2. the Hippo Pathway Effector TAZ Regulates Ferroptosis in Renal Cell Carcinoma

Molecular Mechanisms of TAZ-regulating Ferroptosis in Cancer Cells and tRNA

Fragment in Erythrocytes

Wen-Hsuan Yang

Department of Biochemistry Duke University

Date:______Approved:

______Jen-Tsan Ashley Chi, Supervisor

______Kate Meyer

______Margarethe Kuehn

______Perry Blackshear

Dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Biochemistry in the Graduate School of Duke University

2019

ABSTRACT

Molecular Mechanisms of TAZ-regulating Ferroptosis in Cancer Cells and tRNA

Fragment in Erythrocytes

Wen-Hsuan Yang

Department of Biochemistry Duke University

Date:______Approved:

______Jen-Tsan Ashley Chi, Supervisor

______Kate Meyer

______Margarethe Kuehn

______Perry Blackshear

An abstract of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Biochemistry in the Graduate School of Duke University

2019

Abstract

Here, I sought to determine the molecular mechanisms of the cellular response to stresses in two contexts. In the first part of my thesis, I focus on how ferroptosis, a lipid oxidative stress-induced cell death, can be regulated by cell density via an evolutionarily conserved pathway effector. In the second part, I focus on the transcriptional response of red blood cells (RBCs) during the refrigerated storage.

Ferroptosis is a novel form of programmed cell death characterized by the accumulation of lipid peroxidation. It can be induced by the oxidative stress caused by starvation of cystine, inhibition of glutathione peroxidase 4, or activation of NADPH oxidase(s). The canonical ferroptosis inducer, erastin, is a small molecule which triggers oxidative stress by inhibiting the cystine-glutamate transporter (xCT) and thus reduces intracellular cysteine level and glutathione biosynthesis. Recent studies indicate ferroptosis may have therapeutic potential toward cancer. However, much remains unknown about the determinants of ferroptosis susceptibility. We observed that vulnerability to the ferroptosis of cancer cells is highly influenced by cell confluency.

Since cell density can be sensed by the evolutionarily conserved Hippo pathway effectors, YAP/TAZ, we hypothesize if these Hippo pathway effectors are involved in erastin-induced ferroptosis response. My data show that TAZ, instead of YAP, is abundantly expressed in both renal and ovarian cancer cells and undergoes density-

dependent nuclear/cytosolic translocation. TAZ removal confers ferroptosis resistance, while overexpression of constitutively active form of TAZ, TAZS89A, sensitizes cells to ferroptosis. Similarly, I found that a lower TAZ level in the recurrent ovarian cancer is responsible for reduced ferroptosis susceptibility of these cells. I further investigated the mechanisms by which TAZ regulates ferroptosis. and found that TAZ regulates ferroptosis through EMP1-NOX4 axis in renal cancers and ANGPTL4-NOX2 axis in ovarian cancers. The relevance of the Hippo pathway effector with ferroptosis suggests that ferroptosis-inducing agents may be used to target the TAZ-activated tumors.

The second part of my dissertation investigated the molecular mechanisms of transcriptome changes inside RBCs during ex vivo storages. RBCs are the major component of blood transfusions, one of the most common procedures in the hospital. In addition, some athletes utilize blood transfusion of stored RBCs to increase athletic performance, a practice banned by the world anti-doping agency. Currently, RBCs can be stored for up to 42 days at ~4°C before transfusion. However, transfusion with RBCs after long storage duration may correlate with a poorer prognosis compared with fresh

RBCs and results in increased morbidity and mortality. To recognize the undesirable effects of prolonged RBC storage on transfusion recipients, it is critical to understand storage-associated RBC changes. To this end, our lab has previously identified a variety of RNA species in mature RBCs and profiled the miRNA changes that occur in RBCs at

different time intervals during in vitro storage. This profiling demonstrates that the abundance of most RBC miRNAs did not change significantly during the 42 days of refrigerated storage, indicating extremely long decay half-lives. Unexpectedly, miR-720, a cleavage product of tRNAThr, increased dramatically in the first two weeks and persisted during storage. Furthermore, I present evidence for a role of angiogenin in tRNA cleavage to generate miR-720 during RBC storage. The dramatic increase in miR-

720 may be used to monitor transfused RBCs in clinical patients, athletes performing blood doping, and other settings. Additionally, the increase in miR-720 levels in the stored RBC may potentially contribute to the cellular and clinical phenotypes associated with storage lesions.

Taken together, these studies on how human cells respond to stresses have the potentials as guidance for cancer patients toward ferroptosis-inducing chemotherapeutics or provide a novel way of detecting blood doping or understanding the RBC storage lesions.

Dedication

To my parents, my brother, my husband, and my nephews for all their love, support, and encouragement.

vii

Contents

Abstract ...... iv

Dedication ...... vii

List of Tables ...... xiii

List of Figures ...... xiv

List of Abbreviations ...... xix

1. Introduction ...... 1

1.1 Ferroptosis ...... 3

1.2 Hippo signaling pathway ...... 4

2. The Hippo Pathway Effector TAZ Regulates Ferroptosis in Renal Cell Carcinoma ...... 6

2.1 Introduction ...... 6

2.2 Methods ...... 9

2.2.1 Materials and reagents ...... 9

2.2.2 Cell culture and transfection ...... 9

2.2.3 siRNA-mediated gene knockdown ...... 10

2.2.4 Cell viability assays ...... 10

2.2.5 Western blot analysis ...... 11

2.2.6 RNA isolation and quantitative real-time PCR ...... 11

2.2.7 Microarray ...... 12

2.2.8 Generation of patient-derived xenograft (PDX) and cell lines ...... 13

2.2.9 Immunofluorescence staining ...... 14

viii

2.2.10 Chromatin immunoprecipitation (ChIP) analysis ...... 14

2.2.11 Lipid ROS assay using flow cytometry ...... 16

2.2.12 Statistical analyses ...... 16

2.3 Results ...... 17

2.3.1 Cell density affects the sensitivity of RCC cell lines to erastin-induced ferroptosis ...... 17

2.3.2 TAZ regulates sensitivity to erastin-induced ferroptosis ...... 29

2.3.3 EMP1 is a direct target gene of TAZ that regulates ferroptosis sensitivity ...... 38

2.3.4 EMP1 regulates ferroptosis through NOX4 ...... 49

2.4 Discussion ...... 58

3. A TAZ-ANGPTL4-NOX2 axis regulates ferroptotic cell death and chemoresistance in epithelial ovarian cancer ...... 63

3.1 Introduction ...... 64

3.1.1 The impact of OVCA and the critical need for novel therapeutics ...... 64

3.1.2 Ferroptosis as a novel cell death involving lipid peroxidation...... 65

3.2 Methods ...... 67

3.2.1 Materials and reagents ...... 67

3.2.2 Cell culture and transfection ...... 68

3.2.3 siRNA-mediated gene knockdown ...... 68

3.2.4 RNA isolation and quantitative real-time PCR ...... 69

3.2.5 Western blot analysis ...... 69

3.2.6 Cell viability and cytotoxicity assays ...... 70

3.2.7 Enzyme-linked immunosorbent assay ...... 70

3.2.8 ChIP analysis ...... 71

3.2.9 Lipid ROS assay using flow cytometry ...... 72

3.2.10 Statistical analyses ...... 73

3.2.11 Data availability ...... 73

3.3 Results ...... 73

3.3.1 Ovarian cancer cells are sensitive to cystine deprivation ...... 73

3.3.2 Cell density affects the sensitivity of OVCAs to erastin-induced ferroptosis .. 76

3.3.3 TAZ regulates sensitivity to erastin-induced ferroptosis ...... 80

3.3.4 Resistance to ferroptosis following treatment with carboplatin in vivo ...... 83

3.3.5 ANGPTL4 is a direct target gene of TAZ that regulates sensitivity to ferroptosis ...... 86

3.3.6 Differential ANGPTL4 expressions underlie the ferroptosis sensitivities of CAOV2 pair cells ...... 93

3.3.7 ANGPTL4 regulates ferroptosis through NOX2 ...... 95

3.4 Discussion ...... 103

4. Angiogenin-mediated tRNA cleavage as a novel feature of stored red blood cells .... 107

4.1 Introduction ...... 107

4.1.1 Stored red blood cells ...... 108

4.1.2 Use of stored red blood cells in athletic doping ...... 111

4.1.3 Functionally relevant RNA in RBCs ...... 112

4.1.4 Transfer RNA-related fragments (tRFs) ...... 114

4.1.5 Candidate nucleases that generate tRNA fragments ...... 117

4.2 Methods ...... 122

4.2.1 Blood collection and RNA purification ...... 122

4.2.2 Northern blot ...... 122

4.2.3 In vitro RNA cleavage assays ...... 123

4.2.4 Western blot analysis and immunodepletion ...... 124

4.2.5 Statistical analysis ...... 125

4.3 Results ...... 126

4.3.1 NanoString analysis of the RBC miRNAs during storage ...... 126

4.3.2 Small RNA northern blots indicate the putative tRNA origins of miR-720 .... 128

4.3.3 RBC lysates contain cleavage activities for synthetic tRNAThr(TGT) ...... 130

4.3.4 Angiogenin contributes to tRNA cleavage and increased miR-720 during RBC storage ...... 132

4.4 Discussion ...... 137

4.4.1 Summary of the results ...... 137

4.4.2 RBCs as a unique cellular context for studying RNA metabolism ...... 138

4.4.3 Stress-associated nuclease angiogenin contributes to the storage-associated miR-720 increase ...... 139

4.4.4 Putative functional role of miR-720 associated with transfusion ...... 140

4.4.5 The potential of storage transcriptome signatures for detecting blood doping ...... 141

5. Conclusions ...... 143

5.1 Hippo pathway effectors TAZ regulates ferroptosis ...... 143 xi

5.2 Molecular identification of tRNA cleavage activity in stored erythrocytes ...... 148

References ...... 151

Biography ...... 174

xii

List of Tables

Table 1 TCGA gene analysis of NOX family gene expressions in OVCA ...... 96

Table 2 The sequence(s) of the synthetic RNAs and northern blot probe(s) ...... 130

xiii

List of Figures

Figure 1 Cell density regulates the sensitivity of RCC to erastin-induced ferroptosis ..... 17

Figure 2 EC50 of RCC4 at high and low cell densities ...... 18

Figure 3 Cell death assay by SYTOX Green staining ...... 19

Figure 4 Crystal violet staining of RCC4 ...... 19

Figure 5 Crystal violet staining of low/high densities and quantification ...... 20

Figure 6 RT-qPCR of CHAC1 mRNA expression ...... 21

Figure 7 Density regulates the sensitivity of 293T cells to ferroptosis ...... 22

Figure 8 Density regulates the sensitivity of PDX cells to ferroptosis ...... 22

Figure 9 Erastin-induced cell death is rescued by ferroptosis inhibitor ...... 23

Figure 10 Cell density regulates the sensitivity of RCC to ferroptosis induced by GPX4 inhibitor, RSL3 ...... 23

Figure 11 TAZ is highly expressed in renal cells ...... 25

Figure 12 TAZ is highly expressed in PDX renal tumor ...... 25

Figure 13 Compensatory mechanism among the YAP and TAZ ...... 26

Figure 14 Western blot analysis of YAP/TAZ grown at low vs. high density...... 27

Figure 15 Confocal immunofluorescence images of TAZ ...... 27

Figure 16 TAZ mRNA (WWTR1) expression in cell lines ...... 28

Figure 17 TAZ mRNA (WWTR1) expression in TCGA ...... 29

Figure 18 TAZ knockdown in RCC4 ...... 30

Figure 19 TAZ regulates sensitivity to erastin-induced ferroptosis ...... 30

Figure 20 Interaction between cell density and TAZ ...... 31 xiv

Figure 21 Individual siRNAs targeting TAZ in RCC4 cells ...... 32

Figure 22 Individual siRNAs targeting TAZ in 786O cells ...... 32

Figure 23 Individual siRNAs targeting TAZ in PDX 13-789 cells ...... 33

Figure 24 TAZ also regulates erastin-induced ferroptosis in breast cancer cells ...... 34

Figure 25 Overexpression of TAZ sensitizes RCC4 cells to erastin-induced ferroptosis .. 35

Figure 26 TAZ knockdown decreases tumorsphere under 3D culture ...... 36

Figure 27 Mice xenograft models of TAZ silencing in response to erastin ...... 37

Figure 28 Microarray analysis of gene expression upon siTAZ ...... 38

Figure 29 Venn diagram of siTAZ and cystine deprivation ...... 40

Figure 30 Screen of candidate genes ...... 41

Figure 31 EMP1 gene expression is down-regulated upon siTAZ ...... 42

Figure 32 Gene correlation analysis of TCGA...... 43

Figure 33 Gene correlation analysis of CCLE ...... 43

Figure 34 siRNAs targeting EMP1 reduces erastin-induced ferroptosis ...... 44

Figure 35 EMP1 knockdown reduces ferroptosis sensitivity ...... 45

Figure 36 EMP1 regulates erastin-induced ferroptosis sensitivity ...... 46

Figure 37 Overexpression EMP1 sensitizes RCC4 cells to ferroptosis ...... 47

Figure 38 Genetic interaction between TAZ and EMP1 ...... 47

Figure 39 EMP1 is the direct target of TAZ ...... 48

Figure 40 knockdown of EMP1 down-regulates NOX4 ...... 49

Figure 41 Overexpression of EMP1 up-regulates NOX4 ...... 50

Figure 42 Higher amount of NOX4 protein correlates with EMP1 or TAZS89A overexpression ...... 51

Figure 43 GSH assay upon siTAZ ...... 52

Figure 44 Inhibition of NOX4 reduces erastin-induced ferroptosis...... 53

Figure 45 NOX4 knockdown reduces erastin-induced ferroptosis...... 53

Figure 46 NOX4 overexpression sensitizes RCC4 cells to erastin treatment ...... 54

Figure 47 Genetic interaction between EMP1 and NOX4 ...... 55

Figure 48 EMP1 regulates NOX4 through p38 ...... 56

Figure 49 Inhibition of TAZ, EMP1, or NOX4 abolishes elevated lipid-ROS by erastin treatment ...... 57

Figure 50 The proposed molecular mechanism ...... 58

Figure 51 A schematic model illustrating the ferroptosis regulated by TAZ-EMP1-NOX4 axis ...... 59

Figure 52 Ovarian cancer cells are sensitive to cystine deprivation ...... 74

Figure 53 Ferroptosis inhibitors rescue cystine deprivation-induced cell death ...... 75

Figure 54 Erastin-induced cell death in OVCA is rescued by ferroptosis inhibitor ...... 76

Figure 55 Cell density regulates the sensitivity of CAOV2 to ferroptosis ...... 77

Figure 56 ferroptosis sensitivity is regulated by cell densities ...... 77

Figure 57 Cell density regulates the sensitivity of TOV-21G to ferroptosis ...... 78

Figure 58 TAZ is highly expressed in OVCA cells ...... 79

Figure 59 TAZ localization and abundance are regulated by cell densities ...... 80

Figure 60 TAZ regulates sensitivity to erastin-induced ferroptosis in CAOV2 cells ...... 81

Figure 61 Individual siRNAs targeting TAZ in CAOV2 cells ...... 81 xvi

Figure 62 TAZ regulates sensitivity to erastin-induced ferroptosis in TOV-21G cells ..... 82

Figure 63 Regrowth of chemo-residual tumor cells are more resistant to ferroptosis ...... 84

Figure 64 Regrowth of chemo-residual tumor cells have less abundant TAZ proteins ... 85

Figure 65 Overexpression of TAZ sensitizes CAOV2R cells to ferroptosis ...... 86

Figure 66 Integrated gene analysis ...... 87

Figure 67 TAZ regulates ferroptosis, not through GPSM3 ...... 88

Figure 68 The correlation between WWTR1 (TAZ) and ANGPTL4 ...... 89

Figure 69 ANGPTL4 regulates ferroptosis sensitivity in CAOV2 cells ...... 90

Figure 70 ANGPTL4 regulates ferroptosis sensitivity in TOV-21G cells ...... 91

Figure 71 Validation of ANGPTL4-regulated ferroptosis sensitivity in CAOV2 cells ...... 91

Figure 72 Overexpression of ANGPTL4 sensitizes ferroptosis sensitivity ...... 92

Figure 73 ANGPTL4 is a direct target gene of TAZ ...... 93

Figure 74 Differential ANGPTL4 expressions among CAOV2 and CAOV2R cells ...... 94

Figure 75 Supplying ANGPTL4 protein sensitizes CAOV2R cells to erastin...... 95

Figure 76 lower NOX2 protein level in CAOV2R cells ...... 96

Figure 77 NOX inhibition decreases ferroptosis sensitivity in CAOV2 cells ...... 97

Figure 78 NOX2 inhibition decreases ferroptosis sensitivity in CAOV2 cells ...... 98

Figure 79 Knockdown of NOX2 decreases ferroptosis sensitivity in CAOV2 cells ...... 98

Figure 80 Knockdown of NOX2 decreases ferroptosis sensitivity in TOV-21G cells ...... 99

Figure 81 Interaction between chemo-residual tumor cells and NOX2 inhibition ...... 100

Figure 82 Interaction between TAZ and NOX ...... 100

xvii

Figure 83 TAZ, ANGPTL4, or NOX2 knockdown abolishes elevated lipid-ROS by erastin treatment ...... 102

Figure 84 Schematic representing the model of TAZ-regulated ferroptosis through ANGPTL4-NOX2 ...... 102

Figure 85: Pipeline for analysis of RBC miRNAs during storage ...... 127

Figure 86: Heat map of miRNA expression of RBCs ...... 128

Figure 87: Cloverleaf structure of tRNAThr(TGT) ...... 129

Figure 88: Small RNA northern blot of RBC RNA ...... 130

Figure 89: Northern blot of RBC lysate ...... 131

Figure 90: Northern blot of heat-inactivated RBC lysate...... 132

Figure 91: Western blots of RBC lysate ...... 133

Figure 92: In vitro biochemical study of nuclease(s) ...... 134

Figure 93: Angiogenin plays a role in tRNA cleavage ...... 135

Figure 94: Angiogenin contributes to the “miR-720” increase in stored RBC ...... 136

Figure 95 Hippo pathway effectors, YAP/TAZ, regulates ferroptosis ...... 146

xviii

List of Abbreviations

2,3-DPG 2,3-diphosphoglycerate

ABT autologous blood transfusion

ACSL4 acyl-CoA synthetase long-chain family member 4

ANG angiogenin

ANGPTL4 angiopoietin-Like 4

ATP adenosine triphosphate

CCLE cancer cell line encyclopedia

ChIP chromatin immunoprecipitation

DNMT3A DNA methyltransferase 3 alpha dsRNA double-stranded RNA

EC50 half maximal effective concentration

ELISA enzyme-linked immunosorbent assay

EMP1 epithelial membrane protein 1

EMT epithelial-mesenchymal transition eNOS endothelial nitric oxide synthase

EV extracellular vesicle

Fer-1 ferrostatin-1

FRET fluorescence resonance energy transfer

xix

FSC forward scatter

GeLC-MS gel electrophoresis liquid chromatography-mass spectrometry

GPSM3 G protein signaling modulator 3

GPX4 glutathione peroxidase 4

GSH glutathione

HMVEC human microvascular endothelial cells

HPLC high-performance liquid chromatography

HRP horseradish peroxidase

IP immunoprecipitation

LNA locked nucleic acid miRBase the microRNA database miRNA microRNA mRNA messenger RNA

MV microvesicle

NO nitric oxide

NOX NADPH oxidase nt nucleotide

OVCA ovarian cancer

PARP poly (ADP-ribose) polymerase

PDX patient-derived xenograft

PNK polynucleotide kinase qRT-PCR quantitative reverse transcription polymerase chain reaction

RACE rapid amplification of cDNA ends

RBC red blood cell

RCC renal cell carcinoma

RISC RNA-induced silencing complex

RNase ribonuclease

RNH1 ribonuclease/angiogenin inhibitor 1

ROS reactive oxygen species

SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis siNT siRNA non-targeting

SLC7A11/xCT solute carrier family 7 member 11

SSC side scatter

STR short tandem repeat

TAZ transcriptional coactivator with PDZ-binding motif

TCGA the cancer genome atlas

TEAD transcriptional enhanced associate domain

xxi

TF transcriptional factor(s)

TFRC transferrin receptor

TGIRT thermostable group II intron-encoded reverse transcriptase tRF tRNA-related fragment tRNA transfer ribonucleic acid

UTR untranslated region

WADA world anti-doping agency

WWTR1 WW domain-containing transcription regulator 1

YAP1 yes-associated protein 1

ZEB zinc finger E-Box binding homeobox

Z-VAD Z-VAD-FMK

xxii

1. Introduction

Mammalian cells face a variety of environmental challenges during physiological adaptation, medical procedures, or different pathophysiological disease processes. These challenges impose significant stresses upon the mammalian cells to survive and maintain homeostasis. Therefore, mammalian cells have evolved various mechanisms to sense and adapt to these stresses and maintain systemic homeostasis. Some of these adaptive mechanisms toward stresses are antioxidants, scavenging enzymes, DNA damage response, and unfolded protein response (Galluzzi et al., 2018). When restoring back to cell homeostasis is impossible, cells will trigger programmed cell death as a protection for the whole organism.

Oxidative stress contributes to not only physiological conditions but also pathological conditions by elevated intracellular levels of reactive oxygen species (ROS).

The ROS can be produced by normal cell functions such as mitochondrial metabolism or from environmental threats such as ionizing radiation and hypoxia that result in damages to DNA, protein, and lipids (Schieber and Chandel, 2014).

Increased levels of ROS and deregulated redox biology are the common characteristics of cancer cells (Liou and Storz, 2010). This hallmark of elevated ROS in cancer cells can be triggered by higher metabolic activity for tumorigenesis and progression, abnormal cellular signaling such as increased activities of oncogenes or

decreased activities of tumor suppressor genes, as well as interactions with the tumor microenvironment (Kumari et al., 2018). Therefore, several chemotherapeutic agents were developed by increasing the excessive amount of intracellular ROS to induce cell deaths of tumor cells (Pelicano et al., 2004).

On the other hand, normal cells also cope with oxidative stress not only in vivo but also ex vivo. For example, stored red blood cells for transfusion use undergo oxidative stress even during hypothermic storage conditions, which is a cause for RBC storage lesions (Yoshida et al., 2019).

In my dissertation, I provide two examples that mammalian cells response to distinct environmental challenges. In the first part (Chapter 2, 3), I elucidate the link between the Hippo pathway effector TAZ, triggers with ferroptosis by extreme oxidative stress in renal cell carcinoma and ovarian cancer cells. (Ferroptosis and Hippo signaling pathway are further discussed below. In the second part (Chapter 4), I present how the red blood cells respond to the ex vivo storage conditions from the angel of small

RNA biology. Collectively, these stories present examples by which mammalian cells incorporate their environmental challenges to dictate their responses and fate decisions between life and death. Overall, my dissertation research has advanced our understanding of how mammalian cells response toward stresses through the regulation of ferroptosis and RNA processing.

1.1 Ferroptosis

Ferroptosis is a recently defined form of programmed cell death (Dixon et al.,

2012) characterized by the accumulation of lipid peroxidation (Yang and Stockwell,

2016). Ferroptosis is morphologically, genetically, and biochemically distinct from other cell deaths (Dixon et al., 2012) and has been reported with several pathological relevance such as neurodegenerative diseases (Alzheimer’s, Huntington’s and Parkinson’s diseases), ischemic stroke, ischemia-reperfusion injury, fibrosis (Stockwell et al., 2017).

On the other hand, ferroptosis may have therapeutic potential toward cancer (Cramer et al., 2017). Ferroptosis can be induced by erastin (Dolma et al., 2003), a small molecule that inhibits the glutamate-cystine antiporter system, xCT, resulting in the redox imbalance by decreasing intracellular glutathione levels and accumulation of lipid-based reactive oxygen species (ROS). Therefore, ferroptosis can also be induced by the removal of cystine (limiting component for glutathione synthesis). In addition, lipid ROS can be accumulated by either impaired detoxification of lipid peroxidation via reducing the expression of glutathione peroxidase 4 (GPX4) (Yang et al., 2014) or by the generation of superoxide and hydrogen peroxide involving upregulation of NADPH oxidases (NOXs)

(Dixon et al., 2012). Many studies have identified various genetic determinants of ferroptosis involved in the GSH/lipid metabolisms (Xie et al., 2016), oncogenic somatic mutations, regulation of iron levels (Chen et al., 2019) and process of epithelial-

mesenchymal transitions (Viswanathan et al., 2017). However, much remains unknown about the genetic determinants and underlying mechanisms of ferroptosis to select tumors that are most likely to respond to these ferroptosis-inducing agents and predict potential resistant mechanisms against such an approach.

1.2 Hippo signaling pathway

The Hippo signaling is an evolutionarily conserved pathway that plays critical roles in cell proliferation, tissue homeostasis, organ size control, differentiation, development, chemoresistance, and metastasis (Yu et al., 2015). It consists of upstream adaptor proteins and a core kinase cascade which inhibits downstream effectors.

Upstream adaptor proteins activate phosphorylation of core kinases and ultimately repress downstream effectors. The canonical kinase cascade of the Mammalian Hippo pathway involves MST1/2 (serine/threonine kinases mammalian STE20-like protein kinase 1/2; homologs of Drosophila Hippo) conjugated with SAV1 (Salvador homolog 1) and LATS1/2 (large tumor suppressor homolog 1/2) conjugated with MOB1 (Mob kinase activator 1).

The two Hippo pathway effectors, YAP (Yes Associated Protein 1, encoded by

YAP1) and TAZ (transcriptional coactivator with PDZ-binding motif, encoded by

WWTR1), response to the externally environmental conditions and stimuli (e.g., cell-cell contact, cell mechanics, cell polarity, and cellular metabolism) as well as intracellular

signals by interacting with proteins from other signaling pathways (e.g., GPCR, EGF,

Wnt, Notch, BMP/TGFβ). YAP and TAZ are coactivators of several transcription factors such as TEAD 1-4, Smad, RUNX1/2, p63/p73 proteins, which trigger expressions of genes that determine cell fate and behavior during embryonic development and oncogenesis (Hsiao et al., 2016; Lin et al., 2017; Zhao et al., 2007). Dysregulated Hippo pathway and high activities of YAP and TAZ have been observed in many cancers

(Zanconato et al., 2016). YAP and TAZ share a 60 % protein sequence similarity.

Frequently, YAP and TAZ are redundant and complementary; however, they are distinct in development, structure, and physiology suggesting they may have different regulations and functions (Varelas, 2014).

YAP/TAZ activities are regulated by their phosphorylation and intracellular localization. When Hippo pathway is “ON”, YAP/TAZ are phosphorylated, cytosolically retained and subjected to proteasomal degradation by β-TRCP (β-transducin repeat- containing E3 ubiquitin-protein ligase); When Hippo pathway is “OFF”, YAP/TAZ become dephosphorylated and translocate into the nuclei to associate with transcriptional factors to drive gene expression regulating cell proliferation, differentiation, and migration.

2. The Hippo Pathway Effector TAZ Regulates Ferroptosis in Renal Cell Carcinoma

Portions of this Chapter are reproduced from the Cell Reports:

Cell Reports, 2019, Sep 3;28(10):2501-2508.e4. DOI: 10.1016/j.celrep.2019.07.107

2.1 Introduction

Ferroptosis, a novel form of iron-dependent programmed cell death, is morphologically, genetically, and biochemically distinct from other cell deaths (Dixon et al., 2012) since it is characterized by the accumulation of lipid peroxidation products

(Yang and Stockwell, 2016). Ferroptosis can be induced by erastin (Dolma et al., 2003), which inhibits the glutamate-cystine antiporter system, xCT, resulting in the redox imbalance by decreasing intracellular glutathione levels and accumulation of lipid-based reactive oxygen species (ROS). Lipid ROS can be accumulated by either impaired detoxification of lipid peroxidation via reducing the expression of glutathione peroxidase 4 (GPX4) (Yang et al., 2014) or by the generation of superoxide and hydrogen peroxide involving upregulation of NADPH oxidases (NOXs) (Dixon et al., 2012).

Among different NOXs, NOX4 is highly expressed in the kidney as an important source of renal ROS (Gorin et al., 2005; Sedeek et al., 2013). In addition, inhibition of NOX4 has been reported to reduce cystine deprivation-induced cell death and lipid ROS, suggesting its essential role in ferroptosis (Poursaitidis et al., 2017).

Renal cell carcinoma (RCC) is the 9th most common cancer in men and 14th in women with approximately 338,000 new cases each year worldwide (Medina-Rico et al.,

2018). Despite recent advances in the treatment of RCC using anti-angiogenic agents or immune checkpoint inhibition, the median overall survival rate for patients with the advanced or metastatic RCC remains unsatisfactory (Hsieh et al., 2017). Thus, there is an unmet and urgent need to develop novel therapeutic approaches and targets to improve the treatments for RCC. Among different tumor cells, RCC cells are particularly susceptible to ferroptosis (Dixon et al., 2012; Yang et al., 2014), suggesting that inducing ferroptosis by various means may have therapeutic potential for RCC. However, much remains unknown about the molecular mechanisms of ferroptosis regulation in RCC to rationally select tumors that are sensitive to ferroptosis to optimize the therapeutic potential of this novel approach.

In our current study, we report that RCC sensitivity to erastin and ferroptosis is significantly affected by cell density. RCC cells grown at low density were highly susceptible to erastin-induced ferroptosis, while the same RCC cells grown at high confluency become relatively resistant to erastin-induced ferroptosis. The Hippo pathway and its regulators have been established as molecular sensors that detect cell density and regulate the density-dependent proliferation of cancer cells (Mori et al.,

2014; Zhao et al., 2007). The two Hippo pathway effectors, YAP (Yes Associated Protein

1, encoded by YAP1) and TAZ (transcriptional coactivator with PDZ-binding motif, encoded by WWTR1), are coactivators of TEAD transcription factors, which trigger expressions of genes that determine cell fate and behaviour during embryonic development and oncogenesis (Hsiao et al., 2016; Lin et al., 2017; Zhao et al., 2007).

YAP/TAZ activities are regulated by their phosphorylation and intracellular localization.

In high cell density, YAP/TAZ are phosphorylated, cytosolically retained and subjected to proteasomal degradation; in low cell density, YAP/TAZ become dephosphorylated and translocate into the nuclei to associate with transcriptional factors to drive gene expression regulating cell proliferation, differentiation, and migration (Hsiao et al., 2016;

Zhao et al., 2007).

Here, we have established the role of cell density and TAZ as novel determinants of ferroptosis of RCC. In addition, we found that TAZ regulates ferroptosis through the induction of epithelial membrane protein 1 (EMP1), which in turn activates NADPH oxidase 4 (NOX4) to enhance ROS, lipid peroxidation and ferroptosis. Thus, these data support the role of TAZ in regulating ferroptosis through EMP1-NOX4 and that inducing ferroptosis can be a novel therapy for RCC and other TAZ-activated tumors.

2.2 Methods

2.2.1 Materials and reagents

Erastin was obtained from the Duke University Small Molecule Synthesis

Facility. The following antibodies were used in this study: YAP/TAZ (#8418, Cell

Signaling Technology), phosphor-YAP/TAZ (S127/S89) (#4911, Cell Signaling

Technology), beta-actin (sc-130301, Santa Cruz), beta-tubulin (#86298, Cell Signaling

Technology), V5-tag (MA5-15253, ThermoFisher Scientific), NOX4 (ab133303, Abcam),

GPX4 (ab125066, Abcam). Horseradish peroxidase (HRP)-labeled goat secondary antibody conjugates were purchased from Cell Signaling Technology. Plasmids are obtained from Addgene (TAZS89A #52084; pLX304 #25890) and DNASU (EMP1

#HsCD00442695).

2.2.2 Cell culture and transfection

RCC cell lines (RCC4 and 786O) were kind gifts from Dr. Denise Chan

(Department of Radiation Oncology, University of California, San Francisco), which were further authenticated by DDC (DNA Diagnostics Center) Medical using the short tandem repeat method in November 2015. HEK-293T was acquired from the Duke Cell

Culture Facility. All cells were cultured in Dulbecco’s Modified Eagle Medium (11995-

DMEM, ThermoFisher Scientific) with 10% heat-inactivated Fetal Bovine Serum

(HyCloneTM FBS, GE Healthcare Life Sciences #SH30070.03HI) in a humidified

incubator at 37°C and 5% CO2. Transfections were performed according to the manufacturer’s instructions with TransIT-LT1 transfection reagent (Mirus Bio) or

RNAiMax transfection reagent (ThermoFisher Scientific).

2.2.3 siRNA-mediated gene knockdown

All human siRNAs were purchased from Dharmacon or Qiagen: Non-targeting control, siNT (Qiagen SI03650318); siTAZ (Dharmacon M-016083); siYAP (Dharmacon

M-012200); siEMP1 (Dharmacon M-010507); siNOX4 (Dharmacon M-010194); siANGPTL4 (Dharmacon M-007807); siGPSM3 (Dharmacon M-013906); siSGPL1

(Dharmacon M-008747); siSMAGP (Dharmacon M-015427); siSMG5 (Dharmacon M-

014023); siTAZ#1 (Qiagen SI00111230); siTAZ#2 (Qiagen SI00111237); siTAZ#3 (Qiagen

SI00111216); siTAZ #4 (Dharmacon D-016083-01); siTAZ #5 (Dharmacon D-016083-02); siEMP1#1 (Dharmacon D-010507-02); siEMP1#2 (Qiagen SI04260711); siTAZ#3 (Qiagen

SI04327862); siEMP1 #4 (Dharmacon D-010507-04). siNOX4#1 (Qiagen SI02642500); siNOX4#2 (Qiagen SI02642507). The knockdown efficacy of the siRNAs to reduce target mRNA and protein were validated by RT-qPCR and/or western blots.

2.2.4 Cell viability assays

After being seeded at the ratio of 2500 cells per 96 well and transfected with siRNAs for 2 days, the cells were further treated with erastin for an additional 24 h. Cell viability was evaluated using the crystal violet staining or the CellTiterGloTM

luminescent cell viability assay kit (Promega) according to the manufacturer’s instructions. The CellTiterGloTM kit is based on quantitation of the ATP present indicating metabolically active cells to determine the number of living cells in the culture.

2.2.5 Western blot analysis

For immunoblotting, cells were washed with ice-cold phosphate-buffered saline

(PBS), lysed in RIPA buffer (Sigma), supplemented with protease inhibitor (Roche) and

PhosSTOP phosphatase inhibitor cocktail (Roche). Proteins were quantified by BCA protein assay (ThermoFisher Scientific). Equal amounts of proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to the polyvinylidene difluoride (PVDF) membrane (Millipore). The membranes were blocked with 5% non-fat milk or BSA and then probed with indicated antibodies following by HRP-conjugated secondary antibodies. The immuno-signals were achieved by the Amersham ECL prime western blotting detection reagent (GE Healthcare Life

Sciences RPN2232) and detected by a Bio-Rad ChemiDocTM Imaging System.

2.2.6 RNA isolation and quantitative real-time PCR

Total RNAs of cultured cells were extracted by using the RNeasy mini kit

(Qiagen #74104) with DNase I treatment (Qiagen #79254) and the cDNAs were synthesized from 1μg of the RNA template using SuperScriptTM II Reverse Transcriptase

(ThermoFisher Scientific #18064) with random hexamers following protocols from the manufactures. The levels of gene expression were measured by quantitative PCR (qPCR) with power SYBR green PCR mix (Applied Biosystems, ThermoFisher Scientific).

β-actin-F': GGGGTGTTGAAGGTCTCAAA; β-actin-R': GGCATCCTCACCCTGAAGTA;

TAZ-F': TGCTACAGTGTCCCCACAAC; TAZ-R': GAAACGGGTCTGTTGGGGAT;

YAP-F': CAACTCCAACCAGCAGCAAC; YAP-R': TTGGTAACTGGCTACGCAGG;

EMP1-F': GTGCTGGCTGTGCATTCTTG; EMP1-R': CCGTGGTGATACTGCGTTCC;

NOX4-F': CAGATGTTGGGGCTAGGATTG; NOX4-R': GAGTGTTCGGCACATGGGTA;

GPX4-F': GAGGCAAGACCGAAGTAAACTAC; GPX4-R':

CCGAACTGGTTACACGGGAA.

2.2.7 Microarray

RCC4 cells exposed to knockdown control, siNT, or knockdown TAZ, siTAZ for two days and then treated with 1 μM erastin for 7 hours. RNAs were extracted by

RNeasy mini kit (Qiagen), labeled, and hybridized to Affymetrix U133A 2.0 arrays. The microarray data have been deposited into NCBI GEO with accession number:

GSE121689 (cDNA microarray for siTAZ). The intensities of Affymetrix probes were normalized by the Robust Multi-array Average (RMA) method and zero transformation

(Δlog2) against the control group, siNT (DMSO). Then, the probe sets that varied by 20.8- fold in at least two samples were selected for hierarchical clustering.

2.2.8 Generation of patient-derived xenograft (PDX) and cell lines

Renal cell carcinoma (RCC) tissue sample (13-789) was collected under a Duke

IRB-approved protocol (Pro000022289) and all participants provided written informed consent to participate in the study. The patient-derived xenograft (PDX) model of 13-789 was then generated as described previously (Kim et al., 2012a; Uronis et al., 2012), and the in vivo PDX generation was performed in accordance with the animal guidelines and with the approval of the Institutional Animal Care and Use Committee (IACUC) at the Duke

University Medical Center. Briefly, to generate PDXs, the tissue sample was washed in phosphate-buffered saline (PBS), dissected into small pieces (<2 mm), and injected into the flanks of 8-10-week-old JAX NOD.CB17-PrkdcSCID/J mice obtained from the Duke

University Rodent Genetic and Breeding Core. The matched PDX cell line (13-789) was then generated from the PDX as followed. Once the PDX tumors reached a size of >1000 mm3, tumors were harvested, homogenized and grown in 10 cm2 tissue culture dishes in cell culture media (DMEM media, 10% fetal bovine serum (FBS), 10 U/ml penicillin and streptomycin) at 37°C and 5% CO2. Clonal populations of each cell line were then obtained by isolating a single clone using trypsinization of the clone sealed off from the dish by an

O ring. Finally, the 13-789 cell line was authenticated using the Duke University DNA

Analysis Facility Human cell line authentication (CLA) service by analyzing DNA

samples from each individual cell line for polymorphic short tandem repeat (STR) markers using the GenePrint 10 kit from Promega (Madison, WI, USA).

2.2.9 Immunofluorescence staining

Cells were cultured on chamber slide (ThermoFisher Scientific #177437) to appropriate density, fixed with 4% formaldehyde for 10 min and then wash three times with PBS. After blocking in 5% BSA and permeabilized with 0.1% Saponin for 1 hour, slides were incubated with the TAZ antibody (BD Biosciences #560235) 1:100 diluted in

1% BSA with 0.1% Saponin overnight. After washing with PBS, slides were incubated with Alexa Fluor 488-conjugated secondary antibodies (1:200, ThermoFisher Scientific

#A11001) and Alexa Fluor 568-conjugated phalloidin (ThermoFisher Scientific #A12380) for 1 hour. The slides were then washed and mounted with SlowFade Glod antifade mountant with DAPI ((ThermoFisher Scientific #S36938). Images were acquired using a

Leica TCS SP8 confocal microscope equipped with a 40X objective.

2.2.10 Chromatin immunoprecipitation (ChIP) analysis

ChIP-qPCR experiment was carried out according to the Myers Lab ChIP-seq protocol (Johnson et al., 2007). Briefly, RCC4 cells were incubated in a cross-linking solution (1% formaldehyde) at room temperature for 10 min and then added 0.125M final concentration of glycine to stop cross-linking. The cells were then washed with cold

PBS and suspended in Farnham lysis buffer (5mM PIPES pH8.0, 85mM KCl and 0.5%

NP-40) with freshly added protease inhibitor. The lysate was subsequently passed through a 20-gauge needle 20 times to breaks cells while keeping intact nuclei. After centrifugation, the pellet was resuspended with RIPA buffer with freshly added protease inhibitor. Chromatin fragmentation was performed by sonication using the

Bioruptor (Diagenode) high speed for 30 min (30 seconds ON, 30 seconds OFF). Proteins were immunoprecipitated in PBS/BSA buffer using TAZ antibody (Cell signaling

#70148) or control antibody, rabbit IgG (Cell signaling #2729) which have been conjugated to DynabeadsTM protein G magnetic beads (Thermo Fisher Scientific

#10004D) at 4°C for 2 hr. The antibody-chromatin complexes were washed with LiCl wash buffer for 5 times and then washed with TE buffer. The crosslinking was reversed by incubation with elution buffer (1%SDS, 0.1M NaHCO3) at 65°C overnight followed by incubated with RNase A and proteinase K. DNA was recovered by using QIAquick

PCR purification kit (Qiagen #28104). Precipitated DNA was analyzed by qPCR using primers targeting TEAD-binding sites found in CTGF (Stein et al., 2015) and EMP1 promoter regions as well as negative control chromatin 14 (Stein et al., 2015). Primer sequences are listed below.

CTGF-F’ GCCAATGAGCTGAATGGAGT; CTGF-R’ CAATCCGGTGTGAGTTGATG;

EMP1-F’ TTTGGTGAGGAGGAATGGGC; EMP1-R’ CAGTCAAGAGTGGCTGGGAG;

Ch14-F’ GTGGGCCTTTGGAATATCCT; Ch14-R’ GACCTTGGCTGTGTTGTCCT.

2.2.11 Lipid ROS assay using flow cytometry

Lipid ROS levels were determined using 10 µM of C11-BODIPY dye (D3861,

ThermoFisher Scientific) with the positive control, cumene hydroperoxide, according to the manufacturer’s instructions. Cells were seeded and treated with siRNAs in six-well plates for two days, then the culture medium was replaced with 1µM erastin treatment overnight. The next day, the medium was replaced with 10 µM C11-BODIPY-containing medium for 1 hour. Later, the cells were harvested by trypsin and washed three times with ice-cold PBS followed by resuspending in PBS plus 1% BSA. The amount of ROS within cells was examined by flow cytometry analysis (FACSCantoTM II, BD Biosciences).

2.2.12 Statistical analyses

Graphs were drawn with GraphPad Prism Software and the statistical analyses were performed using either GraphPad or Microsoft Excel software packages. Data were analyzed using the unpaired Student’s t-test and expressed as mean ±SEM. P values less than 0.05 were considered significant (*< 0.05; **<0.01; ***<0.001).

2.3 Results

2.3.1 Cell density affects the sensitivity of RCC cell lines to erastin- induced ferroptosis

RCCs are susceptible to ferroptosis induced by erastin and cystine deprivation

(Tang et al., 2016; Yang et al., 2014). While investigating a renal cancer cell line, RCC4, we observed that the erastin sensitivity was significantly affected by cell density. To characterize this observation further, we systemically examined erastin-induced death of

RCC4 grown at different densities. When grown at low cell density (<50% confluency),

RCC4 cells were highly sensitive to erastin-induced ferroptosis as the morphological change of cell rounding shown under light microscopy. In contrast, when the same

RCC4 cells were grown at high cell density (>80% confluency), they became significantly less sensitive to erastin (Figure 1).

Figure 1 Cell density regulates the sensitivity of RCC to erastin-induced ferroptosis

Bright-field images of RCC4 cells cultured in low/high densities treated with the indicated concentrations of erastin. Scale bar, 200 µm.

Next, the sensitivity of RCC4 to erastin under high and low cell density conditions was quantified by CellTiter-GloTM assay which determined cellular viability by cellular ATP levels. Consistently, RCC4 grown at low density were highly sensitive to erastin-induced cell death. In contrast, high cell density was found to mitigate the erastin-induced cell death with more than double the EC50 at both day 1 and day 3

(Figure 2).

Figure 2 EC50 of RCC4 at high and low cell densities

Cell viability assay by CelltiterGlo after RCC4 cells of low/high cell densities treated with indicated concentrations of erastin for 1 day or 3 days to determine the EC50 of RCC4 at high and low cell density (The data are shown from one representative experiment with three biological replicates, and the data were reproduced from at least three independent experiments; mean ± SEM; two-way ANOVA).

The cell density-dependent sensitivity of erastin-induced cell death was further validated by SYTOX green, a direct cell death assay based on released DNA upon cell

death (Figure 3). Crystal violet staining also confirmed the reduced erastin sensitivity of

RCC4 grown at high cell density (Figure 4).

Figure 3 Cell death assay by SYTOX Green staining

Flow cytometry of RCC4 grown at low or high densities treated with DMSO or 4µM erastin for 16 h. Representative data of two independent experiments.

Figure 4 Crystal violet staining of RCC4

Crystal violet staining of RCC4, cultured at low or high densities, were treated with vehicle or 1µM erastin. Raw data and normalization of OD590 are shown.

To exclude the possibility that cell death is due to the availability of erastin to the erastin target xCT under high and low cell densities, we conducted two additional 19

experiments. First, we seeded the same number of cells in a larger or smaller area to represent low and high cell densities. Crystal violet staining and quantification (Figure

5) revealed that the RCC4 seeded at high cell density were more resistant to erastin treatment. Second, the upregulation of CHAC1 (ChaC, cation transport regulator homolog 1) gene has been reported to be a useful biomarker for inhibition of system Xc- by erastin (Dixon et al., 2014). After erastin treatment, we found comparable CHAC1 upregulation when cells were seeded at either high or low densities by RT-qPCR (Figure

6).

Figure 5 Crystal violet staining of low/high densities and quantification

RCC4 cells cultured in low/high densities were treated with DMSO or 2µM erastin for 1 day and then stained with crystal violet and quantified by extracting with 10% acetic acid. Quantitative analysis of crystal violet stain of RCC4 cells at low and high cell density after normalization to DMSO group using a microplate reading at OD 590 nm (n=4; mean ± SEM; Student’s t-test).

Figure 6 RT-qPCR of CHAC1 mRNA expression

CHAC1 mRNA expression levels in the erastin-treated RCC4 grown at low or high density were determined by RT-qPCR. Data are represented as mean±SEM, n=3 after normalized to the DMSO control of low density.

Similarly, we found cell density-dependent erastin sensitivity in human embryonic kidney cells, 293T (Figure 7), as well as in an early passaged patient-derived xenograft (PDX) RCC cell line, 13-789 (Figure 8). To rule out the possibility that cell density alters the mode of cell death induced by erastin, we found that the erastin- induced cell death in low cell density can be rescued by ferroptosis inhibitor, ferrostatin-

1, but not caspase inhibitor, Z-VAD-FMK (Figure 9). In addition, the ferroptosis death triggered by GPX4 inhibitor, RSL3, in RCC4 cells can be also affected by cell death

(Figure 10) as reported by a pre-print from an independent research group (Panzilius et

al., 2018). Together, these data provide compelling evidence for the ability of cell density and confluency to regulate the ferroptosis sensitivity of RCC.

Figure 7 Density regulates the sensitivity of 293T cells to ferroptosis

Cell viability assay by CelltiterGlo after 293T cells grown at low vs. high (2500/10,000 cells per 96 well) cell densities treated with indicated concentrations of erastin. (The data are shown from one representative experiment with three biological replicates, and the data were reproduced from two independent experiments; mean ± SEM; two-way ANOVA; ***p < 0.001).

Figure 8 Density regulates the sensitivity of PDX cells to ferroptosis

Cell viability assay by CelltiterGlo after PDX 13-789 cells were grown at low, medium or high (2250/4500/9000 cells per 96 well) cell densities and treated with indicated concentrations of erastin for 72h. (n=3; mean ± SEM; two-way ANOVA; ***p < 0.001).

Figure 9 Erastin-induced cell death is rescued by ferroptosis inhibitor

Cell viability was determined by CelltiterGlo and bright-field images of RCC4 cells treated with 1µM erastin, 20µM Z-VAD-FMK (Z-vad), or 1µM ferrostatin-1 (Fer-1). Scale bar, 200 µm.

Figure 10 Cell density regulates the sensitivity of RCC to ferroptosis induced by GPX4 inhibitor, RSL3

Cell viability was determined by CelltiterGlo after RCC4 cells were grown at low or high cell densities, then treated with indicated concentrations of RSL3. (n=3; mean ± SEM; two- way ANOVA; ***p < 0.001).

The Hippo pathway effectors, YAP/TAZ, mediate cell density-dependent responses (Mori et al., 2014; Zhang et al., 2018; Zhao et al., 2007). Thus, we investigated the potential role of the YAP/TAZ in the ferroptosis of RCC. First, we determined the expression of YAP and TAZ proteins in two RCC cell lines (RCC4 and 786O) (Figure 11), and PDX 13-789 cell line (Figure 12). By using the breast cancer cell line, MDA-MB-231 as a control for YAP expression, we found that TAZ, but not YAP, protein is the predominantly expressed co-activator in RCC. To verify that the activated TAZ expression in RCC 13-789 PDX tissues also occurred in vivo, we performed cytosolic and nuclear fractionations of PDX tissues. As shown in Figure 12, the combination of significant nuclear TAZ expression with a low level of phosphorylation indicated that

TAZ is activated in the RCC 13-789 PDX model. In addition, silencing TAZ in RCC4 and

293T increased the YAP expression, suggesting a compensatory mechanism among the

YAP and TAZ (Figure 13).

Figure 11 TAZ is highly expressed in renal cells

Western blot analysis of YAP and TAZ in renal carcinoma cell lines, RCC4, and 786-O cells, as well as in the breast cancer cell line, MDA-MB-231. Representative data of at least three independent experiments.

Figure 12 TAZ is highly expressed in PDX renal tumor

Western blot measuring YAP and TAZ abundance and phospho-TAZ (Ser89) in RCC PDX 13-789 cells of cytosolic/nucleic extractions. Histone H3 serves as a nuclear marker; β-actin serves as a cytosolic marker. The Ponceau S staining blot of RCC PDX cytosolic/nucleic extractions as the loading control.

Figure 13 Compensatory mechanism among the YAP and TAZ

Western blot of YAP and TAZ abundance in RCC4 and 293T cell lines treated with siRNA non-targeting (siNT) control or siRNAs targeting TAZ (siTAZ) or YAP (siYAP).

Next, we compared the expression levels of TAZ between cells grown at different densities. We found that RCC4 grown at higher cell density expressed a lower level of

TAZ protein (Figure 14), consistent with the cytosolic sequestration and degradation of

TAZ during high cell density (Zhao et al., 2007). In addition, when RCC4 and PDX 13-

789 cells were shifted from high density to low density in vitro, TAZ translocated from the cytosol to nuclei based on immunofluorescence (Figure 15).

Figure 14 Western blot analysis of YAP/TAZ grown at low vs. high density

Whole-cell lysate of RCC4 cells seeded in low or high cell densities. GAPDH serves as a loading control. Representative data of at least two independent experiments.

Figure 15 Confocal immunofluorescence images of TAZ

Confocal immunofluorescence images of TAZ in RCC4 cells or RCC PDX, 13-789 tumor cells grown at low vs. high density. DAPI staining of cell nuclei and rhodamine-phalloidin staining for the F-actin cytoskeleton to mark cell boundaries. Merged images with nuclear stain, DAPI, are also shown. Scale bars, 50 μm.

Collectively, these results indicate that TAZ is the predominant Hippo effector in

RCC, and its subcellular localization is regulated by cell density. Consistent with the ferroptosis sensitivity of RCC, RCC tumors and cells express the highest level of WWTR1 mRNA (encoding TAZ) from the analysis of the Cancer Cell Line Encyclopedia (CCLE)

(Figure 16) and The Cancer Genome Atlas (TCGA) (Figure 17). Therefore, we focused on

TAZ as the main Hippo effector which regulates ferroptosis in RCC cells.

Figure 16 TAZ mRNA (WWTR1) expression in cell lines

TAZ mRNA (WWTR1) expression in cell lines from the CCLE dataset (https://portals.broadinstitute.org/ccle) with the kidney cells indicated by the red arrow.

Figure 17 TAZ mRNA (WWTR1) expression in TCGA

TAZ mRNA (WWTR1) expression in TCGA dataset with the renal tumors indicated by the red arrow. Data is retrieved from The Human Protein Atlas (https://www.proteinatlas.org/).

2.3.2 TAZ regulates sensitivity to erastin-induced ferroptosis

To test the possibility that TAZ regulates cell susceptibility to erastin, we silenced

TAZ expression by siRNA to reduce TAZ protein in the RCC4 cell line (Figure 18). TAZ silencing significantly reduced erastin-induced cell death based on CellTiter-GloTM assay and crystal violet staining (Figure 19). We further find that knockdown of TAZ in the low-density RCC4 cells conferred ferroptosis resistance to similar degrees as RCC4 grown at high cell density (Figure 20), suggesting that TAZ activation contributes significantly to the density-dependent ferroptosis sensitivity. 29

Figure 18 TAZ knockdown in RCC4

RCC4 cells were treated with siRNA control (siNT), or siRNA targeting TAZ (siTAZ) for two days. Cells were harvested for western blot analysis to check knockdown efficiency.

Figure 19 TAZ regulates sensitivity to erastin-induced ferroptosis

(A) Cell viability was determined by CelltiterGlo after 24 hours of indicated dosage of erastin treatment. (n=3; mean ± SEM; two-way ANOVA; ***p < 0.001). (B) Crystal violet staining of RCC4 treated with 1µM erastin for 2 days after incubating with siRNA control, siNT, or silencing of TAZ, siTAZ for 1 day. Representative data from one of two independent studies.

Figure 20 Interaction between cell density and TAZ

RCC4 cells were treated with siRNA control (siNT), or siRNA targeting TAZ (siTAZ) for 24 h. Later, cell viability was determined by CelltiterGlo after 24 h treatment of 1µM erastin. Data are represented as mean±SEM, n=3 by comparing to the DMSO controls of each group (two-way ANOVA; *p < 0.05; ***p < 0.001, ns: not significant).

In addition, knockdown of TAZ in RCC4 by multiple independent siRNAs all reduced sensitivity to erastin (Figure 21). To determine if similar results were observed in other RCC cells, we silenced TAZ expression in both 786O and 13-789 cells. We found that TAZ knockdown also reduced the erastin sensitivity in these cell lines (Figure 22 and Figure 23). Furthermore, we found that knockdown TAZ by pooled or individual siRNAs in MDA-MB-231, a breast cancer cell line, also conferred resistance to erastin by both crystal violet staining and CellTiter-GloTM assays (Figure 24). Therefore, the regulation of ferroptosis by TAZ also extends to multiple RCC and breast cancer cells.

Figure 21 Individual siRNAs targeting TAZ in RCC4 cells

(A) The relative viability of cells transfected with multiple individual siRNAs targeting TAZ (#1 to #5), or siRNA control (siNT) for two days in renal carcinoma cells, RCC4 cells. Cell viability was then determined by CelltiterGlo after 24 h treatment of 2.5μM erastin. Data are represented as mean SEM and by comparing to the DMSO controls of each group (one-way ANOVA; *p < 0.05; **p<0.01; ***p < 0.001). (B-C) The knockdown efficiency siRNAs targeting TAZ was confirmed by RT-qPCR (B) and western blot (C).

Figure 22 Individual siRNAs targeting TAZ in 786O cells 32

(A) The relative viability of cells transfected with multiple individual siRNAs targeting TAZ (#1 to #3), siTAZ (pooled), or siRNA control (siNT) for two days in 786O cells. Cell viability was then determined by CelltiterGlo after 24 h treatment of 1μM erastin. Data are represented as mean SEM and by comparing to the DMSO controls of each group (one- way ANOVA; *p < 0.05; **p<0.01; ***p < 0.001). (B-C) The knockdown efficiency siRNAs targeting TAZ was confirmed by RT-qPCR (B) and western blot (C).

Figure 23 Individual siRNAs targeting TAZ in PDX 13-789 cells

(A) The relative viability of cells transfected with multiple individual siRNAs targeting TAZ (#1 to #2), siTAZ (pooled), or siRNA control (siNT) for two days in RCC patient- derived xenograft (PDX) 13-789 cells. Cell viability was then determined by CelltiterGlo after 24 h treatment of 16μM of erastin. Data are represented as mean SEM and by comparing to the DMSO controls of each group (one-way ANOVA; *p < 0.05; **p<0.01; ***p < 0.001). (B-C) The knockdown efficiency siRNAs targeting TAZ was confirmed by RT-qPCR (B) and western blot (C).

Figure 24 TAZ also regulates erastin-induced ferroptosis in breast cancer cells

(A-B) MDA-MB-231 cells were treated with siRNA control (siNT) or siRNA targeting TAZ (siTAZ) for two days. The transfected cells were treated by indicated concentrations of erastin for 24 hours and cell viability was determined by crystal violet staining (A) or by CelltiterGlo (n=3; two-way ANOVA; **p < 0.01) (B). The knockdown efficiency of transfected cells was checked by western blots (C). (D-E) MDA-MB-231 cells were treated with two additional siRNAs targeting TAZ (confirmed by Western blots (E)) and treated with 2.5μM erastin for 24 hours and their viability was determined by CelltiterGlo. Data are represented as mean±SEM and by comparing to the DMSO controls of each group (n=2; one-way ANOVA; *p < 0.05).

Conversely, we tested whether forced activation of TAZ would also affect erastin sensitivities. The expression of a constitutively active form of TAZ, TAZS89A (Lei et al.,

2008), significantly increased the sensitivity of RCC4 to erastin (Figure 25). Collectively, these data strongly indicate that the activation status of TAZ regulates the sensitivities to erastin-induced ferroptosis under different cell density.

Figure 25 Overexpression of TAZ sensitizes RCC4 cells to erastin-induced ferroptosis

RCC4 cells were transfected with pLX304 control vector or TAZS89A plasmid and then treated with 4µM erastin for 24 hours. (A) Cell viability was then determined by CelltiterGo. Data are represented as mean±SEM, n=3 by comparing to the DMSO controls. (Student’s t-test; ***p < 0.001) (B) Cells were also harvested for western blot analysis of TAZ protein.

To characterize the in vivo relevance of TAZ-regulated ferroptosis, we determined the erastin sensitivity of 786O grown in 3D cell culture by Matrigel and found that erastin decreased the sphere size in control cells, but not in TAZ knockdown

cells (Figure 26). In addition, we determine how the TAZ knockdown affected the response of 786-O xenografts to erastin treatments in vivo (Figure 27). After the 786-O cells were injected subcutaneously to establish tumors, the mice were treated with erastin. We found that erastin administration significantly reduced the tumor growth in the control group, but not in the TAZ knockdown group (Figure 27). Together, these data support the relevance of TAZ to the in vivo response of xenograft and tumorspheres to erastin-induced ferroptosis.

Figure 26 TAZ knockdown decreases tumorsphere under 3D culture

(A) Western blot confirmed the knockdown efficiency of shRNA targeting TAZ in 786O cells. (B-C) 3D Matrigel culture of 786O cells which have stable integration of shRNA targeting TAZ or control, pLKO.1 were treated with DMSO or 0.5µM erastin for 7-10 days before the measurement of colony diameters. (n=56 each group; two-way ANOVA; ***p < 0.001, ns: not significant).

Figure 27 Mice xenograft models of TAZ silencing in response to erastin

(A-B) 786O cells with or without TAZ knockdown were injected subcutaneously to establish tumors until 120-150 mm3 before starting erastin treatments. The tumor volume was measured for 20 days during the oral gavage of ORA-Plus® or 0.1 ml of erastin (4 mg/ml) twice daily. Data are presented for four to five animals per group.

2.3.3 EMP1 is a direct target gene of TAZ that regulates ferroptosis sensitivity

Next, we sought to identify the mechanism by which TAZ regulates erastin sensitivity in the RCC cell lines. TAZ is a transcriptional coactivator that affects cellular phenotypes through transcriptional regulation of its target genes.

Figure 28 Microarray analysis of gene expression upon siTAZ

(A) RT-qPCR validates the knockdown efficiency of siTAZ for the RNA samples for microarray analysis. Data are represented as mean SEM, n=3 biological replicates after normalized to the DMSO controls (mock). (B) Gene expression profiles were analyzed by U133A2.0 microarray in RCC4 cells after treating with 1µM erastin for 7 hours following silencing TAZ for two days.

Based on the assumption that TAZ silencing may repress genes essential for ferroptosis, we focused on genes that were both repressed during TAZ knockdown

(Figure 28) and essential for the cystine-deprived death of RCC4 (RNAi screen) (Ding, unpublished data).

To accomplish this, we first determined the transcriptional response of RCC4 to knockdown of TAZ. RNA was isolated from RCC4 transfected with either control (siNT) or TAZ-targeting siRNAs (siTAZ) with or without erastin treatment in triplicates and hybridized to Affymetrix U133A2.0 microarrays (Figure 28). The transcriptional responses of the TAZ knockdown were then derived by zero-transformation against the average of the DMSO-treated siNT control samples as previously performed (Keenan et al., 2015; Tang et al., 2017a). Next, we identified the candidate genes which were both repressed during TAZ knockdown (microarrays) and found to be essential for the cystine-deprived death of RCC4 (RNAi screen) (Figure 29).

Figure 29 Venn diagram of siTAZ and cystine deprivation

Venn diagram showing genes that are downregulated upon siTAZ (176 genes) and siRNA screen of genes that are resistant to cystine deprivation (388 genes).

From these comparisons, we identified eleven candidate genes (Figure 29), including TAZ. After excluding TAZ and two other genes encoding only one subunit of multi-component protein complexes, we set up a mini-screen to prioritize the eight remaining candidate genes. First, we used qRT-PCR to validate the mRNA downregulation upon TAZ removal (Figure 30A). Second, we performed functional assays to determine if the silencing of each candidate gene in RCC4 would confer protection against ferroptosis (Figure 30B).

Figure 30 Screen of candidate genes

(A) Validating knockdown efficiency of siRNAs targeting candidate genes by qRT-PCR in RCC4 cells. (B) Functional assay of candidate genes by siRNA-mediated knockdown. The cell viability data were determined by celltiterGlo and showed as the ratio of siRNA(s) to siNT control after normalization of its own 2µM erastin to DMSO, that is, siRNAerastin/DMSO/siNTerastin/DMSO.

Based on these criteria, the epithelial membrane protein 1 (EMP1) emerged as the most promising candidate. The EMP1 mRNA was significantly downregulated upon

TAZ silencing (Figure 31).

Figure 31 EMP1 gene expression is down-regulated upon siTAZ

RT-qPCR validates that EMP1 is downregulated when TAZ is silenced by siRNAs. Data are represented as mean±SEM, n=3 after normalized to the DMSO controls (mock); one- way ANOVA; ***p < 0.001, ns: not significant).

In addition, the expression EMP1 mRNA was strongly correlated with the expression of CTGF and CYR61 mRNAs, two well-known YAP/TAZ target genes (Choi et al., 2015; Zhao et al., 2008), in TCGA (Figure 32) and CCLE (Figure 33).

Figure 32 Gene correlation analysis of TCGA

The correlation between the expression of indicated genes in the TCGA kidney cancer datasets. Data is retrieved from GEPIA (http://gepia.cancer-pku.cn/) with Spearman correlation analysis. TPM: Transcripts Per Kilobase Million.

Figure 33 Gene correlation analysis of CCLE

The correlation between the expression of indicated genes in the CCLE dataset. Data were downloaded and re-graphed by GraphPad PRISM with Spearman correlation analysis. 43

Most importantly, the EMP1 silencing by multiple siRNAs protected RCC4 from erastin-induced death (Figure 34A) and reduced the mRNA (Figure 34B) and protein expression (Figure 34C). EMP1 knockdown also conferred ferroptosis resistance in 786O cells (Figure 35) and MDA-MB-231 cells by both cell viability and cell death assays

(Figure 36). Therefore, the downregulation of EMP1 may contribute to the ferroptosis resistance conferred by the TAZ silencing in multiple cancer cell lines.

Figure 34 siRNAs targeting EMP1 reduces erastin-induced ferroptosis

Treatment of siRNA control, non-targeting (siNT), multiple individual siRNAs targeting EMP1 (#1 to #4) for two days in renal carcinoma cells, RCC4 cells. (A) Cell viability was determined by CelltiterGlo after 24 hours of 1µM erastin treatment. Data are represented

as mean SEM, n=3 by comparing to the DMSO controls. one-way ANOVA; ***p < 0.001. (B-C) The successful knockdown efficiency of siRNAs targeting EMP1 by qRT-PCR (B) and western blots (C) in RCC4 cells.

Figure 35 EMP1 knockdown reduces ferroptosis sensitivity

Treatment of siRNA control, non-targeting (siNT), or siEMP1 (pooled), for two days in renal carcinoma cells, 786O cells. (A) Cell viability was determined by CelltiterGlo after 24 hours of 1µM erastin treatment. Data are represented as mean±SEM, n=3 by comparing to the DMSO controls. Student’s t-test; **p < 0.01. (B) Validation of knockdown efficiency of siRNAs targeting EMP1 by qRT-PCR.

Figure 36 EMP1 regulates erastin-induced ferroptosis sensitivity

MDA-MB-231 cells were treated with siNT or individual siRNAs targeting EMP1 (#1 to #4) for two days following by indicated concentrations of erastin treatment. Data are represented as mean±SEM, n=3 by normalized to the DMSO controls of each group. (A) Cell viability was determined by CelltiterGlo (n=3; two-way ANOVA; ***p < 0.001). (B) Cytotoxicity was determined by CytoTox-Fluor assay after treatment of 8µM erastin (n=4; one-way ANOVA; ***p < 0.001).

Next, we examined whether overexpression of EMP1 may sensitize ferroptosis after TAZ knockdown. First, we observed that overexpression of V5-tag EMP1 significantly sensitized RCC4 to erastin (Figure 37). Next, RCC4 stably expressing control vector (pLX304) or V5-tag EMP1 were transfected with either control (siNT) or

TAZ-targeting siRNAs (siTAZ) followed by erastin treatment. Overexpression of EMP1 significantly reduced the ferroptosis protection conferred by TAZ knockdown (Figure

38). Collectively, these data are consistent with the concept that the repression of EMP1 genetically works downstream of TAZ to regulate ferroptosis.

Figure 37 Overexpression EMP1 sensitizes RCC4 cells to ferroptosis

(A) Crystal violet staining of RCC4 cells that are stably overexpressed control vector, pLX304 or pLX304-EMP1-V5 (EMP1) after treating with 3µM erastin for 1 day. (B) Validation of EMP1 overexpression by V5 tag antibody.

Figure 38 Genetic interaction between TAZ and EMP1

After stably overexpressing with EMP1, RCC4 cells were treated with siRNAs targeting TAZ. Later, cell viabilities were determined by CelltiterGlo after 24 hours of 2 µM erastin treatment. Data are represented as mean±SEM, n=3 after normalized to the DMSO controls (two-way ANOVA; *p < 0.05; ***p < 0.001).

Previous ChIP-seq studies have indicated that the regulatory regions of EMP1 were physically associated with YAP/TAZ/TEAD complexes (Stein et al., 2015; Walko et al., 2017; Zanconato et al., 2015). To validate that EMP1 is a direct target gene of TAZ, we performed ChIP-qPCR using an antibody specific for endogenous TAZ protein. As shown in Figure 39, we found that CTGF, a well-known YAP/TAZ target gene (Zhao et al., 2008), was highly enriched in the TAZ pull-down.

Figure 39 EMP1 is the direct target of TAZ

EMP1 is the direct target of TAZ. From RCC4 lysates, ChIP-qPCR with TAZ antibody (CST #70148) validates that TAZ binds to EMP1 promoter. ChIP-qPCR of the CTGF promoter serves as a positive control, while Ch14 serves as a negative control. Data are shown as three technical replicates, mean+SEM, from one representative experiment out of four biological repeats.

Similarly, EMP1 was also significantly enriched in the TAZ pull-down, indicating that the EMP1 promoter is directly associated with TAZ-containing complexes. Together, these data support the binding of TAZ to the promoter of EMP1 as a potential direct target gene and its downregulation upon TAZ silencing contributes to the ferroptosis resistance phenotypes.

2.3.4 EMP1 regulates ferroptosis through NOX4

To investigate the mechanistic link between EMP1 and ferroptosis, we sought to determine whether EMP1 would affect the levels of GPX4 or NOX4, two key regulators of lipid peroxidation and ferroptosis (Dixon et al., 2012; Poursaitidis et al., 2017; Yang et al., 2014). We found that EMP1 silencing decreases the mRNA expression of NOX4, but not GPX4 (Figure 40). Conversely, overexpression of EMP1 increases the mRNA level of

NOX4, but not GPX4 (Figure 41).

Figure 40 knockdown of EMP1 down-regulates NOX4

(A) siRNA-mediated knockdown of the EMP1 gene reduces the mRNA level of NOX4. (B- C) qRT-PCR of EMP1 (B) and GPX4 (C) mRNA level upon siRNA-mediated knockdown of EMP1 (n=3; mean ± SEM; Student’s t-test; *p < 0.05; ***p < 0.001; ns: not significant).

Figure 41 Overexpression of EMP1 up-regulates NOX4

(A) Overexpression of EMP1 increases NOX4 mRNA level. (B-C) qRT-PCR of EMP1 (B) and GPX4 (C) mRNA level upon EMP1 overexpression (n=3; mean ± SEM; Student’s t test; **p < 0.01; ***p < 0.001; ns: not significant).

Consistently, overexpression of EMP1 also increases NOX4 protein level (Figure 42A), but not GPX4 protein level (Figure 42B). Consistent with its ability to induce EMP1,

TAZS89A also increased the level of NOX4 protein (Figure 42C), but not GPX4 (Figure

42B). Collectively, these data suggest that EMP1 may regulate ferroptosis by affecting the levels and activities of NOX4. Therefore, we further test the role of NOX4 in the cell-

density-dependent ferroptosis mediated by TAZ and EMP1.

Figure 42 Higher amount of NOX4 protein correlates with EMP1 or TAZS89A overexpression

(A) V5 antibody indicated EMP1-V5 overexpression. The western blots of NOX4 were quantified, normalized to the β-tubulin, expressed relative to the control vector, and showed as below numbers. Representative data from one of at least three independent studies. (B) Western blot of GPX4 protein level when overexpression of TAZS89A or EMP1 (C) Western blot of NOX4 protein level when overexpression of TAZS89A.

To rule out that elevated NOX4 may affect GSH, we measured the GSH levels when the control and TAZ-knockdown cells were exposed to erastin. We found that erastin treatments significantly reduced the GSH levels (Figure 43), as previously reported (Yang et al., 2014). However, the TAZ knockdown did not affect the GSH levels before or after erastin treatments (Figure 43).

Figure 43 GSH assay upon siTAZ

The GSH (glutathione) levels were determined after TAZ knockdown followed by 1µM erastin treatment for 1 day and normalized to the siNT DMSO (n=7; mean ± SEM; two- way ANOVA; ***p < 0.001).

Furthermore, NOX1/NOX4 inhibitor, GKT136901 (Laleu et al., 2010) protected

RCC4 from ferroptosis (Figure 44). Since NOX1 was not expressed at detectable levels in

RCC cell lines (Gregg et al., 2014), GKT136901 probably mediated the ferroptosis protection through NOX4.

Figure 44 Inhibition of NOX4 reduces erastin-induced ferroptosis

Cell viability was determined by CelltiterGlo after 24 hours of indicated concentrations of erastin with NOX4 inhibitor, 10 µM GKT136901 (mean SEM, n=3, two-way ANOVA***p < 0.001).

Consistently, the silencing of NOX4 by siRNAs also conferred ferroptosis resistance (Figure 45). Conversely, overexpression of NOX4 lets the cells more sensitive to erastin treatment determined by both cell viability and cell death assays (Figure 46).

Figure 45 NOX4 knockdown reduces erastin-induced ferroptosis

(A) Cell viability was determined by CelltiterGlo after 24 hours of indicated concentrations of erastin with silencing NOX4 by two individual siRNAs. mean±SEM, n=3, Data are represented by comparing to the DMSO controls; two-way ANOVA; ***p < 0.001. (B-C) The knockdown of NOX4 by siRNAs was validated by qRT-PCR (B) and western blot (C).

Figure 46 NOX4 overexpression sensitizes RCC4 cells to erastin treatment

(A) Validation of NOX4 overexpression by western blot. (B-C) NOX4 overexpression sensitizes RCC4 cells to erastin treatment (4 µM) determined by CelltiterGlo (B) or CytoTox-Fluor assay (C); n=3; mean ± SEM; Student’s t-test; ***p < 0.001.

To investigate the genetic interaction between EMP1 and NOX4, we exposed

EMP1-overexpressing RCC4 to the NOX4 inhibitor, GKT136901. We found that the

EMP1 overexpression increased erastin sensitivity, but this increased sensitivity was abolished by GKT136901 (Figure 47). These data indicate the role of elevated NOX4 in the ferroptosis-sensitizing effects of EMP1 over-expression.

Figure 47 Genetic interaction between EMP1 and NOX4

Cell viabilities were determined by CelltiterGlo after RCC4 cells were treated with EMP1 overexpression for 1 day following 20µM NOX4 inhibitor, GKT136901, and 8µM erastin for 1 day. Data are represented as mean±SEM, n=3 after normalized to the DMSO controls; two-way ANOVA; **p < 0.01; ***p < 0.001.

In addition, as p38 MAPK has been reported to regulate NOX4 reciprocally

(Dougherty et al., 2017; Huang et al., 2018; Park et al., 2010; Peng et al., 2013), we examined the role of p38 activity in the NOX4 induction by EMP1. We found that EMP1 expression increases the phosphorylation and activation of p38 (Figure 48A).

Furthermore, EMP1 silencing reduced the p38 phosphorylation (Figure 48B). In addition, p38 inhibitor, SB203580, abrogated the increase in NOX4 protein associated with EMP1 expression (Figure 48C).

Figure 48 EMP1 regulates NOX4 through p38

(A-B) Western blot analysis of phospho-p38 (Thr180/Tyr182), total p38, and control β- tubulin upon overexpression of EMP1 (A) or siRNA-mediated silencing of EMP1 (B). (C) Western blot analysis of NOX4 upon overexpression of EMP1 with the treatment of p38 inhibitor, SB203580. Quantification of the ratio of NOX4 to β-tubulin was normalized to DMSO control.

Figure 49 Inhibition of TAZ, EMP1, or NOX4 abolishes elevated lipid-ROS by erastin treatment

(A) The negative (unstained) and positive controls (cumene hydroperoxide) for C11- BODIPY flow cytometry. (B) Lipid ROS in RCC4 cells treated with knockdown TAZ or EMP1 as well as inhibition of NOX4 was assessed by flow cytometry using C11-BODIPY. Representative data from one of five independent studies.

Finally, because the accumulation of lipid-based reactive oxygen species (lipid

ROS) is crucial for ferroptosis (Dixon et al., 2012), we measured the lipid ROS by C11-

BODIPY staining combined with flow cytometry detection. The dye senses lipid peroxidation by oxidation of the polyunsaturated butadienyl portion resulting in the fluorescence shift from ~590 nm to ~510 nm (Pap et al., 1999). As shown in Figure 49, the

increased accumulation of lipid ROS induced by erastin is abolished by knockdown of

TAZ, EMP1 mRNA as well as NOX4 inhibitor, GKT136901.

Figure 50 The proposed molecular mechanism

A schematic model illustrating the erastin-induced ferroptosis regulated by the TAZ- EMP1-NOX4 axis.

Taken together, we propose a signaling mechanism (Figure 50), in which TAZ is a cell-density-dependent determinant of ferroptosis sensitivity through affecting the levels

EMP1, which in turn regulates NOX4, lipid peroxidation and ferroptotic death of RCC.

2.4 Discussion

Here, we provide evidence that a non-genetic factor, cell density, contributes to the regulation of ferroptosis sensitivities. Similar observations are also reported recently

(Panzilius et al., 2018; Wu et al., 2019). Importantly, we have elucidated a molecular

mechanism by which cell density regulates ferroptosis in RCC through the Hippo pathway effector, TAZ, which affects the levels of EMP1, NOX4 and lipid peroxidation

(Figure 51).

Figure 51 A schematic model illustrating the ferroptosis regulated by TAZ-EMP1- NOX4 axis

As the Hippo pathway and YAP/TAZ has been implicated in sensing cell density, the activation status of TAZ contributes significantly to density-dependent ferroptosis. Furthermore, we found that TAZ directly binds to the EMP1 promoter region and regulates the EMP1 expression. The higher EMP1 protein, in turn, regulates ferroptosis through elevating NOX4 and resulting lipid peroxidation. Higher levels of

NOX4 has been reported to enhance lipid peroxidation and sensitize cells to ferroptosis

(Poursaitidis et al., 2017). Consistent with our findings, sub-confluent colonic epithelial cells have higher ROS production than cells in confluent cells (Perner et al., 2003).

Collectively, our results suggest that TAZ activation may sensitize cells to ferroptosis and be predictive of sensitivity to ferroptosis-inducing agents.

YAP/TAZ and other components of the Hippo pathway integrate a wide variety of non-genetic factors, such as cell density, mechanical properties and metabolic status

(Zanconato et al., 2016). Therefore, the regulation of ferroptosis susceptibility by TAZ suggests that these non-genetic factors may play important roles in ferroptosis sensitivities. For example, both epithelial-mesenchymal transition and fibrosis are prominent features of RCC. These changes may create the “stiff” extracellular environment known to activate the YAP/TAZ, which may promote the sensitivities to ferroptosis. In addition, YAP/TAZ is also regulated by metabolic pathways. A previous study has shown the essential role of glutamine metabolism in ferroptosis (Gao et al.,

2015) but the mechanisms remain unclear. YAP/TAZ has been reported to be activated by O-GlcNAcylation (Peng et al., 2017), a reversible modification that is highly sensitive to glucose or glutamine levels (Chen et al., 2018). Therefore, it is tempting to hypothesize that nutrient-sensing O-GlcNAcylation of YAP/TAZ may contribute to the metabolic regulation of ferroptosis by glutamine metabolisms.

Several studies have implicated Hippo pathways in various types of RCC (Cao et al., 2014; Schutte et al., 2014; Sourbier et al., 2018). However, most of these papers have focused on the YAP without rigorously testing the expression and function of TAZ. Our results suggest the expression and functional role of TAZ in the RCC cell lines and early passage tumor cells. In addition, YAP can be induced in RCC4 upon TAZ silencing.

Therefore, it will be important to clearly distinguish between these two highly similar but distinct Hippo effectors, especially when some detection reagents and methods may not reliably distinguish between YAP and TAZ.

Our results may also have significant therapeutic implications. While inducing ferroptosis may have significant anti-tumor potential, much remains unknown to select tumors that may best respond to these ferroptosis-inducing agents. Our results may suggest that TAZ-activated tumors may be particularly responsive to ferroptosis- inducing therapies. In addition, while many tumor cells initially respond to various therapeutic agents, chemo-resistant persister cells eventually emerge and cause significant morbidity and mortality. Recently, these persister tumor cells are found to express a low level of GPX4 and be highly sensitive to ferroptosis (Hangauer et al., 2017;

Viswanathan et al., 2017). The drug resistance and persister cells are also known to be associated with YAP/TAZ activation (Lin et al., 2015; Zanconato et al., 2016) and EMP1 expression (Jain et al., 2005). Our results suggest that a GPX4-independent regulatory

link between chemo-resistance and ferroptosis through TAZ-EMP1-NOX4 activation.

Therefore, TAZ-EMP1-NOX4 and ZEB-GPX4 may represent two distinct and independent pathways linking the chemo-resistance with ferroptosis, a possibility that needs to be rigorously tested in a large panel of drug-resistant persister cells in the future.

Besides cancer, ferroptosis also has emerging roles in other pathophysiological processes and human diseases, such as neurotoxicity (Skouta et al., 2014), acute renal failure (Angeli et al., 2014), cardiac injury (Gao et al., 2015) and ischemia-reperfusion injury (Linkermann et al., 2014). Therefore, modulating ferroptosis may have therapeutic potentials. In addition, the ability of NOX4 inhibitors to abolish ferroptosis suggests the potential for developing combinational therapies. Currently, several oral NOX4 inhibitors are under preclinical studies (Borbély et al., 2010) and may be tested for their efficacy in mitigating these ferroptosis-associated diseases.

3. A TAZ-ANGPTL4-NOX2 axis regulates ferroptotic cell death and chemoresistance in epithelial ovarian cancer

Portions of this Chapter are reproduced from the Molecular Cancer Research:

Molecular Cancer Research, 2019, DOI: 10.1158/1541-7786.MCR-19-0691

Ovarian cancer (OVCA) is the deadliest gynecologic cancer. Despite recent advances, clinical outcomes remain poor, necessitating novel therapeutic approaches. To investigate metabolic susceptibility, we performed nutrigenetic screens on a panel of clear cell and serous OVCA cells and identified cystine addiction and vulnerability to ferroptosis, a novel form of regulated cell death. Our results may have therapeutic potential, but little is known about the determinants of ferroptosis susceptibility in

OVCA. We found that vulnerability to ferroptosis in OVCA is enhanced by lower cell confluency. Since the Hippo pathway effectors YAP/TAZ are recognized as sensors of cell density, and TAZ is the predominant effector in the tested OVCA cell lines, we investigated the role of TAZ in OVCA ferroptosis. TAZ removal confers ferroptosis resistance, while TAZS89A overexpression sensitizes cells to ferroptosis. In addition, we found that lower TAZ level in chemoresistant recurrent OVCA is responsible for reduced ferroptosis susceptibility. The integrative genomic analysis identified

ANGPTL4 as a direct TAZ target gene that regulates ferroptosis by activating NOX2.

Collectively, cell density-regulated ferroptosis in OVCA is mediated by TAZ through

the regulation of the ANGPTL4-NOX2 axis, suggesting therapeutic potentials for

OVCAs and other TAZ-activated tumors.

3.1 Introduction

3.1.1 The impact of OVCA and the critical need for novel therapeutics

Epithelial ovarian cancer is the deadliest gynecologic cancer and takes the lives of approximately 150,000 women every year worldwide (Lheureux et al., 2019). The symptoms of ovarian cancer are vague and often attributed to other more common ailments. As a result, the correct diagnosis usually only occurs after cancer has spread beyond the ovaries. Standard therapy involves surgical debulking followed by chemotherapy with a platinum-taxane doublet (Aletti et al., 2007). While many patients initially respond favorably to this combined treatment (Ozols, 2006), most patients relapse with a recurrent disease that is often resistant to platinum and taxane drugs.

Other chemotherapeutic options are used, mainly in an effort to prolong survival. platinum-PARP inhibitor combinations have been proven to be beneficial for OVCA regardless of the BRCA1/2 mutation status (Mirza et al., 2016; Shen et al., 2019).

However, the outcomes for most women with OVCA are still unsatisfactory, therefore, novel therapeutic options are still urgently needed.

3.1.2 Ferroptosis as a novel cell death involving lipid peroxidation

One possible therapeutic approach is the induction of ferroptosis, a novel and distinct form of iron-dependent programmed cell death (Dixon et al., 2012; Yang et al.,

2014). Ferroptosis sensitivity is found to be affected by various biological processes, such as loss of p53 (Jiang et al., 2015; Tarangelo et al., 2018; Xie et al., 2017), DNA damage pathway (Chen et al., 2019), or epithelial-mesenchymal transition (EMT) (Viswanathan et al., 2017), which are often dysregulated in OVCA. Ferroptosis can be induced by the small molecule, erastin (Dolma et al., 2003), or by cystine deprivation (Poursaitidis et al.,

2017); both approaches reduce cystine import and result in a redox imbalance by reducing intracellular glutathione levels. Glutathione is a critical cofactor for glutathione peroxidase (GPX4), an enzyme that resolves the accumulation of lipid-based reactive oxygen species (ROS) (Yang and Stockwell, 2016). Therefore, ferroptosis and lipid peroxidation can also be induced by chemical or genetic inhibition of GPX4 (Yang et al.,

2014). A previous study indicated that the low level of GPX4 in drug-resistant cancer cells, mediated by the epithelial-mesenchymal transition (EMT) regulator: ZEB1, may enhance ferroptosis sensitivity, implicating GPX4 as a major determinant of ferroptosis

(Viswanathan et al., 2017). On the other hand, accumulation of lipid-based ROS and ferroptosis can also be induced by the generation of superoxide and hydrogen peroxide upon upregulation of NADPH oxidases (NOXs) (Dixon et al., 2012).

In our current study, we performed a nutrigenetic screen and show that most

OVCA cell lines are addicted to cystine and sensitive to ferroptosis. Furthermore, we found that OVCA ferroptosis susceptibility is significantly affected by cell density in a way that is highly analogous to our study of the RCC’s ferroptosis as detailed in Chapter

2 (Yang et al., 2019b). Low density, but not high-density OVCA cells, were highly susceptible to erastin-induced ferroptosis. The density-dependent phenotypes of cancer cells are sensed and regulated by the evolutionarily conserved Hippo pathway (Mori et al., 2014; Zhao et al., 2007) converging into two transcriptional co-activators, YAP (yes- associated protein 1) and TAZ (transcriptional coactivator with PDZ-binding motif).

YAP/TAZ activities are regulated by their phosphorylation and intracellular localization.

When grown at high cell density, YAP/TAZ are phosphorylated, retained in the cytosol and are subjected to proteasomal degradation. Upon shifting cells to low cell density,

YAP/TAZ become dephosphorylated and translocate into the nucleus to associate with transcriptional factors that together drive gene expression regulating cell proliferation, differentiation, and migration (Hsiao et al., 2016; Zhao et al., 2007). Here, we have established the role of cell density and TAZ in regulating OVCA ferroptosis. In addition, we found that TAZ regulates erastin-induced ferroptosis through the induction of

ANGPTL4, which in turn activates NOX2 and ferroptosis. Thus, these data support the role of TAZ in regulating ferroptosis through ANGPTL4-NOX2 and that inducing

ferroptosis may be a novel therapeutic strategy for OVCA and other TAZ-activated tumors.

3.2 Methods

3.2.1 Materials and reagents

The following antibodies were used in this section: YAP/TAZ (#8418, Cell

Signaling Technology, 1:1000), beta-tubulin (#86298, Cell Signaling Technology, 1:2000), vinculin (sc-73614, Santa Cruz, 1:2000), V5 tag (#13202, Cell Signaling Technology,

1:2000), H3 (#4499, Cell Signaling Technology, 1:2000), GAPDH (sc-25778, Santa Cruz,

1:2000), ANGPTL4 (#40-9800, ThermoFisher Scientific, 1:1000), NOX2 (sc-130543, Santa

Cruz, 1:1000), anti-rabbit IgG, horseradish peroxidase (HRP)-linked antibody (#7074,

Cell Signaling Technology, 1:2000-1:4000) and anti-mouse IgG, HRP-linked

Antibody (#7072, Cell Signaling Technology, 1:2000-1:4000). Plasmids were obtained from Addgene (TAZS89A #52084; ANGPTL4-V5 #102446). The NOX2 inhibitor, gp91 dstat, was purchased from Eurogentec (cat #: AS-63818) and recombinant human

ANGPTL4 protein was purchased from Novus Biologicals (4487-AN). VAS2870

(Calbiochem-492000), GKT136901 (Calbiochem-534032), Ferrostatin-1 (SML0583) and liproxstatin-1 (SML1414) were purchased from Sigma. Erastin was obtained from the

Duke University Small Molecule Synthesis Facility.

3.2.2 Cell culture and transfection

Cell lines are maintained as part of the cell line repository within the Division of

Reproductive Sciences at Duke University. STR profilings at the Duke University DNA

Analysis Facility were performed each time when frozen stocks are prepared. They were last genotyped and mycoplasma-free tested on 6/5/18. For nutrigenetic screens, all

OVCA cells were cultured in a humidified incubator at 37°C and 5% CO2 using custom- made Dulbecco’s Modified Eagle Medium (11995-DMEM, ThermoFisher Scientific) with

10% heat-inactivated and dialyzed Fetal Bovine Serum (HyCloneTM FBS, GE Healthcare

Life Sciences #SH30070.03HI) with the indicated amino acid removed as previously described (Tang et al., 2017a). Transfections were performed according to the manufacturer’s instructions with TransIT-LT1 transfection reagent (Mirus Bio) or

RNAiMax transfection reagent (ThermoFisher Scientific).

3.2.3 siRNA-mediated gene knockdown

All human siRNAs were purchased from Dharmacon or Qiagen: Non-targeting control, siNT (Qiagen SI03650318); siTAZ (Dharmacon M-016083); siANGPTL4

(Dharmacon M-007807); siNOX2 (Dharmacon M-011021); ; siTAZ#1 (target sequence:

AGA CAT GAG ATC CAT CAC TAA); siTAZ#2 (target sequence: ACA GTA GTA CCA

AAT GCT TTA); siANGPTL4 #1 (target sequence: CTG CGA ATT CAG CAT CTG

CAA); siANGPTL4 #2 (target sequence: CAC CAT GTT GAT CCA GCC CAT); siNOX2

#1 (target sequence: GAA GAC AAC TGG ACA GGA A); siNOX2 #2 (target sequence:

GAA ACT ACC TAA GAT AGC G). Knockdown efficacy was validated by RT-qPCR and/or western blots.

3.2.4 RNA isolation and quantitative real-time PCR

Total RNAs from cultured cells were extracted using the RNeasy mini kit

(Qiagen #74104) with DNase I treatment (Qiagen #79254). cDNAs were synthesized from

1μg of total RNA using SuperScriptTM II Reverse Transcriptase (ThermoFisher Scientific

#18064) with random hexamers following protocols from the manufacturer. The levels of gene expression were measured by quantitative PCR (qPCR) with a power SYBR green

PCR mix (Applied Biosystems, ThermoFisher Scientific #4367659). Primers used included (listed 5’ to 3’): β-actin-F': GGG GTG TTG AAG GTC TCA AA; β-actin-R': GGC

ATC CTC ACC CTG AAG TA; TAZ-F': TGC TAC AGT GTC CCC ACA AC; TAZ-R':

GAA ACG GGT CTG TTG GGG AT; ANGPTL4-F': GGC TCA GTG GAC TTC AAC CG;

ANGPTL4-R': CCG TGA TGC TAT GCA CCT TCT; NOX2-F': TGG AGT TGT CAT CAC

GCT GTG; NOX2-R': CTG CCC ACG TAC AAT TCG TTC; 18S-F': CTG GAT ACC GCA

GCT AGG AA; 18S-R': CCC TCT TAA TCA TGG CCT CA.

3.2.5 Western blot analysis

For immunoblotting, please refer to the previous Chapter 2.2.5. In short, cells were collected and quantified by BCA protein assay. After separating by SDS-PAGE, the

proteins were transferred to PVDF membranes that were further blocked with 5% non-fat milk or BSA and probed with indicated antibodies following by HRP-conjugated secondary antibodies. The signals were developed and detected by Amersham ECL prime western blotting detection reagent and the Bio-Rad ChemiDocTM Imaging System.

3.2.6 Cell viability and cytotoxicity assays

After seeded and transfected with siRNAs for two days, cells (~60-70% confluence) were further treated for another 24 hours with erastin. Cell viability was evaluated using crystal violet staining or the CelltiterGlo luminescent cell viability assay kit (Promega G7571) according to the manufacturer’s instructions. The CelltiterGlo luminescent cell viability assay is based on quantitation of the ATP present, which is an indicator of metabolically active cells, to determine the number of living cells in the culture. The cytotoxicity and cell death of treated cells were determined by CytoTox-

Fluor™ Cytotoxicity Assay (Promega G9260) according to the manufacturer’s instructions.

3.2.7 Enzyme-linked immunosorbent assay

Quantification of secreted ANGPTL4 in the culture media from cells following a two-day incubation was performed by Human ANGPTL4 ELISA kit (RAB0017, Sigma

Aldrich) according to the manufacturer’s instruction.

3.2.8 ChIP analysis

ChIP-qPCR experiment was carried out according to the Myers Lab ChIP-seq protocol (Johnson et al., 2007). Briefly, OVCA cells were incubated in a cross-linking solution (1% formaldehyde) at room temperature for 10 min and then glycine was added at a final concentration of 0.125M to stop cross-linking. The cells were then washed with cold PBS and suspended in Farnham lysis buffer (5mM PIPES pH8.0, 85mM KCl and

0.5% NP-40) with freshly added protease inhibitor. The lysate was subsequently passed through a 20-gauge needle 20 times to breaks cells while keeping intact nuclei. After centrifugation, the pellet was resuspended with RIPA buffer with freshly added protease inhibitor. Chromatin fragmentation was performed by sonication using the

Bioruptor (Diagenode) at high speed for 30 min (30 seconds ON, 30 seconds OFF).

Proteins were immunoprecipitated in PBS/BSA buffer using TAZ antibody (Cell

Signaling Technology #70148) or control antibody, rabbit IgG (Cell Signaling Technology

#2729) which were then conjugated to DynabeadsTM protein G magnetic beads (Thermo

Fisher Scientific #10004D) at 4°C for 2 hr. The antibody-chromatin complexes were washed five times with LiCl wash buffer and then were washed with TE buffer. The crosslinking was reversed by incubation with elution buffer (1%SDS, 0.1M NaHCO3) at

65°C overnight followed by incubation with RNase A and proteinase K. DNA was recovered by using the QIAquick PCR purification kit (Qiagen #28104). Precipitated

DNA was analyzed by qPCR using primers targeting TEAD-binding sites found in

CTGF, a known target gene for YAP/TAZ (Stein et al., 2015) and ANGPTL4 promoter regions. Primers targeting chromosome 10 was used as a negative control for potential non-specific binding (Stein et al., 2015). Primer sequences are as follows (listed 5’ to 3’):

CTGF-F': GCC AAT GAG CTG AAT GGA GT; CTGF-R': CAA TCC GGT GTG AGT

TGA TG; ANGPTL4-F': GTC TCC CAC GGT TCG TAG AG; ANGPTL4-R': TAT AAG

TTG GGT GCG GAG TGG; Ch10-F': ACC AAC ACT CTT CCC TCA GC; Ch10-R': TTA

TTT TGG TTC AGG TGG TTG A.

3.2.9 Lipid ROS assay using flow cytometry

Lipid ROS levels were determined using 10 µM of C11-BODIPY dye (D3861,

ThermoFisher Scientific) according to the manufacturer’s instructions. Cells were seeded and treated with siRNAs in six-well plates for two days, then the culture medium was replaced with 10 µM erastin treatment for 18 hours. The next day, the medium was replaced with 10 µM C11-BODIPY-containing medium for 1 hour. Later, the cells were harvested by trypsin and washed three times with ice-cold PBS followed by resuspending in PBS plus 1% BSA. The amount of ROS within cells was examined by flow cytometry analysis (FACSCantoTM II, BD Biosciences).

3.2.10 Statistical analyses

The results were evaluated with two-tailed Student’s t-test or ANOVA (one- or two-way) analyses using GraphPad Prism version 8.0.1 (GraphPad Software).

Statistically significant differences were set to P < 0.05 between experimental groups (*<

0.05; **<0.01; ***<0.001). The statistical data were represented as mean ± SEM. The number of biological replicates is listed in each figure.

3.2.11 Data availability

RNA-seq for transcriptional profiles of CAOV2 primary and CAOV2 recurrent cells (GSE133663) and CAOV2 ovarian cancer cells with TAZ silencing (GSE133664) have been deposited in the NCBI Genome Expression Omnibus (GEO) under

SuperSeries GSE 133673.

3.3 Results

3.3.1 Ovarian cancer cells are sensitive to cystine deprivation

Many tumor cells have dysregulated metabolisms that render them addicted to certain nutrients, such as amino acids (Keenan and Chi, 2015; Vander Heiden, 2011). To reveal such nutrient addiction in OVCAs, we established a nutrigenetic screen by determining the phenotypic response upon the removal of glucose or each of the 15 individual amino acids usually included in Dulbecco's Modified Eagle Medium (Keenan

and Chi, 2015; Tang et al., 2015a; Tang et al., 2016). The nutrigenetic screen was applied to a panel of eight OVCA cell lines including four cell lines of clear cell subtype (TOV-

21G, ES-2, RMG-2, RMG-V) and four cell lines of serous subtype (OVCA432, OVCA429,

OVCA420, 41M). The screen revealed that the removal of cystine, the dimeric form of cysteine, dramatically reduced the viability of most tested clear cell and serous subtypes of OVCA cells (Figure 52).

Figure 52 Ovarian cancer cells are sensitive to cystine deprivation

Normalized cell viabilities of the indicated ovarian cancer cells after deprivation of individual amino acids or glucose. Error bars represent mean ± SEM (n=3, biological replicates).

Glutamine or glucose deprivation also decreased cell viability, especially in the clear cell subtype, reminiscent of the glutamine addiction of basal-type breast cancer cells (Kung et al., 2011). Knowing that cystine deprivation triggers ferroptosis (Dixon et

al., 2012), we determined whether ferroptosis inhibitors can rescue the cell death triggered by cystine deprivation. Both liproxstatin-1 and ferrostatin-1, two ferroptosis inhibitors, significantly rescued cystine deprivation-induced cell death of TOV-21G cells, indicating that cystine deprivation indeed induces ferroptosis (Figure 53). We further confirm that the erastin-induced cell death can be rescued by ferrostatin-1, but not apoptosis inhibitor, Z-VAD-FMK (Figure 54). Therefore, we used erastin, a canonical inducer of ferroptosis, in our subsequent studies.

Figure 53 Ferroptosis inhibitors rescue cystine deprivation-induced cell death

TOV-21G cells were seeded in 10% (20 µM of cystine) or 20% (40 µM of cystine) of full DMEM media supplied with or without either 200 nM liproxstatin-1 (A) or 2 µM ferrostatin-1 (B). After 2 days, the cell viabilities were determined by CelltiterGlo and normalized to the signal of full DMEM media (200 µM of cystine) of each group. (n=3 per group; mean ± SEM; two-way ANOVA).

Figure 54 Erastin-induced cell death in OVCA is rescued by ferroptosis inhibitor

Cell viability of CAOV2 was determined by CelltiterGlo after cells were treated with 16 µM erastin, 20µM Z-VAD-FMK (Z-VAD), or 1µM ferrostatin-1 (Fer-1).

3.3.2 Cell density affects the sensitivity of OVCAs to erastin-induced ferroptosis

While investigating ferroptosis in OVCA cells, we observed that cell density affects the ferroptosis sensitivity of CAOV2, a serous ovarian cancer cell line. When

CAOV2 cells were plated at low density, they were highly sensitive to erastin based on

CelltiterGlo assay (Figure 55) and crystal violet staining (Figure 56). In contrast, when plated at higher densities, CAOV2 cells became much less sensitive to erastin-induced death (Figure 55 and Figure 56).

Figure 55 Cell density regulates the sensitivity of CAOV2 to ferroptosis

The relative cell viabilities of CAOV2 cells grown at low/medium/high (4000/8000/16000 cells per 96 well) cell densities after treated with indicated concentrations of erastin. Data are represented as mean ± SEM with three biological replicates per group; two-way ANOVA; ***p < 0.001.

Figure 56 ferroptosis sensitivity is regulated by cell densities

(A) Crystal violet staining of CAOV2 cultured at low/medium/high (1.6x105/3.2x105/6.4x105 cells per 6 well) cell densities and treated with 16 µM erastin. (B) Quantitative analysis of crystal violet stain of CAOV2 cells at low, medium and high cell densities using a microplate reading at OD 590 nm (n=4; mean ± SEM; two-way ANOVA; *p < 0.05; **p < 0.01; ns: not significant).

To exclude the possibility that such cell-density-sensitive ferroptosis may result from different levels of available erastin and target (xCT) per cell under lower or higher cell density, the same number of TOV-21G cells were seeded in larger or smaller areas to recreate lower or higher cell densities. We found that the same number of cells plated at lower cell density was consistently more sensitive to ferroptosis (Figure 57). Therefore, cell confluency impacts ferroptosis sensitivity.

Figure 57 Cell density regulates the sensitivity of TOV-21G to ferroptosis

Crystal violet staining of a half-million TOV21G cells cultured in larger/smaller areas representing low/high densities treated with DMSO or 2µM erastin.

Since the two closely related Hippo pathway paralogues, YAP/TAZ, sense and mediate cell density-dependent responses (Mori et al., 2014; Zhang et al., 2018; Zhao et

al., 2007), we investigated the potential role of YAP/TAZ in regulating ferroptosis in

OVCAs. Among the two paralogues (YAP and TAZ), western blots revealed that TAZ is the predominantly expressed protein in both TOV-21G (clear cell type) and CAOV2 cells

(serous type) (Figure 58). In contrast, YAP is the predominant protein in the breast cancer cells MDA-MB-231, which serves as a control for YAP expression and detection

(Figure 58).

Figure 58 TAZ is highly expressed in OVCA cells

Western blot analysis of YAP and TAZ proteins in ovarian cancer (OVCA) cells (TOV- 21G, CAOV2, and CAOV2R) and one breast cancer cell line (MDA-MB-231). Representative data of at least three independent experiments.

We further performed cytosolic and nuclear fractionations of CAOV2 cells under low or high cell densities. Consistently, TAZ is the main protein whose nuclear levels were elevated when grown in low cell density (Figure 59). Therefore, we focused on the potential role of TAZ as the major effector of the Hippo pathway in regulating density- dependent ferroptosis in OVCAs.

Figure 59 TAZ localization and abundance are regulated by cell densities

Western blot measuring YAP and TAZ protein levels in the cytosolic (Cyto) /nucleic (Nuc) fractions of CAOV2 extracts when grown at low (L) or high (H) cell density. Histone H3 serves as a nuclear marker; GAPDH serves as a cytosolic marker.

3.3.3 TAZ regulates sensitivity to erastin-induced ferroptosis

To test the possibility that TAZ regulates susceptibility of OVCA cells to erastin, we first reduced TAZ expression (>85% knockdown) by using siRNAs in CAOV2 cells

(Figure 60). We found that knockdown of TAZ reduced erastin-induced cell death based on CelltiterGlo assays (P ≤ 0.001; Figure 60). In addition, knockdown of TAZ in CAOV2 by two more independent siRNAs also reduced sensitivity to erastin (Figure 61).

Figure 60 TAZ regulates sensitivity to erastin-induced ferroptosis in CAOV2 cells

(A) The relative cell viabilities of CAOV2 cells after silencing of control (siNT) or TAZ (siTAZ) for 2 days before they were treated with indicated dosages of erastin for 1 day. (n=3; mean ± SEM; two-way ANOVA; ***p < 0.001). (B-C) The knockdown efficiency of siRNAs targeting TAZ was confirmed by RT-qPCR and western blots in CAOV2 cells n=3; mean ± SEM; Student’s t-test; ***p < 0.001.

Figure 61 Individual siRNAs targeting TAZ in CAOV2 cells

(A) The relative cell viabilities of CAOV2 cells after silencing of control (siNT) or two individual TAZ-targeting siRNAs (siTAZ #1 and siTAZ #2) for 2 days before they were treated with indicated dosages of erastin for 1 day. (n=3; mean ± SEM; two-way ANOVA; ***p < 0.001) (B-C) The knockdown efficiency of siRNAs targeting TAZ was confirmed by

RT-qPCR and western blots in CAOV2 cells (n=3; mean ± SEM; one-way ANOVA; **p < 0.01; ***p < 0.001).

To determine if TAZ also regulates ferroptosis in TOV-21G cells, TAZ was silenced in TOV-21G cells by siRNAs which conferred ferroptosis resistance (Figure 62).

Collectively, these data strongly indicate that the TAZ levels regulate the sensitivity of

OVCA cells to erastin-induced ferroptosis.

Figure 62 TAZ regulates sensitivity to erastin-induced ferroptosis in TOV-21G cells

(A) The relative cell viabilities of TOV-21G cells after silencing of control (siNT) or TAZ (siTAZ) for 2 days before they were treated by indicated dosages of erastin treatment for 1 day. (n=3; mean ± SEM; two-way ANOVA; ***p < 0.001) (B-C) The knockdown efficiency of siRNAs targeting TAZ was confirmed by RT-qPCR and western blots in TOV-21G cells. n=3; mean ± SEM; Student’s t-test; *p < 0.05; ***p < 0.001.

3.3.4 Resistance to ferroptosis following treatment with carboplatin in vivo

We performed our experiments in CAOV2 because we have established paired

CAOV2 xenograft cells from a mouse model of ovarian cancer that mimics the course of primary and recurrent OVCA after chemotherapy (Huang et al., 2019). In short, the

CAOV2 cells were stably transduced with a construct containing GFP and luciferase

(pGreenFire plasmid constructs, SBI) to generate CAOV2-GFP/LUC cells, allowing us to monitor tumor growth using in vivo bioluminescence imaging. CAOV2-GFP/LUC cells

(3.5x10^5 per mouse) were injected intraperitoneally into female nude mice with tumor formation monitored weekly using the IVIS® in vivo imaging system. Once tumor formation was evident, we initiated treatment with carboplatin (60mg/kg intraperitoneally twice a week for two weeks) and monitored for changes in tumor volume based on the luciferase flux. In this model, tumor signal was reduced be below detection levels following carboplatin treatment, but then recurred approximately 2-4 weeks after carboplatin treatment had stopped. Therefore, we refer to the residual tumor cells that eventually emerged as CAOV2R (recurrent). We then compared cystine dependency and erastin sensitivity between the CAOV2 and CAOV2R cells and found that CAOV2R is more resistant to ferroptosis based on crystal violet staining, cytotoxicity assay and CelltiterGlo assay (Figure 63).

Figure 63 Regrowth of chemo-residual tumor cells are more resistant to ferroptosis

(A) Crystal violet staining of CAOV2 and CAOV2R cells when culture with full or cystine- deprived (Cys-) media. Representative data of at least three independent experiments. (B) Cytotoxicity was determined by CytoTox-Fluor assay after treatment of 10% (20 µM of cystine) of full media (n=3; two-way ANOVA; ***p < 0.001; ns: not significant). (C) Crystal violet staining of CAOV2 and CAOV2R cells when cultured with DMSO or 15µM erastin. Representative data of at least three independent experiments. (D) The relative cell viability of CAOV2 and CAOV2R cells when they were treated with the indicated dosage of erastin (n=3; mean ± SEM; two-way ANOVA; ***p < 0.001). 84

Since TAZ regulates ferroptosis sensitivity, we compared TAZ protein expression in CAOV2 and CAOV2R cells. We found that CAOV2R exhibits a lower level of TAZ protein (Figure 64).

Figure 64 Regrowth of chemo-residual tumor cells have less abundant TAZ proteins

Western blot analysis of TAZ protein levels in CAOV2 and CAOV2R cells. Representative data of at least three independent experiments.

Next, we overexpressed the constitutively active form of TAZ, TAZS89A (Lei et al., 2008), in the CAOV2R cells and found that increased TAZ sensitizes the CAOV2R cells to erastin (Figure 65). These findings suggest that comparisons of CAOV2 vs.

CAOV2R can be used to elucidate the molecular mechanisms of TAZ-regulated ferroptosis in OVCA.

Figure 65 Overexpression of TAZ sensitizes CAOV2R cells to ferroptosis

(A) Validation of TAZS89A overexpression in CAOV2R cells. Representative data of at least three independent experiments. (B) The relative cell viabilities of CAOV2R cells after the transfection of control vector or constitutively active TAZ, TAZS89A, following by treating with indicated dosages of erastin. Data are represented as mean SEM, n=3 biological replicates; two-way ANOVA; ***p < 0.001.

3.3.5 ANGPTL4 is a direct target gene of TAZ that regulates sensitivity to ferroptosis

TAZ is a transcriptional coactivator that affects cellular phenotypes through regulating the expression of target genes upon association with transcriptional factors such as TEAD. Assuming that knockdown of TAZ may repress the genes essential for ferroptosis, we determined the transcriptional response of CAOV2 to knockdown of

TAZ by RNA-seq (Figure 66A; GSE133673). Next, we integrated these TAZ-affected genes with the genes that were downregulated in CAOV2R when compared to CAOV2

(Figure 66B) to identify approximately 1100 candidate genes. 86

Figure 66 Integrated gene analysis

(A) RT-qPCR validates the knockdown efficiency of siTAZ for RNA-seq. n=3; mean ± SEM; Student’s t-test; **p < 0.01. (B) Venn diagram showing genes that are downregulated upon TAZ knockdown in CAOV2 cells (3370 +1179 genes) and downregulated genes (2540 +1179 genes) in CAOV2R cells compared to CAOV2 cells by RNA-seq. (C) Venn diagram showing genes that are downregulated upon siTAZ in CAOV2 cells by RNA-seq (blue color), downregulated genes in recurrent cells compared to CAOV2 cells by RNA-seq (red color), genome-wide siRNA screen of cystine deprivation in renal cell carcinoma, RCC4 (green color) and downregulated genes upon siTAZ in RCC4 cells by microarray (yellow color).

To identify the TAZ-regulated genes that are essential for ferroptosis across different cells, we further narrowed the candidate gene list by comparing to the gene lists identified from our renal cell studies through both RNAi screen and TAZ-affected genes (Yang et al., 2019b) (Figure 66C). Among the three most likely candidate genes,

ATP6V1B2 was not pursued because it was only one subunit of the multi-component protein complex. The knockdown of GPSM3 did not confer erastin resistance (Figure 67).

Figure 67 TAZ regulates ferroptosis, not through GPSM3

(A) RT-qPCR validates the knockdown efficiency of siGPSM3 for RNA-seq. n=5; mean ± SEM; Student’s t-test; ***p < 0.001. (B) The relative cell viabilities of CAOV2 cells after knockdown of GPSM3 following by treating with indicated dosages of erastin. Data are represented as mean±SEM, n=4 biological replicates.

Therefore, these integrative analyses reveal that Angiopoietin-Like 4 (ANGPTL4) emerged as the most promising candidate gene. First, ANGPTL4 mRNA was downregulated upon TAZ knockdown, which was confirmed by RT-qPCR (Figure 68A).

Second, there is a significant correlation between the expression of ANGPTL4 and TAZ

(encoded by WWTR1) in the TCGA ovarian tumor dataset. The expression of WWTR1 is also consistently correlated with CTGF (Zhao et al., 2008) and CYR61 (Choi et al., 2015), two canonical YAP/TAZ target genes (Figure 68B-D). Third, the knockdown of

ANGPTL4 in both serous subtype (CAOV2) and clear cell subtype (TOV-21G) of OVCA

cells protects cells from erastin-induced death (Figure 69 and Figure 70). This protective effect was further validated by using two more independent siRNAs targeting

ANGPTL4 in CAOV2 cells (Figure 71). Consistently, overexpression of V5-tag ANGPTL4 sensitizes TOV-21G cells to ferroptosis (Figure 72). Collectively, these data indicate that

ANGPTL4 is a prominent regulator of TAZ-mediated ferroptosis sensitivity in OVCA.

Figure 68 The correlation between WWTR1 (TAZ) and ANGPTL4

(A) Validation of downregulated ANGPTL4 mRNA level by RT-qPCR upon TAZ knockdown. (n=3; mean ± SEM; Student’s t-test; *p < 0.05) (B-D) The correlation between

the expression of indicated genes in the TCGA ovarian cancer datasets. Data is retrieved from GEPIA (http://gepia.cancer-pku.cn/) with Pearson correlation analysis. TPM: Transcripts Per Kilobase Million.

Figure 69 ANGPTL4 regulates ferroptosis sensitivity in CAOV2 cells

(A) The relative cell viabilities of CAOV2 cells after knockdown of ANGPTL4 following by treating with indicated dosages of erastin. Data are represented as mean SEM, n=3 biological replicates. (B) Check for knockdown efficiency of siANGPTL4 by western blots in CAOV2 cells.

Figure 70 ANGPTL4 regulates ferroptosis sensitivity in TOV-21G cells

(A) The relative cell viabilities of TOV-21G cells after knockdown of ANGPTL4 following by treating with indicated dosages of erastin. Data are represented as mean SEM, n=3 biological replicates. (B) Check for knockdown efficiency of siANGPTL4 by western blots in TOV-21G cells.

Figure 71 Validation of ANGPTL4-regulated ferroptosis sensitivity in CAOV2 cells

(A) The relative cell viabilities of CAOV2 cells after silencing of ANGPTL4 by two individual siRNAs for 2 days before treated with indicated dosages of erastin for 1 day. Data are represented as mean SEM, n=3 biological replicates. (B) Check for knockdown efficiency of siANGPTL4 by western blots in CAOV2 cells.

Figure 72 Overexpression of ANGPTL4 sensitizes ferroptosis sensitivity

(A) The relative cell viabilities of TOV-21G cells after overexpression of V5-tagged ANGPTL4 and treatment of indicated dosages of erastin. Data are represented as mean SEM, n=3 biological replicates. (B) Validation of ANGPTL4-V5 overexpression in TOV- 21G cells

We analyzed previous ChIP-seq studies and found the regulatory regions of

ANGPTL4 were physically associated with YAP/TAZ/TEAD complexes (Stein et al.,

2015; Walko et al., 2017; Zanconato et al., 2015). To further validate that ANGPTL4 is a direct target gene of TAZ, we performed ChIP-qPCR using an antibody specific for endogenous TAZ protein. As shown in Figure 73, we found that the promoter region of

ANGPTL4, similar to the CTGF (positive control), was enriched in the TAZ pull-down, indicating that the ANGPTL4 promoter is directly associated with TAZ. Thus, ANGPTL4 is a direct downstream target gene of TAZ that may contribute to TAZ-regulated ferroptosis.

Figure 73 ANGPTL4 is a direct target gene of TAZ

The relative levels of TAZ-associated genomic DNA in the indicated promoters associated with endogenous TAZ protein using ChIP-qPCR with TAZ antibody in CAOV2 protein lysate. CTGF promoter serves as a positive control for a TAZ target gene, while Ch10 serves as a negative control.

3.3.6 Differential ANGPTL4 expressions underlie the ferroptosis sensitivities of CAOV2 pair cells

We next determined whether the expression of ANGPTL4 could explain the different ferroptosis sensitivity between CAOV2 and CAOV2R cells. From the results of

RT-qPCR and western blot, we found CAOV2R has less ANGPTL4 expression at both mRNA and protein levels (Figure 74). Since ANGPTL4 encodes a secreted protein, we also examined the level of extracellular ANGPTL4 in the culture media by enzyme- linked immunosorbent assay (ELISA). Consistently, extracellular ANGPTL4 proteins are less abundant in the CAOV2R cells as compared to the CAOV2 cells (Figure 74). To test the possibility that relative resistance of CAOV2R to ferroptosis may be due to lower expression of ANGPTL4, we determined whether the ferroptosis sensitivity of CAOV2R can be affected by the addition of recombinant ANGPTL4 proteins in the media. As shown in Figure 75, soluble ANGPTL4 sensitizes CAOV2R cells to ferroptosis.

Figure 74 Differential ANGPTL4 expressions among CAOV2 and CAOV2R cells

(A) The relative levels of ANGPTL4 mRNA were measured by RT-qPCR in CAOV2 and CAOV2R cells. (n=3; mean ± SEM; Student’s t-test; *p < 0.05) (B) Western blot of ANGPTL4 and β-tubulin proteins in CAOV2 and CAOV2R cells. Representative data of at least three independent experiments. (C) ELISA of extracellular ANGPTL4 proteins in the culture

media of CAOV2 and CAOV2R cells incubated for 2 days (n=4, mean ± SEM; Student’s t- test; *p < 0.05).

Figure 75 Supplying ANGPTL4 protein sensitizes CAOV2R cells to erastin

Relative cell viability assay of CAOV2 and CAOV2R cells when treated with recombinant ANGPTL4 protein in media followed by 8 µM erastin and then normalized to DMSO control group (n=3, mean ± SEM; two-way ANOVA; *p < 0.05; **p < 0.01).

3.3.7 ANGPTL4 regulates ferroptosis through NOX2

With regard to the mechanistic link between ANGPTL4 and ferroptosis, it is interesting to note that ANGPTL4 has been reported to activate the NADPH oxidase

NOX1 in the keratinocyte carcinoma cells (Terada and Nwariaku, 2011; Zhu et al., 2011).

The NADPH oxidases are recognized to generate reactive oxygen species (ROS), promote lipid peroxidation and ferroptosis (Dixon et al., 2012). Therefore, we investigated the expression levels of all seven members of the NOX protein family 95

(NOX1-5 and DUOX1-2) in OVCA. Among different NOXs, we found that NOX2 is the most abundantly expressed member from the analysis of RNA-seq data from the TCGA dataset (Table 1). We also noticed NOX2 protein was lower in CAOV2R when compared to CAOV2 cells (Figure 76).

Table 1 TCGA gene analysis of NOX family gene expressions in OVCA

Data were retrieved from The Human Protein Atlas (https://www.proteinatlas.org/). FPKM: Fragments Per Kilobase Million.

Figure 76 lower NOX2 protein level in CAOV2R cells

Western blot of NOX2 and β-tubulin proteins in CAOV2 and CAOV2R cells.

Therefore, we focused on the NOX2 protein. First, we found a pan NOX inhibitor, VAS2870, protected CAOV2 cells from ferroptosis (Figure 77A). Furthermore, another NOX inhibitor, GKT136901, also protects cells from ferroptosis (Figure 77B).

Moreover, treatment using a NOX2 specific inhibitor, gp91dstat peptide, enhances ferroptosis resistance in the CAOV2 cells (Figure 78). In addition, the silencing of NOX2 by siRNA also conferred ferroptosis resistance in both CAOV2 (Figure 79) and TOV-21G cells (Figure 80).

Figure 77 NOX inhibition decreases ferroptosis sensitivity in CAOV2 cells

(A) The relative cell viability assays of erastin-treated CAOV2 cells when combined with NOX inhibition by (A) 20 µM pan-NOX inhibitor VAS2870; (B) 20 µM NOX inhibitor, GKT136901; n=3, mean ± SEM; two-way ANOVA; ***p < 0.001.

Figure 78 NOX2 inhibition decreases ferroptosis sensitivity in CAOV2 cells

The relative cell viability assays of erastin-treated CAOV2 cells when combined with specific NOX2 inhibitor, gp91dstat (33 µg/ml). n=3, mean ± SEM; two-way ANOVA; ***p < 0.001.

Figure 79 Knockdown of NOX2 decreases ferroptosis sensitivity in CAOV2 cells

(A) The relative cell viability assays of erastin-treated CAOV2 cells when combined with NOX2 knockdown by pooled siRNAs (siNOX2). n=3, mean ± SEM; two-way ANOVA; ***p < 0.001. (B-C) Knockdown efficiency check of siNOX2 by RT-qPCR and western blot in CAOV2 cells.

Figure 80 Knockdown of NOX2 decreases ferroptosis sensitivity in TOV-21G cells

(A) The relative cell viability assays of erastin-treated TOV-21G cells when combined with NOX2 knockdown by pooled siRNAs (siNOX2) and two independent NOX2-targeting siRNAs (siNOX2 #1 and #2). n=3, mean ± SEM; two-way ANOVA; ***p < 0.001. (B-C) Knockdown efficiency check of siNOX2 by RT-qPCR and western blot in TOV-21G cells.

To examine the mechanism of TAZ-ANGPTL4-NOX2, we determined the erastin-induced death of CAOV2 and CAOV2R cells with the NOX2 inhibitor, gp91dstat.

We found that inhibiting NOX2 rendered CAOV2 cells more resistant to ferroptosis, but not CAOV2R cells (Figure 81). Meanwhile, NOX inhibitor GKT136901 also abolished the ferroptosis-sensitizing effects of the over-expression of TAZS89A in CAOV2R cells

(Figure 82). Therefore, the NOX2 activities are essential for the ferroptosis regulation by

TAZ.

Figure 81 Interaction between chemo-residual tumor cells and NOX2 inhibition

The relative cell viabilities of CAOV2 or CAOV2R cells were determined by CelltiterGlo after 24 hours of 8 µM erastin with or without NOX2 inhibitor, gp91dstat (33 µg/ml). Data are represented as mean±SEM, n=3 after normalized to the DMSO controls (two-way ANOVA; *p < 0.05; **p < 0.01; ***p < 0.001; ns: not significant).

Figure 82 Interaction between TAZ and NOX

The relative cell viabilities of CAOV2R expressing control vector or TAZS89A were determined by CelltiterGlo after 24 hours of 10 µM erastin with or without 20 µM NOX

100

inhibitor, GKT136901. Data are represented as mean±SEM, n=6 after normalized to the DMSO controls (two-way ANOVA; ***p < 0.001). Finally, we measured the lipid-based reactive oxygen species (lipid ROS), the hallmark of ferroptosis, by C11-BODIPY staining and validated that knockdown of TAZ,

ANGPTL4, or NOX2 reduced the erastin-induced lipid peroxidation (Figure 83). Taken together, we propose a signaling mechanism (Figure 84) by which TAZ is a cell-density- dependent determinant of ferroptosis sensitivity in OVCA through regulating levels of

ANGPTL4, which in turn regulates NOX2 activity and ferroptotic death. The TAZ-

ANGPTL4-COX2 was significantly lower in the recurrent OVCA which render these cells more resistant to ferroptosis.

101

Figure 83 TAZ, ANGPTL4, or NOX2 knockdown abolishes elevated lipid-ROS by erastin treatment

Figure 84 Schematic representing the model of TAZ-regulated ferroptosis through ANGPTL4-NOX2

102

The proposed molecular mechanism on how TAZ-regulated ferroptosis sensitivity through ANGPTL4-NOX2 axis in ovarian cancer cells.

3.4 Discussion

Here, we employed nutrigenetic screens (Tang et al., 2017b; Tang et al., 2016) and identified cystine addiction and ferroptosis susceptibility of OVCA cells, implying that ferroptosis-inducing agents may hold therapeutic potential for OVCA. However, little is known about the process and genetic determinants of ferroptosis in OVCA other than being briefly mentioned in two recent papers (Basuli et al., 2017; Sato et al., 2018) without mechanistic investigation. By studying how cell density regulates ferroptosis sensitivity, as described for other cancer types in preprints (Panzilius et al., 2018; Yang et al., 2019b), we have now elucidated how TAZ affects ferroptosis through the regulation of ANGPTL4 and NOX2 in OVCA. This study enhances our understanding of the role of

TAZ and how it regulates ferroptosis in OVCA.

YAP/TAZ and other components of the Hippo signaling pathway are important in the oncogenesis, progression and migration of OVCA cells (Hall et al., 2010; Zhang et al., 2011). In our experiments, TAZ, rather than YAP, is the dominant effector in the tested OVCA cells. While the role of YAP was the first recognized, many studies have also supported the role of TAZ in OVCA. For example, increased expression of TAZ mRNA is correlated with poor prognosis and TAZ manipulation affects migration, proliferation, treatment response and EMT of OVCA (Chen et al., 2016; Jeong et al., 2013; 103

Lopez-Guerrero et al., 2015; Tan et al., 2015). Therefore, our findings suggest that inducing ferroptosis will be an effective strategy for eradicating TAZ-activated tumors which are particularly aggressive and resistant to current standard treatments. A recent study has found that YAP promotes ferroptosis by upregulating several ferroptosis modulators, including ACSL4 and TFRC (Wu et al., 2019). In the future, it will be interesting to test whether other regulators of the Hippo pathway also play a role in regulating ferroptosis.

ANGPTL4 is a member of the angiopoietin family and the members of which act as regulators of lipid and glucose metabolism (Hato et al., 2008; Santulli, 2014).

ANGPTL4 also plays a role in tumor biology; ANGPTL4 is upregulated in several human cancers associated with metastasis and poor outcome (Baba et al., 2017; Liao et al., 2016). In addition, ANGPTL4 is induced by various oncogenic pathways including

ERK (Zhu et al., 2016), TGFβ (Padua et al., 2008), hypoxia (Murata et al., 2009) and

EGF(Liao et al., 2016) to promote angiogenesis, invasion, and metastasis (Tanaka et al., 2015). Especially relevant is a study that ANGPTL4 stimulates oncogenic ROS and anoikis resistance through the activation of NADPH oxidases (NOX) (Terada and

Nwariaku, 2011). Our current study reveals that ANGPTL4 is a direct transcriptional target of TAZ, consistent with the repression of ANGPTL4 by the TAZ/YAP inhibitor, verteporfin (Sun and Ying, 2015). Based on the previous studies, ANGPTL4 is

104

expected to increase ROS and predispose cells to signal that induces ferroptosis.

Therefore, our findings implicate high ANGPTL4 levels and anoikis resistance as novel and highly relevant oncogenic properties that can be targeted by triggering ferroptosis. Since anoikis resistance is essential for tumor metastasis (Kim et al.,

2012b; Simpson et al., 2008), our results also imply that ferroptosis may target tumor cells at different stages of tumor progression. Anoikis resistance is one of the phenotypic changes that occur during EMT (Huang et al., 2013; Smit et al., 2009). Our findings are also in agreement with other studies showing increased ferroptosis sensitivity during EMT that is in part due to the regulation of GPX4 by ZEB1

(Viswanathan et al., 2017).

We have found that the carboplatin-treated CAOV2R cells are less sensitive to ferroptosis and have a lower level of TAZ. These results seemingly contradict previous reports on the ferroptosis sensitivity of persister cells (Viswanathan et al.,

2017) due to increased GPX4 expression. However, our results are consistent with the supplemental data in (Tanaka et al., 2015) that show cells re-growing at two-months, similar to the timeframe of carboplatin-treated CAOV2R cells in our studies, are much less responsive to GPX4 inhibitor RSL3-induced cell death. Of course, the discrepancy between results may also be due to the different cell lines used, different cancer drugs or different ferroptosis inducers in these two studies.

105

Our results may have significant therapeutic implications for patients with

OVCA. While inducing ferroptosis may have substantial anti-tumor potential, it is not clear which tumors would respond to ferroptosis-inducing agents. Our results indicate that TAZ-activated tumors may be particularly sensitive to various ferroptosis-inducing therapies. Thus, TAZ activation and the high expression of its canonical target genes may serve as predictive biomarkers for response to ferroptosis-inducing therapeutics.

This is similar to the use of YAP as a marker of sensitivity in OVCA to MSC2363318A, a dual AKT and P70S6K inhibitor (Previs et al., 2017). In addition, it is important to consider the compounds or therapeutic agents that would be most effective for inducing ferroptosis. While erastin and various GPX4 inhibitors can induce ferroptosis, their toxicity and stability may limit in vivo application and translation potential. One promising agent is the recombinant human cyst(e)inase that can trigger ferroptosis by depleting plasma cystine (Cramer et al., 2017). Since its anti-tumor efficacy and in vivo safety have been demonstrated in murine models for multiple tumors, it may have the potentials to trigger in vivo ferroptosis of OVCA and improve clinical outcomes.

106

4. Angiogenin-mediated tRNA cleavage as a novel feature of stored red blood cells

Portions of this Chapter are reproduced from the British Journal of Haematology:

British Journal of Haematology, 2019, 185, 752–806

4.1 Introduction

Red blood cell (RBC) transfusion is one of the most common procedures in hospitalized patients. However, the transfusion of RBCs stored for extended durations correlates with a higher risk for death and complications. Such poor clinical outcomes in patients receiving these older stored blood units may due to “storage lesion”, which refers to the collectively adverse changes that occur in RBCs during storage. Thus, it is critical to elucidate the molecular mechanisms of storage lesion in stored RBCs to prevent harm to the recipients who receive older RBC units. Another application of stored RBCs is to increase oxygen-carrying capacity and enhance athletic performance; this is a form of blood doping called autologous blood transfusion (ABT) banned by the world anti-doping agency. As such, the identification of biomarkers of stored RBCs will prove important for the development of fast and reliable methods for detecting ABT in athletes. Importantly, our lab and others have found that RBCs contain abundant and diverse species of miRNAs that are stable and have functional roles, which could be ideal biomarkers. In order to identify strategies to improve storage conditions and

107

identify biomarkers in ABT, we profiled the global miRNA changes during six weeks of in vitro storage using the Nanostring platform. While the abundance of almost all miRNAs did not change significantly, there was a dramatic 10-fold increase in miR-720, which is a 3’ fragment from tRNAThr, after two weeks of storage, which remained steady for an additional four weeks. However, it is unclear what causes elevated RBC miR-720 during in vitro storage. We hypothesize that the increase of RBC miR-720 during storage is due to increased activity of one or more nucleases that cleaves tRNAThr. To identify the nucleases responsible for the increase in miR-720 during storage, we established tRNAThr cleavage assays and identify the nuclease responsible for miR-720 induction during storage. By completing these independent and complementary experiments, we have a better understanding of the mechanisms of storage lesion from the angle of small RNAs.

4.1.1 Stored red blood cells

Blood transfusion is a routine medical procedure with more than 85 million units of blood collected worldwide annually (García-Roa et al., 2017). In the United States, around 14 million units of stored RBCs are administered to patients each year (Whitaker et al., 2015). Currently, RBC concentrates in blood banks with contemporary preservative solutions that can be stored between 1°C to 6°C for as long as 42 days (six weeks) before transfusion. These storage conditions should enable <1% hemolysis

108

during storage and ≥75% in vivo cell survival at 24 post-transfusion (D’Alessandro et al.,

2010).

Many studies indicate that RBC quality decreases proportionally as storage time increases (Berezina et al., 2002; Kim‐Shapiro et al., 2011; Tinmouth et al., 2006). During storage, RBCs undergo a series of biochemical and biomechanical changes that reduce

RBC survival and function. These changes are collectively referred to as “storage lesion”(Wolfe, 1985) and seem to be associated with poor clinical outcomes. These storage-associated changes include morphological alteration from a discoid to spherocytic phenotype and biochemical alternations such as decreased concentrations of adenosine triphosphate (ATP), 2,3-diphsophoglycerate (2,3-DPG), and glutathione

(GSH). Other characteristics of storage lesion are decreased oxygen delivery capacity, acidosis, altered cation homeostasis (loss of intracellular potassium and accumulation of sodium within the cytoplasm), phosphatidylserine exposure to the outer membrane leaflet, and oxidative damage resulting in impaired membrane health (Bennett-Guerrero et al., 2007; Berezina et al., 2002; D'Alessandro et al., 2015; Kim‐Shapiro et al., 2011).

Therefore, transfusion of RBCs after prolonged storage may correlate with a higher risk of death and more clinical complications for recipients. Other studies have found that transfusion of RBCs after prolonged storage has negative effects, including postoperative infections, renal failure, clotting disorders, and increased morbidity and

109

mortality, for recipients, especially traumatized, post-operative, and critically ill patients

(Koch et al., 2008; Marik and Corwin, 2008; Offner et al., 2002; Tinmouth et al., 2006;

Wang et al., 2012; Weinberg et al., 2008). However, statistical limitations such as lack of randomized double-blind clinical prospective trials and standard methods between laboratories have made the analysis of this difficult (Hess, 2011). Therefore, understanding the molecular mechanisms during storage lesion is important to appreciate any contributory role in these pathologies with transfusion. In addition, it is becoming clear that current standards for routine storage of RBCs with a six-week expiration date may not be adequate for all stored RBCs. RBCs may deteriorate more quickly in some units and thus present risks before 42 days, while other RBCs may be stored for even longer periods that can significantly expand the supplies of stored RBC

(D’Alessandro et al., 2010; Dern et al., 1966; Van 't Erve et al., 2015; Weisenhorn et al.,

2016).

In addition, stored RBCs generate a large number of extracellular vesicles (EVs) through direct budding from the cell membrane (Rubin et al., 2008). For example, at the end of the storage period, the concentration of microvesicles (MVs), a type of EVs, increases more than 20-fold (~63,000 MVs/μL) compared to initial value (~3,000 MVs/μL)

(Gao et al., 2013; Rubin et al., 2008). The MVs range from 50 nm to 1000 nm and contain lipids, proteins (such as band 3 protein, glycophorins, and actin), and RNA species

110

(Greenwalt et al., 1984; Mantel et al., 2013). Blood MVs have been reported to contain miRNAs (Hunter et al., 2008). In addition, miRNAs in EVs can functionally regulate target genes via fusion with recipient cells (Mittelbrunn et al., 2011; Montecalvo et al.,

2012; Setzer et al., 2006). This regulation is achieved by the accumulation of miRNA- coupled RNA-induced silencing complex (RISC) in these EVs (Gibbings et al., 2009), and the enrichment of proteins AGO2 and GW182, which support canonical microRNA- mRNA interactions. In sum, understanding the molecular mechanisms of stored RBCs that contribute to pathologies in recipients is important especially giving the use of blood transfusions in many settings.

4.1.2 Use of stored red blood cells in athletic doping

Some athletes utilize stored RBCs for autologous blood transfusion (ABT) to enhance the sport performance, which is banned by the world anti-doping agency

(WADA). Athletes may use substances such as growth hormones or recombinant human erythropoietin (rHuEPO), as well as allogeneic or autologous blood transfusions to rapidly enhance sports performance; all these procedures are banned by WADA (Thevis et al., 2014). With advances in chromatographic-mass spectrometry and immunological methodologies to detect performance-enhancing agents, some athletes favor blood doping, which increases RBC count and thus oxygen-carrying capacity. While allogeneic blood transfusion can be detected by identifying antigens of distinct RBC populations, a

111

direct and reliable method is not yet available to detect ABTs (Salamin et al., 2016).

Current approaches to indirectly detect ABT include monitoring hematological parameters that are recorded in the Athlete Biological Passport as well as detecting biologic responses to blood transfusion. Responses that indicate ABT abuse include decreased concentration of erythropoietin, increased concentration of hepcidin

(Leuenberger et al., 2016) (an iron regulatory peptide secreted by the liver to target membrane-bound ferroportin for degradation and decreased iron export), increased inflammation markers, and increased circulating miRNAs from pulmonary and liver tissues (Leuenberger et al., 2013). However, these parameters either have narrow detection windows or can be triggered by exposure to altitude or infection (Salamin et al., 2016). Thus, the development of fast, reliable, and direct techniques for monitoring

ABT use is needed.

4.1.3 Functionally relevant RNA in RBCs

One unique approach to understand storage lesion and detect ABT is by genomic analyses of the RBC transcriptome. RBCs are terminally differentiated cells with the main function of oxygen-transport. Mature RBCs do not have nuclei; during terminal differentiation of RBCs, extrusion of the nuclei leads to anucleate cells. Therefore, circulating RBCs were once thought to lack RNA expression, which was supported by the barely detectable signals using RNA-binding dyes such as methylene blue and

112

thiazole orange (Lee et al., 1986). However, our group and others found that human erythrocytes contain diverse and abundant RNA species, including non-coding RNAs such as Y1, Y4, miRNAs; tRNAs; and mRNAs (Chen et al., 2008; Doss et al., 2015;

Hamilton, 2010; O'Brien and Harley, 1990; Rathjen et al., 2006; Sangokoya et al., 2010a).

These erythrocytic miRNAs play a role in anemia (Sangokoya et al., 2010b) and malaria resistance (LaMonte et al., 2012) in sickle cell disease. In addition, distinct changes of miRNA profiles have been found in RBCs during in vitro blood storage (Kannan and

Atreya, 2010).

miRNAs are a class of single-stranded non-coding RNA with 19 to 25 nucleotides derived from transcripts and processed by the RNase III endonucleases Drosha and

Dicer. Mature miRNA is loaded onto the RNA-induced silencing complex (RISC), thus promoting target mRNA cleavage and decay or translation repression (Friedman et al.,

2009; Sarachana et al., 2015). Approximately 30% of human protein-coding genes are regulated by miRNAs (Lewis et al., 2005). Moreover, miRNAs regulate host-pathogen interactions in mammals (Berkhout and Haasnoot, 2006; LaMonte et al., 2012).

Dysregulated miRNA levels are found not only in cancer, but also in metabolic, cardiovascular, and neurodegenerative disorders (Furer et al., 2010; Im and Kenny,

2012). Since the first discovery of miRNA, lin-4 in C. elegans (Lee et al., 1993), 2588

113

miRNAs from 1881 miRNA precursors (pre-miRNAs) have been identified in the human genome (Kozomara and Griffiths-Jones, 2014).

miR-720 was first annotated as a miRNA in the miRNA database (miRBase).

However, in 2010, Schopman et al. indicated that miR-720 is a fragment of threonine tRNA (Schopman et al., 2010), and thus miR-720 was subsequently removed from miRBase. In physiological conditions, miR-720 has been associated with epithelial development in keratinocytes (Chikh et al., 2011) as well as dental pulp cell differentiation (Hara et al., 2013). Upregulated miR-720 has been reported to correlate with tumor progression in a variety of tumors such as breast, cervical, colorectal, and bladder cancers, as well as renal cell carcinoma, myeloma, and melanoma (Das et al.,

2016; Jones et al., 2012; Nonaka et al., 2015; Park et al., 2014; Ragusa et al., 2012; Sand et al., 2013; Tang et al., 2015b; Wang et al., 2015a; Yi et al., 2010). In addition, miR-720 has been observed in EVs from breast cancer cells (Guzman et al., 2015). Furthermore, upregulated expression of miR-720 is found in chronic hepatitis B virus (HBV)-specific

CD8+ T cells (Wang et al., 2015a). These findings suggest that miR-720 has important functional roles and can be a therapeutic target.

4.1.4 Transfer RNA-related fragments (tRFs)

miR-720 is postulated to be a fragment of threonine tRNA (Schopman et al.,

2010). tRNAs themselves are cleavage products from precursor tRNAs (pre-tRNA). Pre-

114

tRNAs are transcribed by RNA polymerase III and contain extra sequences at their 5’ and 3’ termini called leader and trailer sequences, respectively. During tRNA maturation, the leader sequence is removed by RNase P and the trailer sequence is removed by RNase Z (Frank and Pace, 1998; Maraia and Lamichhane, 2011).

Subsequently, the non-templated “CCA” sequence is added to the 3’ ends of the trailer- free tRNAs by CCA-adding enzyme, a tRNA nucleotidyltransferase (Martin and Keller,

2007) to form the mature tRNAs. Mature tRNAs serve as an adaptor for protein synthesis and contain 70 to 90 nucleotides that form cloverleaf secondary structures and

L-shaped tertiary structures. Despite their well-known function for translation, increasing evidence shows that tRNAs can be further processes and serve as the source of tRNA related-/derived- fragments (tRFs) with functional roles. With the advances of recent sequencing technology, many studies have identified tRFs as the second most abundant small RNAs after miRNAs (Cole et al., 2009). The fragments are found in several studies by high-throughput sequencing analysis (Cole et al., 2009; Lee et al.,

2009; Li et al., 2012) and documented in a tRF database (Kumar et al., 2015). These fragments are broadly classified into two main types based on the mapped positions on the primary or mature tRNA: tRNA halves or smaller tRNA-tRFs. However, the nomenclature system is not yet finalized. tRNA halves are 31 to 40 bases in length that is generated by cleavage in the anticodon loops of mature tRNAs. tRNA halves are also

115

called stress-induced tRNA fragments or tRNA-derived stress-induced RNAs (tiRNAs), even though they have also been identified under non-stress conditions (Nowacka et al.,

2013). On the other hand, tRFs are much smaller fragments that can be divided into three species: tRF-5, tRF-3, and tRF-1. The 5’ tRFs and 3’ tRFs are generated from the 5’ and 3’ ends of mature tRNAs coupling with CCA post-transcriptional modification after cleavages in the D loop or the TψC loop, respectively. tRF-1 fragments correspond to precursor tRNA 3’ trailers.

Stress-induced cleavage of tRNA was first discovered in Tetrahymena thermophila with cleavage near the anticodon loop as a starvation response (Lee and Collins, 2005).

Later, this phenomenon was described in bacteria (Haiser et al., 2008), fungi (Jochl et al.,

2008; Thompson and Parker, 2009a), plants (Thompson et al., 2008), and mammalian cells (Kawaji et al., 2008; Thompson et al., 2008) in response to a range of stress stimuli, such as oxidative stress, UV irradiation, and heat/cold shock. These stress-induced 5’- tRNA halves promote stress granule formation (Emara et al., 2010) and mediate translational inhibition by associating with Y-box-binding protein (YB-1), which displaces eIF4G from RNA as well as eIF4E/G/A from the m7G cap (Evdokimova et al.,

2006). In addition, a new type of tRNA halves generated by angiogenin (ANG), called

Sex HOrmone-dependent TRNA-derived RNAs (SHOT-RNAs) (Honda et al., 2015), exists in sex hormone-dependent breast and prostate cancer cell lines; 5’-SHOT-RNAs

116

contribute to tumor cell proliferation. A recent study showed that tRNA fragments in sperm regulate transcript expression, and the levels of these tRNA fragments are influenced by the paternal diet (Sharma et al., 2016).

These tRNA-related small RNAs are not simply non-specific degradation products of tRNAs, rather, they are likely generated from tRNA processing for specific functional roles (Kawaji et al., 2008; Pederson, 2010). This view is supported by some functional roles of these tRNA fragments (Shigematsu et al., 2014). Also, the fragments are not random; the predominant reads of these fragments correspond to the 5’ or 3’ ends, while far fewer reads correspond to the anticodon stem-loop regions of tRNAs. In addition, the abundance of tRNA fragments does not correlate with the prevalence of their parent tRNAs, and their abundance varies between cell lines suggesting specific expression of tRNA-related small RNAs. Lastly, tRNA fragments also have precisely defined ends suggesting specific cleavage and further processing.

4.1.5 Candidate nucleases that generate tRNA fragments

Dicer

Human Dicer is a 220-kDa multi-domain protein with an N-terminal helicase domain, a PAZ (Piwi-Argonaute-Zwille) domain, a double-stranded RNA-binding domain, and two catalytic RNase III domains (Lau et al., 2012; Macrae et al., 2006). Only one Dicer is found in vertebrates and nematodes, while insects have Dicer-1 and Dicer-2

117

and plants have four Dicer-like (DCL) genes (Mukherjee et al., 2013). Dicer cleaves pre- miRNAs, which are hairpin structures with 50-70 nucleotides, or double-stranded RNAs into miRNAs and siRNAs, respectively. Dicer plays a critical role in the production of small regulatory RNAs, inflammation, chromatin structure remodeling, and apoptotic

DNA degradation. In addition, Dicer is reportedly responsible for tRF biogenesis (Cole et al., 2009; Haussecker et al., 2010; Maute et al., 2013). For example, CU1276, a 22- nucleotide tRF from the 5’ end of tRNAGly, is Dicer-dependent and associates with Ago proteins to repress endogenous RPA1 gene by targeting 3’ untranslated regions (3’

UTRs) (Maute et al., 2013). Taken together, Dicer represents a potential nuclease target for the generation of miR-720 in erythrocytes.

Angiogenin

ANG is a secreted basic ribonuclease that was originally purified from tumor cell-conditioned medium as an angiogenic factor (Fett et al., 1985; Shapiro et al., 1987).

ANG is a 14-kDa protein with 123 amino acids that belongs to the RNase A superfamily.

Despite sharing sequence homology (35% identity) to RNase A (Strydom et al., 1985),

ANG shows different enzymatic activities toward RNA transcripts; ANG shows no significant activity to RNase A substrates such as yeast RNA, poly C, and poly U, but it displays ribonucleolytic activity toward some rRNAs and tRNAs (Shapiro et al., 1986).

In addition, ANG inhibits protein translation, with tRNA degradation found in Xenopus

118

oocytes and rabbit reticulocyte lysates (Saxena et al., 1992). ANG mRNA and/or protein levels are elevated in cancers (Etoh et al., 2000; Li et al., 1994), and correlate with tumor progression and metastasis, as well as poor prognosis (Shimoyama et al., 1996). ANG has involved the development of a variety of human tumor types in athymic mice

(Olson et al., 1995). Furthermore, ANG is up-regulated by hypoxia (Hartmann et al.,

1999) and inflammatory stimuli (Olson et al., 1998). In summary, angiogenin is a stress- activated ribonuclease that selectively cleaves tRNA and inhibits protein synthesis (Fu et al., 2009; Ivanov et al., 2011; Yamasaki et al., 2009).

ANG binds to ribonuclease/angiogenin inhibitor 1 (RNH1) with an extraordinary affinity (Ki <1 fM). RNH1, also called placental ribonuclease inhibitor (PRI), is a leucine- rich repeat protein of 50 kDa. The localization and activity of ANG are controlled by

RNH1 (Pizzo et al., 2013). That is, RNH1 associates with cytosolic ANG, but not nuclear

ANG during normal growth conditions, and while in stress conditions, RNH1 associates with nuclear, but not cytosolic, ANG. ANG promotes angiogenesis in the nucleus; while in the cytoplasm ANG cleaves tRNA at the anticodon loop and produces tRFs when not bound to RNH1. However, RNH1 is sensitive to oxidation and has broad specificity for the RNase A superfamily; thus, it is less likely to be developed as anti-tumor therapy

(Blackburn et al., 1977; Lee and Vallee, 1993). Meanwhile, monoclonal antibodies and antisense oligonucleotides to ANG effectively inhibit colon, prostate, breast, lung, and

119

fibroblast tumors (Olson et al., 2001; Olson et al., 1995; Piccoli et al., 1998). Since ANG cleaves the anticodon loops of mature tRNAs, RNH1 has been reported to mitigate ANG cleavage (Yamasaki et al., 2009). In all, the nuclease activity toward the tRNA of ANG enables it as a candidate for the regulation of miR-720.

RNase T2

In S. cerevisiae, the RNase T2 orthologue Rny1 cleaves tRNA transcripts during oxidative stress (Thompson and Parker, 2009a). RNase T2 is the only member of the

Rh/T2/S family ribonucleases in humans (Thorn et al., 2012). Moreover, RNase T2 is a secreted glycoprotein that can suppress tumor and metastasis (Acquati et al., 2005;

Smirnoff et al., 2006). T2 family ribonucleases are transferase-type RNases as RNase T1 and RNase A family proteins (Deshpande and Shankar, 2002). Nevertheless, they are distinguished from the RNase A and RNase T1 protein families based on three characteristics. First, T2 ribonucleases are ubiquitously distributed in organisms across kingdoms, including bacteria, plants, protozoans, animals, and even viruses, while

RNase T1 ribonucleases only exist in bacteria and fungi and RNase A ribonucleases primarily occur in animals. Next, the optimal pH for the enzyme activity of RNase T2 proteins is between 4 and 5. By contrast, RNase T1 and RNase A family proteins prefer alkaline (pH 7 to 8) or weakly acidic pH (pH 6.5 to 7). Third, RNase T2 ribonucleases cleave the 3'-end of all four bases, but prefer the 3'-end of adenine, which is in contrast

120

with RNase A and RNaseT1 families, which tend to be specific for pyrimidine or guanosine, respectively (Luhtala and Parker, 2010). Based on these characteristics, we consider RNase T2 as a candidate nuclease.

121

4.2 Methods

4.2.1 Blood collection and RNA purification

All studies were approved by the Institutional Review Board at Emory

University. Following informed consent, whole blood from healthy adults was drawn into CPD-ADSOL [AS-1] collection set, and processed into leukodepleted packed RBC units following standard procedures. Units were stored at 1-6ºC in a monitored refrigerator for 42 days. 5 mL samples were removed for testing at one day, 3day (d), 7d,

10d, 14d, 28d, 36d or 42d (expiration). Removal of aliquots was accomplished anaerobically via a sterile-docking device fitted with a valve to ensure there was no re- entry of air, or other contamination, into the RBC unit. RNA was isolated from RBCs using the mirVana™ miRNA Isolation Kit (Thermo Fisher).

4.2.2 Northern blot

RNA samples from RBCs collected at different storage times were size- fractionated on a 15% TBE-urea polyacrylamide gel (Bio-Rad #3450091), transferred to a nylon membrane (GE Healthcare RPN303B) and blocked with ExpressHybTM

Hybridization solution (Clontech #636831). The probes used for the northern blot were miR-720 Linked Nucleic Acid (LNA) detection probe (5′-TGGAGGCCCCAGCGAGA-3′)

(Exiqon #21363-00); tRNA-Thr-TGT (5′- TGGAGGCCCC AGCGAGATTT

GAACTCGCGA CCCCTGGTTT ACAAGACCAG TGCTCTAACC CCTGAGCTAT

122

AGGGCC-3′); tRNA-Ala-GCA (5′- TGGTGGAGGA TGCGGGCATC GATCCCGCTA

CCTCTCGCAA GCAAAGCGAG CGCTCTACCA TTTGAGCTAA TCCCCC-3′); tRNA-

Tyr-UAA (5′- TGGTCCTTCG AGCCGGAATC GAACCAGCGA CCTAAGGATC

TACAGTCCTC CGCTCTACCA GCTGAGCTAT CGAAGG-3′). The probes were labeled with γ-P32 by T4 ligase (New England Biolabs M0201) and purified by a G25 column (GE

Healthcare #27532501). The control 18mer miR-720 RNAs (5′-

AUCUCGCUGGGGCCUCCA-3′ were synthesized by IDT (Integrated DNA

Technologies) with RNase Free HPLC purification. The Ambion® Decade™ Marker

(Thermo AM7778), DynaMarker® Prestain Marker for small RNA (BioDynamics

Laboratory Inc. DM253), and microRNA marker (New England Biolabs N2102) were used as indicators.

4.2.3 In vitro RNA cleavage assays

Synthetic tRNA-Thr-TGT RNA (5′-GGCCCTATAG CTCAGGGGTT

AGAGCACTGG TCTTGTAAAC CAGGGGTCGC GAGTTCAAAT CTCGCTGGGG

CCTCCA-3′) was synthesized by IDT (Integrated DNA Technologies) using RNase-Free

HPLC purification. After incubating the synthetic tRNA-Thr-TGT RNAs with RBC lysates at 37 °C for 3h, 5 μL of the mixture was separated in the TBE-urea polyacrylamide gel, then detected with the miR-720 probe. In addition, the RBC lysates were first heat-inactivated by incubating at 95 °C for 5 min, then incubated with

123

synthetic tRNA-Thr-TGT RNAs. For experiments with different candidate nucleases, the synthetic tRNA-Thr-TGT RNA was incubated with either increased concentrations of the recombinant Dicer proteins (Genlantis #T510002, at 0.5 unit/μL, 0, 0.4, 2, 4 μL) or angiogenin proteins (Research and Diagnostic Systems, Inc. #265-AN, at 0, 4, 10, 40

μg/mL) in final volume of 20 μl at 37 °C for 3h, and analyzed by northern blot with the miR-720 probe. DNase I (M0303) and RNase H (M0297) were purchased from New

England Biolabs.

4.2.4 Western blot analysis and immunodepletion

Proteins were prepared by hypotonic lysis of RBCs and quantified using the

Pierce BCA assay (Thermo Fisher Scientific). Proteins were separated proteins in polyacrylamide gels (PAGE) were transferred to PVDF membranes using the semi-dry transfer system. Antibodies against angiogenin(ANG) (Santa Cruz sc-74528), Dicer (Cell

Signaling #3363), RNH1 (Santa Cruz sc-365783), and 4.1R (Santa Cruz sc-166759) were used in western blot analysis in accordance with the manufacturers’ instructions. Signals were visualized with the ECLTM Prime Western Blotting System (GE Healthcare

RPN2232) by exposure to films or ChemiDocTM Imaging Systems (Bio-Rad). For immunodepletion, the RBC lysates were incubated with anti-mIgG (Santa Cruz sc-2025) or anti-ANG (Thermo Fisher Scientific #14-9762-80) prebound to DynabeadsTM protein G

(Thermo Fisher Scientific 10004D) for 1.5 h at 4°C with rotating. Bound proteins were

124

eluted from the magnetic beads with loading buffer by boiling. After three rounds of immunodepletion, the supernatants of the immunodepleted RBC lysates were used for western blots and cleavage of synthetic tRNA-Thr-TGT RNAs. Before RNA extraction, the synthetic cel-miR-254 RNAs were added. Before RNA extraction, the synthetic cel- miR-254 RNAs were added. The cleavage levels were determined by TaqMan miRNA qPCR assays.

4.2.5 Statistical analysis

P values less than 0.05 were considered significant (*< 0.05; **<0.01; ***<0.001).

125

4.3 Results

4.3.1 NanoString analysis of the RBC miRNAs during storage

We sought to define the transcriptomic changes of RBCs during standard blood bank storage conditions with a focus on miRNAs. Aliquots of three leukodepleted AS-1 packed RBC units were sampled at day(s) 1, 3, 7, 10, 14, 28, 36 and 42 for RNA isolation during storage (Figure 85) (Sangokoya et al., 2010a). During RNA extraction, non-human spike-in miRNAs (cel-miR-254 (C. elegans) and osa-miR-442 (O. sativa)) were added to the samples to allow for normalization. Day 1 is a reasonable starting point since this is the typical earliest time to use the store RBC for transfusion. Given that the RBC transcriptome is enriched in small-sized RNA (Chen et al., 2017; Doss et al., 2015), we profiled the miRNA content using the NanoString nCounter human miRNA v2 expression assay. The resulting data were first normalized based on the spike-in foreign miRNA controls. Then, we normalized the changes of all miRNAs against the day one sample for each individual donor by zero-transformation (Tang et al., 2017a; Tang et al.,

2016). These storage-associated changes were then arranged by similarity via hierarchical clustering as shown in the overall heatmap (Figure 86). This analysis revealed that the expression levels of most miRNAs did not significantly change during

42 days of storage (Figure 86), indicating the surprisingly long half-life of most miRNAs.

However, there are consistent storage-associated changes, including two increased

126

(miR-720 and miR-33a-5p) and two reduced miRNAs (miR-563 and miR-582-5p) (Figure

86). Among these 4 storage-affected RNAs, miR-720 induction was the most dramatic and consistent change. Therefore, we have further validated and investigated miR-720 induction during storage.

Figure 85: Pipeline for analysis of RBC miRNAs during storage

Leukodepleted RBC units from three different individuals were incubated at 4 °C for the indicated number of days in a single bag. Aliquots of RBCs were removed from the bag, centrifuged, and total RNA (including small RNAs) was isolated from the packed cells. Spiked-in nonhuman miRNAs were added during different steps of the RNA isolation process to control for miRNA input for the NanoString platform.

127

Figure 86: Heat map of miRNA expression of RBCs

Select RBC miRNAs are upregulated or downregulated during RBC unit storage from three individuals. miRNA ratios were normalized to Day 1 of the storage of each individual. Data Credit: Jennifer Doss.

4.3.2 Small RNA northern blots indicate the putative tRNA origins of miR-720

Considering that NanoString profiling data identified the robust induction of miR-720, we sought to investigate this miRNA further. miR-720 has been reported as a prognostic biomarker, and to have a functional role in several distinct biological

128

contexts, including tumor invasion and T lymphocyte activation (Hara et al., 2013; Li et al., 2014; Nonaka et al., 2015; Shinozuka et al., 2013; Wang et al., 2015a). miR-720 was first annotated as a miRNA in the miRNA database (miRBase). However, in 2010,

Schopman et al. indicated that miR-720 is likely a cleavage fragment of threonine tRNA

(Schopman et al., 2010). Thus, miR-720 was subsequently removed from miRBase.

However, we still refer to this RNA fragment as miR-720 in the following section.

To further verify miR-720 increase during storage, we conducted small RNA northern blots using RNA extracted from stored RBCs at different time points of storage.

The miR-720 LNA probe (Figure 87; Table 2) identified a band near 18 nt, corresponding to the expected size of miR-720, in stored RBC samples using a synthetic 18 nt miR-720 as size marker (Figure 88; Table 2).

Figure 87: Cloverleaf structure of tRNAThr(TGT)

Full-length probe (blue) and miR-720 probe (orange) are illustrated. Figures were modified from Schopman et al.

129

Table 2 The sequence(s) of the synthetic RNAs and northern blot probe(s)

Figure 88: Small RNA northern blot of RBC RNA

RNA from RBCs stored for 4 and 28 days was probed with the miR-720 LNA probe. Synthetic 18mer miR-720 as a size marker.

4.3.3 RBC lysates contain cleavage activities for synthetic tRNAThr(TGT)

To understand the basis for increased miR-720 levels, we hypothesized that RBC lysates may contain nuclease activity affecting tRNAThr(TGT) cleavage. When RBC lysates

130

were incubated with synthetic tRNAThr(TGT), we observed cleavage of the tRNA into RNA fragments with the expected sizes of miR-720 (Figure 89), suggesting the presence of cleavage activities in the RBC lysates. To identify the nature of the cleavage activities, we incubated the synthetic tRNAThr(TGT) with RBC lysates before and after heat inactivation.

Heat inactivation abolished the cleavage activities (Figure 90), suggesting cleavage activity requires heat-sensitive molecules, such as proteins.

Figure 89: Northern blot of RBC lysate

Northern blot of RBC lysate incubated with synthetic tRNAThr(TGT) shows a generation of “miR-720”-sized fragments.

131

Figure 90: Northern blot of heat-inactivated RBC lysate

Heat inactivation (HI) abolishes the cleavage activity in RBC lysates. Northern blot data show heat inactivation from two different donors with a miR-720 probe.

4.3.4 Angiogenin contributes to tRNA cleavage and increased miR-720 during RBC storage

Based on literature (Anderson and Ivanov, 2014; Wang et al., 2015b), we considered angiogenin and Dicer as candidate nucleases that could cleave tRNA-Thr-

TGT into miR-720. Angiogenin is a well-known stress-induced nuclease responsible for tRNA cleavage (Ivanov et al., 2011; Thompson and Parker, 2009b; Yamasaki et al., 2009).

Angiogenin is often bound by the ribonuclease/angiogenin inhibitor 1 (RNH1), which inhibits its activity (Shapiro and Vallee, 1987). In addition, Dicer was suggested to be essential for miR-720 biogenesis (Wang et al., 2015b) and has been shown to contribute to tRNA cleavage (Cole et al., 2009).

132

To investigate the role of angiogenin and Dicer in RBC lysates, we first confirmed their presence via western blots (Figure 91). Next, we incubated the synthetic tRNA-Thr-TGT with increasing amounts of either recombinant Dicer or angiogenin protein. Interestingly, angiogenin, in a dose-dependent manner, resulted in cleavage of the synthetic tRNA into a product corresponding to the expected size of miR-720 (Figure

92A). In contrast, Dicer incubation generated a fragment of ~25 nt, different from the expected size of miR-720. Recombinant DNase I and RNase H failed to cleave synthetic tRNA (Figure 92B). Finally, the cleavage products of tRNA-Thr-TGT fragments, generated by either angiogenin or RBC lysates, share a ~18 nt fragment (Figure 93), corresponding to the expected size of miR-720.

Figure 91: Western blots of RBC lysate

133

Western blots of the indicated proteins in RBC lysate during storage. Ponceau S staining and 4.1R serve as loading controls.

Figure 92: In vitro biochemical study of nuclease(s)

(A) In vitro study of cleavage activity of candidate recombinant nuclease(s). Synthetic tRNAThr(TGT) was incubated with increased dosage of recombinant Dicer proteins (Genlantis #T510002, at 0.5 unit/μL, 0, 0.4, 2, 4 μL in a final volume of 20 μL) or recombinant angiogenin (ANG) proteins (Research and Diagnostic Systems, Inc. #265-

134

AN, at 0, 4, 10, 40 μg/mL) and then detected by northern blot with the miR-720 probe. (B) tRNAThr(TGT) RNA was incubated without or with DNase I (M0303, New England Biolabs) and RNase H (M0297, New England Biolabs) or ANG protein. Cleavage is not mediated by DNase I or RNase H. Data are representative of at least two independent experiments.

Figure 93: Angiogenin plays a role in tRNA cleavage

The comparison of cleavage between angiogenin and RBC lysates toward tRNAThr(TGT).

To test whether angiogenin may contribute to increased miR-720 during RBC storage, recombinant angiogenin was added to RBCs during ex vivo storage. The addition of angiogenin led to further enhancement of the miR-720 induction (Figure

94A). Conversely, we removed angiogenin from RBC lysate via immunodepletion, which led to a significant reduction in tRNA cleavage activity (Figure 94B). These data

135

indicate that angiogenin contributes to tRNA cleavage to generate the miR-720 fragment during RBC storage.

Figure 94: Angiogenin contributes to the “miR-720” increase in stored RBC

136

(A) The level of “miR-720” increases in angiogenin spike-in stored RBCs. Vehicle control or recombinant angiogenin protein (2 pg/ml) was added to RBCs every week. The RBC RNAs will be used to determine the “miR-720” levels. (B) Immunodepletion of angiogenin (ANG) from RBC lysate reduces the tRNAThr(TGT) cleavage activities. Mouse IgG (anti- mIgG) serves as the appropriate isotype control for the angiogenin antibody (anti-ANG). Data are shown as mean ± SEM of technical duplicates and are representative of three independent experiments; **p<0.01, ***p<0.001. 4.4 Discussion

4.4.1 Summary of the results

In this study, we performed extensive miRNA profiling of RBCs during 42 days of ex vivo storage using conditions routinely used in modern blood banks. Unexpectedly, the levels of most erythrocyte miRNAs remained unchanged over this storage interval, indicating their remarkable stability over the extended period of storage. However, a few miRNAs showed significant changes during storage. One especially dramatic and consistent change was an increase in miR-720, the level of which increased significantly after two weeks of storage. Small RNA northern blots indicated that the increased miR-

720 levels resulted from the specific cleavage of its tRNA precursor during storage. In addition, we provide evidence that angiogenin may contribute to miR-720 generation during RBC storage. Together, our results indicate that a specific RNA processing occurs in RBCs during refrigerated storage. Several studies have used various profiling methods to quantify miRNA changes during storage (Kannan and Atreya, 2010;

Sarachana et al., 2015). While these studies have identified some miRNAs with modest changes, these studies did not find a dramatic increase in miR-720 observed in our 137

studies. The differences may be caused by several factors. First, we employed the nCounter miRNA Assay (NanoString) which includes >800 mature and pre-miRNAs found in miRBase v18. This version of the nCounter miRNA assays includes miR-720 based on its earlier designation as a microRNA in the miRBase. The most recent version does not contain miR-720 as it is removed from miRBase (Schopman et al., 2010). In addition, two foreign miRNAs were added as normalization controls for all samples to control for global RNA changes or variations in the processing.

4.4.2 RBCs as a unique cellular context for studying RNA metabolism

While the transcription and processing of miRNAs are well-defined, miRNA decay and degradation are much less well understood. Part of the difficulties lies in the dynamic transcription and highly regulated multi-step processing that may complicate analysis. Several studies on miRNA decay kinetics(Chatterjee and Grosshans, 2009;

Ruegger and Grosshans, 2012; Tu et al., 2015; Zhang et al., 2012) indicate miRNAs have extremely long half-lives. Mature RBCs lack nuclei and novel RNA synthesis and are often stored ex vivo for up to 42 days in blood banks prior to blood transfusion. Thus, mature RBCs provide a unique cellular context to evaluate miRNA turnover without the confounding factor of novel synthesis from nuclei. Our results indicate a surprisingly long half-life of most miRNAs, showing limited changes in their levels during storage.

Such an extended half-life of miRNAs is consistent with other studies (Chatterjee and

138

Grosshans, 2009; Ruegger and Grosshans, 2012; Tu et al., 2015; Zhang et al., 2012), and likely contributes to the persistence of miRNAs long after RBC enucleation.

4.4.3 Stress-associated nuclease angiogenin contributes to the storage- associated miR-720 increase

Angiogenin is a well-recognized nuclease for tRNA cleavage in response to various stresses (Ivanov et al., 2011; Yamasaki et al., 2009). These resulting tRNA fragments repress translation. One of these studies was performed with rabbit reticulocyte lysates (Saxena et al., 1992). Our results further indicate the role of angiogenin toward tRNAThr for miR-720 increase during RBC storage. The distinct and novel features of our findings are: First, since there is a limited active translation in the mature RBC, the increased numbers of miR-720 may exert their functional effects in other adjacent cells instead of RBC (discussed in the next section). Second, there seems to be significant specificity for tRNAThr, but not two other tested tRNAs. The source of such specificity is not yet known. Finally, it is not clear how angiogenin becomes activated during storage. Several possibilities may be worthy of further investigation. For example, certain plasma proteins or factors and proteins may inhibit angiogenin. The removal of these factors during the washing steps of RBC processing may allow the angiogenin activation during ex vivo storage. One potent angiogenin RNH) (Pizzo et al.,

2013) was found to gradually decline during storage may contribute to the angiogenin activation. Angiogenin is usually activated by stresses and its activation during storage 139

may indicate significant cellular stresses not currently appreciated. These stresses may be caused by a large number of physiological and metabolic changes (Bennett-Guerrero et al., 2007; Roback et al., 2014) that occur during storage. Understanding the basis of the angiogenin activation may allow us to improve the storage conditions with the potential to reduce storage lesions.

4.4.4 Putative functional role of miR-720 associated with transfusion

In physiological conditions, miR-720 has been associated with epithelial development in keratinocytes(Chikh et al., 2011) as well as dental pulp cell differentiation(Hara et al., 2013). Upregulated miR-720 has been reported to correlate with tumor progression in a variety of tumors such as breast, cervical, colorectal, and bladder cancers, as well as renal cell carcinoma, myeloma, and melanoma (Das et al.,

2016; Jones et al., 2012; Nonaka et al., 2015; Park et al., 2014; Ragusa et al., 2012; Sand et al., 2013; Tang et al., 2015b; Wang et al., 2015a; Yi et al., 2010). These findings suggest that miR-720 has important functional roles. Interestingly, miR-720 is often found to be enriched in extracellular vesicles (EVs) (Guzman et al., 2015). In addition, EVs are produced during RBC storage (Danesh et al., 2014). It is possible that increased miR-720 in EVs of stored RBCs may be transferred to other cells to affect gene expression and phenotypes. Additionally, miR-720 has been shown to exert potent functional roles in many biological settings (Hara et al., 2013; Li et al., 2014; Shinozuka et al., 2013; Tang et

140

al., 2015a; Wang et al., 2015a). MiR-720 in storage-associated exosomes may contribute to the cellular phenotypes of storage lesions. Should this be the case, inhibition of miR-720 may hold the potential for abolishing certain putative adverse effects of stored RBCs.

4.4.5 The potential of storage transcriptome signatures for detecting blood doping

In addition to offering therapeutic benefits, blood transfusion is sometimes used to enhance athletic performance to increase red cell mass and oxygen-delivering capacity. Though this practice is officially banned by the World Anti-Doping Agency, no effective methods have been developed to detect this practice. This is especially challenging for the detection of transfusion of athletes’ own blood, known as autologous blood transfusion (ABT). Current approaches to indirectly detect ABT include monitoring hematological parameters that are recorded in the athlete's biological passport, as well as detecting markers of the biologic responses to blood transfusion.

However, these parameters either have narrow detection windows or can be triggered by exposure to altitude or infection (Salamin et al., 2016). Therefore, the dramatic increase in miR-720 may serve as an intrinsic and reliable biomarker to identify stored

RBCs and detect ABT. Given that the increase of miR-720 in stored RBCs is irreversible

141

and persistent throughout the life span of stored RBCs, such a signature may provide a sensitive and reliable biomarker of blood doping.

142

5. Conclusions

5.1 Hippo pathway effectors TAZ regulates ferroptosis

Portions of this Chapter are reproduced from the review for Molecular & Cellular

Oncology: Molecular & Cellular Oncology, 2019, accepted

Ferroptosis is a recently defined form of programmed cell death (Dixon et al.,

2012) characterized by the accumulation of lipid peroxidation. The relevance of ferroptosis for various human diseases is now beginning to be appreciated (Stockwell et al., 2017). The lipid peroxidation results from oxidative stresses generated by NADPH oxidase(s) (NOXs) that is often repaired by glutathione peroxidase 4 (GPX4) using the glutathione as a co-factor. Therefore, ferroptosis can be induced by the removal of cystine (limiting component for glutathione synthesis), inhibition of GPX4, or activation of NOXs. The canonical ferroptosis inducer, erastin, is an inhibitor of cystine-glutamate transporter (xCT) that reduces cystine import and depletes glutathione (Dixon et al.,

2012). While ferroptosis may have therapeutic potential toward cancer (Cramer et al.,

2017), much remains unknown about the genetic determinants and underlying mechanisms of ferroptosis to select tumors that are most likely to respond to these ferroptosis-inducing agents and predict potential resistant mechanisms against such approach.

143

Many studies have identified various genetic determinants of ferroptosis involved in the GSH/lipid metabolisms (Xie et al., 2016), oncogenic somatic mutations, regulation of iron levels (Chen et al., 2019) and process of epithelial-mesenchymal transitions (Viswanathan et al., 2017). In the tumor microenvironment, there are also many non-genetic factors that may impact tumor progression, metastasis, and treatment response. These factors include physical, chemical and mechanic stresses, such as tissue hypoxia, lactic acidosis, nutrient deprivation, osmotic pressure, tissue tension, and stiffness. Unlike cell-intrinsic genetic factors, it is not clear whether these non-genetic factors also affect ferroptosis. Recently, we and other independent research groups have observed that vulnerability to ferroptosis is highly influenced by a non-genetic factor, cell density (Wu et al., 2019; Yang et al., 2019a; Yang et al., 2019b). While the renal cell carcinoma and ovarian tumor cells are highly sensitive to ferroptosis when grown at low density; they become highly resistant to ferroptosis when grown in confluent conditions.

Since the cell density-dependent phenotypes can be sensed and regulated by the evolutionarily conserved Hippo pathway effectors, YAP (Yes-associated protein 1) /TAZ

(transcriptional coactivator with PDZ-binding motif), we wondered if Hippo pathway effectors are involved in such density-dependent ferroptosis response. The activities and functions of YAP/TAZ are regulated by the state of phosphorylation and intracellular localization. That is, in high cell density, YAP/TAZ are phosphorylated, restricted in the

144

cytosol, and subjected to proteasome-mediated degradation. On the other hand,

YAP/TAZ are dephosphorylated and localized in the nucleus to associate with transcriptional factors, such as TEAD (transcriptional enhanced associate domain) family proteins, to drive expressions of proliferation and metastasis genes.

Our data show that TAZ, instead of YAP, is abundantly expressed in both renal and ovarian cancer cells and undergoes density-dependent nuclear/cytosolic translocation (Yang et al., 2019a; Yang et al., 2019b). TAZ removal confers ferroptosis resistance, while overexpression of constitutively active form of TAZ, TAZS89A, sensitizes cells to ferroptosis. We have found that ovarian cancers including both clear cell and serous subtypes are sensitive to ferroptosis induced by cystine deprivation. In addition, we found that lower TAZ level in the recurrent ovarian cancer is responsible for reduced ferroptosis susceptibility. We further investigated the mechanisms and found that TAZ regulates ferroptosis through epithelial membrane protein 1 (EMP1)-

NADPH oxidase 4 (NOX4) axis in renal cancers (Yang et al., 2019b) and Angiopoietin- like 4 (ANGPTL4)- NADPH oxidase 2 (NOX2) axis in ovarian cancers (Yang et al.,

2019a). An independent study also identified that YAP as a novel regulator of ferroptosis sensitivity in mesothelioma (Wu et al., 2019). Collectively, these three studies have shown the relevance of Hippo pathway effectors for ferroptosis and suggest that ferroptosis-inducing agents may be used to target the YAP/TAZ-activated tumors.

145

Our studies (Yang et al., 2019a; Yang et al., 2019b) have provided novel insights into ferroptosis, a novel form of cell death, firstly reporting on the link between the

Hippo pathway effector TAZ and ferroptosis sensitivities in renal and ovarian cancers.

These studies have several implications for the fields of ferroptosis and cancer biology.

Figure 95 Hippo pathway effectors, YAP/TAZ, regulates ferroptosis

146

In response to various non-genetic environmental cues, Hippo pathway effectors, YAP/TAZ, regulates ferroptosis by affecting the expression of different sets of tissue- specific targets genes encoding lipid/iron metabolism and ROS production by NOXs.

First, our results indicate a model (Figure 95) in which outside environmental cues in the tumor microenvironments are sensed by the Hippo pathway and its effectors, resulting in the translocation of YAP or TAZ to associate with TEAD transcriptional factor and regulate the transcriptional outputs. The YAP/TAZ target genes encode proteins involved in the signaling relay and amplification to affect lipid/iron metabolism (ACSL4: acyl-CoA synthetase long-chain family member 4/ TFRC: transferrin receptor) or reactive oxygen species (ROS) productions (NOX2/NOX4) as part of the adaptive responses to changes in tumor microenvironments. These Hippo- driven changes affect life vs. death decisions when the cancer cells are placed under ferroptosis-inducing conditions, such as erastin or cystine deprivation. Second, while

Hippo pathway effectors, YAP/TAZ, regulate ferroptosis in multiple biological contexts, the specific effectors, transcriptional targets and ferroptosis executors may be different in each cell type. For example, NOXs are involved in the TAZ-regulated ferroptosis in both renal cell carcinoma and ovarian cancers. However, different NOXs are involved in renal vs. ovarian cancer cells. NOX4 is the predominant NOX family protein in the renal cells and NOX2 is the highest-expressed transcripts in ovarian cancer patients. Third, since Hippo pathway effectors are involved in the proliferation and metastasis of many

147

human tumors, our results suggest that the YAP/TAZ activation status may predict the sensitivity to ferroptosis-inducing agents. Furthermore, the Hippo pathway integrates a wide variety of non-genetic factors, such as mechanical properties and tissue stiffness

(Zanconato et al., 2016). Therefore, our findings may suggest many non-genetic factors may also regulate ferroptosis sensitivities in the setting of a “stiff” tumor environment known to activate the YAP/TAZ and promote ferroptosis. Finally, since the higher activity of YAP/TAZ promotes invasion and metastases and often associate with poor prognosis, our findings suggest that triggering ferroptosis may be valuable in combinational therapy for these tumors which tend to become resistant to standard treatments. Taken together, including ferroptosis in the current cancer therapeutics may improve the response rate and clinical outcomes of patients, especially with YAP/TAZ- activated tumors.

5.2 Molecular identification of tRNA cleavage activity in stored erythrocytes

Red blood cells (RBCs) are the major component of blood transfusions, one of the most common procedures in the hospital. Annually, more than 85 million units of RBCs are transfused worldwide (Yoshida et al., 2019). In addition, some athletes utilize blood transfusion to increase athletic performance, a practice banned by the world anti-doping agency. Currently, RBCs can be stored for up to 42 days at ~4°C before transfusion.

148

However, transfusion with RBCs after long storage duration correlates with a poorer prognosis compared with fresh RBCs and results in increased morbidity and mortality.

To recognize the undesirable effects of prolonged RBC storage on transfusion recipients, it is critical to understand storage-associated RBC changes. To this end, our lab has previously identified a variety of RNA species in mature RBCs. We speculate that changes in the RBC transcriptome may offer insight into RNA biology, storage lesion, and novel means to detect illegal autologous blood transfusions in athletes. Therefore, our lab has profiled the miRNA changes that occur in RBCs at different time intervals during in vitro storage. Here, we found that the abundance of almost all miRNAs does not change significantly during storage. However, miR-720, a cleavage product of threonine tRNA, dramatically increased (~10-fold) after two weeks of storage and remained elevated for at least an additional four weeks of storage. This increase of miR-

720 was verified by qRT-PCR and Northern blot. This remarkable increase of miR-720 may be used as a biomarker to monitor the quality of stored RBCs for transfusion as well as to develop an anti-doping test for athletes to determine if they had autologous blood transfusions. However, the factors that induce increased levels of miR-720 are not yet understood. Here, we demonstrated that the increased miR-720 level results from the cleavage of tRNAThr by angiogenin in the stored RBCs. Angiogenin is a nuclease recognized for tRNA cleavage in response to various stresses (Ivanov et al., 2011;

149

Yamasaki et al., 2009). The activation of angiogenin during storage may indicate significant cellular stresses not currently appreciated. Understanding the basis of the angiogenin activation may allow us to improve the storage conditions with the potential to reduce storage lesions. In addition to offering therapeutic benefits, our study has the potentials to develop as a detection method for autologous blood transfusion (ABT), a blood doping. Given that the increase of miR-720 in stored RBCs is irreversible and persistent throughout the life span of stored RBCs, the dramatic increase of miR-720 may serve as an intrinsic and reliable biomarker to identify the storage lesion of stored

RBCs and a sensitive and reliable biomarker for blood doping.

150

References

Acquati, F., Possati, L., Ferrante, L., Campomenosi, P., Talevi, S., Bardelli, S., Margiotta, C., Russo, A., Bortoletto, E., Rocchetti, R., et al. (2005). Tumor and metastasis suppression by the human RNASET2 gene. Int J Oncol 26, 1159-1168.

Aletti, G.D., Gallenberg, M.M., Cliby, W.A., Jatoi, A., and Hartmann, L.C. (2007). Current management strategies for ovarian cancer. Mayo Clin Proc 82, 751-770.

Anderson, P., and Ivanov, P. (2014). tRNA fragments in human health and disease. FEBS Lett 588, 4297-4304.

Angeli, J.P.F., Schneider, M., Proneth, B., Tyurina, Y.Y., Tyurin, V.A., Hammond, V.J., Herbach, N., Aichler, M., Walch, A., and Eggenhofer, E. (2014). Inactivation of the ferroptosis regulator Gpx4 triggers acute renal failure in mice. Nature cell biology 16, 1180.

Baba, K., Kitajima, Y., Miyake, S., Nakamura, J., Wakiyama, K., Sato, H., Okuyama, K., Kitagawa, H., Tanaka, T., Hiraki, M., et al. (2017). Hypoxia-induced ANGPTL4 sustains tumour growth and anoikis resistance through different mechanisms in scirrhous gastric cancer cell lines. Sci Rep 7, 11127.

Basuli, D., Tesfay, L., Deng, Z., Paul, B., Yamamoto, Y., Ning, G., Xian, W., McKeon, F., Lynch, M., Crum, C.P., et al. (2017). Iron addiction: a novel therapeutic target in ovarian cancer. Oncogene 36, 4089-4099.

Bennett-Guerrero, E., Veldman, T.H., Doctor, A., Telen, M.J., Ortel, T.L., Reid, T.S., Mulherin, M.A., Zhu, H., Buck, R.D., Califf, R.M., et al. (2007). Evolution of adverse changes in stored RBCs. Proc Natl Acad Sci U S A 104, 17063-17068.

Berezina, T.L., Zaets, S.B., Morgan, C., Spillert, C.R., Kamiyama, M., Spolarics, Z., Deitch, E.A., and Machiedo, G.W. (2002). Influence of storage on red blood cell rheological properties. Journal of surgical research 102, 6-12.

Berkhout, B., and Haasnoot, J. (2006). The interplay between virus infection and the cellular RNA interference machinery. FEBS Lett 580, 2896-2902.

Blackburn, P., Wilson, G., and Moore, S. (1977). Ribonuclease inhibitor from human placenta. Purification and properties. J Biol Chem 252, 5904-5910.

151

Borbély, G., Szabadkai, I., Horváth, Z., Markó, P., Varga, Z., Breza, N., Baska, F., Vántus, T., Huszár, M., Geiszt, M., et al. (2010). Small-Molecule Inhibitors of NADPH Oxidase 4. Journal of Medicinal Chemistry 53, 6758-6762.

Cao, J.J., Zhao, X.M., Wang, D.L., Chen, K.H., Sheng, X., Li, W.B., Li, M.C., Liu, W.J., and He, J. (2014). YAP is overexpressed in clear cell renal cell carcinoma and its knockdown reduces cell proliferation and induces cell cycle arrest and apoptosis. Oncol Rep 32, 1594-1600.

Chatterjee, S., and Grosshans, H. (2009). Active turnover modulates mature microRNA activity in Caenorhabditis elegans. Nature 461, 546-549.

Chen, G., Xie, J., Huang, P., and Yang, Z. (2016). Overexpression of TAZ promotes cell proliferation, migration and epithelial-mesenchymal transition in ovarian cancer. Oncol Lett 12, 1821-1825.

Chen, P.-H., Chi, J.-T., and Boyce, M. (2018). Functional crosstalk among oxidative stress and O-GlcNAc signaling pathways. Glycobiology 28, 556-564.

Chen, P.-H., Hong, J., and Chi, J.-T. (2017). Discovery, Genomic Analysis, and Functional Role of the Erythrocyte RNAs. Current Pathobiology Reports, 1-6.

Chen, P.-H., Wu, J., Ding, C.-K.C., Lin, C.-C., Pan, S., Bossa, N., Xu, Y., Yang, W.-H., Mathey-Prevot, B., and Chi, J.-T. (2019). Kinome screen of ferroptosis reveals a novel role of ATM in regulating iron metabolism. Cell Death & Differentiation, 1.

Chen, S.-Y., Wang, Y., Telen, M.J., and Chi, J.-T. (2008). The Genomic Analysis of Erythrocyte microRNA Expression in Sickle Cell Diseases. PLOS One 3, e2360.

Chikh, A., Matin, R.N.H., Senatore, V., Hufbauer, M., Lavery, D., Raimondi, C., Ostano, P., Mello-Grand, M., Ghimenti, C., Bahta, A., et al. (2011). iASPP/p63 autoregulatory feedback loop is required for the homeostasis of stratified epithelia. Embo Journal 30, 4261-4273.

Choi, H.-J., Zhang, H., Park, H., Choi, K.-S., Lee, H.-W., Agrawal, V., Kim, Y.-M., and Kwon, Y.-G. (2015). Yes-associated protein regulates endothelial cell contact-mediated expression of angiopoietin-2. Nature Communications 6, 6943.

152

Cole, C., Sobala, A., Lu, C., Thatcher, S.R., Bowman, A., Brown, J.W., Green, P.J., Barton, G.J., and Hutvagner, G. (2009). Filtering of deep sequencing data reveals the existence of abundant Dicer-dependent small RNAs derived from tRNAs. RNA 15, 2147-2160.

Cramer, S.L., Saha, A., Liu, J., Tadi, S., Tiziani, S., Yan, W., Triplett, K., Lamb, C., Alters, S.E., Rowlinson, S., et al. (2017). Systemic depletion of L-cyst(e)ine with cyst(e)inase increases reactive oxygen species and suppresses tumor growth. Nat Med 23, 120-127.

D'Alessandro, A., Kriebardis, A.G., Rinalducci, S., Antonelou, M.H., Hansen, K.C., Papassideri, I.S., and Zolla, L. (2015). An update on red blood cell storage lesions, as gleaned through biochemistry and omics technologies. Transfusion 55, 205-219.

D’Alessandro, A., Liumbruno, G., Grazzini, G., and Zolla, L. (2010). Red blood cell storage: the story so far. Blood Transfusion 8, 82-88.

Danesh, A., Inglis, H.C., Jackman, R.P., Wu, S., Deng, X., Muench, M.O., Heitman, J.W., and Norris, P.J. (2014). Exosomes from red blood cell units bind to monocytes and induce proinflammatory cytokines, boosting T-cell responses in vitro. Blood 123, 687- 696.

Das, S.G., Romagnoli, M., Mineva, N.D., Barillé-Nion, S., Jézéquel, P., Campone, M., and Sonenshein, G.E. (2016). miR-720 is a downstream target of an ADAM8-induced ERK signaling cascade that promotes the migratory and invasive phenotype of triple-negative breast cancer cells. Breast Cancer Research : BCR 18, 40.

Dern, R.J., Gwinn, R.P., and Wiorkowski, J.J. (1966). Studies on the preservation of human blood. I. Variability in erythrocyte storage characteristics among healthy donors. J Lab Clin Med 67, 955-965.

Deshpande, R.A., and Shankar, V. (2002). Ribonucleases from T2 family. Critical Reviews in Microbiology 28, 79-122.

Dixon, S.J., Lemberg, K.M., Lamprecht, M.R., Skouta, R., Zaitsev, E.M., Gleason, C.E., Patel, D.N., Bauer, A.J., Cantley, A.M., Yang, W.S., et al. (2012). Ferroptosis: an iron- dependent form of nonapoptotic cell death. Cell 149, 1060-1072.

Dixon, S.J., Patel, D.N., Welsch, M., Skouta, R., Lee, E.D., Hayano, M., Thomas, A.G., Gleason, C.E., Tatonetti, N.P., Slusher, B.S., et al. (2014). Pharmacological inhibition of cystine–glutamate exchange induces endoplasmic reticulum stress and ferroptosis. eLife 3, e02523. 153

Dolma, S., Lessnick, S.L., Hahn, W.C., and Stockwell, B.R. (2003). Identification of genotype-selective antitumor agents using synthetic lethal chemical screening in engineered human tumor cells. Cancer Cell 3, 285-296.

Doss, J.F., Corcoran, D.L., Jima, D.D., Telen, M.J., Dave, S.S., and Chi, J.-T. (2015). A comprehensive joint analysis of the long and short RNA transcriptomes of human erythrocytes. BMC Genomics 16, 952.

Dougherty, J.A., Kilbane Myers, J., Khan, M., Angelos, M.G., and Chen, C.-A. (2017). Dual-specificity phosphatase 4 overexpression in cells prevents hypoxia/reoxygenation- induced apoptosis via the upregulation of eNOS. Frontiers in cardiovascular medicine 4, 22.

Emara, M.M., Ivanov, P., Hickman, T., Dawra, N., Tisdale, S., Kedersha, N., Hu, G.F., and Anderson, P. (2010). Angiogenin-induced tRNA-derived Stress-induced RNAs Promote Stress-induced Stress Granule Assembly. Journal of Biological Chemistry 285, 10959-10968.

Etoh, T., Shibuta, K., Barnard, G.F., Kitano, S., and Mori, M. (2000). Angiogenin Expression in Human Colorectal Cancer: The Role of Focal Macrophage Infiltration. Clinical Cancer Research 6, 3545-3551.

Evdokimova, V., Ovchinnikov, L.P., and Sorensen, P.H. (2006). Y-box binding protein 1: providing a new angle on translational regulation. Cell Cycle 5, 1143-1147.

Fett, J.W., Strydom, D.J., Lobb, R.R., Alderman, E.M., Bethune, J.L., Riordan, J.F., and Vallee, B.L. (1985). Isolation and characterization of angiogenin, an angiogenic protein from human carcinoma cells. Biochemistry 24, 5480-5486.

Frank, D.N., and Pace, N.R. (1998). Ribonuclease P: unity and diversity in a tRNA processing ribozyme. Annu Rev Biochem 67, 153-180.

Friedman, R.C., Farh, K.K., Burge, C.B., and Bartel, D.P. (2009). Most mammalian mRNAs are conserved targets of microRNAs. Genome Res 19, 92-105.

Fu, H.J., Feng, J.J., Liu, Q., Sun, F., Tie, Y., Zhu, J., Xing, R.Y., Sun, Z.X., and Zheng, X.F. (2009). Stress induces tRNA cleavage by angiogenin in mammalian cells. Febs Letters 583, 437-442.

154

Furer, V., Greenberg, J.D., Attur, M., Abramson, S.B., and Pillinger, M.H. (2010). The role of microRNA in rheumatoid arthritis and other autoimmune diseases. Clinical Immunology 136, 1-15.

Galluzzi, L., Yamazaki, T., and Kroemer, G. (2018). Linking cellular stress responses to systemic homeostasis. Nature Reviews Molecular Cell Biology 19, 731-745.

Gao, M., Monian, P., Quadri, N., Ramasamy, R., and Jiang, X. (2015). Glutaminolysis and transferrin regulate ferroptosis. Molecular cell 59, 298-308.

Gao, Y., Lv, L., Liu, S., Ma, G., and Su, Y. (2013). Elevated levels of thrombin‐generating microparticles in stored red blood cells. Vox sanguinis 105, 11-17.

García-Roa, M., Del Carmen Vicente-Ayuso, M., Bobes, A.M., Pedraza, A.C., González- Fernández, A., Martín, M.P., Sáez, I., Seghatchian, J., and Gutiérrez, L. (2017). Red blood cell storage time and transfusion: current practice, concerns and future perspectives. Blood transfusion = Trasfusione del sangue 15, 222-231.

Gibbings, D.J., Ciaudo, C., Erhardt, M., and Voinnet, O. (2009). Multivesicular bodies associate with components of miRNA effector complexes and modulate miRNA activity. Nat Cell Biol 11, 1143-1149.

Gorin, Y., Block, K., Hernandez, J., Bhandari, B., Wagner, B., Barnes, J.L., and Abboud, H.E. (2005). Nox4 NAD(P)H oxidase mediates hypertrophy and fibronectin expression in the diabetic kidney. J Biol Chem 280, 39616-39626.

Greenwalt, T.J., Bryan, D.J., and Dumaswala, U.J. (1984). Erythrocyte Membrane Vesiculation and Changes in Membrane Composition during Storage in Citrate‐ Phosphate‐Dextrose‐Adenine‐1. Vox sanguinis 47, 261-270.

Gregg, J.L., Turner, R.M., Chang, G., Joshi, D., Zhan, Y., Chen, L., and Maranchie, J.K. (2014). NADPH Oxidase NOX4 Supports Renal Tumorigenesis by Promoting the Expression and Nuclear Accumulation of HIF2α. Cancer research 74, 3501-3511.

Guzman, N., Agarwal, K., Asthagiri, D., Yu, L., Saji, M., Ringel, M.D., and Paulaitis, M.E. (2015). Breast Cancer-Specific miR Signature Unique to Extracellular Vesicles includes “microRNA-like” tRNA Fragments. Molecular cancer research : MCR 13, 891-901.

155

Haiser, H.J., Karginov, F.V., Hannon, G.J., and Elliot, M.A. (2008). Developmentally regulated cleavage of tRNAs in the bacterium Streptomyces coelicolor. Nucleic Acids Res 36, 732-741.

Hall, C.A., Wang, R., Miao, J., Oliva, E., Shen, X., Wheeler, T., Hilsenbeck, S.G., Orsulic, S., and Goode, S. (2010). Hippo Pathway Effector Yap Is an Ovarian Cancer Oncogene. Cancer Research 70, 8517-8525.

Hamilton, A.J. (2010). MicroRNA in erythrocytes. Biochemical Society transactions 38, 229-231.

Hangauer, M.J., Viswanathan, V.S., Ryan, M.J., Bole, D., Eaton, J.K., Matov, A., Galeas, J., Dhruv, H.D., Berens, M.E., Schreiber, S.L., et al. (2017). Drug-tolerant persister cancer cells are vulnerable to GPX4 inhibition. Nature 551, 247-250.

Hara, E.S., Ono, M., Eguchi, T., Kubota, S., Pham, H.T., Sonoyama, W., Tajima, S., Takigawa, M., Calderwood, S.K., and Kuboki, T. (2013). miRNA-720 Controls Stem Cell Phenotype, Proliferation and Differentiation of Human Dental Pulp Cells. Plos One 8.

Hartmann, A., Kunz, M., Köstlin, S., Gillitzer, R., Toksoy, A., Bröcker, E.-B., and Klein, C.E. (1999). Hypoxia-induced Up-Regulation of Angiogenin in Human Malignant Melanoma. Cancer research 59, 1578-1583.

Hato, T., Tabata, M., and Oike, Y. (2008). The role of angiopoietin-like proteins in angiogenesis and metabolism. Trends Cardiovasc Med 18, 6-14.

Haussecker, D., Huang, Y., Lau, A., Parameswaran, P., Fire, A.Z., and Kay, M.A. (2010). Human tRNA-derived small RNAs in the global regulation of RNA silencing. RNA 16, 673-695.

Hess, J.R. (2011). Improving the predictive value of red blood cell storage trials: lessons from the Biomedical Excellence for Safer Transfusion (BEST) Collaborative Trial 41. Transfusion 51 Suppl 1, 34S-37S.

Honda, S., Loher, P., Shigematsu, M., Palazzo, J.P., Suzuki, R., Imoto, I., Rigoutsos, I., and Kirino, Y. (2015). Sex hormone-dependent tRNA halves enhance cell proliferation in breast and prostate cancers. Proceedings of the National Academy of Sciences 112, E3816-E3825.

156

Hsiao, C., Lampe, M., Nillasithanukroh, S., Han, W., Lian, X., and Palecek, S.P. (2016). Human pluripotent stem cell culture density modulates YAP signaling. Biotechnology Journal 11, 662-675.

Hsieh, J.J., Purdue, M.P., Signoretti, S., Swanton, C., Albiges, L., Schmidinger, M., Heng, D.Y., Larkin, J., and Ficarra, V. (2017). Renal cell carcinoma. Nature Reviews Disease Primers 3, 17009.

Huang, R.Y., Wong, M.K., Tan, T.Z., Kuay, K.T., Ng, A.H., Chung, V.Y., Chu, Y.S., Matsumura, N., Lai, H.C., Lee, Y.F., et al. (2013). An EMT spectrum defines an anoikis- resistant and spheroidogenic intermediate mesenchymal state that is sensitive to e- cadherin restoration by a src-kinase inhibitor, saracatinib (AZD0530). Cell Death Dis 4, e915.

Huang, Y., Cai, G.-q., Peng, J.-P., and Shen, C. (2018). Glucocorticoids induce apoptosis and matrix metalloproteinase-13 expression in chondrocytes through the NOX4/ROS/p38 MAPK pathway. The Journal of steroid biochemistry and molecular biology 181, 52-62.

Huang, Z., Kondoh, E., Visco, Z., Baba, T., Matsumura, N., Dolan, E., Whitaker, R.S., Konishi, I., Fujii, S., and Berchuck, A. (2019). Targeting dormant ovarian cancer cells in vitro and in an in vivo model of platinum resistance. bioRxiv, 716464.

Hunter, M.P., Ismail, N., Zhang, X., Aguda, B.D., Lee, E.J., Yu, L., Xiao, T., Schafer, J., Lee, M.-L.T., and Schmittgen, T.D. (2008). Detection of microRNA expression in human peripheral blood microvesicles. PloS one 3, e3694.

Im, H.I., and Kenny, P.J. (2012). MicroRNAs in neuronal function and dysfunction. Trends in Neurosciences 35, 325-334.

Ivanov, P., Emara, M.M., Villen, J., Gygi, S.P., and Anderson, P. (2011). Angiogenin- Induced tRNA Fragments Inhibit Translation Initiation. Molecular Cell 43, 613-623.

Jain, A., Tindell, C.A., Laux, I., Hunter, J.B., Curran, J., Galkin, A., Afar, D.E., Aronson, N., Shak, S., Natale, R.B., et al. (2005). Epithelial membrane protein-1 is a biomarker of gefitinib resistance. Proceedings of the National Academy of Sciences of the United States of America 102, 11858-11863.

157

Jeong, G.O., Shin, S.H., Seo, E.J., Kwon, Y.W., Heo, S.C., Kim, K.H., Yoon, M.S., Suh, D.S., and Kim, J.H. (2013). TAZ mediates lysophosphatidic acid-induced migration and proliferation of epithelial ovarian cancer cells. Cell Physiol Biochem 32, 253-263.

Jiang, L., Kon, N., Li, T., Wang, S.J., Su, T., Hibshoosh, H., Baer, R., and Gu, W. (2015). Ferroptosis as a p53-mediated activity during tumour suppression. Nature 520, 57-62.

Jochl, C., Rederstorff, M., Hertel, J., Stadler, P.F., Hofacker, I.L., Schrettl, M., Haas, H., and Huttenhofer, A. (2008). Small ncRNA transcriptome analysis from Aspergillus fumigatus suggests a novel mechanism for regulation of protein synthesis. Nucleic Acids Res 36, 2677-2689.

Johnson, D.S., Mortazavi, A., Myers, R.M., and Wold, B. (2007). Genome-wide mapping of in vivo protein-DNA interactions. Science 316, 1497-1502.

Jones, C.I., Zabolotskaya, M.V., King, A.J., Stewart, H.J.S., Horne, G.A., Chevassut, T.J., and Newbury, S.F. (2012). Identification of circulating microRNAs as diagnostic biomarkers for use in multiple myeloma. British journal of cancer 107, 1987-1996.

Kannan, M., and Atreya, C. (2010). Differential profiling of human red blood cells during storage for 52 selected microRNAs. Transfusion 50, 1581-1588.

Kawaji, H., Nakamura, M., Takahashi, Y., Sandelin, A., Katayama, S., Fukuda, S., Daub, C.O., Kai, C., Kawai, J., Yasuda, J., et al. (2008). Hidden layers of human small RNAs. BMC Genomics 9, 157.

Keenan, M.M., and Chi, J.T. (2015). Alternative fuels for cancer cells. Cancer J 21, 49-55.

Keenan, M.M., Liu, B., Tang, X., Wu, J., Cyr, D., Stevens, R.D., Ilkayeva, O., Huang, Z., Tollini, L.A., Murphy, S.K., et al. (2015). ACLY and ACC1 Regulate Hypoxia-Induced Apoptosis by Modulating ETV4 via alpha-ketoglutarate. PLoS Genet 11, e1005599.

Kim‐Shapiro, D.B., Lee, J., and Gladwin, M.T. (2011). Storage lesion: role of red blood cell breakdown. Transfusion 51, 844-851.

Kim, M.K., Osada, T., Barry, W.T., Yang, X.Y., Freedman, J.A., Tsamis, K.A., Datto, M., Clary, B.M., Clay, T., Morse, M.A., et al. (2012a). Characterization of an oxaliplatin sensitivity predictor in a preclinical murine model of colorectal cancer. Molecular cancer therapeutics 11, 1500-1509.

158

Kim, Y.N., Koo, K.H., Sung, J.Y., Yun, U.J., and Kim, H. (2012b). Anoikis resistance: an essential prerequisite for tumor metastasis. Int J Cell Biol 2012, 306879.

Koch, C.G., Li, L., Sessler, D.I., Figueroa, P., Hoeltge, G.A., Mihaljevic, T., and Blackstone, E.H. (2008). Duration of red-cell storage and complications after cardiac surgery. New England Journal of Medicine 358, 1229-1239.

Kozomara, A., and Griffiths-Jones, S. (2014). miRBase: annotating high confidence microRNAs using deep sequencing data. Nucleic Acids Research 42, D68-D73.

Kumar, P., Mudunuri, S.B., Anaya, J., and Dutta, A. (2015). tRFdb: a database for transfer RNA fragments. Nucleic Acids Res 43, D141-145.

Kumari, S., Badana, A.K., and Malla, R. (2018). Reactive oxygen species: a key constituent in cancer survival. Biomarker insights 13, 1177271918755391.

Kung, H.N., Marks, J.R., and Chi, J.T. (2011). Glutamine synthetase is a genetic determinant of cell type-specific glutamine independence in breast epithelia. PLoS Genet 7, e1002229.

Laleu, B., Gaggini, F., Orchard, M., Fioraso-Cartier, L., Cagnon, L., Houngninou- Molango, S., Gradia, A., Duboux, G., Merlot, C., Heitz, F., et al. (2010). First in class, potent, and orally bioavailable NADPH oxidase isoform 4 (Nox4) inhibitors for the treatment of idiopathic pulmonary fibrosis. J Med Chem 53, 7715-7730.

LaMonte, G., Philip, N., Reardon, J., Lacsina, J.R., Majoros, W., Chapman, L., Thornburg, C.D., Telen, M.J., Ohler, U., Nicchitta, C.V., et al. (2012). Translocation of sickle cell erythrocyte microRNAs into Plasmodium falciparum inhibits parasite translation and contributes to malaria resistance. Cell Host Microbe 12, 187-199.

Lau, P.W., Guiley, K.Z., De, N., Potter, C.S., Carragher, B., and MacRae, I.J. (2012). The molecular architecture of human Dicer. Nat Struct Mol Biol 19, 436-440.

Lee, F.S., and Vallee, B.L. (1993). Structure and action of mammalian ribonuclease (angiogenin) inhibitor. Prog Nucleic Acid Res Mol Biol 44, 1-30.

Lee, L.G., Chen, C.H., and Chiu, L.A. (1986). Thiazole orange: a new dye for reticulocyte analysis. Cytometry 7, 508-517.

159

Lee, R.C., Feinbaum, R.L., and Ambros, V. (1993). The C. elegans heterochronic gene lin- 4 encodes small RNAs with antisense complementarity to lin-14. Cell 75, 843-854.

Lee, S.R., and Collins, K. (2005). Starvation-induced cleavage of the tRNA anticodon loop in Tetrahymena thermophila. J Biol Chem 280, 42744-42749.

Lee, Y.S., Shibata, Y., Malhotra, A., and Dutta, A. (2009). A novel class of small RNAs: tRNA-derived RNA fragments (tRFs). Genes Dev 23, 2639-2649.

Lei, Q.Y., Zhang, H., Zhao, B., Zha, Z.Y., Bai, F., Pei, X.H., Zhao, S., Xiong, Y., and Guan, K.L. (2008). TAZ promotes cell proliferation and epithelial-mesenchymal transition and is inhibited by the hippo pathway. Mol Cell Biol 28, 2426-2436.

Leuenberger, N., Barras, L., Nicoli, R., Robinson, N., Baume, N., Lion, N., Barelli, S., Tissot, J.D., and Saugy, M. (2016). Hepcidin as a new biomarker for detecting autologous blood transfusion. Am J Hematol 91, 467-472.

Leuenberger, N., Schumacher, Y.O., Pradervand, S., Sander, T., Saugy, M., and Pottgiesser, T. (2013). Circulating microRNAs as biomarkers for detection of autologous blood transfusion. PLoS One 8, e66309.

Lewis, B.P., Burge, C.B., and Bartel, D.P. (2005). Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell 120, 15-20.

Lheureux, S., Gourley, C., Vergote, I., and Oza, A.M. (2019). Epithelial ovarian cancer. The Lancet 393, 1240-1253.

Li, D.C., Bell, J., Brown, A., and Berry, C.L. (1994). The observation of angiogenin and basic fibroblast growth factor gene expression in human colonic adenocarcinomas, gastric adenocarcinomas, and hepatocellular carcinomas. J Pathol 172, 171-175.

Li, L.Z., Zhang, C.Z., Liu, L.L., Yi, C., Lu, S.X., Zhou, X., Zhang, Z.J., Peng, Y.H., Yang, Y.Z., and Yun, J.P. (2014). miR-720 inhibits tumor invasion and migration in breast cancer by targeting TWIST1. Carcinogenesis 35, 469-478.

Li, Z., Ender, C., Meister, G., Moore, P.S., Chang, Y., and John, B. (2012). Extensive terminal and asymmetric processing of small RNAs from rRNAs, snoRNAs, snRNAs, and tRNAs. Nucleic Acids Res 40, 6787-6799.

160

Liao, Y.H., Chiang, K.H., Shieh, J.M., Huang, C.R., Shen, C.J., Huang, W.C., and Chen, B.K. (2016). Epidermal growth factor-induced ANGPTL4 enhances anoikis resistance and tumour metastasis in head and neck squamous cell carcinoma. Oncogene 36, 2228.

Lin, K.C., Park, H.W., and Guan, K.L. (2017). Regulation of the Hippo Pathway Transcription Factor TEAD. Trends Biochem Sci 42, 862-872.

Lin, L., Sabnis, A.J., Chan, E., Olivas, V., Cade, L., Pazarentzos, E., Asthana, S., Neel, D., Yan, J.J., Lu, X., et al. (2015). The Hippo effector YAP promotes resistance to RAF- and MEK-targeted cancer therapies. Nat Genet 47, 250-256.

Linkermann, A., Skouta, R., Himmerkus, N., Mulay, S.R., Dewitz, C., De Zen, F., Prokai, A., Zuchtriegel, G., Krombach, F., Welz, P.S., et al. (2014). Synchronized renal tubular cell death involves ferroptosis. Proc Natl Acad Sci U S A 111, 16836-16841.

Liou, G.-Y., and Storz, P. (2010). Reactive oxygen species in cancer. Free radical research 44, 479-496.

Lopez-Guerrero, J.A., Pecharroman, A.G., Palacios, J., Lana, E.M.C., Romero, I., Hardisson, D., Vera-Sempere, F., Illueca, C., Vieites, B.Â.a., Garcia, A., et al. (2015). Clinicopathological Significance of Taz and VGLL1 Expression in Early Stage Ovarian Cancer. A study by Spanish Group for Ovarian Cancer Research (GEICO). Journal of Clinical Oncology 33, e16583-e16583.

Luhtala, N., and Parker, R. (2010). T2 Family ribonucleases: ancient enzymes with diverse roles. Trends in Biochemical Sciences 35, 253-259.

Macrae, I.J., Zhou, K., Li, F., Repic, A., Brooks, A.N., Cande, W.Z., Adams, P.D., and Doudna, J.A. (2006). Structural basis for double-stranded RNA processing by Dicer. Science 311, 195-198.

Mantel, P.-Y., Hoang, A.N., Goldowitz, I., Potashnikova, D., Hamza, B., Vorobjev, I., Ghiran, I., Toner, M., Irimia, D., and Ivanov, A.R. (2013). Malaria-infected erythrocyte- derived microvesicles mediate cellular communication within the parasite population and with the host immune system. Cell host & microbe 13, 521-534.

Maraia, R.J., and Lamichhane, T.N. (2011). 3' processing of eukaryotic precursor tRNAs. Wiley Interdiscip Rev RNA 2, 362-375.

161

Marik, P.E., and Corwin, H.L. (2008). Efficacy of red blood cell transfusion in the critically ill: a systematic review of the literature. Critical care medicine 36, 2667-2674.

Martin, G., and Keller, W. (2007). RNA-specific ribonucleotidyl transferases. RNA-Publ RNA Soc 13, 1834-1849.

Maute, R.L., Schneider, C., Sumazin, P., Holmes, A., Califano, A., Basso, K., and Dalla- Favera, R. (2013). tRNA-derived microRNA modulates proliferation and the DNA damage response and is down-regulated in B cell lymphoma. Proc Natl Acad Sci U S A 110, 1404-1409.

Medina-Rico, M., Ramos, H.L., Lobo, M., Romo, J., and Prada, J.G. (2018). Epidemiology of renal cancer in developing countries: Review of the literature. Canadian Urological Association Journal 12, E154.

Mirza, M.R., Monk, B.J., Herrstedt, J., Oza, A.M., Mahner, S., Redondo, A., Fabbro, M., Ledermann, J.A., Lorusso, D., Vergote, I., et al. (2016). Niraparib Maintenance Therapy in Platinum-Sensitive, Recurrent Ovarian Cancer. N Engl J Med 375, 2154-2164.

Mittelbrunn, M., Gutiérrez-Vázquez, C., Villarroya-Beltri, C., González, S., Sánchez- Cabo, F., González, M.Á., Bernad, A., and Sánchez-Madrid, F. (2011). Unidirectional transfer of microRNA-loaded exosomes from T cells to antigen-presenting cells. Nature communications 2, 282.

Montecalvo, A., Larregina, A.T., Shufesky, W.J., Stolz, D.B., Sullivan, M.L., Karlsson, J.M., Baty, C.J., Gibson, G.A., Erdos, G., and Wang, Z. (2012). Mechanism of transfer of functional microRNAs between mouse dendritic cells via exosomes. Blood 119, 756-766.

Mori, M., Triboulet, R., Mohseni, M., Schlegelmilch, K., Shrestha, K., Camargo, F.D., and Gregory, R.I. (2014). Hippo signaling regulates microprocessor and links cell-density- dependent miRNA biogenesis to cancer. Cell 156, 893-906.

Mukherjee, K., Campos, H., and Kolaczkowski, B. (2013). Evolution of animal and plant dicers: early parallel duplications and recurrent adaptation of antiviral RNA binding in plants. Mol Biol Evol 30, 627-641.

Murata, M., Yudo, K., Nakamura, H., Chiba, J., Okamoto, K., Suematsu, N., Nishioka, K., Beppu, M., Inoue, K., Kato, T., et al. (2009). Hypoxia upregulates the expression of angiopoietin-like-4 in human articular chondrocytes: role of angiopoietin-like-4 in the

162

expression of matrix metalloproteinases and cartilage degradation. J Orthop Res 27, 50- 57.

Nonaka, R., Miyake, Y., Hata, T., Kagawa, Y., Kato, T., Osawa, H., Nishimura, J., Ikenaga, M., Murata, K., Uemura, M., et al. (2015). Circulating miR-103 and miR-720 as novel serum biomarkers for patients with colorectal cancer. International Journal of Oncology 47, 1097-1102.

Nowacka, M., Strozycki, P.M., Jackowiak, P., Hojka-Osinska, A., Szymanski, M., and Figlerowicz, M. (2013). Identification of stable, high copy number, medium-sized RNA degradation intermediates that accumulate in plants under non-stress conditions. Plant molecular biology 83, 191-204.

O'Brien, C.A., and Harley, J.B. (1990). A subset of hY RNAs is associated with erythrocyte Ro ribonucleoproteins. The EMBO journal 9, 3683.

Offner, P.J., Moore, E.E., Biffl, W.L., Johnson, J.L., and Silliman, C.C. (2002). Increased rate of infection associated with transfusion of old blood after severe injury. Archives of Surgery 137, 711-717.

Olson, K.A., Byers, H.R., Key, M.E., and Fett, J.W. (2001). Prevention of human prostate tumor metastasis in athymic mice by antisense targeting of human angiogenin. Clin Cancer Res 7, 3598-3605.

Olson, K.A., Fett, J.W., French, T.C., Key, M.E., and Vallee, B.L. (1995). Angiogenin antagonists prevent tumor growth in vivo. Proc Natl Acad Sci U S A 92, 442-446.

Olson, K.A., Verselis, S.J., and Fett, J.W. (1998). Angiogenin is regulated in vivo as an acute phase protein. Biochem Biophys Res Commun 242, 480-483.

Ozols, R.F. (2006). Systemic therapy for ovarian cancer: current status and new treatments. Semin Oncol 33, S3-11.

Padua, D., Zhang, X.H., Wang, Q., Nadal, C., Gerald, W.L., Gomis, R.R., and Massague, J. (2008). TGFbeta primes breast tumors for lung metastasis seeding through angiopoietin-like 4. Cell 133, 66-77.

Panzilius, E., Holstein, F., Bannier-Hélaouët, M., von Toerne, C., Koenig, A.-C., Hauck, S.M., Ganz, H.M., Angeli, J.P.F., Conrad, M., and Scheel, C.H. (2018). A cell-density dependent metabolic switch sensitizes breast cancer cells to ferroptosis. bioRxiv, 417949.

163

Pap, E.H.W., Drummen, G.P.C., Winter, V.J., Kooij, T.W.A., Rijken, P., Wirtz, K.W.A., Op den Kamp, J.A.F., Hage, W.J., and Post, J.A. (1999). Ratio-fluorescence microscopy of lipid oxidation in living cells using C11-BODIPY581/591. FEBS Letters 453, 278-282.

Park, I.H., Kang, J.H., Lee, K.S., Nam, S., Ro, J., and Kim, J.H. (2014). Identification and clinical implications of circulating microRNAs for estrogen receptor-positive breast cancer. Tumor Biology 35, 12173-12180.

Park, S., Ahn, J.-Y., Lim, M.-J., Kim, M.-H., Yun, Y.-S., Jeong, G., and Song, J.-Y. (2010). Sustained expression of NADPH oxidase 4 by p38 MAPK-Akt signaling potentiates radiation-induced differentiation of lung fibroblasts. Journal of molecular medicine 88, 807-816.

Pederson, T. (2010). Regulatory RNAs derived from transfer RNA? Rna 16, 1865-1869.

Pelicano, H., Carney, D., and Huang, P. (2004). ROS stress in cancer cells and therapeutic implications. Drug resistance updates 7, 97-110.

Peng, C., Zhu, Y., Zhang, W., Liao, Q., Chen, Y., Zhao, X., Guo, Q., Shen, P., Zhen, B., Qian, X., et al. (2017). Regulation of the Hippo-YAP Pathway by Glucose Sensor O- GlcNAcylation. Mol Cell 68, 591-604 e595.

Peng, H., Li, W., Seth, D.M., Nair, A.R., Francis, J., and Feng, Y. (2013). (Pro) renin receptor mediates both angiotensin II-dependent and-independent oxidative stress in neuronal cells. PloS one 8, e58339.

Perner, A., Andresen, L., Pedersen, G., and Rask-Madsen, J. (2003). Superoxide production and expression of NAD(P)H oxidases by transformed and primary human colonic epithelial cells. Gut 52, 231-236.

Piccoli, R., Olson, K.A., Vallee, B.L., and Fett, J.W. (1998). Chimeric anti-angiogenin antibody cAb 26-2F inhibits the formation of human breast cancer xenografts in athymic mice. Proc Natl Acad Sci U S A 95, 4579-4583.

Pizzo, E., Sarcinelli, C., Sheng, J., Fusco, S., Formiggini, F., Netti, P., Yu, W., D'Alessio, G., and Hu, G.F. (2013). Ribonuclease/angiogenin inhibitor 1 regulates stress-induced subcellular localization of angiogenin to control growth and survival. J Cell Sci 126, 4308-4319.

164

Poursaitidis, I., Wang, X., Crighton, T., Labuschagne, C., Mason, D., Cramer, S.L., Triplett, K., Roy, R., Pardo, O.E., Seckl, M.J., et al. (2017). Oncogene-Selective Sensitivity to Synchronous Cell Death following Modulation of the Amino Acid Nutrient Cystine. Cell Rep 18, 2547-2556.

Previs, R.A., Armaiz-Pena, G.N., Ivan, C., Dalton, H.J., Rupaimoole, R., Hansen, J.M., Lyons, Y., Huang, J., Haemmerle, M., Wagner, M.J., et al. (2017). Role of YAP1 as a Marker of Sensitivity to Dual AKT and P70S6K Inhibition in Ovarian and Uterine Malignancies. JNCI: Journal of the National Cancer Institute 109.

Ragusa, M., Statello, L., Maugeri, M., Majorana, A., Barbagallo, D., Salito, L., Sammito, M., Santonocito, M., Angelica, R., Cavallaro, A., et al. (2012). Specific alterations of the microRNA transcriptome and global network structure in colorectal cancer after treatment with MAPK/ERK inhibitors. Journal of Molecular Medicine-Jmm 90, 1421- 1438.

Rathjen, T., Nicol, C., McConkey, G., and Dalmay, T. (2006). Analysis of short RNAs in the malaria parasite and its red blood cell host. FEBS Lett 580, 5185-5188.

Roback, J.D., Josephson, C.D., Waller, E.K., Newman, J.L., Karatela, S., Uppal, K., Jones, D.P., Zimring, J.C., and Dumont, L.J. (2014). Metabolomics of ADSOL (AS-1) red blood cell storage. Transfus Med Rev 28, 41-55.

Rubin, O., Crettaz, D., Canellini, G., Tissot, J.D., and Lion, N. (2008). Microparticles in stored red blood cells: an approach using flow cytometry and proteomic tools. Vox sanguinis 95, 288-297.

Ruegger, S., and Grosshans, H. (2012). MicroRNA turnover: when, how, and why. Trends Biochem Sci 37, 436-446.

Salamin, O., De Angelis, S., Tissot, J.D., Saugy, M., and Leuenberger, N. (2016). Autologous Blood Transfusion in Sports: Emerging Biomarkers. Transfusion Medicine Reviews 30, 109-115.

Sand, M., Skrygan, M., Sand, D., Georgas, D., Gambichler, T., Hahn, S.A., Altmeyer, P., and Bechara, F.G. (2013). Comparative microarray analysis of microRNA expression profiles in primary cutaneous malignant melanoma, cutaneous malignant melanoma metastases, and benign melanocytic nevi. Cell and Tissue Research 351, 85-98.

165

Sangokoya, C., LaMonte, G., and Chi, J.T. (2010a). Isolation and characterization of microRNAs of human mature erythrocytes. Methods in molecular biology (Clifton, NJ) 667, 193-203.

Sangokoya, C., Telen, M.J., and Chi, J.T. (2010b). microRNA miR-144 modulates oxidative stress tolerance and associates with anemia severity in sickle cell disease. Blood 116, 4338-4348.

Santulli, G. (2014). Angiopoietin-like proteins: a comprehensive look. Front Endocrinol (Lausanne) 5, 4.

Sarachana, T., Kulkarni, S., and Atreya, C.D. (2015). Evaluation of small noncoding RNAs in ex vivo stored human mature red blood cells: changes in noncoding RNA levels correlate with storage lesion events. Transfusion 55, 2672-2683.

Sato, M., Kusumi, R., Hamashima, S., Kobayashi, S., Sasaki, S., Komiyama, Y., Izumikawa, T., Conrad, M., Bannai, S., and Sato, H. (2018). The ferroptosis inducer erastin irreversibly inhibits system xc− and synergizes with cisplatin to increase cisplatin’s cytotoxicity in cancer cells. Scientific Reports 8, 968.

Saxena, S.K., Rybak, S.M., Davey, R.T., Youle, R.J., and Ackerman, E.J. (1992). Angiogenin is a cytotoxic, tRNA-specific ribonuclease in the RNase A superfamily. Journal of Biological Chemistry 267, 21982-21986.

Schieber, M., and Chandel, N.S. (2014). ROS function in redox signaling and oxidative stress. Curr Biol 24, R453-462.

Schopman, N.C., Heynen, S., Haasnoot, J., and Berkhout, B. (2010). A miRNA-tRNA mix-up: tRNA origin of proposed miRNA. RNA Biol 7, 573-576.

Schutte, U., Bisht, S., Heukamp, L.C., Kebschull, M., Florin, A., Haarmann, J., Hoffmann, P., Bendas, G., Buettner, R., Brossart, P., et al. (2014). Hippo signaling mediates proliferation, invasiveness, and metastatic potential of clear cell renal cell carcinoma. Transl Oncol 7, 309-321.

Sedeek, M., Nasrallah, R., Touyz, R.M., and Hébert, R.L. (2013). NADPH Oxidases, Reactive Oxygen Species, and the Kidney: Friend and Foe. Journal of the American Society of Nephrology 24, 1512-1518.

166

Setzer, F., Oberle, V., Bläss, M., Möller, E., Russwurm, S., Deigner, H.-P., Claus, R.A., Bauer, M., Reinhart, K., and Lösche, W. (2006). Platelet-derived microvesicles induce differential gene expression in monocytic cells: a DNA microarray study. Platelets 17, 571-576.

Shapiro, R., Riordan, J.F., and Vallee, B.L. (1986). Characteristic ribonucleolytic activity of human angiogenin. Biochemistry 25, 3527-3532.

Shapiro, R., Strydom, D.J., Olson, K.A., and Vallee, B.L. (1987). Isolation of angiogenin from normal human plasma. Biochemistry 26, 5141-5146.

Shapiro, R., and Vallee, B.L. (1987). Human placental ribonuclease inhibitor abolishes both angiogenic and ribonucleolytic activities of angiogenin. Proc Natl Acad Sci U S A 84, 2238-2241.

Sharma, U., Conine, C.C., Shea, J.M., Boskovic, A., Derr, A.G., Bing, X.Y., Belleannee, C., Kucukural, A., Serra, R.W., and Sun, F. (2016). Biogenesis and function of tRNA fragments during sperm maturation and fertilization in mammals. Science 351, 391-396.

Shen, J., Zhao, W., Ju, Z., Wang, L., Peng, Y., Labrie, M., Yap, T.A., Mills, G.B., and Peng, G. (2019). PARPi Triggers the STING-Dependent Immune Response and Enhances the Therapeutic Efficacy of Immune Checkpoint Blockade Independent of BRCAness. Cancer research 79, 311-319.

Shigematsu, M., Honda, S., and Kirino, Y. (2014). Transfer RNA as a source of small functional RNA. J Mol Biol Mol Imaging 1.

Shimoyama, S., Gansauge, F., Gansauge, S., Negri, G., Oohara, T., and Beger, H.G. (1996). Increased Angiogenin Expression in Pancreatic Cancer Is Related to Cancer Aggressiveness. Cancer research 56, 2703-2706.

Shinozuka, E., Miyashita, M., Mizuguchi, Y., Akagi, I., Kikuchi, K., Makino, H., Matsutani, T., Hagiwara, N., Nomura, T., Uchida, E., et al. (2013). SnoN/SKIL modulates proliferation through control of hsa-miR-720 transcription in esophageal cancer cells. Biochemical and Biophysical Research Communications 430, 101-106.

Simpson, C.D., Anyiwe, K., and Schimmer, A.D. (2008). Anoikis resistance and tumor metastasis. Cancer Lett 272, 177-185.

167

Skouta, R., Dixon, S.J., Wang, J., Dunn, D.E., Orman, M., Shimada, K., Rosenberg, P.A., Lo, D.C., Weinberg, J.M., and Linkermann, A. (2014). Ferrostatins inhibit oxidative lipid damage and cell death in diverse disease models. Journal of the American Chemical Society 136, 4551-4556.

Smirnoff, P., Roiz, L., Angelkovitch, B., Schwartz, B., and Shoseyov, O. (2006). A recombinant human RNASET2 glycoprotein with antitumorigenic and antiangiogenic characteristics. Cancer 107, 2760-2769.

Smit, M.A., Geiger, T.R., Song, J.Y., Gitelman, I., and Peeper, D.S. (2009). A Twist-Snail axis critical for TrkB-induced epithelial-mesenchymal transition-like transformation, anoikis resistance, and metastasis. Mol Cell Biol 29, 3722-3737.

Sourbier, C., Liao, P.J., Ricketts, C.J., Wei, D., Yang, Y., Baranes, S.M., Gibbs, B.K., Ohanjanian, L., Spencer Krane, L., Scroggins, B.T., et al. (2018). Targeting loss of the Hippo signaling pathway in NF2-deficient papillary kidney cancers. Oncotarget 9, 10723-10733.

Stein, C., Bardet, A.F., Roma, G., Bergling, S., Clay, I., Ruchti, A., Agarinis, C., Schmelzle, T., Bouwmeester, T., Schübeler, D., et al. (2015). YAP1 Exerts Its Transcriptional Control via TEAD-Mediated Activation of Enhancers. PLOS Genetics 11, e1005465.

Stockwell, B.R., Friedmann Angeli, J.P., Bayir, H., Bush, A.I., Conrad, M., Dixon, S.J., Fulda, S., Gascon, S., Hatzios, S.K., Kagan, V.E., et al. (2017). Ferroptosis: A Regulated Cell Death Nexus Linking Metabolism, Redox Biology, and Disease. Cell 171, 273-285.

Strydom, D.J., Fett, J.W., Lobb, R.R., Alderman, E.M., Bethune, J.L., Riordan, J.F., and Vallee, B.L. (1985). Amino acid sequence of human tumor derived angiogenin. Biochemistry 24, 5486-5494.

Sun, H., and Ying, M. (2015). Abstract 4444: Small molecule drug Verteporfin inhibits TAZ/YAP-driven signaling and tumorigenicity of breast cancer cells. Cancer Research 75, 4444-4444.

Tan, G., Cao, X., Dai, Q., Zhang, B., Huang, J., Xiong, S., Zhang, Y., Chen, W., Yang, J., and Li, H. (2015). A novel role for microRNA-129-5p in inhibiting ovarian cancer cell proliferation and survival via direct suppression of transcriptional co-activators YAP and TAZ. Oncotarget 6, 8676-8686.

168

Tanaka, J., Irie, T., Yamamoto, G., Yasuhara, R., Isobe, T., Hokazono, C., Tachikawa, T., Kohno, Y., and Mishima, K. (2015). ANGPTL4 regulates the metastatic potential of oral squamous cell carcinoma. J Oral Pathol Med 44, 126-133.

Tang, X., Ding, C.K., Wu, J., Sjol, J., Wardell, S., Spasojevic, I., George, D., McDonnell, D.P., Hsu, D.S., Chang, J.T., et al. (2017a). Cystine addiction of triple-negative breast cancer associated with EMT augmented death signaling. Oncogene 36, 4235-4242.

Tang, X., Ding, C.K., Wu, J., Sjol, J., Wardell, S., Spasojevic, I., George, D., McDonnell, D.P., Hsu, D.S., Chang, J.T., et al. (2017b). Cystine addiction of triple-negative breast cancer associated with EMT augmented death signaling. Oncogene 36, 4379.

Tang, X., Keenan, M.M., Wu, J., Lin, C.-A., Dubois, L., Thompson, J.W., Freedland, S.J., Murphy, S.K., and Chi, J.-T. (2015a). Comprehensive Profiling of Amino Acid Response Uncovers Unique Methionine-Deprived Response Dependent on Intact Creatine Biosynthesis. PLOS Genetics 11, e1005158.

Tang, X., Wu, J., Ding, C.K., Lu, M., Keenan, M.M., Lin, C.C., Lin, C.A., Wang, C.C., George, D., Hsu, D.S., et al. (2016). Cystine Deprivation Triggers Programmed Necrosis in VHL-Deficient Renal Cell Carcinomas. Cancer research 76, 1892-1903.

Tang, Y.L., Lin, Y., Li, C., Hu, X.W., Liu, Y., He, M.Y., Luo, J., Sun, G.H., Wang, T., Li, W.X., et al. (2015b). MicroRNA-720 promotes in vitro cell migration by targeting Rab35 expression in cervical cancer cells. Cell and Bioscience 5.

Tarangelo, A., Magtanong, L., Bieging-Rolett, K.T., Li, Y., Ye, J., Attardi, L.D., and Dixon, S.J. (2018). p53 Suppresses Metabolic Stress-Induced Ferroptosis in Cancer Cells. Cell Rep 22, 569-575.

Terada, L.S., and Nwariaku, F.E. (2011). Escaping Anoikis through ROS: ANGPTL4 controls integrin signaling through Nox1. Cancer Cell 19, 297-299.

Thevis, M., Kuuranne, T., Geyer, H., and Schanzer, W. (2014). Annual banned-substance review: analytical approaches in human sports drug testing. Drug Test Anal 6, 164-184.

Thompson, D.M., Lu, C., Green, P.J., and Parker, R. (2008). tRNA cleavage is a conserved response to oxidative stress in eukaryotes. RNA 14, 2095-2103.

169

Thompson, D.M., and Parker, R. (2009a). The RNase Rny1p cleaves tRNAs and promotes cell death during oxidative stress in Saccharomyces cerevisiae. J Cell Biol 185, 43-50.

Thompson, D.M., and Parker, R. (2009b). Stressing Out over tRNA Cleavage. Cell 138, 215-219.

Thorn, A., Steinfeld, R., Ziegenbein, M., Grapp, M., Hsiao, H.H., Urlaub, H., Sheldrick, G.M., Gartner, J., and Kratzner, R. (2012). Structure and activity of the only human RNase T2. Nucleic Acids Res 40, 8733-8742.

Tinmouth, A., Fergusson, D., Yee, I.C., and Hébert, P.C. (2006). Clinical consequences of red cell storage in the critically ill. Transfusion 46, 2014-2027.

Tu, B., Liu, L., Xu, C., Zhai, J., Li, S., Lopez, M.A., Zhao, Y., Yu, Y., Ramachandran, V., Ren, G., et al. (2015). Distinct and cooperative activities of HESO1 and URT1 nucleotidyl transferases in microRNA turnover in Arabidopsis. PLoS Genet 11, e1005119.

Uronis, J.M., Osada, T., McCall, S., Yang, X.Y., Mantyh, C., Morse, M.A., Lyerly, H.K., Clary, B.M., and Hsu, D.S. (2012). Histological and molecular evaluation of patient- derived colorectal cancer explants. PLoS One 7, e38422.

Van 't Erve, T.J., Wagner, B.A., Martin, S.M., Knudson, C.M., Blendowski, R., Keaton, M., Holt, T., Hess, J.R., Buettner, G.R., Ryckman, K.K., et al. (2015). The heritability of hemolysis in stored human red blood cells. Transfusion 55, 1178-1185.

Vander Heiden, M.G. (2011). Targeting cancer metabolism: a therapeutic window opens. Nat Rev Drug Discov 10, 671-684.

Varelas, X. (2014). The Hippo pathway effectors TAZ and YAP in development, homeostasis and disease. Development 141, 1614-1626.

Viswanathan, V.S., Ryan, M.J., Dhruv, H.D., Gill, S., Eichhoff, O.M., Seashore-Ludlow, B., Kaffenberger, S.D., Eaton, J.K., Shimada, K., Aguirre, A.J., et al. (2017). Dependency of a therapy-resistant state of cancer cells on a lipid peroxidase pathway. Nature 547, 453- 457.

Walko, G., Woodhouse, S., Pisco, A.O., Rognoni, E., Liakath-Ali, K., Lichtenberger, B.M., Mishra, A., Telerman, S.B., Viswanathan, P., Logtenberg, M., et al. (2017). A genome-

170

wide screen identifies YAP/WBP2 interplay conferring growth advantage on human epidermal stem cells. Nature Communications 8, 14744.

Wang, D., Sun, J., Solomon, S.B., Klein, H.G., and Natanson, C. (2012). Transfusion of older stored blood and risk of death: a meta‐analysis. Transfusion 52, 1184-1195.

Wang, X., Kuang, Y.T., Shen, X.C., Zhou, H., Chen, Y., Han, Y., Yuan, B., Zhou, J., Zhao, H., Zhi, Q.M., et al. (2015a). Evaluation of miR-720 prognostic significance in patients with colorectal cancer. Tumor Biology 36, 719-727.

Wang, Y., Zhang, Z., Ji, D., Chen, G.F., Feng, X., Gong, L.L., Guo, J., Li, Z.W., Chen, C.F., Zhao, B.B., et al. (2015b). Regulation of T cell function by microRNA-720. Sci Rep 5, 12159.

Weinberg, J.A., McGwin Jr, G., Marques, M.B., Cherry III, S.A., Reiff, D.A., Kerby, J.D., and Rue III, L.W. (2008). Transfusions in the less severely injured: does age of transfused blood affect outcomes? Journal of Trauma and Acute Care Surgery 65, 794-798.

Weisenhorn, E.M., van, T.E.T.J., Riley, N.M., Hess, J.R., Raife, T.J., and Coon, J.J. (2016). Multi-omics Evidence for Inheritance of Energy Pathways in Red Blood Cells. Molecular & cellular proteomics : MCP 15, 3614-3623.

Whitaker, B., Rajbhandary, S., and Harris, A. (2015). The 2013 AABB blood collection, utilization, and patient blood management survey report. American Association of Blood Banks.

Wolfe, L.C. (1985). The membrane and the lesions of storage in preserved red cells. Transfusion 25, 185-203.

Wu, J., Minikes, A.M., Gao, M., Bian, H., Li, Y., Stockwell, B.R., Chen, Z.N., and Jiang, X. (2019). Intercellular interaction dictates cancer cell ferroptosis via NF2-YAP signalling. Nature.

Xie, Y., Hou, W., Song, X., Yu, Y., Huang, J., Sun, X., Kang, R., and Tang, D. (2016). Ferroptosis: process and function. Cell Death Differ 23, 369-379.

Xie, Y., Zhu, S., Song, X., Sun, X., Fan, Y., Liu, J., Zhong, M., Yuan, H., Zhang, L., Billiar, T.R., et al. (2017). The Tumor Suppressor p53 Limits Ferroptosis by Blocking DPP4 Activity. Cell Rep 20, 1692-1704.

171

Yamasaki, S., Ivanov, P., Hu, G.F., and Anderson, P. (2009). Angiogenin cleaves tRNA and promotes stress-induced translational repression. Journal of Cell Biology 185, 35-42.

Yang, W.-H., Huang, Z., Wu, J., Ding, C.-K.C., Murphy, S.K., and Chi, J.-T. (2019a). A TAZ-ANGPTL4-NOX2 axis regulates ferroptotic cell death and chemoresistance in epithelial ovarian cancer. Molecular Cancer Research, molcanres.0691.2019.

Yang, W.H., Ding, C.C., Sun, T., Rupprecht, G., Lin, C.C., Hsu, D., and Chi, J.T. (2019b). The Hippo Pathway Effector TAZ Regulates Ferroptosis in Renal Cell Carcinoma. Cell Rep 28, 2501-2508.e2504.

Yang, W.S., SriRamaratnam, R., Welsch, M.E., Shimada, K., Skouta, R., Viswanathan, V.S., Cheah, J.H., Clemons, P.A., Shamji, A.F., Clish, C.B., et al. (2014). Regulation of Ferroptotic Cancer Cell Death by GPX4. Cell 156, 317-331.

Yang, W.S., and Stockwell, B.R. (2016). Ferroptosis: Death by Lipid Peroxidation. Trends in Cell Biology 26, 165-176.

Yi, Z., Fu, Y., Zhao, S., Zhang, X., and Ma, C. (2010). Differential expression of miRNA patterns in renal cell carcinoma and nontumorous tissues. J Cancer Res Clin Oncol 136, 855-862.

Yoshida, T., Prudent, M., and D'Alessandro, A. (2019). Red blood cell storage lesion: causes and potential clinical consequences. Blood transfusion = Trasfusione del sangue 17, 27-52.

Yu, F.-X., Zhao, B., and Guan, K.-L. (2015). Hippo pathway in organ size control, tissue homeostasis, and cancer. Cell 163, 811-828.

Zanconato, F., Cordenonsi, M., and Piccolo, S. (2016). YAP/TAZ at the Roots of Cancer. Cancer Cell 29, 783-803.

Zanconato, F., Forcato, M., Battilana, G., Azzolin, L., Quaranta, E., Bodega, B., Rosato, A., Bicciato, S., Cordenonsi, M., and Piccolo, S. (2015). Genome-wide association between YAP/TAZ/TEAD and AP-1 at enhancers drives oncogenic growth. Nature cell biology 17, 1218.

Zhang, Q., Han, X., Chen, J., Xie, X., Xu, J., Zhao, Y., Shen, J., Hu, L., Xu, P., and Song, H. (2018). Yes-associated protein (YAP) and transcriptional coactivator with PDZ-binding

172

motif (TAZ) mediate cell density–dependent proinflammatory responses. Journal of Biological Chemistry 293, 18071-18085.

Zhang, X., George, J., Deb, S., Degoutin, J.L., Takano, E.A., Fox, S.B., group, A.S., Bowtell, D.D.L., and Harvey, K.F. (2011). The Hippo pathway transcriptional coactivator, YAP, is an ovarian cancer oncogene. Oncogene 30, 2810.

Zhang, Z., Qin, Y.W., Brewer, G., and Jing, Q. (2012). MicroRNA degradation and turnover: regulating the regulators. Wiley Interdiscip Rev RNA 3, 593-600.

Zhao, B., Wei, X., Li, W., Udan, R.S., Yang, Q., Kim, J., Xie, J., Ikenoue, T., Yu, J., Li, L., et al. (2007). Inactivation of YAP oncoprotein by the Hippo pathway is involved in cell contact inhibition and tissue growth control. Genes & Development 21, 2747-2761.

Zhao, B., Ye, X., Yu, J., Li, L., Li, W., Li, S., Yu, J., Lin, J.D., Wang, C.-Y., and Chinnaiyan, A.M. (2008). TEAD mediates YAP-dependent gene induction and growth control. Genes & development 22, 1962-1971.

Zhu, P., Tan, M.J., Huang, R.L., Tan, C.K., Chong, H.C., Pal, M., Lam, C.R., Boukamp, P., Pan, J.Y., Tan, S.H., et al. (2011). Angiopoietin-like 4 protein elevates the prosurvival intracellular O2(-):H2O2 ratio and confers anoikis resistance to tumors. Cancer Cell 19, 401-415.

Zhu, X., Guo, X., Wu, S., and Wei, L. (2016). ANGPTL4 Correlates with NSCLC Progression and Regulates Epithelial-Mesenchymal Transition via ERK Pathway. Lung 194, 637-646.

173

Biography

Wen-Hsuan Yang graduated from National Taiwan University in Taipei, Taiwan with a degree of Bachelor of Science in Life Science and a certificate in Biotechnology in

May 2010. She then obtained her master's degree in Biochemistry at National Taiwan

University under the mentorship of Dr. Ching-Jin Chang in May 2012. In 2014, she enrolled in the Biochemistry program at Duke University. In January 2017, she joined the lab of Dr. Jen-Tsan Ashley Chi to pursue and obtain her doctoral degree in

Biochemistry.

Publications Yang, W. H., Ding, C. K., Sun, T., Rupprecht, G., Lin, C. C., Hsu, D. Chi, J. T. (2019) The Hippo Pathway Effector TAZ Regulates Ferroptosis in Renal Cell Carcinoma. Cell Reports 28(10):2501-8.e4.

Yang, W. H., Huang, Z., Wu, J., Ding, C. K., Murphy S. K., Chi, J. T. (2019) A TAZ- ANGPTL4-NOX2 axis regulates ferroptotic cell death and chemoresistance in epithelial ovarian cancer. Molecular Cancer Research DOI: 10.1158/1541-7786.MCR-19-0691

Yang, W. H., Chi, J. T. (2019) Hippo pathway effectors YAP/TAZ as novel determinants of ferroptosis. Molecular & Cellular Oncology accepted

Park, H. S., Eldridge, W. J., Yang, W. H., Crose, M., Ceballos, S., Roback, J. D., Chi, J. T., Wax A. (2019) Quantitative phase imaging of erythrocytes under microfluidic constriction in a high refractive index medium reveals water content changes. Nature Microsystems & Nanoengineering accepted

Chen, P. H., Wu, J., Ding, C. K., Lin, C. C., Pan, S., Bossa, N., Xu, Y., Yang, W. H., Mathey-Prevot, B., Chi, J. T. (2019) Kinome screen of ferroptosis reveals a novel role of ATM in regulating iron metabolism Cell Death and Differentiation Jul 18:1.

174

Yang, W. H.*, Doss, J. F.*, Walzer, K. A.*, McNulty, S. M., Wu, J., Roback, J. D., & Chi, J. T. (2019). Angiogenin‐mediated tRNA cleavage as a novel feature of stored red blood cells. British journal of haematology, 185(4), 760-764.

White, P. J., McGarrah, R. W., Grimsrud, P. A., Tso, S. C., Yang, W. H., Haldeman, J. M., ... & Hannou, S. A. (2018). The BCKDH Kinase and Phosphatase Integrate BCAA and Lipid Metabolism via Regulation of ATP-Citrate Lyase. Cell metabolism, 27(6), 1281- 1293.

Lin, N.Y.*, Lin, T.Y.*, Yang, W.H.*, Wang, S.C., Wang, K.T., Su, Y.L., Jiang, Y.W., Chang, G.D., and Chang, C.J. (2012). Differential expression and functional analysis of the tristetraprolin family during early differentiation of 3T3-L1 preadipocytes. Int J Biol Sci 8, 761-777.

Yeh, P.A.*, Yang, W.H.*, Chiang, P.Y.*, Wang, S.C., Chang, M.S., and Chang, C.J. (2012). Drosophila eyes absent is a novel mRNA target of the tristetraprolin (TTP) protein DTIS11. Int J Biol Sci 8, 606-619.

* denotes co-first authors

175