti E, Dick JE. 2012. Hematopoiesis: a human perspective. Cell Stem Cell 10:120-136.

4. Gonzalez-Mejia ME, Doseff AI. 2009. Regulation of monocytes and macrophages cell fate. Front Biosci (Landmark Ed) 14:2413-2431.

5. Akira S, Takeda K. 2004. Toll-like receptor signalling. Nat Rev Immunol 4:499- 511.

6. Fahy RJ, Doseff AI, Wewers MD. 1999. Spontaneous human monocyte apoptosis utilizes a caspase-3-dependent pathway that is blocked by endotoxin and is independent of caspase-1. J Immunol 163:1755-1762.

7. Kelley TW, Graham MM, Doseff AI, Pomerantz RW, Lau SM, Ostrowski MC, Franke TF, Marsh CB. 1999. Macrophage colony-stimulating factor promotes cell survival through Akt/protein kinase B. J Biol Chem 274:26393-26398.

8. Goyal A, Wang Y, Graham MM, Doseff AI, Bhatt NY, Marsh CB. 2002. Monocyte survival factors induce Akt activation and suppress caspase-3. Am J Respir Cell Mol Biol 26:224-230.

9. Roca H, Varsos ZS, Sud S, Craig MJ, Ying C, Pienta KJ. 2009. CCL2 and interleukin-6 promote survival of human CD11b+ peripheral blood mononuclear cells and induce M2-type macrophage polarization. J Biol Chem 284:34342- 34354.

10. Cole JE, Georgiou E, Monaco C. 2010. The expression and functions of toll-like receptors in atherosclerosis. Mediators Inflamm 2010:393946.

11. Huang Q, Yang J, Lin Y, Walker C, Cheng J, Liu ZG, Su B. 2004. Differential regulation of interleukin 1 receptor and Toll-like receptor signaling by MEKK3. Nat Immunol 5:98-103.

257

12. Takaesu G, Surabhi RM, Park KJ, Ninomiya-Tsuji J, Matsumoto K, Gaynor RB. 2003. TAK1 is critical for IκB kinase-mediated activation of the NF-κB pathway. J Mol Biol 326:105-115.

13. Scheidereit C. 2006. IκB kinase complexes: gateways to NF-κB activation and transcription. Oncogene 25:6685-6705.

14. Lawrence T. 2009. The nuclear factor NF-κB pathway in inflammation. Cold Spring Harb Perspect Biol 1:a001651.

15. Sun SC. 2012. The noncanonical NF-κB pathway. Immunol Rev 246:125-140.

16. Kreuz S, Siegmund D, Scheurich P, Wajant H. 2001. NF-κB inducers upregulate cFLIP, a cycloheximide-sensitive inhibitor of death receptor signaling. Mol Cell Biol 21:3964-3973.

17. Ueda A, Okuda K, Ohno S, Shirai A, Igarashi T, Matsunaga K, Fukushima J, Kawamoto S, Ishigatsubo Y, Okubo T. 1994. NF-κB and Sp1 regulate transcription of the human monocyte chemoattractant protein-1 gene. J Immunol 153:2052-2063.

18. O'Connell RM, Taganov KD, Boldin MP, Cheng G, Baltimore D. 2007. MicroRNA-155 is induced during the macrophage inflammatory response. Proc Natl Acad Sci U S A 104:1604-1609.

19. Bala S, Marcos M, Kodys K, Csak T, Catalano D, Mandrekar P, Szabo G. 2011. Up-regulation of microRNA-155 in macrophages contributes to increased tumor necrosis factor α (TNFα) production via increased mRNA half-life in alcoholic liver disease. J Biol Chem 286:1436-1444.

20. Tili E, Michaille JJ, Cimino A, Costinean S, Dumitru CD, Adair B, Fabbri M, Alder H, Liu CG, Calin GA, Croce CM. 2007. Modulation of miR-155 and miR- 125b levels following lipopolysaccharide/TNF-α stimulation and their possible roles in regulating the response to endotoxin shock. J Immunol 179:5082-5089.

21. Li X, Tian F, Wang F. 2013. Rheumatoid arthritis-associated microRNA-155 targets SOCS1 and upregulates TNF-α and IL-1β in PBMCs. Int J Mol Sci 14:23910-23921.

22. Rajaram MV, Ni B, Morris JD, Brooks MN, Carlson TK, Bakthavachalu B, Schoenberg DR, Torrelles JB, Schlesinger LS. 2011. Mycobacterium tuberculosis lipomannan blocks TNF biosynthesis by regulating macrophage MAPK-activated protein kinase 2 (MK2) and microRNA miR-125b. Proc Natl Acad Sci U S A 108:17408-17413.

258

23. Tak PP, Firestein GS. 2001. NF-κB: a key role in inflammatory diseases. J Clin Invest 107:7-11.

24. Cohen J. 2002. The immunopathogenesis of sepsis. Nature 420:885-891.

25. Elixhauser A, Friedman B, Stranges E. 2006. Septicemia in U.S. Hospitals, 2009: Statistical Brief #122. Healthcare Cost and Utilization Project (HCUP) Statistical Briefs, Rockville (MD)).

26. Harirforoosh S, Asghar W, Jamali F. 2013. Adverse effects of nonsteroidal antiinflammatory drugs: an update of gastrointestinal, cardiovascular and renal complications. J Pharm Pharm Sci 16:821-847.

27. O'Sullivan AW, Wang JH, Redmond HP. 2009. NF-κB and p38 MAPK inhibition improve survival in endotoxin shock and in a cecal ligation and puncture model of sepsis in combination with antibiotic therapy. J Surg Res 152:46-53.

28. Scheibel M, Klein B, Merkle H, Schulz M, Fritsch R, Greten FR, Arkan MC, Schneider G, Schmid RM. 2010. IκBβ is an essential co-activator for LPS- induced IL-1β transcription in vivo. J Exp Med 207:2621-2630.

29. DeNardo DG, Coussens LM. 2007. Inflammation and breast cancer. Balancing immune response: crosstalk between adaptive and innate immune cells during breast cancer progression. Breast Cancer Res 9:212.

30. Ferrara F, Schiffer CA. 2013. Acute myeloid leukaemia in adults. Lancet 381:484-495.

31. WHO. 2014. The top 10 Causes of death. The World Health Organization Facts Sheet N°310.

32. Siegel R, Ma J, Zou Z, Jemal A. 2014. Cancer statistics, 2014. CA Cancer J Clin 64:9-29.

33. Voss OH, Kim S, Wewers MD, Doseff AI. 2005. Regulation of monocyte apoptosis by the protein kinase Cδ-dependent phosphorylation of caspase-3. J Biol Chem 280:17371-17379.

34. Voss OH, Batra S, Kolattukudy SJ, Gonzalez-Mejia ME, Smith JB, Doseff AI. 2007. Binding of caspase-3 prodomain to heat shock protein 27 regulates monocyte apoptosis by inhibiting caspase-3 proteolytic activation. J Biol Chem 282:25088-25099.

35. Lu Y, Chen GQ. 2011. Effector caspases and leukemia. Int J Cell Biol

259

2011:738301.

36. Iida H, Towatari M, Tanimoto M, Morishita Y, Kodera Y, Saito H. 1997. Overexpression of cyclin E in acute myelogenous leukemia. Blood 90:3707-3713. 37. Fulda S. 2009. Tumor resistance to apoptosis. Int J Cancer 124:511-515.

38. Igney FH, Krammer PH. 2002. Death and anti-death: tumour resistance to apoptosis. Nat Rev Cancer 2:277-288.

39. Yang J, Liu X, Bhalla K, Kim CN, Ibrado AM, Cai J, Peng TI, Jones DP, Wang X. 1997. Prevention of apoptosis by Bcl-2: release of cytochrome c from mitochondria blocked. Science 275:1129-1132.

40. Du C, Fang M, Li Y, Li L, Wang X. 2000. Smac, a mitochondrial protein that promotes cytochrome c-dependent caspase activation by eliminating IAP inhibition. Cell 102:33-42.

41. Srinivasula SM, Ahmad M, Fernandes-Alnemri T, Alnemri ES. 1998. Autoactivation of procaspase-9 by Apaf-1-mediated oligomerization. Mol Cell 1:949-957.

42. Shiozaki EN, Chai J, Shi Y. 2002. Oligomerization and activation of caspase-9, induced by Apaf-1 CARD. Proc Natl Acad Sci U S A 99:4197-4202.

43. Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Alnemri ES, Wang X. 1997. Cytochrome c and dATP-dependent formation of Apaf-1/caspase-9 complex initiates an apoptotic protease cascade. Cell 91:479-489.

44. Gyrd-Hansen M, Meier P. 2010. IAPs: from caspase inhibitors to modulators of NF-κB, inflammation and cancer. Nat Rev Cancer 10:561-574.

45. Fernandez-Luna JL. 2008. Regulation of pro-apoptotic BH3-only proteins and its contribution to cancer progression and chemoresistance. Cell Signal 20:1921- 1926.

46. Youle RJ, Strasser A. 2008. The BCL-2 protein family: opposing activities that mediate cell death. Nat Rev Mol Cell Biol 9:47-59.

47. Fulda S, Debatin KM. 2006. Extrinsic versus intrinsic apoptosis pathways in anticancer chemotherapy. Oncogene 25:4798-4811.

48. Muzio M, Stockwell BR, Stennicke HR, Salvesen GS, Dixit VM. 1998. An induced proximity model for caspase-8 activation. J Biol Chem 273:2926-2930.

49. Juo P, Woo MS, Kuo CJ, Signorelli P, Biemann HP, Hannun YA, Blenis J. 1999.

260

FADD is required for multiple signaling events downstream of the receptor Fas. Cell Growth Differ 10:797-804.

50. Scaffidi C, Schmitz I, Krammer PH, Peter ME. 1999. The role of c-FLIP in modulation of CD95-induced apoptosis. J Biol Chem 274:1541-1548.

51. Yi X, Yin XM, Dong Z. 2003. Inhibition of Bid-induced apoptosis by Bcl-2. tBid insertion, Bax translocation, and Bax/Bak oligomerization suppressed. J Biol Chem 278:16992-16999.

52. DeSantis C, Ma J, Bryan L, Jemal A. 2014. Breast cancer statistics, 2013. CA Cancer J Clin 64:52-62.

53. Bombonati A, Sgroi DC. 2011. The molecular pathology of breast cancer progression. J Pathol 223:307-317.

54. Yersal O, Barutca S. 2014. Biological subtypes of breast cancer: Prognostic and therapeutic implications. World J Clin Oncol 5:412-424.

55. Badve S, Dabbs DJ, Schnitt SJ, Baehner FL, Decker T, Eusebi V, Fox SB, Ichihara S, Jacquemier J, Lakhani SR, Palacios J, Rakha EA, Richardson AL, Schmitt FC, Tan PH, Tse GM, Weigelt B, Ellis IO, Reis-Filho JS. 2011. Basal- like and triple-negative breast cancers: a critical review with an emphasis on the implications for pathologists and oncologists. Mod Pathol 24:157-167.

56. Brufsky AM. 2014. Current Approaches and Emerging Directions in HER2- resistant Breast Cancer. Breast Cancer (Auckl) 8:109-118.

57. Kim IS, Baek SH. 2010. Mouse models for breast cancer metastasis. Biochem Biophys Res Commun 394:443-447.

58. Diaz-Cruz ES, Cabrera MC, Nakles R, Rutstein BH, Furth PA. 2010. BRCA1 deficient mouse models to study pathogenesis and therapy of triple negative breast cancer. Breast Dis 32:85-97.

59. Xu X, Wagner KU, Larson D, Weaver Z, Li C, Ried T, Hennighausen L, Wynshaw-Boris A, Deng CX. 1999. Conditional mutation of Brca1 in mammary epithelial cells results in blunted ductal morphogenesis and tumour formation. Nat Genet 22:37-43.

60. Brodie SG, Xu X, Qiao W, Li WM, Cao L, Deng CX. 2001. Multiple genetic changes are associated with mammary tumorigenesis in Brca1 conditional knockout mice. Oncogene 20:7514-7523.

61. Couch FJ, DeShano ML, Blackwood MA, Calzone K, Stopfer J, Campeau L,

261

Ganguly A, Rebbeck T, Weber BL. 1997. BRCA1 mutations in women attending clinics that evaluate the risk of breast cancer. N Engl J Med 336:1409-1415.

62. Stewart TA, Pattengale PK, Leder P. 1984. Spontaneous mammary adenocarcinomas in transgenic mice that carry and express MTV/myc fusion genes. Cell 38:627-637.

63. Muller WJ, Sinn E, Pattengale PK, Wallace R, Leder P. 1988. Single-step induction of mammary adenocarcinoma in transgenic mice bearing the activated c-neu oncogene. Cell 54:105-115.

64. Fantozzi A, Christofori G. 2006. Mouse models of breast cancer metastasis. Breast Cancer Res 8:212.

65. Guy CT, Cardiff RD, Muller WJ. 1992. Induction of mammary tumors by expression of polyomavirus middle T oncogene: a transgenic mouse model for metastatic disease. Mol Cell Biol 12:954-961.

66. Lin EY, Jones JG, Li P, Zhu L, Whitney KD, Muller WJ, Pollard JW. 2003. Progression to malignancy in the polyoma middle T oncoprotein mouse breast cancer model provides a reliable model for human diseases. Am J Pathol 163:2113-2126.

67. Solinas G, Germano G, Mantovani A, Allavena P. 2009. Tumor-associated macrophages (TAM) as major players of the cancer-related inflammation. J Leukoc Biol 86:1065-1073.

68. Gordon S, Taylor PR. 2005. Monocyte and macrophage heterogeneity. Nat Rev Immunol 5:953-964.

69. Murray PJ, Allen JE, Biswas SK, Fisher EA, Gilroy DW, Goerdt S, Gordon S, Hamilton JA, Ivashkiv LB, Lawrence T, Locati M, Mantovani A, Martinez FO, Mege JL, Mosser DM, Natoli G, Saeij JP, Schultze JL, Shirey KA, Sica A, Suttles J, Udalova I, van Ginderachter JA, Vogel SN, Wynn TA. 2014. Macrophage activation and polarization: nomenclature and experimental guidelines. Immunity 41:14-20.

70. Martinez FO. 2011. Regulators of macrophage activation. Eur J Immunol 41:1531-1534.

71. Li M, Knight DA, L AS, Smyth MJ, Stewart TJ. 2013. A role for CCL2 in both tumor progression and immunosurveillance. Oncoimmunology 2:e25474.

72. Yu W, Chen J, Xiong Y, Pixley FJ, Yeung YG, Stanley ER. 2012. Macrophage proliferation is regulated through CSF-1 receptor tyrosines 544, 559, and 807. J

262

Biol Chem 287:13694-13704.

73. Pierce JH, Di Marco E, Cox GW, Lombardi D, Ruggiero M, Varesio L, Wang LM, Choudhury GG, Sakaguchi AY, Di Fiore PP, et al. 1990. Macrophage- colony-stimulating factor (CSF-1) induces proliferation, chemotaxis, and reversible monocytic differentiation in myeloid progenitor cells transfected with the human c-fms/CSF-1 receptor cDNA. Proc Natl Acad Sci U S A 87:5613-5617.

74. Sanchez-Martin L, Estecha A, Samaniego R, Sanchez-Ramon S, Vega MA, Sanchez-Mateos P. 2011. The chemokine CXCL12 regulates monocyte- macrophage differentiation and RUNX3 expression. Blood 117:88-97.

75. Lewis JS, Landers RJ, Underwood JC, Harris AL, Lewis CE. 2000. Expression of vascular endothelial growth factor by macrophages is up-regulated in poorly vascularized areas of breast carcinomas. J Pathol 192:150-158.

76. Qian BZ, Li J, Zhang H, Kitamura T, Zhang J, Campion LR, Kaiser EA, Snyder LA, Pollard JW. 2011. CCL2 recruits inflammatory monocytes to facilitate breast-tumour metastasis. Nature 475:222-225.

77. Boimel PJ, Smirnova T, Zhou ZN, Wyckoff J, Park H, Coniglio SJ, Qian BZ, Stanley ER, Cox D, Pollard JW, Muller WJ, Condeelis J, Segall JE. 2012. Contribution of CXCL12 secretion to invasion of breast cancer cells. Breast Cancer Res 14:R23.

78. Goswami S, Sahai E, Wyckoff JB, Cammer M, Cox D, Pixley FJ, Stanley ER, Segall JE, Condeelis JS. 2005. Macrophages promote the invasion of breast carcinoma cells via a colony-stimulating factor-1/epidermal growth factor paracrine loop. Cancer Res 65:5278-5283.

79. Strachan DC, Ruffell B, Oei Y, Bissell MJ, Coussens LM, Pryer N, Daniel D. 2013. CSF1R inhibition delays cervical and mammary tumor growth in murine models by attenuating the turnover of tumor-associated macrophages and enhancing infiltration by CD8 T cells. Oncoimmunology 2:e26968.

80. Lin EY, Li JF, Gnatovskiy L, Deng Y, Zhu L, Grzesik DA, Qian H, Xue XN, Pollard JW. 2006. Macrophages regulate the angiogenic switch in a mouse model of breast cancer. Cancer Res 66:11238-11246.

81. Franklin RA, Liao W, Sarkar A, Kim MV, Bivona MR, Liu K, Pamer EG, Li MO. 2014. The cellular and molecular origin of tumor-associated macrophages. Science 344:921-925.

82. Luo Y, Zhou H, Krueger J, Kaplan C, Lee SH, Dolman C, Markowitz D, Wu W, Liu C, Reisfeld RA, Xiang R. 2006. Targeting tumor-associated macrophages as a

263

novel strategy against breast cancer. J Clin Invest 116:2132-2141.

83. Fend L, Accart N, Kintz J, Cochin S, Reymann C, Le Pogam F, Marchand JB, Menguy T, Slos P, Rooke R, Fournel S, Bonnefoy JY, Preville X, Haegel H. 2013. Therapeutic effects of anti-CD115 monoclonal antibody in mouse cancer models through dual inhibition of tumor-associated macrophages and osteoclasts. PLoS One 8:e73310.

84. Lin EY, Nguyen AV, Russell RG, Pollard JW. 2001. Colony-stimulating factor 1 promotes progression of mammary tumors to malignancy. J Exp Med 193:727- 740.

85. Lin EY, Li JF, Bricard G, Wang W, Deng Y, Sellers R, Porcelli SA, Pollard JW. 2007. Vascular endothelial growth factor restores delayed tumor progression in tumors depleted of macrophages. Mol Oncol 1:288-302.

86. Mitchem JB, Brennan DJ, Knolhoff BL, Belt BA, Zhu Y, Sanford DE, Belaygorod L, Carpenter D, Collins L, Piwnica-Worms D, Hewitt S, Udupi GM, Gallagher WM, Wegner C, West BL, Wang-Gillam A, Goedegebuure P, Linehan DC, DeNardo DG. 2013. Targeting tumor-infiltrating macrophages decreases tumor-initiating cells, relieves immunosuppression, and improves chemotherapeutic responses. Cancer Res 73:1128-1141.

87. Cortez-Retamozo V, Etzrodt M, Newton A, Rauch PJ, Chudnovskiy A, Berger C, Ryan RJ, Iwamoto Y, Marinelli B, Gorbatov R, Forghani R, Novobrantseva TI, Koteliansky V, Figueiredo JL, Chen JW, Anderson DG, Nahrendorf M, Swirski FK, Weissleder R, Pittet MJ. 2012. Origins of tumor-associated macrophages and neutrophils. Proc Natl Acad Sci U S A 109:2491-2496.

88. Yang J, Zhang L, Yu C, Yang XF, Wang H. 2014. Monocyte and macrophage differentiation: circulation inflammatory monocyte as biomarker for inflammatory diseases. Biomark Res 2:1.

89. Schouppe E, De Baetselier P, Van Ginderachter JA, Sarukhan A. 2012. Instruction of myeloid cells by the tumor microenvironment: Open questions on the dynamics and plasticity of different tumor-associated myeloid cell populations. Oncoimmunology 1:1135-1145.

90. De Veirman K, Van Valckenborgh E, Lahmar Q, Geeraerts X, De Bruyne E, Menu E, Van Riet I, Vanderkerken K, Van Ginderachter JA. 2014. Myeloid- derived suppressor cells as therapeutic target in hematological malignancies. Front Oncol 4:349.

91. Diaz-Montero CM, Salem ML, Nishimura MI, Garrett-Mayer E, Cole DJ, Montero AJ. 2009. Increased circulating myeloid-derived suppressor cells

264

correlate with clinical cancer stage, metastatic tumor burden, and doxorubicin- cyclophosphamide chemotherapy. Cancer Immunol Immunother 58:49-59.

92. Markowitz J, Wesolowski R, Papenfuss T, Brooks TR, Carson WE, 3rd. 2013. Myeloid-derived suppressor cells in breast cancer. Breast Cancer Res Treat 140:13-21.

93. Biswas DK, Shi Q, Baily S, Strickland I, Ghosh S, Pardee AB, Iglehart JD. 2004. NF-κB activation in human breast cancer specimens and its role in cell proliferation and apoptosis. Proc Natl Acad Sci U S A 101:10137-10142.

94. Sovak MA, Bellas RE, Kim DW, Zanieski GJ, Rogers AE, Traish AM, Sonenshein GE. 1997. Aberrant NF-κB/Rel expression and the pathogenesis of breast cancer. J Clin Invest 100:2952-2960.

95. Finco TS, Westwick JK, Norris JL, Beg AA, Der CJ, Baldwin AS, Jr. 1997. Oncogenic Ha-Ras-induced signaling activates NF-κB transcriptional activity, which is required for cellular transformation. J Biol Chem 272:24113-24116.

96. Galang CK, Garcia-Ramirez J, Solski PA, Westwick JK, Der CJ, Neznanov NN, Oshima RG, Hauser CA. 1996. Oncogenic Neu/ErbB-2 increases ets, AP-1, and NF-κB-dependent gene expression, and inhibiting ets activation blocks Neu- mediated cellular transformation. J Biol Chem 271:7992-7998.

97. Romieu-Mourez R, Landesman-Bollag E, Seldin DC, Sonenshein GE. 2002. Protein kinase CK2 promotes aberrant activation of NF-κB, transformed phenotype, and survival of breast cancer cells. Cancer Res 62:6770-6778.

98. Romieu-Mourez R, Landesman-Bollag E, Seldin DC, Traish AM, Mercurio F, Sonenshein GE. 2001. Roles of IKK kinases and protein kinase CK2 in activation of NF-κB in breast cancer. Cancer Res 61:3810-3818.

99. Liu M, Sakamaki T, Casimiro MC, Willmarth NE, Quong AA, Ju X, Ojeifo J, Jiao X, Yeow WS, Katiyar S, Shirley LA, Joyce D, Lisanti MP, Albanese C, Pestell RG. 2010. The canonical NF-κB pathway governs mammary tumorigenesis in transgenic mice and tumor stem cell expansion. Cancer Res 70:10464-10473.

100. Connelly L, Barham W, Onishko HM, Sherrill T, Chodosh LA, Blackwell TS, Yull FE. 2011. Inhibition of NF-κB activity in mammary epithelium increases tumor latency and decreases tumor burden. Oncogene 30:1402-1412.

101. Zhang J, Li Y, Yu M, Chen B, Shen B. 2003. Lineage-dependent NF-κB activation contributes to the resistance of human macrophages to apoptosis. Hematol J 4:277-284.

265

102. Liu H, Ma Y, Pagliari LJ, Perlman H, Yu C, Lin A, Pope RM. 2004. TNF-α- induced apoptosis of macrophages following inhibition of NF-κB: a central role for disruption of mitochondria. J Immunol 172:1907-1915.

103. Mannick EE, Mishra J, Marque J, Clavell M, Miller MJ, Oliver PD. 1997. Inhibitors of NF-κB cause apoptosis in cultured macrophages. Mediators Inflamm 6:225-232.

104. Hagemann T, Lawrence T, McNeish I, Charles KA, Kulbe H, Thompson RG, Robinson SC, Balkwill FR. 2008. "Re-educating" tumor-associated macrophages by targeting NF-κB. J Exp Med 205:1261-1268.

105. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA, Jr., Kinzler KW. 2013. Cancer genome landscapes. Science 339:1546-1558.

106. Stratton MR, Campbell PJ, Futreal PA. 2009. The cancer genome. Nature 458:719-724.

107. Cancer Genome Atlas N. 2012. Comprehensive molecular portraits of human breast tumours. Nature 490:61-70.

108. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS, Thorsen T, Quist H, Matese JC, Brown PO, Botstein D, Lonning PE, Borresen-Dale AL. 2001. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 98:10869-10874.

109. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO, Botstein D. 2000. Molecular portraits of human breast tumours. Nature 406:747-752.

110. Al-Ejeh F, Simpson PT, Sanus JM, Klein K, Kalimutho M, Shi W, Miranda M, Kutasovic J, Raghavendra A, Madore J, Reid L, Krause L, Chenevix-Trench G, Lakhani SR, Khanna KK. 2014. Meta-analysis of the global gene expression profile of triple-negative breast cancer identifies genes for the prognostication and treatment of aggressive breast cancer. Oncogenesis 3:e100.

111. Dutertre M, Vagner S, Auboeuf D. 2010. Alternative splicing and breast cancer. RNA Biol 7:403-411.

112. Venables JP, Klinck R, Bramard A, Inkel L, Dufresne-Martin G, Koh C, Gervais- Bird J, Lapointe E, Froehlich U, Durand M, Gendron D, Brosseau JP, Thibault P, Lucier JF, Tremblay K, Prinos P, Wellinger RJ, Chabot B, Rancourt C, Elela SA.

266

2008. Identification of alternative splicing markers for breast cancer. Cancer Res 68:9525-9531.

113. Shapiro IM, Cheng AW, Flytzanis NC, Balsamo M, Condeelis JS, Oktay MH, Burge CB, Gertler FB. 2011. An EMT-driven alternative splicing program occurs in human breast cancer and modulates cellular phenotype. PLoS Genet 7:e1002218.

114. Lapuk A, Marr H, Jakkula L, Pedro H, Bhattacharya S, Purdom E, Hu Z, Simpson K, Pachter L, Durinck S, Wang N, Parvin B, Fontenay G, Speed T, Garbe J, Stampfer M, Bayandorian H, Dorton S, Clark TA, Schweitzer A, Wyrobek A, Feiler H, Spellman P, Conboy J, Gray JW. 2010. Exon-level microarray analyses identify alternative splicing programs in breast cancer. Mol Cancer Res 8:961- 974.

115. Eswaran J, Horvath A, Godbole S, Reddy SD, Mudvari P, Ohshiro K, Cyanam D, Nair S, Fuqua SA, Polyak K, Florea LD, Kumar R. 2013. RNA sequencing of cancer reveals novel splicing alterations. Sci Rep 3:1689.

116. Dutertre M, Lacroix-Triki M, Driouch K, de la Grange P, Gratadou L, Beck S, Millevoi S, Tazi J, Lidereau R, Vagner S, Auboeuf D. 2010. Exon-based clustering of murine breast tumor transcriptomes reveals alternative exons whose expression is associated with metastasis. Cancer Res 70:896-905.

117. Bemmo A, Dias C, Rose AA, Russo C, Siegel P, Majewski J. 2010. Exon-level transcriptome profiling in murine breast cancer reveals splicing changes specific to tumors with different metastatic abilities. PLoS One 5:e11981.

118. David CJ, Chen M, Assanah M, Canoll P, Manley JL. 2010. HnRNP proteins controlled by c-Myc deregulate pyruvate kinase mRNA splicing in cancer. Nature 463:364-368.

119. David CJ, Manley JL. 2010. Alternative pre-mRNA splicing regulation in cancer: pathways and programs unhinged. Genes Dev 24:2343-2364.

120. Wang ET, Sandberg R, Luo S, Khrebtukova I, Zhang L, Mayr C, Kingsmore SF, Schroth GP, Burge CB. 2008. Alternative isoform regulation in human tissue transcriptomes. Nature 456:470-476.

121. Zhang J, Manley JL. 2013. Misregulation of pre-mRNA alternative splicing in cancer. Cancer Discov 3:1228-1237.

122. Zhou J, Allred DC, Avis I, Martinez A, Vos MD, Smith L, Treston AM, Mulshine JL. 2001. Differential expression of the early lung cancer detection marker, heterogeneous nuclear ribonucleoprotein-A2/B1 (hnRNP-A2/B1) in normal breast

267

and neoplastic breast cancer. Breast Cancer Res Treat 66:217-224.

123. Sueoka E, Sueoka N, Iwanaga K, Sato A, Suga K, Hayashi S, Nagasawa K, Nakachi K. 2005. Detection of plasma hnRNP B1 mRNA, a new cancer biomarker, in lung cancer patients by quantitative real-time polymerase chain reaction. Lung Cancer 48:77-83.

124. Wu S, Sato M, Endo C, Sakurada A, Dong B, Aikawa H, Chen Y, Okada Y, Matsumura Y, Sueoka E, Kondo T. 2003. hnRNP B1 protein may be a possible prognostic factor in squamous cell carcinoma of the lung. Lung Cancer 41:179- 186.

125. Golan-Gerstl R, Cohen M, Shilo A, Suh SS, Bakacs A, Coppola L, Karni R. 2011. Splicing factor hnRNP A2/B1 regulates tumor suppressor gene splicing and is an oncogenic driver in glioblastoma. Cancer Res 71:4464-4472.

126. Miura K, Fujibuchi W, Unno M. 2012. Splice variants in apoptotic pathway. Exp Oncol 34:212-217.

127. Seol DW, Billiar TR. 1999. A caspase-9 variant missing the catalytic site is an endogenous inhibitor of apoptosis. J Biol Chem 274:2072-2076.

128. Krueger A, Schmitz I, Baumann S, Krammer PH, Kirchhoff S. 2001. Cellular FLICE-inhibitory protein splice variants inhibit different steps of caspase-8 activation at the CD95 death-inducing signaling complex. J Biol Chem 276:20633-20640.

129. Piggott L, Omidvar N, Marti Perez S, Eberl M, Clarkson RW. 2011. Suppression of apoptosis inhibitor c-FLIP selectively eliminates breast cancer stem cell activity in response to the anti-cancer agent, TRAIL. Breast Cancer Res 13:R88.

130. Goehe RW, Shultz JC, Murudkar C, Usanovic S, Lamour NF, Massey DH, Zhang L, Camidge DR, Shay JW, Minna JD, Chalfant CE. 2010. hnRNP L regulates the tumorigenic capacity of lung cancer xenografts in mice via caspase-9 pre-mRNA processing. J Clin Invest 120:3923-3939.

131. Havsteen BH. 2002. The biochemistry and medical significance of the flavonoids. Pharmacol Ther 96:67-202.

132. Prasad S, Phromnoi K, Yadav VR, Chaturvedi MM, Aggarwal BB. 2010. Targeting inflammatory pathways by flavonoids for prevention and treatment of cancer. Planta Med 76:1044-1063.

133. Williams RJ, Spencer JP, Rice-Evans C. 2004. Flavonoids: antioxidants or signalling molecules? Free Radic Biol Med 36:838-849.

268

134. Riemersma RA, Rice-Evans CA, Tyrrell RM, Clifford MN, Lean ME. 2001. Tea flavonoids and cardiovascular health. QJM 94:277-282.

135. Crozier A, Jaganath IB, Clifford MN. 2009. Dietary phenolics: chemistry, bioavailability and effects on health. Nat Prod Rep 26:1001-1043.

136. Manach C, Scalbert A, Morand C, Remesy C, Jimenez L. 2004. Polyphenols: food sources and bioavailability. Am J Clin Nutr 79:727-747.

137. McKay DL, Blumberg JB. 2006. A review of the bioactivity and potential health benefits of chamomile tea (Matricaria recutita L.). Phytother Res 20:519-530.

138. Gates MA, Vitonis AF, Tworoger SS, Rosner B, Titus-Ernstoff L, Hankinson SE, Cramer DW. 2009. Flavonoid intake and ovarian cancer risk in a population- based case-control study. Int J Cancer 124:1918-1925.

139. Vargo MA, Voss OH, Poustka F, Cardounel AJ, Grotewold E, Doseff AI. 2006. Apigenin-induced-apoptosis is mediated by the activation of PKCδ and caspases in leukemia cells. Biochem Pharmacol 72:681-692.

140. Pavese JM, Farmer RL, Bergan RC. 2010. Inhibition of cancer cell invasion and metastasis by genistein. Cancer Metastasis Rev 29:465-482.

141. Sak K. 2014. Site-specific anticancer effects of dietary flavonoid quercetin. Nutr Cancer 66:177-193.

142. Tuorkey MJ. 2015. Molecular targets of luteolin in cancer. Eur J Cancer Prev.

143. Gonzalez-Mejia ME, Voss OH, Murnan EJ, Doseff AI. 2010. Apigenin-induced apoptosis of leukemia cells is mediated by a bimodal and differentially regulated residue-specific phosphorylation of heat-shock protein-27. Cell Death Dis 1:e64.

144. Budhraja A, Gao N, Zhang Z, Son YO, Cheng S, Wang X, Ding S, Hitron A, Chen G, Luo J, Shi X. 2012. Apigenin induces apoptosis in human leukemia cells and exhibits anti-leukemic activity in vivo. Mol Cancer Ther 11:132-142.

145. Kim KC, Choi EH, Lee C. 2014. Axl receptor tyrosine kinase is a novel target of apigenin for the inhibition of cell proliferation. Int J Mol Med 34:592-598.

146. Shukla S, Fu P, Gupta S. 2014. Apigenin induces apoptosis by targeting inhibitor of apoptosis proteins and Ku70-Bax interaction in prostate cancer. Apoptosis 19:883-894.

147. Lu HF, Chie YJ, Yang MS, Lu KW, Fu JJ, Yang JS, Chen HY, Hsia TC, Ma CY,

269

Ip SW, Chung JG. 2011. Apigenin induces apoptosis in human lung cancer H460 cells through caspase- and mitochondria-dependent pathways. Hum Exp Toxicol 30:1053-1061.

148. Das S, Das J, Samadder A, Boujedaini N, Khuda-Bukhsh AR. 2012. Apigenin- induced apoptosis in A375 and A549 cells through selective action and dysfunction of mitochondria. Exp Biol Med (Maywood) 237:1433-1448.

149. Cao X, Liu B, Cao W, Zhang W, Zhang F, Zhao H, Meng R, Zhang L, Niu R, Hao X, Zhang B. 2013. Autophagy inhibition enhances apigenin-induced apoptosis in human breast cancer cells. Chin J Cancer Res 25:212-222.

150. Ding J, Polier G, Kohler R, Giaisi M, Krammer PH, Li-Weber M. 2012. Wogonin and related natural flavones overcome tumor necrosis factor-related apoptosis- inducing ligand (TRAIL) protein resistance of tumors by down-regulation of c- FLIP protein and up-regulation of TRAIL receptor 2 expression. J Biol Chem 287:641-649.

151. Oishi M, Iizumi Y, Taniguchi T, Goi W, Miki T, Sakai T. 2013. Apigenin sensitizes prostate cancer cells to Apo2L/TRAIL by targeting adenine nucleotide translocase-2. PLoS One 8:e55922.

152. Kim EY, Yu JS, Yang M, Kim AK. 2013. Sub-toxic dose of apigenin sensitizes HepG2 cells to TRAIL through ERK-dependent up-regulation of TRAIL receptor DR5. Mol Cells 35:32-40.

153. Seo HS, Ju JH, Jang K, Shin I. 2011. Induction of apoptotic cell death by phytoestrogens by up-regulating the levels of phospho-p53 and p21 in normal and malignant estrogen receptor α-negative breast cells. Nutr Res 31:139-146.

154. Shukla S, Gupta S. 2004. Molecular mechanisms for apigenin-induced cell-cycle arrest and apoptosis of hormone refractory human prostate carcinoma DU145 cells. Mol Carcinog 39:114-126.

155. McVean M, Weinberg WC, Pelling JC. 2002. A p21(waf1)-independent pathway for inhibitory phosphorylation of cyclin-dependent kinase p34(cdc2) and concomitant G(2)/M arrest by the chemopreventive flavonoid apigenin. Mol Carcinog 33:36-43.

156. Wu DG, Yu P, Li JW, Jiang P, Sun J, Wang HZ, Zhang LD, Wen MB, Bie P. 2014. Apigenin potentiates the growth inhibitory effects by IKK-β-mediated NF- κB activation in pancreatic cancer cells. Toxicol Lett 224:157-164.

157. Ye Q, Kantonen S, Gomez-Cambronero J. 2013. Serum deprivation confers the MDA-MB-231 breast cancer line with an EGFR/JAK3/PLD2 system that

270

maximizes cancer cell invasion. J Mol Biol 425:755-766.

158. Gao AM, Ke ZP, Wang JN, Yang JY, Chen SY, Chen H. 2013. Apigenin sensitizes doxorubicin-resistant hepatocellular carcinoma BEL-7402/ADM cells to doxorubicin via inhibiting PI3K/Akt/Nrf2 pathway. Carcinogenesis 34:1806- 1814.

159. Masuelli L, Marzocchella L, Quaranta A, Palumbo C, Pompa G, Izzi V, Canini A, Modesti A, Galvano F, Bei R. 2011. Apigenin induces apoptosis and impairs head and neck carcinomas EGFR/ErbB2 signaling. Front Biosci (Landmark Ed) 16:1060-1068.

160. Wang QR, Yao XQ, Wen G, Fan Q, Li YJ, Fu XQ, Li CK, Sun XG. 2011. Apigenin suppresses the growth of colorectal cancer xenografts via phosphorylation and up-regulated FADD expression. Oncol Lett 2:43-47.

161. Chen D, Landis-Piwowar KR, Chen MS, Dou QP. 2007. Inhibition of proteasome activity by the dietary flavonoid apigenin is associated with growth inhibition in cultured breast cancer cells and xenografts. Breast Cancer Res 9:R80.

162. Mafuvadze B, Liang Y, Besch-Williford C, Zhang X, Hyder SM. 2012. Apigenin induces apoptosis and blocks growth of medroxyprogesterone acetate-dependent BT-474 xenograft tumors. Horm Cancer 3:160-171.

163. Shukla S, Gupta S. 2006. Molecular targets for apigenin-induced cell cycle arrest and apoptosis in prostate cancer cell xenograft. Mol Cancer Ther 5:843-852.

164. Shukla S, Bhaskaran N, Babcook MA, Fu P, Maclennan GT, Gupta S. 2014. Apigenin inhibits prostate cancer progression in TRAMP mice via targeting PI3K/Akt/FoxO pathway. Carcinogenesis 35:452-460.

165. Hougee S, Sanders A, Faber J, Graus YM, van den Berg WB, Garssen J, Smit HF, Hoijer MA. 2005. Decreased pro-inflammatory cytokine production by LPS- stimulated PBMC upon in vitro incubation with the flavonoids apigenin, luteolin or chrysin, due to selective elimination of monocytes/macrophages. Biochem Pharmacol 69:241-248.

166. Kang OH, Lee JH, Kwon DY. 2011. Apigenin inhibits release of inflammatory mediators by blocking the NF-κB activation pathways in the HMC-1 cells. Immunopharmacol Immunotoxicol 33:473-479.

167. Palmieri D, Perego P, Palombo D. 2012. Apigenin inhibits the TNFα-induced expression of eNOS and MMP-9 via modulating Akt signalling through oestrogen receptor engagement. Mol Cell Biochem 371:129-136.

271

168. Wang YC, Huang KM. 2013. In vitro anti-inflammatory effect of apigenin in the Helicobacter pylori-infected gastric adenocarcinoma cells. Food Chem Toxicol 53:376-383.

169. Nicholas C, Batra S, Vargo MA, Voss OH, Gavrilin MA, Wewers MD, Guttridge DC, Grotewold E, Doseff AI. 2007. Apigenin blocks lipopolysaccharide-Induced lethality in vivo and proinflammatory cytokines expression by inactivating NF-κB through the suppression of p65 phosphorylation. J Immunol 179:7121-7127.

170. Wang J, Liu YT, Xiao L, Zhu L, Wang Q, Yan T. 2014. Anti-inflammatory effects of apigenin in lipopolysaccharide-induced inflammatory in acute lung injury by suppressing COX-2 and NF-kB pathway. Inflammation 37:2085-2090.

171. Henkels KM, Frondorf K, Gonzalez-Mejia ME, Doseff AL, Gomez-Cambronero J. 2011. IL-8-induced neutrophil chemotaxis is mediated by Janus kinase 3 (JAK3). FEBS Lett 585:159-166.

172. Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, Aros F, Gomez-Gracia E, Ruiz-Gutierrez V, Fiol M, Lapetra J, Lamuela-Raventos RM, Serra-Majem L, Pinto X, Basora J, Munoz MA, Sorli JV, Martinez JA, Martinez-Gonzalez MA. 2013. Primary prevention of cardiovascular disease with a Mediterranean diet. N Engl J Med 368:1279-1290.

173. Tong X, Pelling JC. 2013. Targeting the PI3K/Akt/mTOR axis by apigenin for cancer prevention. Anticancer Agents Med Chem 13:971-978.

174. Sarno S, de Moliner E, Ruzzene M, Pagano MA, Battistutta R, Bain J, Fabbro D, Schoepfer J, Elliott M, Furet P, Meggio F, Zanotti G, Pinna LA. 2003. Biochemical and three-dimensional-structural study of the specific inhibition of protein kinase CK2 by [5-oxo-5,6-dihydroindolo-(1,2-a)quinazolin-7-yl]acetic acid (IQA). Biochem J 374:639-646.

175. Iizumi Y, Oishi M, Taniguchi T, Goi W, Sowa Y, Sakai T. 2013. The flavonoid apigenin downregulates CDK1 by directly targeting ribosomal protein S9. PLoS One 8:e73219.

176. Ahmad R, Raina D, Trivedi V, Ren J, Rajabi H, Kharbanda S, Kufe D. 2007. MUC1 oncoprotein activates the IκB kinase β complex and constitutive NF-κB signalling. Nat Cell Biol 9:1419-1427.

177. Zhou Y, Rajabi H, Kufe D. 2011. Mucin 1 C-terminal subunit oncoprotein is a target for small-molecule inhibitors. Mol Pharmacol 79:886-893.

178. Choudhury D, Ganguli A, Dastidar DG, Acharya BR, Das A, Chakrabarti G. 2013. Apigenin shows synergistic anticancer activity with curcumin by binding at

272

different sites of tubulin. Biochimie 95:1297-1309.

179. Ginhoux F, Jung S. 2014. Monocytes and macrophages: developmental pathways and tissue homeostasis. Nat Rev Immunol 14:392-404.

180. Forget MA, Voorhees JL, Cole SL, Dakhlallah D, Patterson IL, Gross AC, Moldovan L, Mo X, Evans R, Marsh CB, Eubank TD. 2014. Macrophage colony- stimulating factor augments Tie2-expressing monocyte differentiation, angiogenic function, and recruitment in a mouse model of breast cancer. PLoS One 9:e98623.

181. Singh NP, McCoy MT, Tice RR, Schneider EL. 1988. A simple technique for quantitation of low levels of DNA damage in individual cells. Exp Cell Res 175:184-191.

182. Konca K, Lankoff A, Banasik A, Lisowska H, Kuszewski T, Gozdz S, Koza Z, Wojcik A. 2003. A cross-platform public domain PC image-analysis program for the comet assay. Mutat Res 534:15-20.

183. Ren C, Zhang L, Freitas MA, Ghoshal K, Parthun MR, Jacob ST. 2005. Peptide mass mapping of acetylated isoforms of histone H4 from mouse lymphosarcoma cells treated with histone deacetylase (HDACs) inhibitors. J Am Soc Mass Spectrom 16:1641-1653.

184. Smyth GK. 2004. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol 3:Article3.

185. Schmittgen TD, Lee EJ, Jiang J, Sarkar A, Yang L, Elton TS, Chen C. 2008. Real-time PCR quantification of precursor and mature microRNA. Methods 44:31-38.

186. Slinker BK. 1998. The statistics of synergism. J Mol Cell Cardiol 30:723-731.

187. Loscher CJ, Hokamp K, Kenna PF, Ivens AC, Humphries P, Palfi A, Farrar GJ. 2007. Altered retinal microRNA expression profile in a mouse model of retinitis pigmentosa. Genome Biol 8:R248.

188. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L. 2010. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol 28:511-515.

189. Abdulkhalek S, Amith SR, Franchuk SL, Jayanth P, Guo M, Finlay T, Gilmour A, Guzzo C, Gee K, Beyaert R, Szewczuk MR. 2011. Neu1 sialidase and matrix metalloproteinase-9 cross-talk is essential for Toll-like receptor activation and

273

cellular signaling. J Biol Chem 286:36532-36549.

190. Katz Y, Wang ET, Airoldi EM, Burge CB. 2010. Analysis and design of RNA sequencing experiments for identifying isoform regulation. Nat Methods 7:1009- 1015.

191. Rozen SS, H.J. . 1998. Primer3.

192. Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, Haussler D. 2002. The human genome browser at UCSC. Genome Res 12:996-1006.

193. Kuramochi K, Miyano Y, Enomoto Y, Takeuchi R, Ishi K, Takakusagi Y, Saitoh T, Fukudome K, Manita D, Takeda Y, Kobayashi S, Sakaguchi K, Sugawara F. 2008. Identification of small molecule binding molecules by affinity purification using a specific ligand immobilized on PEGA resin. Bioconjug Chem 19:2417- 2426.

194. Jiang H, Wong WH. 2008. SeqMap: mapping massive amount of oligonucleotides to the genome. Bioinformatics 24:2395-2396.

195. Saeed AI, Sharov V, White J, Li J, Liang W, Bhagabati N, Braisted J, Klapa M, Currier T, Thiagarajan M, Sturn A, Snuffin M, Rezantsev A, Popov D, Ryltsov A, Kostukovich E, Borisovsky I, Liu Z, Vinsavich A, Trush V, Quackenbush J. 2003. TM4: a free, open-source system for microarray data management and analysis. Biotechniques 34:374-378.

196. Huang DW, Sherman, B.T., & Lempick A. 2009. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature Protocols 44:44-59.

197. Friend LR, Han SP, Rothnagel JA, Smith R. 2008. Differential subnuclear localisation of hnRNPs A/B is dependent on transcription and cell cycle stage. Biochim Biophys Acta 1783:1972-1980.

198. Mayeda A, Munroe SH, Caceres JF, Krainer AR. 1994. Function of conserved domains of hnRNP A1 and other hnRNP A/B proteins. EMBO J 13:5483-5495.

199. Barnes JD, Arhel NJ, Lee SS, Sharp A, Al-Okail M, Packham G, Hague A, Paraskeva C, Williams AC. 2005. Nuclear BAG-1 expression inhibits apoptosis in colorectal adenoma-derived epithelial cells. Apoptosis 10:301-311.

200. Bhattacharyya R, Wedegaertner PB. 2003. Characterization of G α13-dependent plasma membrane recruitment of p115RhoGEF. Biochem J 371:709-720.

201. Nichols RC, Wang XW, Tang J, Hamilton BJ, High FA, Herschman HR, Rigby

274

WF. 2000. The RGG domain in hnRNP A2 affects subcellular localization. Exp Cell Res 256:522-532.

202. Fehr M, Frommer WB, Lalonde S. 2002. Visualization of maltose uptake in living yeast cells by fluorescent nanosensors. Proc Natl Acad Sci U S A 99:9846-9851.

203. Benesi HA, Hildebrand, J.H. 1949. A Spectrophotometric Investigation of the Interaction of Iodine with Aromatic Hydrocarbons. J Am Chem Soc 71.

204. Zhu M, Rajamani S, Kaylor J, Han S, Zhou F, Fink AL. 2004. The flavonoid baicalein inhibits fibrillation of α-synuclein and disaggregates existing fibrils. J Biol Chem 279:26846-26857.

205. Hostetler G, Riedl K, Cardenas H, Diosa-Toro M, Arango D, Schwartz S, Doseff AI. 2012. Flavone deglycosylation increases their anti-inflammatory activity and absorption. Mol Nutr Food Res 56:558-569.

206. Greten FR, Eckmann L, Greten TF, Park JM, Li ZW, Egan LJ, Kagnoff MF, Karin M. 2004. IKKβ links inflammation and tumorigenesis in a mouse model of colitis-associated cancer. Cell 118:285-296.

207. Santidrian AF, Matsuno-Yagi A, Ritland M, Seo BB, LeBoeuf SE, Gay LJ, Yagi T, Felding-Habermann B. 2013. Mitochondrial complex I activity and NAD+/NADH balance regulate breast cancer progression. J Clin Invest 123:1068- 1081.

208. Shukla S, Gupta S. Apigenin: a promising molecule for cancer prevention. Pharm Res 27:962-978.

209. Van Dross R, Xue Y, Knudson A, Pelling JC. 2003. The chemopreventive bioflavonoid apigenin modulates signal transduction pathways in keratinocyte and colon carcinoma cell lines. J Nutr 133:3800S-3804S.

210. Lord CJ, Ashworth A. 2012. The DNA damage response and cancer therapy. Nature 481:287-294.

211. Nitiss JL. 2009. Targeting DNA topoisomerase II in cancer chemotherapy. Nat Rev Cancer 9:338-350.

212. Pommier Y. 2006. Topoisomerase I inhibitors: camptothecins and beyond. Nat Rev Cancer 6:789-802.

213. Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM. 1998. DNA double- stranded breaks induce histone H2AX phosphorylation on serine 139. J Biol Chem 273:5858-5868.

275

214. Helt CE, Cliby WA, Keng PC, Bambara RA, O'Reilly MA. 2005. Ataxia telangiectasia mutated (ATM) and ATM and Rad3-related protein exhibit selective target specificities in response to different forms of DNA damage. J Biol Chem 280:1186-1192.

215. Bakkenist CJ, Kastan MB. 2003. DNA damage activates ATM through intermolecular autophosphorylation and dimer dissociation. Nature 421:499-506.

216. Kastan MB, Lim DS. 2000. The many substrates and functions of ATM. Nat Rev Mol Cell Biol 1:179-186.

217. Cimprich KA, Cortez D. 2008. ATR: an essential regulator of genome integrity. Nat Rev Mol Cell Biol 9:616-627.

218. Zhou BB, Elledge SJ. 2000. The DNA damage response: putting checkpoints in perspective. Nature 408:433-439.

219. Kinner A, Wu W, Staudt C, Iliakis G. 2008. γ-H2AX in recognition and signaling of DNA double-strand breaks in the context of chromatin. Nucleic Acids Res 36:5678-5694.

220. Lavin MF, Birrell G, Chen P, Kozlov S, Scott S, Gueven N. 2005. ATM signaling and genomic stability in response to DNA damage. Mutat Res 569:123-132.

221. Brodie C, Blumberg PM. 2003. Regulation of cell apoptosis by protein kinase δ. Apoptosis 8:19-27.

222. Reyland ME, Anderson SM, Matassa AA, Barzen KA, Quissell DO. 1999. Protein kinase C δ is essential for etoposide-induced apoptosis in salivary gland acinar cells. J Biol Chem 274:19115-19123.

223. Clavijo C, Chen JL, Kim KJ, Reyland ME, Ann DK. 2007. Protein kinase Cδ- dependent and -independent signaling in genotoxic response to treatment of desferroxamine, a hypoxia-mimetic agent. Am J Physiol Cell Physiol 292:C2150- 2160.

224. Bulavin DV, Saito S, Hollander MC, Sakaguchi K, Anderson CW, Appella E, Fornace AJ, Jr. 1999. Phosphorylation of human p53 by p38 kinase coordinates N-terminal phosphorylation and apoptosis in response to UV radiation. EMBO J 18:6845-6854.

225. Visvardis EE, Tassiou AM, Piperakis SM. 1997. Study of DNA damage induction and repair capacity of fresh and cryopreserved lymphocytes exposed to H2O2 and γ-irradiation with the alkaline comet assay. Mutat Res 383:71-80.

276

226. Matkovic K, Lukinovic-Skudar V, Banfic H, Visnjic D. 2009. The activity of extracellular signal-regulated kinase is required during G2/M phase before metaphase-anaphase transition in synchronized leukemia cell lines. Int J Hematol 89:159-166.

227. Boege F, Straub T, Kehr A, Boesenberg C, Christiansen K, Andersen A, Jakob F, Kohrle J. 1996. Selected novel flavones inhibit the DNA binding or the DNA religation step of eukaryotic topoisomerase I. J Biol Chem 271:2262-2270.

228. Bandele OJ, Osheroff N. 2007. Bioflavonoids as poisons of human topoisomerase IIα and IIβ. Biochemistry 46:6097-6108.

229. Sharma H, Kanwal R, Bhaskaran N, Gupta S. 2014. Plant flavone apigenin binds to nucleic acid bases and reduces oxidative DNA damage in prostate epithelial cells. PLoS One 9:e91588.

230. Wang Y, Xu YS, Yin LH, Xu LN, Peng JY, Zhou H, Kang W. 2013. Synergistic anti-glioma effect of Hydroxygenkwanin and Apigenin in vitro. Chem Biol Interact 206:346-355.

231. Wang IK, Lin-Shiau SY, Lin JK. 1999. Induction of apoptosis by apigenin and related flavonoids through cytochrome c release and activation of caspase-9 and caspase-3 in leukaemia HL-60 cells. Eur J Cancer 35:1517-1525.

232. Janicke RU, Sprengart ML, Wati MR, Porter AG. 1998. Caspase-3 is required for DNA fragmentation and morphological changes associated with apoptosis. J Biol Chem 273:9357-9360.

233. Yamashita N, Kawanishi S. 2000. Distinct mechanisms of DNA damage in apoptosis induced by quercetin and luteolin. Free Radic Res 33:623-633.

234. Liu Y, Deng K, Li J, Liu S, Yao S. 2004. Investigation of double stranded DNA damage induced by quercetin-copper(II) using piezoelectric quartz crystal impedance technique and potentiometric stripping analysis. Biophys Chem 112:69-76.

235. Rahman A, Fazal F, Greensill J, Ainley K, Parish JH, Hadi SM. 1992. Strand scission in DNA induced by dietary flavonoids: role of Cu(I) and oxygen free radicals and biological consequences of scission. Mol Cell Biochem 111:3-9.

236. Fazal F, Rahman A, Greensill J, Ainley K, Hadi SM, Parish JH. 1990. Strand scission in DNA by quercetin and Cu(II): identification of free radical intermediates and biological consequences of scission. Carcinogenesis 11:2005- 2008.

277

237. Yoshida K, Yamaguchi T, Shinagawa H, Taira N, Nakayama KI, Miki Y. 2006. Protein kinase C δ activates topoisomerase IIα to induce apoptotic cell death in response to DNA damage. Mol Cell Biol 26:3414-3431.

238. Lu C, Shi Y, Wang Z, Song Z, Zhu M, Cai Q, Chen T. 2008. Serum starvation induces H2AX phosphorylation to regulate apoptosis via p38 MAPK pathway. FEBS Lett 582:2703-2708.

239. Rakkestad KE, Skaar I, Ansteinsson VE, Solhaug A, Holme JA, Pestka JJ, Samuelsen JT, Dahlman HJ, Hongslo JK, Becher R. 2010. DNA damage and DNA damage responses in THP-1 monocytes after exposure to spores of either Stachybotrys chartarum or Aspergillus versicolor or to T-2 toxin. Toxicol Sci 115:140-155.

240. Kobayashi T, Kuroda J, Ashihara E, Oomizu S, Terui Y, Taniyama A, Adachi S, Takagi T, Yamamoto M, Sasaki N, Horiike S, Hatake K, Yamauchi A, Hirashima M, Taniwaki M. 2010. Galectin-9 exhibits anti-myeloma activity through JNK and p38 MAP kinase pathways. Leukemia 24:843-850.

241. Hsiao PW, Chang CC, Liu HF, Tsai CM, Chiu TH, Chao JI. 2007. Activation of p38 mitogen-activated protein kinase by celecoxib oppositely regulates survivin and γ-H2AX in human colorectal cancer cells. Toxicol Appl Pharmacol 222:97- 104.

242. Reinhardt HC, Aslanian AS, Lees JA, Yaffe MB. 2007. p53-deficient cells rely on ATM- and ATR-mediated checkpoint signaling through the p38MAPK/MK2 pathway for survival after DNA damage. Cancer Cell 11:175-189.

243. Bharti A, Kraeft SK, Gounder M, Pandey P, Jin S, Yuan ZM, Lees-Miller SP, Weichselbaum R, Weaver D, Chen LB, Kufe D, Kharbanda S. 1998. Inactivation of DNA-dependent protein kinase by protein kinase Cδ: implications for apoptosis. Mol Cell Biol 18:6719-6728.

244. Emoto Y, Manome Y, Meinhardt G, Kisaki H, Kharbanda S, Robertson M, Ghayur T, Wong WW, Kamen R, Weichselbaum R, et al. 1995. Proteolytic activation of protein kinase C δ by an ICE-like protease in apoptotic cells. EMBO J 14:6148-6156.

245. Matsuoka S, Ballif BA, Smogorzewska A, McDonald ER, 3rd, Hurov KE, Luo J, Bakalarski CE, Zhao Z, Solimini N, Lerenthal Y, Shiloh Y, Gygi SP, Elledge SJ. 2007. ATM and ATR substrate analysis reveals extensive protein networks responsive to DNA damage. Science 316:1160-1166.

246. Shukla S, Gupta S. 2007. Apigenin-induced cell cycle arrest is mediated by

278

modulation of MAPK, PI3K-Akt, and loss of cyclin D1 associated retinoblastoma dephosphorylation in human prostate cancer cells. Cell Cycle 6:1102-1114.

247. Zheng PW, Chiang LC, Lin CC. 2005. Apigenin induced apoptosis through p53- dependent pathway in human cervical carcinoma cells. Life Sci 76:1367-1379.

248. Takagaki N, Sowa Y, Oki T, Nakanishi R, Yogosawa S, Sakai T. 2005. Apigenin induces cell cycle arrest and p21/WAF1 expression in a p53-independent pathway. Int J Oncol 26:185-189.

249. Lepley DM, Pelling JC. 1997. Induction of p21/WAF1 and G1 cell-cycle arrest by the chemopreventive agent apigenin. Mol Carcinog 19:74-82.

250. Lepley DM, Li B, Birt DF, Pelling JC. 1996. The chemopreventive flavonoid apigenin induces G2/M arrest in keratinocytes. Carcinogenesis 17:2367-2375.

251. Shiohara M, Akashi M, Gombart AF, Yang R, Koeffler HP. 1996. Tumor necrosis factor α: posttranscriptional stabilization of WAF1 mRNA in p53-deficient human leukemic cells. J Cell Physiol 166:568-576.

252. Shoulars K, Rodriguez MA, Thompson T, Markaverich BM. 2010. Regulation of cell cycle and RNA transcription genes identified by microarray analysis of PC-3 human prostate cancer cells treated with luteolin. J Steroid Biochem Mol Biol 118:41-50.

253. Cobrinik D. 1996. Regulatory interactions among E2Fs and cell cycle control proteins. Curr Top Microbiol Immunol 208:31-61.

254. Malumbres M, Barbacid M. 2009. Cell cycle, CDKs and cancer: a changing paradigm. Nat Rev Cancer 9:153-166.

255. Arango D, Parihar A, Villamena FA, Wang L, Freitas MA, Grotewold E, Doseff AI. 2012. Apigenin induces DNA damage through the PKCδ-dependent activation of ATM and H2AX causing down-regulation of genes involved in cell cycle control and DNA repair. Biochem Pharmacol 84:1571-1580.

256. Henson PM. 2005. Dampening inflammation. Nat Immunol 6:1179-1181.

257. Rittirsch D, Flierl MA, Ward PA. 2008. Harmful molecular mechanisms in sepsis. Nat Rev Immunol 8:776-787.

258. Rocha VZ, Libby P. 2009. Obesity, inflammation, and atherosclerosis. Nat Rev Cardiol 6:399-409.

259. Pasparakis M. 2009. Regulation of tissue homeostasis by NF-κB signalling:

279

implications for inflammatory diseases. Nat Rev Immunol 9:778-788.

260. Baltimore D, Boldin MP, O'Connell RM, Rao DS, Taganov KD. 2008. MicroRNAs: new regulators of immune cell development and function. Nat Immunol 9:839-845.

261. Kim VN. 2005. MicroRNA biogenesis: coordinated cropping and dicing. Nat Rev Mol Cell Biol 6:376-385.

262. Louafi F, Martinez-Nunez RT, Sanchez-Elsner T. 2010. MicroRNA-155 targets SMAD2 and modulates the response of macrophages to transforming growth factor-β. J Biol Chem 285:41328-41336.

263. Sugiyama Y, Kakoi K, Kimura A, Takada I, Kashiwagi I, Wakabayashi Y, Morita R, Nomura M, Yoshimura A. 2012. Smad2 and Smad3 are redundantly essential for the suppression of iNOS synthesis in macrophages by regulating IRF3 and STAT1 pathways. Int Immunol 24:253-265.

264. Min M, Peng L, Yang Y, Guo M, Wang W, Sun G. 2014. MicroRNA-155 is involved in the pathogenesis of ulcerative colitis by targeting FOXO3a. Inflamm Bowel Dis 20:652-659.

265. Lin L, Hron JD, Peng SL. 2004. Regulation of NF-κB, Th activation, and autoinflammation by the forkhead transcription factor Foxo3a. Immunity 21:203- 213.

266. Duarte S, Arango D, Parihar A, Hamel P, Yasmeen R, Doseff AI. 2013. Apigenin protects endothelial cells from lipopolysaccharide (LPS)-induced inflammation by decreasing caspase-3 activation and modulating mitochondrial function. Int J Mol Sci 14:17664-17679.

267. Ha SK, Lee P, Park JA, Oh HR, Lee SY, Park JH, Lee EH, Ryu JH, Lee KR, Kim SY. 2008. Apigenin inhibits the production of NO and PGE2 in microglia and inhibits neuronal cell death in a middle cerebral artery occlusion-induced focal ischemia mice model. Neurochem Int 52:878-886.

268. Hostetler GL, Riedl KM, Schwartz SJ. 2012. Endogenous enzymes, heat, and pH affect flavone profiles in parsley (Petroselinum crispum var. neapolitanum) and celery (Apium graveolens) during juice processing. J Agric Food Chem 60:202- 208.

269. Ruggiero T, Trabucchi M, De Santa F, Zupo S, Harfe BD, McManus MT, Rosenfeld MG, Briata P, Gherzi R. 2009. LPS induces KH-type splicing regulatory protein-dependent processing of microRNA-155 precursors in macrophages. FASEB J 23:2898-2908.

280

270. Shih VF, Kearns JD, Basak S, Savinova OV, Ghosh G, Hoffmann A. 2009. Kinetic control of negative feedback regulators of NF-κB/RelA determines their pathogen- and cytokine-receptor signaling specificity. Proc Natl Acad Sci U S A 106:9619-9624.

271. Hoffmann A, Levchenko A, Scott ML, Baltimore D. 2002. The IκB-NF-κB signaling module: temporal control and selective gene activation. Science 298:1241-1245.

272. Hsieh CH, Rau CS, Jeng JC, Chen YC, Lu TH, Wu CJ, Wu YC, Tzeng SL, Yang JC. 2012. Whole blood-derived microRNA signatures in mice exposed to lipopolysaccharides. J Biomed Sci 19:69.

273. Nicholas C. 2009. The Anti-Inflammatory Mechanisms of the Flavonoid Apigenin In Vitro and In Vivo. PhD Dissertation (The Ohio State University, Columbus, OH).

274. Hatada EN, Krappmann D, Scheidereit C. 2000. NF-κB and the innate immune response. Curr Opin Immunol 12:52-58.

275. Austenaa LM, Carlsen H, Hollung K, Blomhoff HK, Blomhoff R. 2009. Retinoic acid dampens LPS-induced NF-κB activity: results from human monoblasts and in vivo imaging of NF-κB reporter mice. J Nutr Biochem 20:726-734.

276. Carlsen H, Moskaug JO, Fromm SH, Blomhoff R. 2002. In vivo imaging of NF- κB activity. J Immunol 168:1441-1446.

277. Nelson G, Paraoan L, Spiller DG, Wilde GJ, Browne MA, Djali PK, Unitt JF, Sullivan E, Floettmann E, White MR. 2002. Multi-parameter analysis of the kinetics of NF-κB signalling and transcription in single living cells. J Cell Sci 115:1137-1148.

278. Chen Y, Gelfond JA, McManus LM, Shireman PK. 2009. Reproducibility of quantitative RT-PCR array in miRNA expression profiling and comparison with microarray analysis. BMC Genomics 10:407.

279. Blade C, Baselga-Escudero L, Salvado MJ, Arola-Arnal A. 2013. miRNAs, polyphenols, and chronic disease. Mol Nutr Food Res 57:58-70.

280. Milenkovic D, Jude B, Morand C. 2013. miRNA as molecular target of polyphenols underlying their biological effects. Free Radic Biol Med 64:40-51.

281. Ohno M, Shibata C, Kishikawa T, Yoshikawa T, Takata A, Kojima K, Akanuma M, Kang YJ, Yoshida H, Otsuka M, Koike K. 2013. The flavonoid apigenin

281

improves glucose tolerance through inhibition of microRNA maturation in miRNA103 transgenic mice. Sci Rep 3:2553.

282. Boesch-Saadatmandi C, Loboda A, Wagner AE, Stachurska A, Jozkowicz A, Dulak J, Doring F, Wolffram S, Rimbach G. 2011. Effect of quercetin and its metabolites isorhamnetin and quercetin-3-glucuronide on inflammatory gene expression: role of miR-155. J Nutr Biochem 22:293-299.

283. Baselga-Escudero L, Blade C, Ribas-Latre A, Casanova E, Suarez M, Torres JL, Salvado MJ, Arola L, Arola-Arnal A. 2014. Resveratrol and EGCG bind directly and distinctively to miR-33a and miR-122 and modulate divergently their levels in hepatic cells. Nucleic Acids Res 42:882-892.

284. Nafisi S, Shadaloi A, Feizbakhsh A, Tajmir-Riahi HA. 2009. RNA binding to antioxidant flavonoids. J Photochem Photobiol B 94:1-7.

285. Schulte LN, Westermann AJ, Vogel J. 2013. Differential activation and functional specialization of miR-146 and miR-155 in innate immune sensing. Nucleic Acids Res 41:542-553.

286. Xiao B, Liu Z, Li BS, Tang B, Li W, Guo G, Shi Y, Wang F, Wu Y, Tong WD, Guo H, Mao XH, Zou QM. 2009. Induction of microRNA-155 during Helicobacter pylori infection and its negative regulatory role in the inflammatory response. J Infect Dis 200:916-925.

287. Imaizumi T, Tanaka H, Tajima A, Yokono Y, Matsumiya T, Yoshida H, Tsuruga K, Aizawa-Yashiro T, Hayakari R, Inoue I, Ito E, Satoh K. 2010. IFN-γ and TNF- α synergistically induce microRNA-155 which regulates TAB2/IP-10 expression in human mesangial cells. Am J Nephrol 32:462-468.

288. Thounaojam MC, Kundu K, Kaushik DK, Swaroop S, Mahadevan A, Shankar SK, Basu A. 2014. MicroRNA 155 regulates Japanese encephalitis virus-induced inflammatory response by targeting Src homology 2-containing inositol phosphatase 1. J Virol 88:4798-4810.

289. Zhou H, Huang X, Cui H, Luo X, Tang Y, Chen S, Wu L, Shen N. 2010. miR-155 and its star-form partner miR-155* cooperatively regulate type I interferon production by human plasmacytoid dendritic cells. Blood 116:5885-5894.

290. Hausser J, Zavolan M. 2014. Identification and consequences of miRNA-target interactions--beyond repression of gene expression. Nat Rev Genet 15:599-612.

291. Wang G, Huang X, Li Y, Guo K, Ning P, Zhang Y. 2013. PARP-1 inhibitor, DPQ, attenuates LPS-induced acute lung injury through inhibiting NF-κB- mediated inflammatory response. PLoS One 8:e79757.

282

292. Nicholas C, Batra S, Vargo MA, Voss OH, Gavrilin MA, Wewers MD, Guttdrige DC, Grotewold E, Doseff AI. 2007. Apigenin blocks lipopolysaccharide-induced lethality in vivo and pro-inflammatory cytokines expression by inactivating NF- κB through the suppression of p65 phosphorylation. J. Immunol. 179:7121-7127.

293. Bhaskaran N, Shukla S, Srivastava JK, Gupta S. 2010. Chamomile: an anti- inflammatory agent inhibits inducible nitric oxide synthase expression by blocking RelA/p65 activity. Int J Mol Med 26:935-940.

294. Arango D, Diosa-Toro M, Rojas-Hernandez LS, Cooperstone JL, Schwartz SJ, Mo X, Jiang J, Schmittgen TD, Doseff AI. 2015. Dietary apigenin reduces LPS- induced expression of miR-155 restoring immune balance during inflammation. Mol Nutr Food Res 59:763-772.

295. Desai KV, Xiao N, Wang W, Gangi L, Greene J, Powell JI, Dickson R, Furth P, Hunter K, Kucherlapati R, Simon R, Liu ET, Green JE. 2002. Initiating oncogenic event determines gene-expression patterns of human breast cancer models. Proc Natl Acad Sci U S A 99:6967-6972.

296. DeNardo DG, Barreto JB, Andreu P, Vasquez L, Tawfik D, Kolhatkar N, Coussens LM. 2009. CD4(+) T cells regulate pulmonary metastasis of mammary carcinomas by enhancing protumor properties of macrophages. Cancer Cell 16:91-102.

297. Erez N, Truitt M, Olson P, Arron ST, Hanahan D. 2010. Cancer-Associated Fibroblasts Are Activated in Incipient Neoplasia to Orchestrate Tumor-Promoting Inflammation in an NF-κB-Dependent Manner. Cancer Cell 17:135-147.

298. Kiriakidis S, Andreakos E, Monaco C, Foxwell B, Feldmann M, Paleolog E. 2003. VEGF expression in human macrophages is NF-κB-dependent: studies using adenoviruses expressing the endogenous NF-κB inhibitor IκBα and a kinase-defective form of the IκB kinase 2. J Cell Sci 116:665-674.

299. Chow A, Zhou W, Liu L, Fong MY, Champer J, Van Haute D, Chin AR, Ren X, Gugiu BG, Meng Z, Huang W, Ngo V, Kortylewski M, Wang SE. 2014. Macrophage immunomodulation by breast cancer-derived exosomes requires Toll-like receptor 2-mediated activation of NF-κB. Sci Rep 4:5750.

300. Yin F, Giuliano AE, Law RE, Van Herle AJ. 2001. Apigenin inhibits growth and induces G2/M arrest by modulating cyclin-CDK regulators and ERK MAP kinase activation in breast carcinoma cells. Anticancer Res 21:413-420.

301. Zhang M, He Y, Sun X, Li Q, Wang W, Zhao A, Di W. 2014. A high M1/M2 ratio of tumor-associated macrophages is associated with extended survival in

283

ovarian cancer patients. J Ovarian Res 7:19.

302. Algars A, Irjala H, Vaittinen S, Huhtinen H, Sundstrom J, Salmi M, Ristamaki R, Jalkanen S. 2012. Type and location of tumor-infiltrating macrophages and lymphatic vessels predict survival of colorectal cancer patients. Int J Cancer 131:864-873.

303. Kong LQ, Zhu XD, Xu HX, Zhang JB, Lu L, Wang WQ, Zhang QB, Wu WZ, Wang L, Fan J, Tang ZY, Sun HC. 2013. The clinical significance of the CD163+ and CD68+ macrophages in patients with hepatocellular carcinoma. PLoS One 8:e59771.

304. Geissmann F, Manz MG, Jung S, Sieweke MH, Merad M, Ley K. 2010. Development of monocytes, macrophages, and dendritic cells. Science 327:656- 661.

305. Danilin S, Merkel AR, Johnson JR, Johnson RW, Edwards JR, Sterling JA. 2012. Myeloid-derived suppressor cells expand during breast cancer progression and promote tumor-induced bone destruction. Oncoimmunology 1:1484-1494.

306. Waldhauer I, Steinle A. 2008. NK cells and cancer immunosurveillance. Oncogene 27:5932-5943.

307. Feng AL, Zhu JK, Sun JT, Yang MX, Neckenig MR, Wang XW, Shao QQ, Song BF, Yang QF, Kong BH, Qu X. 2011. CD16+ monocytes in breast cancer patients: expanded by monocyte chemoattractant protein-1 and may be useful for early diagnosis. Clin Exp Immunol 164:57-65.

308. Schlecker E, Stojanovic A, Eisen C, Quack C, Falk CS, Umansky V, Cerwenka A. 2012. Tumor-infiltrating monocytic myeloid-derived suppressor cells mediate CCR5-dependent recruitment of regulatory T cells favoring tumor growth. J Immunol 189:5602-5611.

309. Hagemann T, Wilson J, Kulbe H, Li NF, Leinster DA, Charles K, Klemm F, Pukrop T, Binder C, Balkwill FR. 2005. Macrophages induce invasiveness of epithelial cancer cells via NF-κB and JNK. J Immunol 175:1197-1205.

310. O'Shaughnessy J. 2005. Extending survival with chemotherapy in metastatic breast cancer. Oncologist 10 Suppl 3:20-29.

311. Kuo CH, Weng BC, Wu CC, Yang SF, Wu DC, Wang YC. 2014. Apigenin has anti-atrophic gastritis and anti-gastric cancer progression effects in Helicobacter pylori-infected Mongolian gerbils. J Ethnopharmacol 151:1031-1039.

312. Fluck MM, Schaffhausen BS. 2009. Lessons in signaling and tumorigenesis from

284

polyomavirus middle T antigen. Microbiol Mol Biol Rev 73:542-563, Table of Contents.

313. Summers SA, Lipfert L, Birnbaum MJ. 1998. Polyoma middle T antigen activates the Ser/Thr kinase Akt in a PI3-kinase-dependent manner. Biochem Biophys Res Commun 246:76-81.

314. Meili R, Cron P, Hemmings BA, Ballmer-Hofer K. 1998. Protein kinase B/Akt is activated by polyomavirus middle-T antigen via a phosphatidylinositol 3-kinase- dependent mechanism. Oncogene 16:903-907.

315. Cui X, Zhang P, Deng W, Oesterreich S, Lu Y, Mills GB, Lee AV. 2003. Insulin- like growth factor-I inhibits progesterone receptor expression in breast cancer cells via the phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin pathway: progesterone receptor as a potential indicator of growth factor activity in breast cancer. Mol Endocrinol 17:575-588.

316. Gu C, Zhang Z, Yu Y, Liu Y, Zhao F, Yin L, Feng Y, Chen X. 2011. Inhibiting the PI3K/Akt pathway reversed progestin resistance in endometrial cancer. Cancer Sci 102:557-564.

317. Dahl J, Jurczak A, Cheng LA, Baker DC, Benjamin TL. 1998. Evidence of a role for phosphatidylinositol 3-kinase activation in the blocking of apoptosis by polyomavirus middle T antigen. J Virol 72:3221-3226.

318. Maroulakou IG, Oemler W, Naber SP, Tsichlis PN. 2007. Akt1 ablation inhibits, whereas Akt2 ablation accelerates, the development of mammary adenocarcinomas in mouse mammary tumor virus (MMTV)-ErbB2/neu and MMTV-polyoma middle T transgenic mice. Cancer Res 67:167-177.

319. Lee WJ, Chen WK, Wang CJ, Lin WL, Tseng TH. 2008. Apigenin inhibits HGF- promoted invasive growth and metastasis involving blocking PI3K/Akt pathway and β4 integrin function in MDA-MB-231 breast cancer cells. Toxicol Appl Pharmacol 226:178-191.

320. Fang J, Xia C, Cao Z, Zheng JZ, Reed E, Jiang BH. 2005. Apigenin inhibits VEGF and HIF-1 expression via PI3K/AKT/p70S6K1 and HDM2/p53 pathways. FASEB J 19:342-353.

321. Srinivas S, Schonthal A, Eckhart W. 1994. Polyomavirus middle-sized tumor antigen modulates c-Jun phosphorylation and transcriptional activity. Proc Natl Acad Sci U S A 91:10064-10068.

322. Rauh MJ, Blackmore V, Andrechek ER, Tortorice CG, Daly R, Lai VK, Pawson T, Cardiff RD, Siegel PM, Muller WJ. 1999. Accelerated mammary tumor

285

development in mutant polyomavirus middle T transgenic mice expressing elevated levels of either the Shc or Grb2 adapter protein. Mol Cell Biol 19:8169- 8179.

323. Way TD, Kao MC, Lin JK. 2005. Degradation of HER2/neu by apigenin induces apoptosis through cytochrome c release and caspase-3 activation in HER2/neu- overexpressing breast cancer cells. FEBS Lett 579:145-152.

324. Way TD, Kao MC, Lin JK. 2004. Apigenin induces apoptosis through proteasomal degradation of HER2/neu in HER2/neu-overexpressing breast cancer cells via the phosphatidylinositol 3-kinase/Akt-dependent pathway. J Biol Chem 279:4479-4489.

325. Li RR, Pang LL, Du Q, Shi Y, Dai WJ, Yin KS. 2010. Apigenin inhibits allergen- induced airway inflammation and switches immune response in a murine model of asthma. Immunopharmacol Immunotoxicol 32:364-370.

326. Wang L, Chen J, Wang B, Wu D, Li H, Lu H, Wu H, Chai Y. 2014. Protective effect of quercetin on lipopolysaccharide-induced acute lung injury in mice by inhibiting inflammatory cell influx. Exp Biol Med (Maywood) 239:1653-1662.

327. Dong J, Zhang X, Zhang L, Bian HX, Xu N, Bao B, Liu J. 2014. Quercetin reduces obesity-associated ATM infiltration and inflammation in mice: a mechanism including AMPKα1/SIRT1. J Lipid Res 55:363-374.

328. Jang JY, Lee JK, Jeon YK, Kim CW. 2013. Exosome derived from epigallocatechin gallate treated breast cancer cells suppresses tumor growth by inhibiting tumor-associated macrophage infiltration and M2 polarization. BMC Cancer 13:421.

329. Caso R, Silvera R, Carrio R, Iragavarapu-Charyulu V, Gonzalez-Perez RR, Torroella-Kouri M. 2010. Blood monocytes from mammary tumor-bearing mice: early targets of tumor-induced immune suppression? Int J Oncol 37:891-900.

330. Medrek C, Ponten F, Jirstrom K, Leandersson K. 2012. The presence of tumor associated macrophages in tumor stroma as a prognostic marker for breast cancer patients. BMC Cancer 12:306.

331. Ma Y, Temkin V, Liu H, Pope RM. 2005. NF-κB protects macrophages from lipopolysaccharide-induced cell death: the role of caspase 8 and receptor- interacting protein. J Biol Chem 280:41827-41834.

332. Epstein RJ. 2004. The CXCL12-CXCR4 chemotactic pathway as a target of adjuvant breast cancer therapies. Nat Rev Cancer 4:901-909.

286

333. Chen L, Xu S, Zeng X, Li J, Yin W, Chen Y, Shao Z, Jin W. 2010. c-myb activates CXCL12 transcription in T47D and MCF7 breast cancer cells. Acta Biochim Biophys Sin (Shanghai) 42:1-7.

334. Ceradini DJ, Kulkarni AR, Callaghan MJ, Tepper OM, Bastidas N, Kleinman ME, Capla JM, Galiano RD, Levine JP, Gurtner GC. 2004. Progenitor cell trafficking is regulated by hypoxic gradients through HIF-1 induction of SDF-1. Nat Med 10:858-864.

335. Schajnovitz A, Itkin T, D'Uva G, Kalinkovich A, Golan K, Ludin A, Cohen D, Shulman Z, Avigdor A, Nagler A, Kollet O, Seger R, Lapidot T. 2011. CXCL12 secretion by bone marrow stromal cells is dependent on cell contact and mediated by connexin-43 and connexin-45 gap junctions. Nat Immunol 12:391-398.

336. Kew RR, Penzo M, Habiel DM, Marcu KB. 2012. The IKKα-dependent NF-κB p52/RelB noncanonical pathway is essential to sustain a CXCL12 autocrine loop in cells migrating in response to HMGB1. J Immunol 188:2380-2386.

337. Madge LA, Kluger MS, Orange JS, May MJ. 2008. Lymphotoxin-α1β2 and LIGHT induce classical and noncanonical NF-κB-dependent proinflammatory gene expression in vascular endothelial cells. J Immunol 180:3467-3477.

338. Forsythe JA, Jiang BH, Iyer NV, Agani F, Leung SW, Koos RD, Semenza GL. 1996. Activation of vascular endothelial growth factor gene transcription by hypoxia-inducible factor 1. Mol Cell Biol 16:4604-4613.

339. Liu LZ, Fang J, Zhou Q, Hu X, Shi X, Jiang BH. 2005. Apigenin inhibits expression of vascular endothelial growth factor and angiogenesis in human lung cancer cells: implication of chemoprevention of lung cancer. Mol Pharmacol 68:635-643.

340. Mirzoeva S, Kim ND, Chiu K, Franzen CA, Bergan RC, Pelling JC. 2008. Inhibition of HIF-1α and VEGF expression by the chemopreventive bioflavonoid apigenin is accompanied by Akt inhibition in human prostate carcinoma PC3-M cells. Mol Carcinog 47:686-700.

341. Fang J, Zhou Q, Liu LZ, Xia C, Hu X, Shi X, Jiang BH. 2007. Apigenin inhibits tumor angiogenesis through decreasing HIF-1α and VEGF expression. Carcinogenesis 28:858-864.

342. Osada M, Imaoka S, Funae Y. 2004. Apigenin suppresses the expression of VEGF, an important factor for angiogenesis, in endothelial cells via degradation of HIF-1α protein. FEBS Lett 575:59-63.

343. Chunhua L, Donglan L, Xiuqiong F, Lihua Z, Qin F, Yawei L, Liang Z, Ge W,

287

Linlin J, Ping Z, Kun L, Xuegang S. 2013. Apigenin up-regulates transgelin and inhibits invasion and migration of colorectal cancer through decreased phosphorylation of AKT. J Nutr Biochem 24:1766-1775.

344. Hu XW, Meng D, Fang J. 2008. Apigenin inhibited migration and invasion of human ovarian cancer A2780 cells through focal adhesion kinase. Carcinogenesis 29:2369-2376.

345. Piantelli M, Rossi C, Iezzi M, La Sorda R, Iacobelli S, Alberti S, Natali PG. 2006. Flavonoids inhibit melanoma lung metastasis by impairing tumor cells endothelium interactions. J Cell Physiol 207:23-29.

346. Castillo-Pichardo L, Martinez-Montemayor MM, Martinez JE, Wall KM, Cubano LA, Dharmawardhane S. 2009. Inhibition of mammary tumor growth and metastases to bone and liver by dietary grape polyphenols. Clin Exp Metastasis 26:505-516.

347. Vantyghem SA, Wilson SM, Postenka CO, Al-Katib W, Tuck AB, Chambers AF. 2005. Dietary genistein reduces metastasis in a postsurgical orthotopic breast cancer model. Cancer Res 65:3396-3403.

348. Lentzsch S, Reichardt P, Weber F, Budach V, Dorken B. 1999. Brain metastases in breast cancer: prognostic factors and management. Eur J Cancer 35:580-585.

349. Koizumi M, Yoshimoto M, Kasumi F, Ogata E. 2003. Comparison between solitary and multiple skeletal metastatic lesions of breast cancer patients. Ann Oncol 14:1234-1240.

350. Mayle A, Luo M, Jeong M, Goodell MA. 2013. Flow cytometry analysis of murine hematopoietic stem cells. Cytometry A 83:27-37.

351. Katz Y, Li F, Lambert NJ, Sokol ES, Tam WL, Cheng AW, Airoldi EM, Lengner CJ, Gupta PB, Yu Z, Jaenisch R, Burge CB. 2014. Musashi proteins are posttranscriptional regulators of the epithelial-luminal cell state. Elife 3:e03915.

352. Jin Y, Yu J, Yu YG. 2002. Identification of hNopp140 as a binding partner for doxorubicin with a phage display cloning method. Chem Biol 9:157-162.

353. Rodi DJ, Janes RW, Sanganee HJ, Holton RA, Wallace BA, Makowski L. 1999. Screening of a library of phage-displayed peptides identifies human bcl-2 as a taxol-binding protein. J Mol Biol 285:197-203.

354. Cooper TA, Wan L, Dreyfuss G. 2009. RNA and disease. Cell 136:777-793.

355. Jonassen I, Collins JF, Higgins DG. 1995. Finding flexible patterns in unaligned

288

protein sequences. Protein Sci 4:1587-1595.

356. de Castro E, Sigrist CJ, Gattiker A, Bulliard V, Langendijk-Genevaux PS, Gasteiger E, Bairoch A, Hulo N. 2006. ScanProsite: detection of PROSITE signature matches and ProRule-associated functional and structural residues in proteins. Nucleic Acids Res 34:W362-365.

357. Aittaleb M, Boguth CA, Tesmer JJ. 2010. Structure and function of heterotrimeric G protein-regulated Rho guanine nucleotide exchange factors. Mol Pharmacol 77:111-125.

358. Vargo MA, Voss OH, Poustka F, Cardounel AJ, Grotewold E, Doseff AI. 2006. Apigenin-induced-apoptosis is mediated by the activation of PKCδ and caspases in leukemia cells. Biochem Pharmacol 72:681-692.

359. Aveic S, Pigazzi M, Basso G. 2011. BAG1: the guardian of anti-apoptotic proteins in acute myeloid leukemia. PLoS One 6:e26097.

360. Wang X, Chen M, Zhou J, Zhang X. 2014. HSP27, 70 and 90, anti-apoptotic proteins, in clinical cancer therapy (Review). Int J Oncol 45:18-30.

361. Qi F, Chen X, Beard DA. 2008. Detailed kinetics and regulation of mammalian NAD-linked isocitrate dehydrogenase. Biochim Biophys Acta 1784:1641-1651.

362. Egger S, Chaikuad A, Klimacek M, Kavanagh KL, Oppermann U, Nidetzky B. 2012. Structural and kinetic evidence that catalytic reaction of human UDP- glucose 6-dehydrogenase involves covalent thiohemiacetal and thioester enzyme intermediates. J Biol Chem 287:2119-2129.

363. Huhta MS, Chen HP, Hemann C, Hille CR, Marsh EN. 2001. Protein-coenzyme interactions in adenosylcobalamin-dependent glutamate mutase. Biochem J 355:131-137.

364. Yang X, Wu D, Du Z, Li R, Chen X, Li X. 2009. Spectroscopy study on the interaction of quercetin with collagen. J Agric Food Chem 57:3431-3435.

365. Iwashina T. 2000. The Structure and Distribution of the Flavonoids in Plants. J Plant Res 113:287-299.

366. Barnett SF, Theiry TA, LeStourgeon WM. 1991. The core proteins A2 and B1 exist as (A2)3B1 tetramers in 40S nuclear ribonucleoprotein particles. Mol Cell Biol 11:864-871.

367. Carson JH, Blondin N, Korza G. 2006. Rules of engagement promote polarity in RNA trafficking. BMC Neurosci 7 Suppl 1:S3.

289

368. Carson JH, Barbarese E. 2005. Systems analysis of RNA trafficking in neural cells. Biol Cell 97:51-62.

369. Huelga SC, Vu AQ, Arnold JD, Liang TY, Liu PP, Yan BY, Donohue JP, Shiue L, Hoon S, Brenner S, Ares M, Jr., Yeo GW. 2012. Integrative genome-wide analysis reveals cooperative regulation of alternative splicing by hnRNP proteins. Cell Rep 1:167-178.

370. Elkon R, Ugalde AP, Agami R. 2013. Alternative cleavage and polyadenylation: extent, regulation and function. Nat Rev Genet 14:496-506.

371. Kandoth C, McLellan MD, Vandin F, Ye K, Niu B, Lu C, Xie M, Zhang Q, McMichael JF, Wyczalkowski MA, Leiserson MD, Miller CA, Welch JS, Walter MJ, Wendl MC, Ley TJ, Wilson RK, Raphael BJ, Ding L. 2013. Mutational landscape and significance across 12 major cancer types. Nature 502:333-339.

372. Johnson JL, de Mejia EG. 2013. Flavonoid apigenin modified gene expression associated with inflammation and cancer and induced apoptosis in human pancreatic cancer cells through inhibition of GSK-3β/NF-κB signaling cascade. Mol Nutr Food Res 57:2112-2127.

373. Oshiumi H, Sasai M, Shida K, Fujita T, Matsumoto M, Seya T. 2003. TIR- containing adapter molecule (TICAM)-2, a bridging adapter recruiting to toll-like receptor 4 TICAM-1 that induces interferon-β. J Biol Chem 278:49751-49762.

374. Doyle SL, Husebye H, Connolly DJ, Espevik T, O'Neill LA, McGettrick AF. 2012. The GOLD domain-containing protein TMED7 inhibits TLR4 signalling from the endosome upon LPS stimulation. Nat Commun 3:707.

375. Palsson-McDermott EM, Doyle SL, McGettrick AF, Hardy M, Husebye H, Banahan K, Gong M, Golenbock D, Espevik T, O'Neill LA. 2009. TAG, a splice variant of the adaptor TRAM, negatively regulates the adaptor MyD88- independent TLR4 pathway. Nat Immunol 10:579-586.

376. Cunningham F, Amode MR, Barrell D, Beal K, Billis K, Brent S, Carvalho-Silva D, Clapham P, Coates G, Fitzgerald S, Gil L, Giron CG, Gordon L, Hourlier T, Hunt SE, Janacek SH, Johnson N, Juettemann T, Kahari AK, Keenan S, Martin FJ, Maurel T, McLaren W, Murphy DN, Nag R, Overduin B, Parker A, Patricio M, Perry E, Pignatelli M, Riat HS, Sheppard D, Taylor K, Thormann A, Vullo A, Wilder SP, Zadissa A, Aken BL, Birney E, Harrow J, Kinsella R, Muffato M, Ruffier M, Searle SM, Spudich G, Trevanion SJ, Yates A, Zerbino DR, Flicek P. 2015. Ensembl 2015. Nucleic Acids Res 43:D662-669.

377. Juan WC, Roca X, Ong ST. 2014. Identification of cis-acting elements and

290

splicing factors involved in the regulation of BIM Pre-mRNA splicing. PLoS One 9:e95210.

378. Wiggins CM, Tsvetkov P, Johnson M, Joyce CL, Lamb CA, Bryant NJ, Komander D, Shaul Y, Cook SJ. 2011. BIM(EL), an intrinsically disordered protein, is degraded by 20S proteasomes in the absence of poly-ubiquitylation. J Cell Sci 124:969-977.

379. Ley R, Balmanno K, Hadfield K, Weston C, Cook SJ. 2003. Activation of the ERK1/2 signaling pathway promotes phosphorylation and proteasome-dependent degradation of the BH3-only protein, Bim. J Biol Chem 278:18811-18816.

380. Qi XJ, Wildey GM, Howe PH. 2006. Evidence that Ser87 of BimEL is phosphorylated by Akt and regulates BimEL apoptotic function. J Biol Chem 281:813-823.

381. Gilley R, Lochhead PA, Balmanno K, Oxley D, Clark J, Cook SJ. 2012. CDK1, not ERK1/2 or ERK5, is required for mitotic phosphorylation of BIMEL. Cell Signal 24:170-180.

382. Davoodi J, Ghahremani MH, Es-Haghi A, Mohammad-Gholi A, Mackenzie A. 2010. Neuronal apoptosis inhibitory protein, NAIP, is an inhibitor of procaspase- 9. Int J Biochem Cell Biol 42:958-964.

383. Li G, Chang H, Zhai YP, Xu W. 2013. Targeted silencing of inhibitors of apoptosis proteins with siRNAs: a potential anti-cancer strategy for hepatocellular carcinoma. Asian Pac J Cancer Prev 14:4943-4952.

384. Fu J, Jin Y, Arend LJ. 2003. Smac3, a novel Smac/DIABLO splicing variant, attenuates the stability and apoptosis-inhibiting activity of X-linked inhibitor of apoptosis protein. J Biol Chem 278:52660-52672.

385. Shi B, Triebe D, Kajiji S, Iwata KK, Bruskin A, Mahajna J. 1999. Identification and characterization of baxε, a novel bax variant missing the BH2 and the transmembrane domains. Biochem Biophys Res Commun 254:779-785.

386. Dales JP, Beaufils N, Silvy M, Picard C, Pauly V, Pradel V, Formisano-Treziny C, Bonnier P, Giusiano S, Charpin C, Gabert J. 2010. Hypoxia inducible factor 1α gene (HIF-1α) splice variants: potential prognostic biomarkers in breast cancer. BMC Med 8:44.

387. Ruas JL, Poellinger L, Pereira T. 2002. Functional analysis of hypoxia-inducible factor-1 α-mediated transactivation. Identification of amino acid residues critical for transcriptional activation and/or interaction with CREB-binding protein. J Biol Chem 277:38723-38730.

291

388. Camats M, Kokolo M, Heesom KJ, Ladomery M, Bach-Elias M. 2009. P19 H-ras induces G1/S phase delay maintaining cells in a reversible quiescence state. PLoS One 4:e8513.

389. Huang MY, Cohen JB. 1997. The alternative H-ras protein p19 displays properties of a negative regulator of p21Ras. Oncol Res 9:611-621.

390. Park SM, Deering RP, Lu Y, Tivnan P, Lianoglou S, Al-Shahrour F, Ebert BL, Hacohen N, Leslie C, Daley GQ, Lengner CJ, Kharas MG. 2014. Musashi-2 controls cell fate, lineage bias, and TGF-β signaling in HSCs. J Exp Med 211:71- 87.

391. Lee JE, Lee JY, Wilusz J, Tian B, Wilusz CJ. 2010. Systematic analysis of cis- elements in unstable mRNAs demonstrates that CUGBP1 is a key regulator of mRNA decay in muscle cells. PLoS One 5:e11201.

392. Koshelev M, Sarma S, Price RE, Wehrens XH, Cooper TA. 2010. Heart-specific overexpression of CUGBP1 reproduces functional and molecular abnormalities of myotonic dystrophy type 1. Hum Mol Genet 19:1066-1075.

393. Dickinson LA, Edgar AJ, Ehley J, Gottesfeld JM. 2002. Cyclin L is an RS domain protein involved in pre-mRNA splicing. J Biol Chem 277:25465-25473.

394. Yang L, Li N, Wang C, Yu Y, Yuan L, Zhang M, Cao X. 2004. Cyclin L2, a novel RNA polymerase II-associated cyclin, is involved in pre-mRNA splicing and induces apoptosis of human hepatocellular carcinoma cells. J Biol Chem 279:11639-11648.

395. Li HL, Huang DZ, Deng T, Zhou LK, Wang X, Bai M, Ba Y. 2012. Overexpression of cyclin L2 inhibits growth and enhances chemosensitivity in human gastric cancer cells. Asian Pac J Cancer Prev 13:1425-1430.

396. Zhuo L, Gong J, Yang R, Sheng Y, Zhou L, Kong X, Cao K. 2009. Inhibition of proliferation and differentiation and promotion of apoptosis by cyclin L2 in mouse embryonic carcinoma P19 cells. Biochem Biophys Res Commun 390:451- 457.

397. Derda R, Tang SK, Li SC, Ng S, Matochko W, Jafari MR. 2011. Diversity of phage-displayed libraries of peptides during panning and amplification. Molecules 16:1776-1803.

398. Kufe DW. 2009. Mucins in cancer: function, prognosis and therapy. Nat Rev Cancer 9:874-885.

292

399. Leng Y, Cao C, Ren J, Huang L, Chen D, Ito M, Kufe D. 2007. Nuclear import of the MUC1-C oncoprotein is mediated by nucleoporin Nup62. J Biol Chem 282:19321-19330.

400. Ahmad R, Raina D, Joshi MD, Kawano T, Ren J, Kharbanda S, Kufe D. 2009. MUC1-C oncoprotein functions as a direct activator of the NF-κB p65 transcription factor. Cancer Res 69:7013-7021.

401. Takeshita M, Ishida Y, Akamatsu E, Ohmori Y, Sudoh M, Uto H, Tsubouchi H, Kataoka H. 2009. Proanthocyanidin from blueberry leaves suppresses expression of subgenomic hepatitis C virus RNA. J Biol Chem 284:21165-21176.

402. Day AJ, Canada FJ, Diaz JC, Kroon PA, McLauchlan R, Faulds CB, Plumb GW, Morgan MR, Williamson G. 2000. Dietary flavonoid and isoflavone glycosides are hydrolysed by the lactase site of lactase phlorizin hydrolase. FEBS Lett 468:166-170.

403. Hwang EY, Huh JW, Choi MM, Choi SY, Hong HN, Cho SW. 2008. Inhibitory effects of gallic acid and quercetin on UDP-glucose dehydrogenase activity. FEBS Lett 582:3793-3797.

404. Zhao M, Ma J, Zhu HY, Zhang XH, Du ZY, Xu YJ, Yu XD. 2011. Apigenin inhibits proliferation and induces apoptosis in human multiple myeloma cells through targeting the trinity of CK2, Cdc37 and Hsp90. Mol Cancer 10:104.

405. Pancetti F, Bosser R, Krehan A, Pyerin W, Itarte E, Bachs O. 1999. Heterogeneous nuclear ribonucleoprotein A2 interacts with protein kinase CK2. Biochem Biophys Res Commun 260:17-22.

406. Bennick A. 2002. Interaction of plant polyphenols with salivary proteins. Crit Rev Oral Biol Med 13:184-196.

407. Le Bourvellec C, Renard CM. 2012. Interactions between polyphenols and macromolecules: quantification methods and mechanisms. Crit Rev Food Sci Nutr 52:213-248.

408. Hou DX, Kumamoto T. 2010. Flavonoids as protein kinase inhibitors for cancer chemoprevention: direct binding and molecular modeling. Antioxid Redox Signal 13:691-719.

409. Li Y, Revalde JL, Reid G, Paxton JW. 2010. Interactions of dietary phytochemicals with ABC transporters: possible implications for drug disposition and multidrug resistance in cancer. Drug Metab Rev 42:590-611.

410. Zhang W, Qiao H, Lv Y, Wang J, Chen X, Hou Y, Tan R, Li E. 2014. Apigenin

293

inhibits enterovirus-71 infection by disrupting viral RNA association with trans- acting factors. PLoS One 9:e110429.

411. Rolo-Naranjo A, Codorniu-Hernandez E, Ferro N. 2010. Quantum chemical associations ligand-residue: their role to predict flavonoid binding sites in proteins. J Chem Inf Model 50:924-933.

412. Boese AD, Codorniu-Hernandez E. 2012. Cross-talk between amino acid residues and flavonoid derivatives: insights into their chemical recognition. Phys Chem Chem Phys 14:15682-15692.

413. Goodarzi H, Najafabadi HS, Oikonomou P, Greco TM, Fish L, Salavati R, Cristea IM, Tavazoie S. 2012. Systematic discovery of structural elements governing stability of mammalian messenger RNAs. Nature 485:264-268.

414. Jing Q, Huang S, Guth S, Zarubin T, Motoyama A, Chen J, Di Padova F, Lin SC, Gram H, Han J. 2005. Involvement of microRNA in AU-rich element-mediated mRNA instability. Cell 120:623-634.

415. Behm-Ansmant I, Rehwinkel J, Doerks T, Stark A, Bork P, Izaurralde E. 2006. mRNA degradation by miRNAs and GW182 requires both CCR4:NOT deadenylase and DCP1:DCP2 decapping complexes. Genes Dev 20:1885-1898.

416. Hallais M, Pontvianne F, Andersen PR, Clerici M, Lener D, Benbahouche Nel H, Gostan T, Vandermoere F, Robert MC, Cusack S, Verheggen C, Jensen TH, Bertrand E. 2013. CBC-ARS2 stimulates 3'-end maturation of multiple RNA families and favors cap-proximal processing. Nat Struct Mol Biol 20:1358-1366.

417. Ge Y, Porse BT. 2014. The functional consequences of intron retention: alternative splicing coupled to NMD as a regulator of gene expression. Bioessays 36:236-243.

418. Chan WK, Bhalla AD, Le Hir H, Nguyen LS, Huang L, Gecz J, Wilkinson MF. 2009. A UPF3-mediated regulatory switch that maintains RNA surveillance. Nat Struct Mol Biol 16:747-753.

419. Mandadi KK, Scholthof KB. 2015. Genome-wide analysis of alternative splicing landscapes modulated during plant-virus interactions in Brachypodium distachyon. Plant Cell 27:71-85.

420. O'Brien K, Matlin AJ, Lowell AM, Moore MJ. 2008. The biflavonoid isoginkgetin is a general inhibitor of Pre-mRNA splicing. J Biol Chem 283:33147-33154.

421. Dutertre M, Chakrama FZ, Combe E, Desmet FO, Mortada H, Polay Espinoza M,

294

Gratadou L, Auboeuf D. 2014. A recently evolved class of alternative 3'-terminal exons involved in cell cycle regulation by topoisomerase inhibitors. Nat Commun 5:3395.

422. Patry C, Bouchard L, Labrecque P, Gendron D, Lemieux B, Toutant J, Lapointe E, Wellinger R, Chabot B. 2003. Small interfering RNA-mediated reduction in heterogeneous nuclear ribonucleoparticule A1/A2 proteins induces apoptosis in human cancer cells but not in normal mortal cell lines. Cancer Res 63:7679-7688.

423. Moran-Jones K, Grindlay J, Jones M, Smith R, Norman JC. 2009. hnRNP A2 regulates alternative mRNA splicing of TP53INP2 to control invasive cell migration. Cancer Res 69:9219-9227.

424. Shao H, Jing K, Mahmoud E, Huang H, Fang X, Yu C. 2013. Apigenin sensitizes colon cancer cells to antitumor activity of ABT-263. Mol Cancer Ther 12:2640- 2650.

425. Lu HF, Chie YJ, Yang MS, Lee CS, Fu JJ, Yang JS, Tan TW, Wu SH, Ma YS, Ip SW, Chung JG. 2010. Apigenin induces caspase-dependent apoptosis in human lung cancer A549 cells through Bax- and Bcl-2-triggered mitochondrial pathway. Int J Oncol 36:1477-1484.

426. Chan LP, Chou TH, Ding HY, Chen PR, Chiang FY, Kuo PL, Liang CH. 2012. Apigenin induces apoptosis via tumor necrosis factor receptor- and Bcl-2- mediated pathway and enhances susceptibility of head and neck squamous cell carcinoma to 5-fluorouracil and cisplatin. Biochim Biophys Acta 1820:1081-1091.

427. Arango D, Morohashi K, Yilmaz A, Kuramochi K, Parihar A, Brahimaj B, Grotewold E, Doseff AI. 2013. Molecular basis for the action of a dietary flavonoid revealed by the comprehensive identification of apigenin human targets. Proc Natl Acad Sci U S A 110:E2153-2162.

428. Patterson G, Day RN, Piston D. 2001. Fluorescent protein spectra. J Cell Sci 114:837-838.

429. Pietta PG. 2000. Flavonoids as antioxidants. J Nat Prod 63:1035-1042.

430. Leung HW, Wu CH, Lin CH, Lee HZ. 2005. Luteolin induced DNA damage leading to human lung squamous carcinoma CH27 cell apoptosis. Eur J Pharmacol 508:77-83.

431. Higa LA, Banks D, Wu M, Kobayashi R, Sun H, Zhang H. 2006. L2DTL/CDT2 interacts with the CUL4/DDB1 complex and PCNA and regulates CDT1 proteolysis in response to DNA damage. Cell Cycle 5:1675-1680.

295

432. Lisby M, Rothstein R, Mortensen UH. 2001. Rad52 forms DNA repair and recombination centers during S phase. Proc Natl Acad Sci U S A 98:8276-8282.

433. Auclair Y, Rouget R, Belisle JM, Costantino S, Drobetsky EA. 2010. Requirement for functional DNA polymerase η in genome-wide repair of UV- induced DNA damage during S phase. DNA Repair (Amst) 9:754-764.

434. Shiomi Y, Hayashi A, Ishii T, Shinmyozu K, Nakayama J, Sugasawa K, Nishitani H. 2012. Two different replication factor C proteins, Ctf18 and RFC1, separately control PCNA-CRL4Cdt2-mediated Cdt1 proteolysis during S phase and following UV irradiation. Mol Cell Biol 32:2279-2288.

435. Christmann M, Tomicic MT, Aasland D, Berdelle N, Kaina B. 2010. Three prime exonuclease I (TREX1) is Fos/AP-1 regulated by genotoxic stress and protects against ultraviolet light and benzo(a)pyrene-induced DNA damage. Nucleic Acids Res 38:6418-6432.

436. O'Driscoll M. 2008. TREX1 DNA exonuclease deficiency, accumulation of single stranded DNA and complex human genetic disorders. DNA Repair (Amst) 7:997- 1003.

437. Wang C, Liu X, Liu Y, Zhang Q, Yao Z, Huang B, Zhang P, Li N, Cao X. 2013. Zinc finger protein 64 promotes Toll-like receptor-triggered proinflammatory and type I interferon production in macrophages by enhancing p65 subunit activation. J Biol Chem 288:24600-24608.

438. Hoffmann MH, Tuncel J, Skriner K, Tohidast-Akrad M, Turk B, Pinol-Roma S, Serre G, Schett G, Smolen JS, Holmdahl R, Steiner G. 2007. The rheumatoid arthritis-associated autoantigen hnRNP-A2 (RA33) is a major stimulator of autoimmunity in rats with pristane-induced arthritis. J Immunol 179:7568-7576.

439. Yu M, Yeh J, Van Waes C. 2006. Protein kinase casein kinase 2 mediates IκB kinase and aberrant NF-κB activation by serum factor(s) in head and neck squamous carcinoma cells. Cancer Res 66:6722-6731.

440. Kato T, Jr., Delhase M, Hoffmann A, Karin M. 2003. CK2 Is a C-Terminal IκB Kinase Responsible for NF-κB Activation during the UV Response. Mol Cell 12:829-839.

441. Parhar K, Morse J, Salh B. 2007. The role of protein kinase CK2 in intestinal epithelial cell inflammatory signaling. Int J Colorectal Dis 22:601-609.

442. Taylor JA, Bren GD, Pennington KN, Trushin SA, Asin S, Paya CV. 1999. Serine 32 and serine 36 of IκBα are directly phosphorylated by protein kinase CKII in vitro. J Mol Biol 290:839-850.

296

443. Wang D, Westerheide SD, Hanson JL, Baldwin AS, Jr. 2000. Tumor necrosis factor α-induced phosphorylation of RelA/p65 on Ser529 is controlled by casein kinase II. J Biol Chem 275:32592-32597.

444. Bird TA, Schooley K, Dower SK, Hagen H, Virca GD. 1997. Activation of nuclear transcription factor NF-κB by interleukin-1 is accompanied by casein kinase II-mediated phosphorylation of the p65 subunit. J Biol Chem 272:32606- 32612.

445. Guha M, Fang JK, Monks R, Birnbaum MJ, Avadhani NG. 2010. Activation of Akt is essential for the propagation of mitochondrial respiratory stress signaling and activation of the transcriptional coactivator heterogeneous ribonucleoprotein A2. Mol Biol Cell 21:3578-3589.

446. Guha M, Pan H, Fang JK, Avadhani NG. 2009. hnRNP A2 Is a Common Transcriptional Coactivator in the Nuclear Transcription Response to Mitochondrial Respiratory Stress. Mol Biol Cell.

447. Leeman JR, Gilmore TD. 2008. Alternative splicing in the NF-κB signaling pathway. Gene 423:97-107.

448. Liu P, Lu M, Tian B, Li K, Garofalo RP, Prusak D, Wood TG, Brasier AR. 2009. Expression of an IKKγ splice variant determines IRF3 and canonical NF-κB pathway utilization in ssRNA virus infection. PLoS One 4:e8079.

449. Guo F, Sigua C, Bali P, George P, Fiskus W, Scuto A, Annavarapu S, Mouttaki A, Sondarva G, Wei S, Wu J, Djeu J, Bhalla K. 2005. Mechanistic role of heat shock protein 70 in Bcr-Abl-mediated resistance to apoptosis in human acute leukemia cells. Blood 105:1246-1255.

450. Zhuang H, Jiang W, Zhang X, Qiu F, Gan Z, Cheng W, Zhang J, Guan S, Tang B, Huang Q, Wu X, Huang X, Jiang W, Hu Q, Lu M, Hua ZC. 2013. Suppression of HSP70 expression sensitizes NSCLC cell lines to TRAIL-induced apoptosis by upregulating DR4 and DR5 and downregulating c-FLIP-L expressions. J Mol Med (Berl) 91:219-235.

451. Takayama S, Sato T, Krajewski S, Kochel K, Irie S, Millan JA, Reed JC. 1995. Cloning and functional analysis of BAG-1: a novel Bcl-2-binding protein with anti-cell death activity. Cell 80:279-284.

452. Roth W, Isenmann S, Nakamura M, Platten M, Wick W, Kleihues P, Bahr M, Ohgaki H, Ashkenazi A, Weller M. 2001. Soluble decoy receptor 3 is expressed by malignant gliomas and suppresses CD95 ligand-induced apoptosis and chemotaxis. Cancer Res 61:2759-2765.

297

453. Pitti RM, Marsters SA, Lawrence DA, Roy M, Kischkel FC, Dowd P, Huang A, Donahue CJ, Sherwood SW, Baldwin DT, Godowski PJ, Wood WI, Gurney AL, Hillan KJ, Cohen RL, Goddard AD, Botstein D, Ashkenazi A. 1998. Genomic amplification of a decoy receptor for Fas ligand in lung and colon cancer. Nature 396:699-703.

454. Rorke S, Murphy S, Khalifa M, Chernenko G, Tang SC. 2001. Prognostic significance of BAG-1 expression in nonsmall cell lung cancer. Int J Cancer 95:317-322.

455. Murphy ME. 2013. The HSP70 family and cancer. Carcinogenesis 34:1181-1188.

456. Hassett MJ, O'Malley AJ, Pakes JR, Newhouse JP, Earle CC. 2006. Frequency and cost of chemotherapy-related serious adverse effects in a population sample of women with breast cancer. J Natl Cancer Inst 98:1108-1117.

298

Appendix A

HnRNPA2 Regulates NF-κB Phosphorylation

HnRNPA2 is the main autoantigen in arthritis rheumatoid promoting expression of pro-inflammatory cytokines such as TNFα [438]. However, the role of hnRNPA2 in inflammation has yet to be uncovered. To determine whether hnRNPA2 regulates the innate immune response in macrophages, we transfected the mouse macrophage cell line

RAW264.7, with an siRNA against hnRNPA2 or a scramble control. Forty-eight hours after transfection, cells were concomitantly treated with 100 ng/ml LPS and DMSO, 100 ng/ml LPS and 25 µM apigenin, PBS and 25 µM apigenin or PBS and DMSO control for

15 min to evaluate phosphorylation of NF-κB-p65Ser536 by western blot and for 4 h to determine TNFα mRNA expression by qRT-PCR. We observed that LPS induces NF-κB- p65 phosphorylation in cells transfected with scramble control (Fig. A1a, lane 6 vs. 5), but had no effect in siRNA-hnRNPA2 cells (Fig. A1a, lane 2 vs. lane 1). Apigenin

treatment decreased the LPS-induced NF-κB-p65 phosphorylation to control levels in

cells transfected with siRNA control (Fig. A1a, lane 7 vs. 6 and 5), and had no effect in cells where hnRNPA2 was silenced (Fig. A1a, lane 3 vs. 2 and 1). Apigenin did not

change NF-κB-p65 phosphorylation in cells silenced with siRNA-hnRNPA2 or scramble

299

control that were not stimulated with LPS (Fig. A1a, lanes 4 and 8). Consistently, silencing of hnRNPA2 decreased by ~2 fold the LPS-induced TNFα expression compared to cells transfected with scramble control and treated with LPS (Fig. A1b, blue bars). Apigenin treatment reduced the expression of TNFα in cells transfected with siRNA-hnRNPA2 or siRNA-control (Fig. A1b, red compared to blue bars). Yet, there was no significant difference between cells where hnRNPA2 was silenced and treated with LPS or LPS and Api (Fig. A1b, red compared to blue bars in siRNA-hnRNPA2).

These results indicate that hnRNPA2 regulates NF-κB phosphorylation during LPS- induced inflammation in macrophages.

Next, we determined whether hnRNPA2 regulates NF-κB phosphorylation in breast cancer cells. For this purpose, MDA-MB-231 cells were transfected with siRNA-control or siRNA-hnRNPA2 and after 48 h cells were treated with 50 µM apigenin or diluent

DMSO for additional 48 h. At this time NF-κB-p65Ser536 phosphorylation was evaluated by Western blot. We observed that NF-κB-p65 subunit is constitutively phosphorylated in breast cancer cells (Fig. A1c lane 1), as previously shown [93, 94]. Apigenin treatment abolished NF-κB phosphorylation in cells silenced with siRNA-hnRNPA2 or siRNA- control (Fig. A1c lanes 2 and 4). Cells silenced with siRNA-hnRNPA2 and treated with

DMSO showed a ~2-fold decreased in NF-κB phosphorylation compared to cells transfected with scramble control and treated with DMSO (Fig. A1c lanes 3 vs. 1).

Altogether, these results suggest that apigenin affects NF-κB phosphorylation in an hnRNPA2-mediated pathway. Yet, hnRNPA2 only partially reduces LPS-induced TNFα expression in macrophages and NF-κB phosphorylation in cancer cells indicating that

300

apigenin may regulate NF-κB activity by more than one mechanism (chapter 7).

301

A !"#$%&,-(./-0+ &+ &+ &+ &+ 4+ 4+ 4+ 4+ !"#$%&'(#$)%*+ 4+ 4+ 4+ 4+ &+ &+ &+ &+ 5):+ &+ 4+ 4+ &+ &+ 4+ 4+ &+ %1"23("(+ &+ &+ 4+ 4+ &+ &+ 4+ 4+ p-p65

hnRNPA2

GAPDH

56(3++ 7+ *+ 8+ 9+ ;+ <+ =+ >+

B 25 TNF * ! 20 DMSO LPS 15 LPS + Api

10 Fold Change

0 siRNA-Control siRNA-hnRNPA2

C !"#$%&,-(./-0+ 4+ 4+ &+ &+ !"#$%&'(#$)%*+ &+ &+4+ 4+ %1"23("(+ &+ 4+ &+ 4+ p-p65

hnRNPA2

GAPDH 56(3++ 7+ *+ 8+ 9+

Figure A1. HnRNPA2 regulates NF-"B-p65 phosphorylation. A. RAW264.7 cells were transfected with siRNA-hnRNPA2 (lanes 1-4) or siRNA-control (lanes 5-8) and after 48 h cells were concomitantly treated with PBS and DMSO (lanes 1 and 5), 100 ng/ml LPS and diluent DMSO (lanes 2 and 6), 100 ng/LPS and 25 µM apigenin (lanes 3 and 7) or PBS and 25 µM apigenin (lanes 4 and 8) for additional 15 min. Cells lysates were resolved by SDS-PAGE and immunoblotted using anti-phospho-NF-"B-p65Ser536, anti-hnRNPA2 and anti-GAPDH antibodies. n = 1. B. RAW264.7 cells were silenced as described in (A) followed by concomitant treatment with PBS and DMSO (black bars), 100 ng/ml LPS and diluent DMSO (blue bars) or 100 ng/LPS and 25 µM apigenin (red bars) for 4 h. RNA was isolated and used to determine TNF# mRNA levels by qRT-PCR. Data represent mean ± SEM, n=3. * p < 0.05, two-tailed t-test. C. MDA- MB-231 cells were transfected with siRNA-control (lanes 1 and 2) or siRNA-hnRNPA2 (lanes 3 and 4). Forty-eight hours after transfections cells were treated diluent DMSO (lanes 1 and 3) or 50 µM apigenin (lanes 2 and 4). Cells lysates were resolved by SDS-PAGE and immunoblotted as described in (A). n = 1. 302

List of Abbreviations

α-SMA Alpha-smooth muscle actin A Adenoma A3SS Alternative 3' splicing site A5SS Alternative 5' splicing site AF Alexa Fluor AFE Alternative first exon AGO1 Argonaut 1 ALE Alternative last exon ALPHA Amplified Luminescent Proximity Homogeneous Assay AML Acute myeloid leukemia APA Alternative polyadenylation APAF1 Apoptotic peptidase activating factor 1 ARHGEF1 Rho guanine nucleotide exchange factor 1 AS Alternative Splicing ATM Ataxia-telangiectasia mutated ATR Ataxia-telangiectasia related ATSS Alternative transcription start site BACH1 BTB And CNC Homology 1 BAD BCL2-Associated Agonist Of Cell Death BAG1 BCL2-Associated Athanogene BAK1 BCL2-Antagonist/Killer 1 BALF Broncho-alveolar lavage fluids BAX BCL2-Associated X Protein BBC3 BCL2 Binding Component 3 BC Breast cancer BCL2 B-Cell CLL/Lymphoma 2 BCL2L1 BCL2-Like 1 BCL2L11 BCL2-Like 11 BF Bayes Factor BID BH3 Interacting Domain Death Agonist BIRC Baculoviral IAP Repeat Containing BMDM Bone marrow derived macrophages BRCA1 Breast Cancer 1 BSA Bovine Serum Albumin CAN Acetonitrile

303

CASP Comet Assay Software Project CCL2 C-C motif ligand 2 CCNA2 Cyclin A1 CCNB1 Cyclin B1 CCNE1 Cyclin E1 CCNL2 Cyclin L2 CCNL2-RI Cyclin L2-retained intron CCR2 Chemokine (C-C Motif) Receptor 2 CD117 Cluster of differentiation 117 CD11b Cluster of differentiation 11b CD16//32 Cluster of differentiation 16/32 CD3 Cluster of differentiation 3 CD4 Cluster of differentiation 8 CD49b Cluster of differentiation 49b CD8 Cluster of differentiation 8 CDC25A Cell Division Cycle 25A CDG Cancer driver genes CDK2 Cyclin dependent kinase 2 CDKN1A Cyclin-Dependent Kinase Inhibitor 1A (p21) CELF1 CUGBP, Elav-Like Family Member 1 CFLAR CASP8 And FADD-Like Apoptosis Regulator (cFLIP) CFP Cyan fluorescent protein CK2 Casein Kinase 2 CLAP Chymostatin, leupeptin, antipain, and pepstatin CMP Common myeloid progenitors COX-2 Cyclooxigenase-2 CPSF3 Cleavage And Polyadenylation Specific Factor 3 CSF1 Colony stimulator factor 1 CSF1R Colony stimulator factor 1 receptor CSTF2 Cleavage Stimulation Factor, 3' Pre-RNA, Subunit 2 CXCL12 CXC chemokine ligand 12 CYC Cytochrome c DAPI 4’,6-diamidino-2-phenylindole DAVID Database for Annotation, Visualization and Integrated Discovery DCFDA 2,7-dichlorofluorescein diacetate DHE Dihydroethidium DISC Death inducing signaling complex DMEM Dulbecco's Modified Eagle Medium DMSO Dimethyl sulfoxide DR5 Decoy receptor 5 DSB Double Strand Breaks DTT Dithiothreitol 304

E2F2 E2F Transcription Factor 2 EC Early Carcinoma ECE Enzyme-treated Celery Extracts ECGC Epigallocatechin gallate ECL Enhanced Chemo-Luminesce ECP Enzyme-treated Celery Pellets EGF Epidermal Growth Factor ELISA Enzyme-linked immuno-absorbent assay EMT Epithelial–mesenchymal transition ER Estrogen receptor ERBB2 V-Erb-B2 Avian Erythroblastic Leukemia Viral Oncogene Homolog 2 FADD Fas-associated Death Domain FASL Fas ligand FBS Fetal bovine serum FC Flow Cytometry FDR false discovery rate FEN1 Flap Structure-Specific Endonuclease 1 FLIP Fluorescent indicator protein FOXO3A Forkhead Box O3 FRET Fluorescence Resonance Energy Transfer FRET Fluorescence resonance energy transfer FVB Friend Virus B-Type G-MDSC Granulocyte-Myeloid derived suppressor cells GADD45A Growth Arrest And DNA-Damage-Inducible, Alpha GEO Gene expression omnibus GFP Green Fluorescent Protein GMP Granulocyte/macrophage progenitors GO Gene Ontology GRD Glycine rich domain GSH Glutathione GST Glutathione S-Transferase H Hyperplasia H&E Hematoxylin and Eosin H2AX Histone 2A, member X H2B Histone H2B HBSS Hank's Balanced Salt Solution HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HER2 Human epidermal growth factor receptor 2 HIF1A Hypoxia-inducible factor 1-alpha HIS Histidine HNRNP Heterogeneous nuclear ribonucleoprotein HNRNPA2 Heterogeneous nuclear ribonucleoprotein A2 305

HPLC High Performance Liquid Chromatography HRAS Harvey Rat Sarcoma Viral Oncogene HRP Horseradish peroxidase HSC Hematopoietic stem cells HSP27 Heat shock protein 27 HTS High-Throughput Screening IAP Inhibitor of apoptosis IDH3 Isocitrate dehydrogenase 3 IF Immunofluorescence IHC Immunohistochemistry IKKα Inhibitor Of Kappa Light Polypeptide Gene Enhancer In B-Cells, Kinase alpha IKKβ Inhibitor Of Kappa Light Polypeptide Gene Enhancer In B-Cells, Kinase Beta IKKγ Inhibitor Of Kappa Light Polypeptide Gene Enhancer In B-Cells, Kinase gamma IMDM Iscove's Modified Dulbecco's Medium IP Immunoprecipitation IPA Ingenuity pathway analysis IRAK IL-1 receptor-associated kinase KL Kit lineage KSL Kit and sca-1 lineage LC Late Carcinoma LC-MS Liquid Chromatography-Mass Spectrometry LMPA Low melting point agarose LPS Lipopolysaccharide LY6C Lymphocyte antigen 6C LY6G Lymphocyte antigen 6G M1 Macrophages M1 M2 Macrophages M2 MAPK Mitogen activated protein kinase MCL1 Myeloid Cell Leukemia 1 MCS Multicloning site MDSC Myeloid derived suppressor cells miR microRNA MISO Mixtures of isoforms mKO Monomeric Kusabira-Orange MMTV Mouse Mammary Tumor Virus MO-MDSC Monocytic-myeloid derived suppressor cells MSI2 Musashi 2 MTOR Mammalian target of rapamycin MUC1 Mucin 1 MXE Mutually exclusive exons MYC Myelocytomatosis Viral Oncogene

306

MYD88 Myeloid Differentiation Primary Response 88 NAD+ Nicotinamide Adenine Dinucleotide NAIP NLR Family, Apoptosis Inhibitory Protein NBT Normal Breast Tissue NF-κB Nuclear Factor kappa B nICPG Normalized In-frame-aligned Counts Per Gene-model NMPA Normal melting point agarose NO Nitric Oxide OG Oncogenes P/S Penicillin/streptomycin PAGE Polyacrylamide Gel Electrophoresis PAMP Pathogen associated molecular patterns PBS Phosphate buffer solution PD-seq Phage Display couple with Illumina sequencing PEGA Polyethyleneglycol-Polyacrylamide Copolymer Beads PEGA Polyethyleneglycol-polyacrylamide copolymer PI Propidium Iodide PI3K Phosphatidylinositol 3-Kinase PIPES Piperazine-N,N′-bis(2-ethanesulfonic acid) PKCδ Protein kinase C delta PMSF Phenyl-methane-sulfonyl-fluoride POLH DNA polymerase eta PR Progesterone receptor PSI Percent Splice Index PyMT Polyoma Virus middle T antigen qRT-PCR Quantitative reverse transcriptase-polymerase chain reaction RBD RNA binding domain RBM3 RNA binding motif 3 RI Retained intron RIN RNA integrity number ROS Reactive Oxygen Species RPMI-1640 Roswell Park Memorial Institute medium-1640 RPS9 Ribosomal protein S9 SCA-1 Stem cell antigen 1 SCID Severe combined immuno-deficient SDS Sodium Dodecyl Sulfate SE Skipped exon SMAD2 Smooth-muscle-actin and MAD-related 2 SRRT Serrate RNA Effector Molecule SRSF5 Serine arginine splicing factor 5 TAK1 TGF-β activated kinase 1 TAM Tumor associated macrophages 307

TFA Trifluoroacetic acid TGF-β Tumor growth factor beta TICAM2 Tir-containing adaptor molecule-2 TLR Toll-like receptor TMED7 Transmembrane Emp24 Protein Transport Domain Containing 7 TNBC Triple negative breast cancer TNFα Tumor necrosis factor alpha TNFRSF10B Tumor Necrosis Factor Receptor Superfamily, Member 10b TNFRSF1A Tumor Necrosis Factor Receptor Superfamily, Member 1A TNFRSF6B Tumor Necrosis Factor Receptor Superfamily, Member 6b, Decoy TRAF6 TNF receptor-associated factor 6 TRAIL TNF-related apoptosis inducing ligand TS Tumor suppressor TSC1 Tuberous sclerosis 1 TUNEL Terminal Uridine Nick-End Labeling TUTR Tandem UTR UDP-glucose Uridine 5'-diphosphoglucose UGDH UDP-glucose-6-dehydrogenase UTR Untranslated region UV Ultraviolet VEGFA Vascular Endothelial Growth Factor A WB Western Blot XIAP X-Linked Inhibitor Of Apoptosis XRCC2 X-Ray Repair Complementing Defective Repair In Chinese Hamster Cells 2 YFP Yellow Fluorescent Protein

308