A Genome Wide Screen in C. Elegans Identifies Cell Non-Autonomous Regulators of Oncogenic Ras Mediated Over-Proliferation DISSER

A genome wide screen in C. elegans identifies cell non-autonomous regulators of oncogenic Ras mediated over-proliferation

DISSERTATION

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

Komal Rambani

Graduate Program in Biomedical Sciences

The Ohio State University

2016

Dissertation Committee: Gustavo Leone, PhD “Advisor" Helen Chamberlin, PhD Gregory Lesinski, PhD Joanna Groden, PhD Jeffrey Parvin, PhD Thomas Ludwig, PhD

Komal Rambani

2016

ABSTRACT

Coordinated proliferative signals from the mesenchymal cells play a crucial role in the regulation of proliferation of epithelial cells during normal development, wound healing and several other normal physiological conditions. However, when epithelial cells acquire a set of malignant mutations, they respond differently to these extrinsic proliferative signals elicited by the surrounding mesenchymal cells. This scenario leads to a pathological signaling microenvironment that enhances abnormal proliferation of mutant epithelial cells and hence tumor growth. Despite mounting evidence that mesenchymal (stromal) cells influence the growth of tumors and cancer progression, it is unclear which specific genes in the mesenchymal cells regulate the molecular signals that promote the over-proliferation of the adjacent mutant epithelial cells. We hypothesized that there are certain genes in the mesenchymal (stromal) cells that regulate proliferation of the adjacent mutant cells. The complexity of various stromal cell types and their interactions in vivo in cancer mouse models and human tumor samples limits our ability to identify mesenchymal genes important in this process. Thus, we took a cross-species approach to use C. elegans vulval development as a model to understand the impact of mesenchymal (mesodermal) cells on the proliferation of epithelial (epidermal) cells. This model includes well-described signaling interactions between mesenchymal cells (anchor cell, gonad and muscle cells) and epidermal cells (vulva precursor cells, VPCs). In this system, three of the six equipotent VPCs normally divide to produce vulval tissue. These

ii divisions result from coordination of Egf and Wnt signals from mesenchymal cells, and lateral Notch signaling between the VPCs. The anchor cell in the gonad (the mesenchymal cells) produces an LIN-3/EGF signal, and the most proximal VPC in the underlying epidermal cell (P6.p) is induced to adopt a vulval fate (termed 1°), and to promote its neighbors (P5.p and P7.p) to adopt a distinct vulval fate (termed 2°) via LIN-

12/Notch. Following signaling, these three cells go through three rounds of invariant cell division to form a symmetric vulva cavity comprised of 22 daughter cells. Abnormal activation of the Egf pathway, such as through introducing gain-of-function (gf) mutations in let-60/Ras, leads to the promotion of vulval fates in the remaining three

VPCs, which then divide inappropriately to develop multiple ectopic vulval protrusions

(the multivulva, or Muv, phenotype). To identify genes that function in mesodermal cells to inhibit the proliferation of Ras-mutant epidermal cells, we genetically engineered C. elegans to develop a model that possess let-60(gf)/Ras mutation mediated over- proliferation (Muv phenotype) and RNA interference (RNAi) competence specific to mesenchymal tissues only. This strain was subjected to a genome-wide RNAi screen using standard methods. Our screen identified 47 genes that, upon RNAi mediated knockdown of a gene in mesodermal cells, suppress over-proliferation of let-60(gf)/Ras epithelial cells in an allele-specific manner. Notably, these genes encode histone proteins, ribosomal proteins, vesicle transport proteins and metabolic factors, rather than secreted molecules. Importantly, candidate genes emerged from this screen are significantly enriched for the conserved genes and signaling-pathways from C. elegans to

iii humans. These results form the basis of a comprehensive understanding of molecular factors in mesenchymal cells that impact oncogenic cell proliferation.

Dedication

This thesis is dedicated to my son, Ritav Das, my husband, my family and God – with

infinite love and gratitude.

My son, you are infinite source of my inspiration and window to look at things through the lens of your curious and creative ideas and questions. May you find your own ways to

discover the world and contribute to solving significant and meaningful problems with

your superb intellect, creativity and curiosity!

Acknowledgments

Research is a challenging and time-consuming endeavor; the written portion is only a small portion of the total work done. For me, the greatest benefit of doing PhD is to learn about a new field of research (cancer) and develop as an independent thinker and problem solver – which have prepared me for my future endeavors. Not only I learned about various thinking and experimental styles to solve scientific problems, I also learned a big deal to work with various personalities in team-oriented projects. I sincerely thank my dissertation committee members and my adivsor, for advice and valuable suggestions to my research and graduate work. Their guidance was instrumental in focusing efforts and completing work presented herein. I am tremendously thankful to BSGP directors, Dr. Groden and Dr. Parvin, and administrators, Amy Lahmers and Lauren Graham, for their incredible mentorship, guidance and support through these years. After Dr. Bloomston moved from OSU, Dr. Parvin and Dr. Groden generously accepted to mentor me as committee members in place of him. When I joined the lab, I had some career goals; I discussed and took steps to achieve them – some succeeded, some did not, irrespective my efforts. I want to pursue translational cancer research even though this project culminated with basic research part only. From this, I have learned several important skills that will benefit in future. All of the BSGP classes were instrumental in gaining broad and rich knowledge on various topics and honing critical thinking; some of them will stay in my memories for long. Several lectures of Professors Tsowin Hai, Larry Kirschner, Thomas Boyd, James Waldman, Deborah Parris, Harold Fisk, James Chen, Raghu Machiraju, and Kun Huang had special impact on me. They brought knowledge, excitement, challenge, thinking on- the-fly and an urge to students to draw rational conclusions along with take-home messages like “Think alternatives” and “correlations are not always causal relationships”. The grant writing class of BSGP is very valuable and was instrumental in success of my

vi fellowship applications. While taking these classes and beyond, I am thankful for great company of my classmates and friends. I am also thankful to all the past and present members of the Leone lab. I thank Raleigh Kladney, Serena Chang, Markus Harrigan, James Dowdle, Elizabeth Brunner, Xing Tang for their contributions to work presented in this thesis. I am grateful to Susan Lutz for her support and help on lab related administrative tasks. I am tremendously thankful to a long list of my friends on BRT 5th floor and my friends across the campus for their company and good times spent with them! A special thanks to Jyotsana, Vijay, Kirana, Jaya, Sriram, Margi, Tarak, Lauren, Sarika, Neelam and Arunima for their unconditional friendship! Above all, I am grateful to my incredible son and husband for their unwavering support throughout these years – without which, I am certain, I would not be writing this thesis today. A special thanks goes to my family members: Ritav, Jayajit, Vikas, Vishal, Arijit, Suruchi, Jasleen, Aratrika, Aariv, Aniket, my mother, my parents-in-law and my extended family for their immense love and support in numerous ways. Without their loving support and understanding, I would not have made through this experience. I am lucky to have delightful and marvelous sister-in-laws, Jasleen, Suruchi and Aratika, who always stand by my side in all my high-n-low times and I cherish their loving company, whenever we talk or meet. I am fortunate to have brothers like Vikas, Vishal and Arijit, whose unconditional love, support and respect makes me proudest women. My darlings – Ritav, Aariv and Aniket – are the sweetest charm of my life whose child-wisdom added to my ability to manage my time efficiently at work. Through these years, I also endured pain of missing those family members whom I would never be able to meet again ever – Sita didi, Onkar Jijaji, baby Anushka and Humpy Dada, my next visit to home would certainly make me realize your absence more than ever!

vii

Vita

2007 ...... M.S., Georgia Institute of Technology, Atlanta, GA

2010 - present ...... Graduate Research Associate, The Ohio State University, Columbus, OH

2013-2015 …………………………………...Pelotonia Graduate Fellow, The Ohio State University, Columbus, OH

Publications

1. Rambani K, Kladney RD, Chang SW, Harrigan ME, Dowdle JE, Tang X, Liu H, Brunner E, Chamberlin HM, and Leone G. A genome-wide screen in C. elegans identifies cell non-autonomous suppressors of oncogenic Ras mediated over- proliferation. in preparation.

2. Liu H*, Dowdle JA*, Khurshid S*, Sullivan NJ*, Bertos N, Rambani K, Mair M, Daniel P, Toth K, Lause M, Harrigan ME, Eiring K, Sullivan C, Chang SW, Kladney RD, Tang X, McElroy J, Lu Y, Tofigh A, Fernandez SA, Parvin JD, Macrae E, Majumder S, Shapiro CL, Yee LD, Hallett M, Ostrowski MC, Park M, Chamberlin HM, and Leone G (*Equal contribution). Discovery of stromal regulatory networks that suppress Ras-sensitized cell proliferation. submitted.

Fields of Study

Major Field: Biomedical Sciences

viii

Table of Contents

Abstract ...... ii

Dedication ...... v

Acknowledgments ...... vi

Vita ...... viii

List of Tables ...... xiii

List of Figures ...... xiiiii

Chapter 1: Introduction ...... 1

1.1 Hallmarks of Cancer...... 1

1.2 Tumor microenvironment...... 2

1.3 Mesodermal-epithelial cell interactions in Tumor microenvironment...... 4

1.4 Mesodermal-epithelial cell interactions in C. elegans vulva development model...... 6

1.5 Role of Ras, Wnt and Notch signaling in vulva development of C. elegans...... 8

1.6 Importance of Screens using C. elegans vulva development phenotype...... 11

1.7 Genetically engineered C. elegans models...... 12

1.8 Organization of thesis: Hypothesis and study overview...... 14

Figures...... 17

Chapter 2: A genetic screen to identify cell non-autonomous suppressors of hyper- proliferation ...... 32

2.1 Introduction...... 32

2.2 Results...... 34

2.3 Discussion...... 37

2.4 Materials and methods ...... 40

Figures...... 45

Chapter 3: Identified cell non-autonomous candidate genes converge on select conserved gene families or molecular complexes ...... 51

3.1 Introduction...... 51

3.2 Results...... 51

3.3 Discussion...... 52

Figures...... 56

Chapter 4: Candidate genes specifically suppress proliferation of let-60 mutant epithelial cells ...... 58

4.1 Introduction...... 58

4.2 Results and Discussion...... 60

4.3 Materials and methods ...... 63

Figures...... 65

Chapter 5: Spatial function of candidate genes in suppression of let-60(gf) mediated hyper-proliferation ...... 70

5.1 Introduction...... 70

5.2 Results and Discussion...... 71

5.3 Materials and methods ...... 73

Figures...... 76

Chapter 6: Concluding remarks and future perspectives ...... 81

References ...... 87

List of Tables

Table 2.1: List of final candidates genes ...... 49

Table 3.1: List of hits arranged in Functional categories ...... 57

Table 4.1: List of candidate genes and their effect in mesodermal-RNAi strains with and without let-60(gf) mutation...... 69

Table 5.1: Most of candidate genes function in cell non-autonomous manner...... 80

xii

List of Figures

Figure 1.1. Stromal component increases with tumor growth...... 17

Figure 1.2. Pten in stromal fibroblasts enhances ErBb2 tumor growth and Table of known stromal regulators ...... 18

Figure 1.3. Conserved Ras signaling pathway between C. elegans and mammals ...... 20

Figure 1.4. C. elegans vulva phenotypes ...... 21

Figure 1.5. C. elegans vulvagenesis and interplay of Ras, Wnt, and Notch pathways .... 23

Figure 1.6. Genetically engineered tissue specific RNAi models of C. elegans...... 25

Figure 1.7. Functional test of the tissue-specific RNAi strains of C. elegans ...... 27

Figure 1.8. let-60(gf) sensitized tissue-specific RNAi C. elegans models and their phenotypes ...... 30

Figure 2.1. A genome wide RNAi screen for cell non-autonomous suppressor genes of oncogenic over-proliferation ...... 45

Figure 2.2. Rescreening and validation of candidate genes ...... 47

Figure 3.1. Candidate genes conservation analysis in Humans ...... 56

Figure 4.1. C. elegans transgenic model and targeted screen ...... 65

Figure 4.2. Results of targeted mini-screen of candidate genes in mesenchymal tissue- specific RNAi strain ...... 67

xiii

Figure 4.3. Comparison of effect of stromal-specific RNAi on suppression of over- proliferation of let-60(gf)-mutant versus suppression of normal let-60(wt) mediated proliferation of VPCs ...... 68

Figure 5.1. C. elegans model and targeted screen in VPC-tissue specific RNAi strain ... 76

Figure 5.2. Results of targeted mini-screen of candidate genes in vulva epithelium tissue- specific RNAi strain ...... 78

Figure 5.3. Comparison of effect of stromal-specific RNAi to VPC-specific RNAi on suppression of over-proliferation of let-60(gf)-mutant VPCs ...... 79

xiv

Chapter 1 Introduction

What are monsters and dragons? Do you know anything about them? - Ritav, 2.5 year old (Unpublished).

1.1 Hallmarks of Cancer

Cancer is a leading cause of death worldwide and accounted for 7.6 million deaths

(around 13% of all deaths) in 2008. This trend is projected to continue to rise to over 13 million in 2030 (International Agency for Research on Cancer (IARC)). In nearly past half century, in the pursuit to understand the etiology of these malignancies, the predominant focus of cancer research was on the oncogenes and tumor suppressor genes in the malignant cancer cell (Malumbres and Barbacid, 2003; Payne and Kemp, 2005;

Lodish H et al., 2000, Lee and Muller, 2010, Hanahan and Wienberg, 2000). Innumerable studies of various human cancers have inexorably linked tumor growth to these genetic lesions in the cancer cells (Watson and Chin, 2013; Watson et al., 2013; Vogelstein et al.,

2013; Pon and Marra, 2015; Martincorena and Campbell, 2015). It is in this elegant but rather single-dimensional context that much of what is known about cancer-cell biology has been revealed. These studies have revealed many critical regulatory signaling pathways that control cell proliferation and growth, apoptosis, cell migration, altered metabolism and other important molecular processes that are often deregulated during cancer progression. In 2000 and 2011, Hanahan and Weinberg summarized these studies

1 and suggested that the most cancers are manifestation of eight essential alterations in cell physiology that collectively dictate malignant growth and are shared across most cancer types: (1) limitless replicative potential, (2) insensitivity to growth-inhibitory

(antigrowth) signals, (3) evasion of programmed cell death (apoptosis), (4) self- sufficiency in growth signals, (5) sustained angiogenesis, (6) tissue invasion and metastasis, (7) avoiding immune destruction and (8) deregulating cellular energetics

(Hanahan and Weinberg 2000, Hanahan and Weinberg 2011). Each of these hallmark characteristics represents successful infringing of anticancer mechanisms that are evolutionarily ingrained into normal development and normal physiological conditions such as wound healing, involution of breast tissue, renewal of uterine tissue periodically in response to hormonal cycle, etc.

1.2 Tumor microenvironment

Histologically, solid tumors are composed of a rich assortment of cell types such as epithelial, fibroblasts, macrophages and other immune cells, rather than just clonal expansion of cancer cells. Just as normal organs consist of various cell types that constantly communicate by means of intercellular signaling to maintain tissue homeostasis, the cells in the tumor microenvironment also constantly communicate via inter-cellular signaling among various cell types (Wiseman and Werb, 2002; Hill and

Van Dyke, 2005; Provenzano et al., 2006; Cunha, 2008; Egeblad et al., 2008; Egeblad and Werb, 2010; Puri and Hebrok, 2010). However, the intercellular signaling in neoplastic growths is deregulated due to residing epithelial (cancer) cells harboring

2 oncogenic lesions in various signaling pathways compared to normal organs. Thus, tumors can be considered as abnormally developing organs with de-regulated intercellular signaling as a result of genetic lesions. The stromal compartment is believed to co-evolve with malignant cells and increases as the tumor grows (Figure 1.1) (Polyak and Campbell, 2009; Josson, and Wang, 2010; Erkan, 2013).

In recent reports, the tumor microenvironment has been shown to play an indispensable role in tumor initiation, maintenance and its contribution to some of the most destructive hallmark characteristics of malignant cells, such as rapid proliferation, invasiveness, metastasis, evading growth suppressors, promoting survival, and evasion of apoptosis and avoiding immune destruction, and reprogrammed energy metabolism

(Hanahan and Folkman, 1996; Bergers and Hanahan, 1999; Orimo et al., 2005; Kalluri and Zeisberg, 2006; Finak et al., 2008; Gabrilovich and Nagaraj, 2009; Erez and

Hanahan, 2010; Franco and Hayward, 2010; Cirri and Chiarugi, 2011; Chow and Merad,

2011; Mantovani and Jaillon, 2011; Porta et al., 2011; Nieman et al., 2011; Beck et al.,

2011; Daldrup-Link et al., 2011; Rattigan et al., 2012; Hanahan & Coussens, 2012;

Martinez-Outschoorn and Lisanti, 2014). Diverse functional contributions of stromal cells significantly promote and contribute to the hallmark properties of the cancers

(Bergers and Hanahan, 1999; Orimo et al., 2005; Kalluri and Zeisberg, 2006; Gabrilovich and Nagaraj, 2009; Franco and Hayward, 2010; Chow and Merad, 2011; Mantovani and

Jaillon, 2011; Porta et al., 2011; Nieman et al., 2011; Beck et al., 2011; Rattigan et al.,

2012; Martinez-Outschoorn and Lisanti, 2014). An emerging concept is that besides cell autonomous potential for over proliferation, mutant cells residing within the tumor

3 microenvironment robustly depend on cell extrinsic signals from stromal cells for maintenance and unlimited proliferation (Gaggioli et al., 2007; DeNardo and Coussens,

2010; Qian and Pollard, 2010; Tan and Rio, 2011; Feitelson et al., 2015). Complex tumor-stromal cells inter-cellular signaling interactions during carcinogenesis are considered to be responsible to instigate pathologic changes that nurture over- proliferation of mutant cells and recruitment of reprogrammed stromal cells and escape of mutant cells from their primary site to locations in distant sites. Currently, not much is known about these signaling interactions.

1.3 Mesodermal-epithelial cell interactions in Tumor microenvironment

It has been now well recognized that stromal cells play a significant role in disease progression (Figure 1.1), however the precise function of genes or signaling networks in stromal cells that render tumor microenvironment conductive to cancer cell proliferation and hence tumor growth remains unknown. To understand biological properties of tumor development and progression at molecular level, detailed understanding of the pathological signaling interactions between cancer cells and their microenvironment is necessary. Thus, investigation of functional contribution of various stromal cells in the tumor microenvironment such as fibroblasts, endothelial cells, immune cells is emerging subject of cancer research today. In recent reports, tumor- stroma functions have been identified only as one gene at a time, and have thus failed to reveal the complexity of interactions that likely exists in inter-cellular signaling between tumor and stromal cells (Figure 1.2) (Bhowmick et al., 2004; Hill et al., 2005; Kim et al.,

2006; Yang et al., 2008; Katajisto et al., 2008; Yauch et al., 2008; Trimboli et al., 2009;

Fu et al., 2011; Sherman et al., 2014). Interestingly, most of these genes are not secreted factors rather are specific genes in stromal cells that regulate various downstream effectors to elicit their effect that supports proliferation and maintenance of tumor cells.

One of these pioneering studies was performed in our laboratory, which has revealed and characterized the role of Pten in stromal cells in emanating proliferative signals to

ErBb2-mutant epithelial cells thereby fostering their tumorigenic potential (Trimboli et al., 2009). These studies were done in mouse models (Figure 1.2B). In summary, ablation of Pten in stromal fibroblasts led to dramatically increased proliferation of

ErBb2-mutant epithelial cells. Subsequent studies demonstrated that stromal Pten ablation resulted in an oncogenic secretome mediated by the Pten - mir-320 - Ets2 signaling axis (Bronisz et al., 2011). Collectively, these results clearly demonstrated that activity of genes in adjacent stromal cells strongly influence the proliferation rate of adjacent mutant epithelial cells in vivo. These studies, along with other emerging reports of the role of stromal genes in promoting tumor growth, led us to hypothesize that there are yet undefined genetic networks in the stromal (mesenchymal) cells that regulate proliferation of the adjacent mutant cells in the tumor microenvironment. Thus, we set out to systematically identify stromal genes at a genome-wide scale in an unbiased manner, that play causal role in the regulation of proliferation of mutant epithelial cells.

Currently, one of the research goals of our laboratory is to identify: Which stromal genes could potentially elicit pathological signals to mutant cells that foster their proliferation and unravel deregulated signaling interactions among stromal and tumor cells?

To unravel the functional genes in a biological process, genetic models and genome-wide screens are powerful tools. For example, genetic screens in S. cerevisiae,

C. elegans and Drosophila and biochemical studies in cell culture systems unraveled components of conserved canonical Ras signaling pathway. In deed, several of the components of this pathway were first identified in C. elegans that later were found to be conserved in their functionality across the species (Figure 1.3) (Lackner and Kim, 1994;

Kornfeld and Horvitz, 1995; Sundaram and Han, 1995; Sundaram and Han, 1996; Gu and

Han, 1998; Li and Guan, 2000; Kao and Sundaram, 2004; Cui and Han, 2006). Towards our research goal to identify causal genes and genetic networks in mesenchymal (stromal) cells that collaborate with the aberrant signaling status of mutant epithelial cells, we took cross species approach to utilize C. elegans vulva development as model system.

1.4 Mesodermal-epithelial cell interactions in C. elegans vulva development model

C. elegans vulva development is a well characterized model system that involves coordinated signaling among mesodermal-epithelial and epithelial-epithelial cells interactions. Primarily three highly conserved signaling pathways, EGFR-Ras-MAPK,

Wnt and Notch, orchestrate these intercellular interactions during vulva development.

Genetic alterations in components of these pathways lead to aberrant (Vulvaless, Vul or

Multivulva, Muv) vulva development (Figure 1.4) (Horvitz and Sulston, 1980; Ferguson and Horvitz, 1985; Beitel and Horvitz, 1990; Han and Sternberg, 1990; Aroian and

Sternberg, 1991; Clark and Horvitz, 1992; Clark and Horvitz, 1993; Han and Sternberg,

1993; Seydoux and Greenwald, 1993; Eisenmann and Kim, 1994).

In the early years, C. elegans researchers performed several chemically induced genome-wide mutational screening studies focused on vulval induction that had revealed many conserved genes that function in the canonical EGFR/Ras pathway that are conserved in humans and are relevant to human biology and disease (reviewed in

Sternberg, 2006) (Figure 1.3). In deed, several seminal discoveries were first made in C. elegans vulva genetic screening studies in identification and delineation of the components of Ras signal transduction pathway, which was later found to be highly conserved between C. elegans and mammals.

Genetic alterations in genes involved in vulval-fate specification exhibit either vulvaless (Vul) phenotype or multivulva (Muv) phenotype. Vul or Muv mutants are defective in the complex intercellular (mesodermal-epithelial cells or epithelial-epithelial cells) signaling interactions. Mutants of these two types (Vul or Muv) led to the identification and characterization of both intercellular (cell-cell) signaling pathways and intracellular signaling cascades that direct vulval cell-fate specification (Horvitz and

Sulston, 1980; Ferguson and Horvitz, 1985; Beitel and Horvitz, 1990; Han and Sternberg,

1990; Aroian and Sternberg, 1991; Clark and Horvitz, 1992; Clark and Horvitz, 1993;

Han and Sternberg, 1993; Seydoux and Greenwald, 1993; Eisenmann and Kim, 1994). In vulvaless (Vul) mutants, none of the vulval precursor cells (P3.p to P8.p) adopts a vulval fate that could lead to division of cells in stereotypic manner as in normal case to produce daughter cells of vulva organ (Sulston and White, 1980; Kimble, 1981; Sternberg and

Horvitz, 1986). In multivulva (Muv) mutants, all six cells (P3.p to P8.p) adopt vulval fates that lead to production of more number of daughter cells than normal vulva case and

7 form more than one vulva cavities (Beitel and Horvitz, 1990; Han and Sternberg, 1990;

Aroian and Sternberg, 1991; Hill and Sternberg, 1992; Katz and Sternberg, 1995)

(Figure 1.4).

1.5 Role of Ras, Wnt and Notch signaling in Normal Vulva development of C. elegans

The normal vulva development process in C. elegans can be broadly described in temporally overlapping developmental stages. In the L1 larva stage, there are initially 11 progenitor cells (P1.p to P11.p) on the ventral side of the nematode (Sulston and Horvitz,

1977). Out of these eleven progenitor cells, six equipotent cells (P3.p to P8.p) in the middle part of the body commit to vulva precursor cell (VPC) fate (also known as vulva equivalence group) and the rest adopt the hypodermal cell fates (Sulston and Horvitz,

1977; Clark et al., 1993; Salser and Kenyon, 1993). It is important for VPCs to maintain their vulva fate and not fuse with the large hyp7 syncytium that forms large part of hypodermis (skin) of the worms. The Ras pathway, Wnt pathway and Notch pathway are required for normal vulval cell-fate specification (Greenwald, 1997; Eisenmann and Kim,

1998; Greenwald and Horvitz, 1983; Gleason and Eisenmann, 2002; Sundaram and Han,

1996; Sundaram, 2004). To retain their VPC fates, these (P3.p to P8.p) cells express lin-

39 gene. lin-39 inhibits expression of eff-1 gene, which is known for promoting cell fusion (Eisenmann et al., 1998; Mohler et al., 2002; Shemer & Podbilewicz, 2002). The

Wnt pathway promotes the expression of the lin-39, a Hox transcription factor, which inhibits fusion of these cells to hypodermis (Eisenmann et al., 1998; Mohler et al., 2002;

Shemer & Podbilewicz, 2002).

Next, three of the six vulva progenitor cells (P5.p, P6.p, P7.p) commit to the vulva fates (1º or 2º). This requires activation of EGFR-Ras pathway and its interplay with the activity of Notch and Wnt pathways. While Wnt pathway promotes induced vulval cell fates (vulva equivalence group) in cooperation with Ras pathway, the Ras and Notch pathways oppose each other to specify 1o and 2o vulval fates, respectively. (Sundaram,

2004). First, LIN-12/Notch influences vulval fates indirectly by fate specification of the anchor cell (AC) at L2/L3 stage (Greenwald, 1983). AC is a specialized cell of gonad that is primary source of the Ras pathway ligand lin-3/EGF (Hill and Sternberg, 1992). The

AC sends the inductive signal, lin-3/EGF, to the underlying VPCs. The P6.p is in the closest physical proximity to the AC and receives maximum dose of lin-3 signal, which binds to EGFR/let-23 receptor and activates canonical EGFR/Ras signaling cascade

(Figure 1.5) (Hill and Sternberg, 1992; Beitel and Horvitz, 1990; Han and Sternberg,

1990; Ferguson and Horvitz, 1987; Moghal and Sternberg, 2003, Tan and Kim, 1999).

The adjacent cells, P5.p and P7.p, which are physically distant from AC compared to

P6.p, receive lesser level of lin-3 signal (Katz et al., 1995; Katz et al., 1996). The Ras pathway promotes the 1o vulval fate and inhibits the 2o vulval fate (Sternberg & Horvitz,

1986; Katz et al., 1996). On the contrary, the Notch pathway down-regulates Ras pathway activity thereby promoting the 2o vulval fate and inhibiting the 1o vulval fate

(Greenwald et al., 1983). Since P6.p receives more lin-3/EGF ligand from AC due to its physical proximity to AC, the Ras pathway is normally most active in this cell, which leads to production of DSL ligands at higher levels compared to the neighboring (P5.p and P7.p) cells. In turn, this leads to greater lateral Notch pathway activity in the

9 neighboring (P5.p and P7.p) cells (Greenwald et al., 1983; Chen & Greenwald, 2004).

Concomitantly, the Ras pathway also leads to internalization of LIN-12/Notch receptors in P6.p leading to diminished activity of Notch pathway in P6.p leading to its 1o vulval cell fate (Shaye and Greenwald, 2002). Consecutively, the Notch pathway activity in P5.p and P7.p diminishes Ras pathway activity in P5.p and P7.p leading to their 2º cells (Yoo and Greenwald, 2004). The remaining cells (P3.p, P4.p, P8.p) receive minimum level of lin-3/EGF signal leading to adoption of 3º fate. The P5.p, P6.p, P7.p cells with acquired

1º or 2º fate divide three times to form symmetrically arranged 22 cells of the vulva cavity. The P5.p and P7.p (2º cell fate) divide 3 times to produce 7 daughter cells each, while P6.p (1º cell fate) divide 3 times to produce 8 daughter cells. The 3º cells divide once and form hypodermal cells. In the final step, during and following the division of cells, 22 cells arrange in seven ring-shaped structure and the fully functional organ is formed after making appropriate connections to vulval muscles, vulval neurons and uterus (Newman & Sternberg, 1996; Burdine and Sternberg, 1998; Inoue et al., 2002;

Sherwood and Sternberg, 2003; Inoue and Sternberg, 2005). The interplay of conserved signaling pathways, the stereotypical phenotype (22 vulva cells forming a symmetric functional organ) and the interactions of mesodermal-epithelial cells make C. elegans vulvagenesis a powerful model to study mesodermal-epithelial cell interactions in the context of mutations (gain of function or loss of function) in the components of these signaling pathways.

1.6 Importance of Screens using C. elegans vulva development phenotype

Ras pathway is highly conserved between C. elegans and humans. Several of the earlier studies to identify conserved signaling pathways, were done in model organisms like S. cerevisiae, C. elegans, Drosophila and cell culture (Huang and Sternberg, 2006;

Simon, 1994; Williams and Roberts, 1994; Troppmair et al., 1994; McKay and Morrison,

2007). Indeed, several components of the Ras signaling pathway were first identified in

C. elegans using genetic screens and later found to have conserved function of their orthologs in other species, including humans (Figure 1.3). Several genetic screens were conducted to identify genes that inhibit or induce vulval cell fates in wild type or sensitized genetic backgrounds, such as suppressors or enhancers of Vul phenotype or

Muv phenotypes caused by mutations in the Ras-pathway components or regulators.

These screens helped not only in identifying genes that play role in vulva development but also in our understanding of conserved signaling pathways at molecular level.

A set of genes that emerged from some of these screens is known as synthetic multivulva (synMuv) genes (Fay and Yochem, 2007). These genes induce Muv phenotype only when knocked down simultaneously, hence their name, synMuv.

Relevant to intercellular signaling interactions, the synMuv genes generally inhibit vulval fates in C. elegans. This inhibition of the vulval fate is demonstrated to be through cell non-autonomous regulation of Ras pathway activity by transcriptional repression of lin-

3/EGF (the Ras pathway ligand) by chromatin remodeling synMuv genes (Cui et al.,

2006). So, with appropriately designed screen, C. elegans vulva model not only can provides insights about the genes contributing to the human carcinogenesis (e.g., gain of

11 function or loss of function mutations in the components of Ras signaling pathway leading to over- or under- proliferation), but we might also learn about the deregulated mesenchymal-epithelial signaling interactions in normal and sensitized backgrounds

(such as gain-of-function Ras mutations) using genetically-engineered versions of this model.

1.7 Genetically engineered C. elegans models

The mesenchymal-epithelial cells signaling interactions, techniques to genetically engineering the model to introduce tissue-specific RNAi, clear phenotypes and above all its importance in identifying signaling pathways and their components, made C. elegans vulvagenesis an attractive model to address our research quest: Which mesenchymal genes might causally regulate proliferation of adjacent mutant epithelial cells? To adapt this model to our queries, past members of our laboratory, genetically-engineered this system in collaboration with Dr. Helen Chamberlin’s laboratory. A brief explanation of the system is given here to introduce the strains that were used to cross with let-60(gf) mutant strain of C. elegans to make experimental strains used in this thesis.

Two tissue-specific RNAi-responsive strains were developed using mesodermal tissues-specific promoters (gonad, AC and muscles) and vulva epithelial tissue-specific promoter (VPC), respectively. The specificity of the promoters was tested first by cloning green fluorescent protein (GFP) under their control and injecting them in C. elegans to visualize expression of GFP (Figure 1.6). GFP was then replaced with a wild type copy of rde-1 gene in these clones. The gene rde-1 encodes a critical component of

Argonaute/PIWI family of proteins that is required for RNA-mediated interference

(RNAi), especially in RNAi induced by exogenous double stranded RNA in C. elegans

(Tabara et al., 1999). A loss of function point mutation in rde-1 renders the nematodes non-responsive to RNAi (Tabara et al., 1999; Qadota et al., 2007). To achieve tissue- specific RNAi effect, colleagues in our lab have reintroduced rde-1 transgenes driven by mesenchymal tissue-specific (anchor cell, somatic gonad and muscle) or vulva precursor cells-specific (VPCs) promoters into an rde-1(-) mutant strain of C. elegans such that

RNAi mediated knock-down of targeted genes occurs exclusively in the rde-1(+) rescued tissues (targeted mesodermal tissues or vulva epithelial tissue specific). The strains were tested functionally for tissue specificity using RNAi against a mesodermal/AC specific gene, Lin-3/EGF or VPC specific gene, lin-39. In wild type strain of C. elegans, RNAi against the lin-3 (mesenchyme-specific gene) and the lin-39 (VPC/epithelial cell-specific gene), knock-down the target genes that lead to vulvaless (Vul) phenotype (Figure 1.7A,

B). In rde-1(-) strain of C. elegans, the RNAi of both lin-3 and lin-39 genes has no effect, as expected, leading to normal vulva formation (Figure 1.7C, D).. In the genetically- engineered, mesenchymal-tissue specific RNAi strain, RNAi against the lin-3 (a mesenchyme-specific gene) leads to vulvaless (Vul) or under-induced vulva phenotype

((Figure 1.7E), but RNAi against lin-39 (a VPC-specific gene) does not have any effect, leading to normal vulva development (Figure 1.7F). This experiments demonstrated mesenchymal tissue-specific RNAi nature of this genetically-engineered strain (Figure

1.7E, F). In contrast, in VPC/epithelial tissue-specific strain, RNAi against the lin-3 (a mesenchyme-specific gene) does not have any effect and leads to normal vulva

13 development (Figure 1.7G), but RNAi against lin-39 (a VPC-specific gene) leads to vulvaless (Vul) or under-induced vulva phenotype (Figure 1.7H). This experiments demonstrated VPC-specific RNAi nature of this genetically-engineered strain (Figure

1.7G, H). ( Liu et al. (submitted)).

To these genetically-engineered tissue-specific RNAi strains, the let-60(gf) and rrf-3(-) strain was crossed to generate the experimental models for this study (Figure

1.8). The gene, rrf-3, encodes an RNA-directed RNA polymerase (RdRP) homolog. Loss of function rrf-3(-) mutants are hypersensitive to somatic RNAi (Simmer et al., 2002).

This allele is generally introduced to enhance the effect of RNAi to help reduce the false negatives in the RNAi based screening studies.

1.8 Organization of thesis: Hypothesis and study overview

The importance of mutant Ras in cancer is unquestionable. While a great amount of knowledge has been revealed about binary switch (active and inactive) like nature of

Ras proteins and activating Ras mutations (Figure 1.8A), the canonical

EGFR/RAS/MAPK pathway and other downstream effectors of Ras and its role in cancers; there remain several questions that are unanswered in the context of signaling interactions of Ras-mutant cells with surrounding cell in the microenvironment. One particular question we are interested in is identifying role of cell non-autonomous genetic networks that collaborate with mutant cells to regulate their proliferation. To understand the role of interactions of mesenchymal cells in the tumor microenvironment with Ras- mutant cells and their malignant properties, our overarching goal is to investigate “cell

14 extrinsic signals” aspects of oncogenic Ras signaling collaborations: the signaling collaborations between mutant cells and the mesenchymal cells in influencing the proliferation of mutant cells. We hypothesized that there are yet undefined genetic networks in the mesenchymal/stromal cells that have causal effect in the regulation of proliferation of the adjacent mutant cells. To test this hypothesis, we took cross-species approach to identify mesenchymal genes that can suppress over-proliferation of adjacent

Ras-mutant cells in vivo at a genome-wide scale.

In Chapter 2, we tested the hypothesis that certain genes in the mesenchymal tissues can suppress over-proliferation of Ras-mutant epithelial cells. To perform genome-wide unbiased screen, we used cross-species approach to utilize C. elegans vulva development as model system. The Genetically engineered C. elegans strains were developed that have mutant let-60(gf)/Ras mutation mediated over-proliferation of epithelial cells (VPCs) and RNAi functionality specific to mesodermal tissues that send instructive signals to the VPCs for proliferating. Using this mesodermal tissue-specific

RNAi strain that also possesses let-60(gf) mutation, we performed a genome wide screen to identify genes in mesodermal (mesenchymal) cells that suppress over-proliferation of the let-60(gf)/Ras mutant epithelial cells. The primary screen hits were rescreened two more times, followed by validation screen done at L4 stage to evaluate the phenotype at cellular level using 100X Nomarski microscopy. These analyses resulted in 47 genes that cell non-autonomously suppressed over-proliferation of let-60(gf) mutant cells.

In Chapter 3, we identify the human orthologs of the genes and discuss the

15 functional categories of the identified cell non-autonomous genes.

In Chapter 4, we evaluated whether these identified cell non-autonomous genes suppress proliferation of only the let-60(gf)-mutant cells or have effect on the proliferation of normal epithelial (VPC) as well.

In Chapter 5, we evaluated whether these genes function exclusively in the cell non-autonomous manner or function in both mesodermal and epithelial cells to exert their effect on the suppression of over-proliferation.

Chapter 6 presents the concluding remarks and future directions for these studies.

We believe in exciting future for follow-up of this project in various directions and in depth studies to reveal role of these genes in mammalian Ras-driven cancer microenvironment.

Normal Tumors

Figure 1.1: Stromal component increases with tumor growth. In normal tissue, the stromal fibroblasts (aSMA, pink) are very few residing within the normal tissue architecture. As the tumor grows, the stromal component with in the tumor also increases.

Images: pancreas or pancreatic tumor sections. Stain: aSMA. All images 10X magnification. (Images: Raleigh Kladney)

NATURE | Vol 461 | 22 October 2009 ARTICLES no expression in cytokeratin-positive epithelial cells, F4/80-positive small nuclei, fine chromatin and abundant eosinophilic cytoplasm18. macrophages and CD31-positive endothelial cells (Fig. 1a and In contrast to non-deleted tumours18,19, Pten stromal-deleted tumours Supplementary Fig. 2a, b). Western blot and PCR assays demon- had a significant amount of stroma surrounding and infiltrating the strated efficient cre-mediated deletion of PtenloxP in stromal fibro- epithelial masses (Fig. 1g). PCR-based and immunohistochemical blasts isolated from Fsp-cre;PtenloxP/loxP mammary glands (Fig. 1b assays confirmed that tumours had intact PtenloxP alleles in the epi- and Supplementary Fig. 3a). Examination of mammary sections by thelial compartment (Supplementary Fig. 5a, b and data not shown). immunohistochemistry (IHC) and immunofluorescence showed Moreover, we used the Rosa26loxP reporter allele to mark genetically deletion of PtenloxP that was confined to stromal fibroblasts, with early epithelial to mesenchymal transition events15 and found no evid- no collateral deletion in epithelial ducts or the adjacent myoepithe- ence of epithelial to mesenchymal transition in tumours that either lium (Fig. 1c and Supplementary Fig. 3b, c). Interestingly, this contained or lacked Pten in stromal fibroblasts (data not shown). resulted in the expansion of the ECM, but did not lead to the trans- Thus, the analysis of the ErbB2 breast-cancer tumour model identified formation of the mammary epithelium (Fig. 1c, e). a potent tumour suppressor role for Pten in stromal fibroblasts of the We then examined the role of stromal Pten on mammary tumori- mammary gland. genesis using an established mouse model of breast cancer, MMTV- ErbB2/neu (ErbB2)16. To avoid possible confounding effects caused Stromal Pten controls ECM and innate immune functions by Pten deletion in mesenchymal cells of other organs, mammary To investigate the tumour suppressive mechanism of Pten action in glands from Fsp-cre;PtenloxP/loxP, ErbB2;PtenloxP/loxP and ErbB2;Fsp- stromal fibroblasts, we profiled the transcriptome of mammary stromal cre;PtenloxP/loxP donors were transplanted into syngeneic wild-type fibroblasts isolated from PtenloxP/loxP and Fsp-cre;PtenloxP/loxP females. recipients17 and tumour development was monitored over the course Details of sample collection, processing of Affymetrix oligo-arrays and of several months. By genetically marking the stroma with the expression data are available in Methods. Briefly, we implemented class Rosa26LoxP reporter allele, we demonstrated that both the epithelium comparison analyses of all probe sets on the Affymetrix mouse genome and its associated stroma were effectively transplanted into host 430 2.0 array to identify genes differentially expressed between the two female mice (Supplementary Fig. 4). Loss of Pten in stromal fibroblasts genetic groups. We also used an unbiased approach similar to gene set dramatically increased the incidence of ErbB2-driven mammary enrichment analysis20 to identify aprioridefined groups of genes that tumours (Fig. 1d–f). By 16 weeks post-transplantation, these lesions were significantly differentially expressed. The analysis of over 14,000 progressed to adenoma, carcinoma in situ and invasive carcinoma mouse genes identified 129 upregulated and 21 downregulated unique (Fig. 1g), and by 26 weeks most females met the criteria for early genes in response to Pten deletion (Supplementary Fig. 6a, b; greater removal due to excessive tumour burden (Fig. 1f). Histological exami- than fourfold at P , 0.001; Supplementary Tables 1 and 2). Reverse nation showed that ErbB2-tumour cells in Pten stromal-deleted transcription followed by quantitative PCR (quantitative RT–PCR) tumours retained their typical oncogene-specific morphology, with assays on a subset of genes confirmed more than 85% of these expression

abFsp-cre;Rosa+/loxP –+ –+: Fsp-cre d Pten Tubulin ErbB2; ErbB2; PtenloxP/loxP Fsp-cre;PtenloxP/loxP

Rosa+/loxP c PtenloxP/loxP Fsp-cre;PtenloxP/loxP -galactosidase Pten β

1.0 cm 1.0 cm lu 6.8 × 4.8 × 1.9 mm 14.4 × 16.7 × 8.3 mm epi Figure 1.2: Pten in stromal fibroblasts enhances ErBb2 tumor growth and Table of lu epi epi str Pten str knowne stromal regulators f str P = 0.051 -galactosidase β 100 10 n = 8 n = 9 (A) List of knownP = 0.024 cell non-autonomous regulators of tumorigenic properties of cancer g ErbB2;PtenloxP/loxP ErbB2;Fsp-cre;PtenloxP/loxP cells.80 These studies were done as a single8 candidate gene at a time in various laboratories. 0% 17% 63% 0 weeks 0 weeks 12 weeks 16 weeks 4.71 ± 2.86 These60 studies clearly demonstrate role6 of genes in stromal cells (listed in column 2) in 40 4

fostering tumorigenicity = 5 of cancer cells (type of cancer studied listed in column 3). The = 12 = 16 n n n

20 load (g) Tumour 2 study marked in red was conducted in our0.73 laboratory. ± 1.02 (continued) 0 0 – + + :ErbB2 18+ + :ErbB2 with tumours (%) Transplants + – + :Fsp-cre – + :Fsp-cre + + + :PtenloxP/loxP + + :PtenloxP/loxP Haematoxylin and eosin Figure 1 | Stromal fibroblast-specific deletion of Pten.a, Wholemount, d, Tumours collected at 26 weeks post-transplantation. e, Tumour X-gal-stained mammary glands from Fsp-cre;Rosa1/loxP and Rosa1/loxP (top, development by 16 weeks in mammary glands with the indicated genotypes. inset) mice. Higher magnification of wholemount gland (bottom left) and a Tumorigenicity was determined by palpation or histological presentation of histological cross section (bottom right); scale bar, 30 mm. lu, Lumen; epi, adenoma/carcinoma at each implantation site and statistically analysed epithelium; str, stroma. b, Representative western blot analysis of mammary using Fisher’s exact test. n, Total number of transplants. f, Total tumour fibroblast lysates derived from 8-week-old PtenloxP/loxP mice with (1)or burden at 26 weeks post-transplantation in mammary glands with the without (2) Fsp-cre. c, Paraffin sections from 8-week-old female mammary indicated genotypes. Values represent mean 6 s.d. Differences were tested glands stained with a Pten-specific antibody; lower panels represent higher using the non-parametric Wilcoxon rank sum test. g, H&E-stained sections magnifications of boxed areas; scale bars: top panels, 200 mm; bottom panels, of mammary glands harvested at time of transplantation (0 weeks) and 30 mm. lu, Lumen; epi, epithelial compartment; str, stromal compartment; indicated times post-transplantation; scale bars, 100 mm. red dotted line indicates the border between the two compartments. 1085 ©2009 Macmillan Publishers Limited. All rights reserved Figure 1.2: Pten in stromal fibroblasts enhances ErBb2 tumor growth and Table of known stromal regulators (continued)

(B) (Left) ErBb2 mediated tumors in mammary glands of mice. Pten is floxed in this model. (Right) After crossing with FSP-cre, a fibroblast specific deletion of floxed Pten was achieved in the ErBb2 mutant mice. Ablation of Pten in fibroblast compartment led to dramatically increased size of tumors. Figures from Trimboli, et al., 2009.

Figure 1.3: Conserved Ras signaling pathway between C. elegans and mammals. The red genes were first identified in C. elegans and later were found to be components of

Ras signaling pathway in other model systems including humans cells. Figure adapted from wormbook.

Figure 1.4: C. elegans vulva phenotypes.

(A) Normal vulva morphology with 22 cells arranged symmetrically at L4 stage.

Magnification 100X. During normal development, Anchor cell (AC) sends inductive signal lin-3/EGF to induce the primary and secondary vulva fates (shown in green) in

VPCs. In addition, Wnt signals from gonad and lateral Notch signaling between VPCs play role in determining their fate (1º or 2º). Then, they divide 3 times to form 22 cells of normal vulva. The cells in the vulva equivalence group (P3.p to P8.p) that do not adopt vulva fate (shown yellow in color) due to reception of less lin-3/EGF signal from AC, adopt 3º fate, divide once and the daughter cells fuse with hyp7 syncium. (continued)

Figure 1.4: C. elegans vulva phenotypes (continued)

(B) Multivulva (Muv) phenotype is generally due to abnormal hyper-activation of Ras signaling pathway (and in some case Wnt- or Notch- pathways), e.g. due to gain-of- function mutations in components of these signaling pathways. This leads to generally more VPCs to adopt vulva fate (1º and 2º) (shown in green) than wild-type scenario and more than one VPC to adopt 1º fate. These cells then lead to more daughter cells than wild-type C. elegans leading to ectopic vulva cavities. Generally, one vulva is functional and the ectopic vulva morph into ventral protrusions that can be seen as multivulva

(Muv) phenotype at adult stage. Image is showing Muv phenotype at L4 stage at 100X magnification. Gain-of-function mutation in let-60/Ras gives rise to Muv phenotype, a mutation that was used in this study to have over-proliferation (Muv) phenotype as baseline for the screen. Suppression of this phenotype to Normal vulva was used as phenotype for screening RNAi effect of genes in the mesenchymal cells.

(C) Vulvaless (Vul) phenotype is generally due to abnormal low activation of Ras signaling pathway (and in some case Wnt- or Notch- pathways), e.g. due to loss-of- function mutations in components of these signaling pathways. This leads to generally none of fewer VPCs to adopt vulva fate (1º and 2º) than wild-type scenario (un-induved

VPCs shown in yellow). These cells adopt 3º fate leading to no vulva formation or underinduced phenotype (partial vulva structure formation). Image is showing Vul phenotype at L4 stage at 100X magnification.

Figure 1.5: C. elegans vulvagenesis and interplay of Ras, Wnt, and Notch pathways.

Cartoons adapted from Green and Sternberg, 2007.

(A) At L1 stage, of 11 cells, six (P3.p to P8.p) adopt vulva equivalence fate. lin-39 activity helps in maintaining their vulva fate status and thus preventing their fusion to the hypodermis. (continued) 23

Figure 1.5: C. elegans vulvagenesis and interplay of Ras, Wnt, and Notch pathways.

(continued)

(B) At L2/L3 stage, anchor cell (AC) sends inductive signal lin-3/EGF. P6.p cell receives more of this signal due to physical proximity to the AC compared to neighboring

P5.p and P7.p cells. Interplay of activated Ras pathway in these cells with Wnt signals from gonad and muscles and lateral Notch signaling between P5.p, P6.p and P7.p help in determining their fates. P6.p adopts 1º fate (green), P5.p and P7.p adopt 2º fate (blue) and

P3.p, P4.p and P8.p adopt 3º fate (yellow).

P7.p give rise to seven daughter cells each and the P6.p gives rise to eight daughter cells, which arrange to a stereotypic symmetric vulva structure (shown in D). The 3º fate cells divide only once. Their daughter cells then fuse with the hyp7 synctium.

(E) Nomarski image of vulva structure at L4 stage. Magnification, 100X.

Figure 1.6: Genetically engineered tissue-specific RNAi models of C. elegans.

(A) Table listing mesenchymal tissue specific and epithelial tissue specific promoters.

The mesodermal tissues are gonad, anchor cell (AC) and muscle cells and the vulva epithelial tissue comprise vulva precursor cells (VPCs). These tissues take part in intercellular communication for vulva development.

(B) Cartoon representation of rde-1(wt) gene under the control of tissue-specific promoters (listed in A). Rescue of rde-1(wt) in the rde-1(-) background worm determines tissue-specific RNAi function. (continued)

Figure 1.6: Genetically engineered tissue specific RNAi models of C. elegans.

(continued)

AC-, gonad-, muscle- and VPC- tissue-specific promoters, respectively (from left to right). Data: Huayang Liu (Liu et al, submitted).

(D) Cartoon representation of mesenchymal tissue specific RNAi strain. The rde-1(wt) gene was cloned under the control of ddr-2p, Acel::∆pes10p and myo-3p promoters individually. These clones were used together in rde-1(-) strain of C. elegans to rescue rde-1(wt) expression in all three mesodermal tissues (gonad, anchor cell, and muscle) such that RNAi functions specifically in these tissues in the resulting strain (shown in green color).

(E) Cartoon representation of vulva epithelium specific RNAi strain. The rde-1(wt) gene was cloned under the control of lin-31p and was used in rde-1(-) strain of C. elegans to rescue rde-1(wt) expression in the vulva precursor cells (VPCs) such that RNAi functions specifically in these cells. (shown in green color).

Figure 1.7: Functional test of the tissue-specific RNAi strains of C. elegans. lin-3 encodes EGF signal that is released by anchor cell (AC), hence, it represents mesenchymal (cell non-autonomous) tissue-specific gene. The gene, lin-39, functions downstream of Ras and Wnt signaling pathways within vulva precursor cells (VPCs), hence it presents VPC specific (cell autonomous) gene. The constructed transgenic strains, “mesodermal-rde” and “vulva-rde” were tested for their tissue-specificity using

RNAi against these two (cell non-autonomous, lin-3 and cell-autonomous, lin-39) genes.

(A) and (B) The rde-1(+) strain has RNAi functioning throughout the worm body. Knock down of lin-3 (cell non-autonomous gene) and lin-39 (cell autonomous gene) using feeding based RNAi shows Vulvaless (Vul) phenotype, as expected. (continued)

Figure 1.7: Functional test of the tissue-specific RNAi strains of C. elegans

(continued)

(C) and (D) The rde-1(-) strain is resistant to RNAi treatment. Knock down of lin-3 (cell non-autonomous gene) and lin-39 (cell autonomous gene) in this strain using feeding based RNAi shows Normal (wild type) vulva phenotype, as it is expected that RNAi will not function in this strain.

(E) In the “mesehchymal-rde” strain, the rde-1(+) is rescued in mesodermal tissues using mesodermal tissue specific promoters (gonad, AC and muscle) in the rde-1(-) strain background. In this strain, RNAi should function in the mesodermal tissues. Knock down of lin-3 using RNAi causes vulvaless (Vul) phenotype, as expected.

(F) In the “mesenchymal-rde” strain, VPCs have rde-1(-) allele, thus RNAi of VPC specific gene should not show any effect on the vulva phenotype. Knock down of lin-39 using RNAi shows normal (wild type) vulva phenotype. These results (showin in (E) and

(F)) confirm mesodermal-tissue specific RNAi effect in this strain.

(G) In the “Vulva-rde” strain, the rde-1(+) is rescued in VPCs using VPC-specific promoter (lin-31p) in the rde-1(-) strain background. In this strain, RNAi should function in VPCs only but not in rest of the tissues. Knock down of lin-3 using RNAi shows

Normal (wild type) vulva phenotype. (continued)

Figure 1.7: Functional test of the tissue-specific RNAi strains of C. elegans

(continued)

(H) In the “Vulva-rde” strain, all cells except VPCs have rde-1(-) allele, thus RNAi of

VPC specific gene should affect the vulva phenotype. Knock down of lin-39 (a VPC specific gene) using RNAi shows underinduced/Vulvaless vulva phenotype. These results

(showin in (G) and (H)) confirm vulva-tissue specific RNAi effect in this strain.

Figure 1.8: let-60(gf) sensitized tissue-specific RNAi C. elegans models and their phenotypes. (continued)

(A) Active and Inactive forms of Ras. Ras-GDP is inactive form, whereas Ras-GTP is active form that can bind to effectors such as Raf. (continued)

Figure 1.8: let-60(gf) sensitized tissue-specific RNAi C. elegans models and their phenotypes (continued)

Gain-of-function (gf) mutations lock Ras in the active state, e.g., let-60(n1046) is a gain- of-function point mutation that locks Ras in constitutively active form (Beitel et al., 1990;

Han and Sternberg, 1990). On the contrary, loss-of-function or dominant-negative (dn) mutations lock Ras in the inactive state (Han and Sternberg, 1991).

(B) Conserved canonical Ras signaling pathway in C. elegans. The gain-of-function mutation in let-60(n1046) leads to constitutively active Ras signaling, leading to Muv phenotype.

(D) Cartoon depicting Muv phenotype of mesenchymal-tissue specific RNAi strain after crossing to let-60(gf) strain.

(E) Cartoon depicting Muv phenotype of vulva-tissue specific RNAi strain after crossing to let-60(gf) strain.

Chapter 2 A genetic screen to identify cell non-autonomous suppressors of hyper-

proliferation

Let’s go for treasure hunt! - Ritav, 4 year old (Unpublished).

2.1 Introduction

As discussed previously, the significant role of stromal cells in cancer progression and even in drug resistance is being recognized (Finak et al., 2008; Beck et al., 2011;

Qian & Pollard, 2010; McMillin, Negri, & Mitsiades, 2013; Errico, 2015; Kottakis and

Bardeesy, 2015; Farmer et al., 2009). Various large-scale studies of tumors from different organ sites highlight the prevalence of oncogenic Ras-mutations as a driver mutations

(Peeters et al., 2015; Dumenil et al., 2015). Abundance and correlation of stromal cells in the Ras-mutant tumors in the disease progression and drug resistance is evident from recent studies (Bacman et al., 2007; Takahashi et al., 2011; Bremnes et al., 2011;

McMillin et al., 2013; Downey et al., 2014; Dumenil et al., 2015; Peeters et al., 2015). A recent focus on tumor microenvironment studies have highlighted role of some genes (on a gene-by-gene basis) in augmenting the tumor size, disease progression, etc. (Figure

1.2). How do stromal cells influence the proliferation of cancer cells? What are the critical genes in stromal that might be playing causal role in enhancing or suppressing the

32 proliferative capacity of malignant cells– a hallmark property of all cancer cells? This literature including the previous studies in our laboratory led to our main hypothesis that stromal signaling plays causative role in regulating proliferation of the malignant epithelial cells.

To date, discovering significant genes that contribute to the pathology from the large-scale gene-expression datasets continues to be challenging. Furthermore, these studies cannot determine direct causal link of stromal signals in mediating or modulating proliferation of cancer cells, which is one of the central focuses of biological inquiry in the tumor microenvironment field. To get insight into such complex problems, the advent and availability of annotated genome-wide screen technologies, conserved signaling pathways across the species and well-characterized genetically tractable model systems provide excellent opportunity for large-scale, unbiased experiments. The practical advantages of using small animal models such as yeast, C. elegans and Drosophila are numerous. For modeling in vivo signaling interactions between the mutant cells and the surrounding mesenchymal (mesodermal) cells, the interplay of conserved signaling pathways in the mesodermal and epithelial cell interactions during the vulvagenesis of C. elegans make it an excellent choice of model system for a genome-wide screening studies. In addition, vulva phenotype, which is tractable at cellular level, in response to oncogenic mutation or naturally operating conserved signaling pathways provides an almost ideal experimental readout for evaluation of the cellular proliferation upon gene- by-gene genetic manipulations in a tissue specific manner.

To identify stromal genes that play role in modulating stroma-originating signals conductive to tumor growth and progression, we took cross species approach of using mesenchymal tissue-specific RNAi to identify cell non-autonomous signaling genes that causatively contribute to the proliferation of Ras-mutant epithelial cells in the vulva development program of C. elegans.

In this chapter, we present results of a genome-wide screen in a genetically engineered C. elegans model system to identify genes in the mesodermal (mesenchymal) cells that suppress the oncogenic proliferation of let-60(gf)/Ras-mutant epithelial cells in vivo.

2.2 Results

A genome-wide RNAi screen to identify cell non-autonomous regulators of oncogenic proliferation

For our screen to identify mesodermal genes that influence epidermal cell proliferation, we developed a new experimental strain of C. elegans, CM2453. This strain comprises a gain-of-function (gf) mutation let-60(n1046) that promotes over-proliferation of vulva precursor cells leading to the Multivulva (Muv) phenotype as well as a transgene to allow mesodermal-tissue specific RNAi (Figure 2.1A). The experimental strain exhibited 80±12% Muv phenotype when fed on control RNAi (non-targeting vector) at

20°C in the 12-well plates as observed by a dissection microscope (Figure 2.1B).

The genome-wide screen was designed for time efficiency and removal of false positives in the subsequent secondary screens (Figure2.1C-D). For the primary screen,

34 we used commercially available well-annotated genome-wide Ahringer RNAi feeding library from Source Biosciences, which is arrayed as single RNAi clones in a set of 52

384-well plates and covers ~ 80% of the C. elegans genome. Briefly, each clone in the library was tested for its effect by counting at least 50 RNAi-treated worms on a gene-by- gene knockdown basis, and Muv phenotype was quantified manually using a dissection microscope. Overall, RNAi-treatment of various genes showed either no effect or varying degree of suppression of Muv phenotype (% number of worms exhibiting Muv phenotype). As a primary screen pass, a stringent cut-off (<30% of Muv phenotype exhibiting worms compared to the non-targeting vector baseline) was used to identify genes with strong RNAi effect. Of the ~18000 RNAi clones (representing over 16000 genes), 231 genes showed reduced (below the cut-off) Muv phenotype compared to the baseline. RNAi clones that caused a reduction of Muv phenotype to 30% or less of non- targeting control were retained and re-tested in secondary screening sessions (Figure

2.2A).

Identification and validation of candidate cell non-autonomous proliferation regulators

The RNAi-carrying bacterial clones of the 231 RNAi hits from the primary screen were then isolated from the library of 384-well plates to construct a mini-RNAi-library.

Then, these 231 clones were tested two more times, and only clones that caused a reduction of the Muv phenotype to at least 40% of the control in two of the three

35 additional trials were retained. This process identified 66 unique candidate genes.

(Figure 2.2A).

At the dissection microscope level, it is not possible to evaluate the effect of

RNAi at the cellular level phenotype, albeit the screen served as a surrogate for suppression of over-proliferation (i.e. Muv phenotype). To understand the impact of the

66 RNAi (screen hits) on VPC proliferation, we evaluated their knockdown effect at the cellular level during the fourth larval (L4) stage using 100X magnification Nomarski imaging. RNAi-treated C. elegans for each of the 66 genes were evaluated for whether they produced a normal vulval development, reduced division (vulvaless or Vul), or other abnormal morphology compared to excessive division (Muv) phenotype of the CM2453 strain. Using these criteria, 47 genes were identified to act as non-autonomous suppressors of let-60(gf)/Ras, and revert ≥40% of worms to a wild type (normal) vulva morphology, relative to the Muv phenotype of the control (Figure 2.2B, C and Table

2.1). This analysis refined the candidate cell non-autonomous suppressors of over- proliferation by eliminating false positives.

Next, from the list of the library ID of the RNAi clones in the candidate list, we identified the corresponding C. elegans genes as annotated by the Ahringer RNAi library.

The identification of the candidate genes was confirmed by sequencing of the RNAi- encoding plasmids isolated from the respective bacterial clones from the RNAi library followed by NCBI BLAST analysis. A well-curated and annotated table of final list of candidate genes is presented in Table 2.1.

2.3 Discussion

Here, we describe an unbiased genome-wide screen for genes discovery that have causal role in the cell non-autonomous suppression of the oncogenic over-proliferation mediated by let-60(gf) mutation. The screen resulted in 47 genes and we present these results as well curated resource (Table 2.1). To our knowledge, it is first genome-wide

RNAi screen to identify cell non-autonomous enhancers of oncogene-mediated over- proliferation. Although, we here present list of genes that only strongly suppress the over-proliferation (Muv) phenotype upon RNAi-mediated knockdown, the genes showing intermediate suppression may represent mild but biologically significant regulators. In addition, since the data collection during the primary screen was unbiased and quantitative in nature for knockdown of each gene, it potentially can be used to identify genes that, upon RNAi-mediated knockdown, enhance the oncogenic over-proliferation mediated by (let-60(gf)) mutation above the baseline.

It is important to consider that false positives and false negatives caused by limitations of the RNAi based genome-wide screens can impact the results of screen. To eliminate false positives, we used blinded and randomized phenotype reading approach for primary and subsequent 2X secondary screens accompanied by stringent cut-off limits for “hit” calling. This approach would have eliminated the non-replicating candidates from the list. Additionally, use of microscopy and imaging of random samples of first 50 worms in a pool of RNAi treated C. elegans would eliminate any errors or bias associated

37 with human scoring. The end point screening using 100X Nomarski imaging at L4 stage

(when vulva phenotype is not visible under the dissection microscope) certainly helped in eliminating false positives and genes with variable RNAi effect across the screening stages (primary, secondary and high–magnification).

On the contrary, our screen was not designed to identify false negatives. At different stages of the screen, we likely have lost candidate genes that are important in the cell non-autonomous suppression (upon RNAi-mediated knockdown), but were not scored due to technical reasons, such as, contamination of the bacterial clones, embryonic or larval lethality, or weak effect of RNAi due to weak concentration of the RNAi-carrying bacteria, phenotype variability, etc. Although it is an unbiased genome-wide screen, it would not be helpful for evaluating certain set of genes that were not scored due to above-mentioned technical reasons. In addition, to cover the complete spectrum of the screen at genome-wide scale, it would require to do screening of additionally available clones of RNAi that were missing in the larger library and are now available as supplement mini-library from the same commercial resource.

Previous whole-body RNAi screens to unravel genes involved in vulval development have reported several classes of cell-autonomous and synthetic multivulva genes involved in the process of proliferation of VPCs (Singh & Han, 1995; Sundaram and Han, 1995; Gu et al., 1998; Li et al., 2000; Kao et al., 2004; Cui, et al., 2006; Fay and

Yochem, 2007). The screens for identifying suppressors of synthetic multivulva genes had recovered genes involved in the processes of chromatin remodeling, protein degradation and tissue-specific regulation of the expression of pgl-1 and lag-2 genes

(Cui, Kim, et al., 2006). Our genome wide screen to identify cell non-autonomous genes is distinct from these screens in several ways. First, our screen is done in a strain in which

RNAi specifically works only in the mesodermal, inductive-signal sending tissues (gonad,

Anchor cell and Muscles). Second, we used (let-60(gf)) as sensitizing background and have recovered genes that suppress let-60(gf) mediated over-proliferation upon RNAi- mediated knockdown. Third, while we used RNAi-feeding technology for the genome- wide screening, the synMuv screens (Ferguson & Horvitz, 1989; Fay and Yochem, 2007;

Thomas and Horvitz, 2003; Ceol and Horvitz, 2004) used EMS mutagen technique to identify suppressors of synMuv genes. RNAi has knock down effect compared to null mutations as in EMS method and RNAi effect starts when larvae start to feed on bacteria.

These two properties of RNAi based screening allow recovery of otherwise embryonic lethal or essential genes, whose null mutations would lead to lethality. Fourth, none of our candidate genes overlap with the candidates from the synMuv genes.

In summary, we have discovered novel role of certain genes in the mesodermal tissues that suppress over-proliferation of the adjacent let-60(gf)-mutant epithelial cells upon RNAi-mediated knockdown. Functional categorization and speculation of their potential roles in abrogating over-proliferating capacity of neighboring mutant cells is presented in chapters 3. These genes deserves a thorough investigation to unravel the molecular understanding of how these genes might be eliciting their cell non-autonomous effect.

2.4 Materials and methods

Genetically engineered tissue-specific RNAi C. elegans models and C. elegans strains used

In this study, we used mesenchymal or VPC tissue-specific RNAi models of C. elegans that also carry let-60(n1046) allele that has gain-of-function mutation in let-60 gene. Briefly, C. elegans strains were constructed using rde-1(ne219) worms in which wild-type rde-1(wt) was cloned to be driven by the mesodermal tissue specific promoters

(ddr-2p (gonad specific), ACEL-∆pes (Anchor cell specific) and myo-3p (Muscle specific)) or vulva epithelial tissue (lin-31p), such that the resulting strains are RNAi competent only in the tissue where rde-1(wt) is function. The mesodermal tissue-specific

RNAi strain is termed CM2073 and the vulva epithelium tissue specific strain is termed

CM2422.

For this screen, we constructed two new strains: CM2453 and CM2611 strains.

For constructing the CM2453 strain, we crossed the CM2073, rde-1(ne219);rrf-

3(pk1417) and let-60(n1046) strains to obtain the desired genotype. The genotyping of let-60(n1046) and rde-1(ne219) alleles was confirmed with sequencing and rrf-3(pk1417) was confirmed by genotyping PCR amplification with genotype specific primers. The presence and the homozygosity of the wild-type rde-1 allele carrying transgene were confirmed by visualizing transgene linked green fluorescent protein (GFP) signal.

For constructing the CM2611 strain, we crossed the CM2422, rde-1(ne219);rrf-

3(pk1417) and MT2124 (let-60(n1046)) strains to obtain the desired genotype. The genotyping of let-60(n1046) and rde-1(ne219) alleles was confirmed with sequencing and

40 rrf-3(pk1417) was confirmed by genotyping PCR amplification with genotype specific primers. The presence and the homozygosity of the wild-type rde-1 allele carrying transgene were confirmed by visualizing transgene linked green fluorescent protein

(GFP) signal.

In this study, The following C. elegans strains were used:

CM2453 rrf-3 (pk1417); let-60(n1046); rde-1(ne219); guIs37[myo-3p::rde-1(+); ddr-

2p::rde-1(+); ACEL-Δpes-10::rde-1(+); sur-5p::gfp; unc-119(+)]. The strain is temperature sensitive.

CM2073 rrf-3(pk1426); rde-1(ne219); guIs37[myo-3p::rde-1(+); ddr-2p::rde-1(+);

ACEL-Δpes-10::rde-1(+); sur-5p::gfp; unc-119(+)]. This strain was previously developed and characterized in our laboratory. (Liu, et al., submitted)

CM2611 rrf-3 (pk1417); let-60(n1046); rde-1(ne219); guIs39[lin-31p::rde-1(+); sur-

5p::gfp; unc-119(+)].

C. Nematode handling and processing

C. elegans strains were cultured and maintained under standard conditions on

NGM plates with E. coli OP50 as a food source (Brenner, 1974). For the screening, the worms were cultured on 10 cm NGM-plates such that they are adult, gravid hermaphrodites on the day of harvest. For each screening session, 9-10 plates of worms were washed to harvest the eggs. The RNAi screen was carried out in 12 well plates on

NGM made with 2x (5 g/L) peptone and supplemented with 1µM IPTG and 50 µg/ml carbenicillin. All experiments were performed at 20°C unless otherwise noted.

Genome-wide screen set-up

The experimental strain (CM2453) was used for the genome-wide screen on clones from the Ahringer RNAi feeding library (Source Bioscience LifeSciences)

(Timmons and Fire, 1998; Fraser et al., 2000; Timmons and Fire, 2001). Each RNAi clone-carrying E. coli strain was cultured in 500ml LB broth supplemented with ampicillin in separate wells of the 96-deep-well plates overnight. Next day, 30µl of each bacterial culture was spread in one well of a 12-well plate containing NGM and left for

12-16 hours at room temperature to allow bacterial growth and induction of RNAi by

IPTG. The experimental strain was grown on OP50 bacteria in 10cm plates to harvest gravid hermaphrodites. The eggs were harvested using a standard bleaching protocol and washed with M9 solution 3 times to remove traces of bleach from eggs. 100-150 of freshly harvested eggs were spotted in each well, and incubated at 20°C to allow hatching of larvae and their growth to young adults (~3.5 days), when the vulva phenotype can be scored using dissection microscope. The RNAi treated C. elegans that had a developmentally delayed phenotype were allowed extra time to grow to adult stage before scoring them. The effect of each RNAi and empty vector was quantified by counting the number of worms exhibiting the Multivulva (Muv) phenotype out of total 50 worms.

Threshold/Cut-offs for “hit” calling

For each plate reader, the suppression of the Muv phenotype was calculated by reduction to 30% or less in the primary screen compared to the empty vector control

42 reference. These RNAi clones were selected to construct a mini-library and re-tested using a high-magnification assay. The RNAi clones of the hits from primary screen were re-tested 2 more times and, for each plate reader, the suppression of the Muv phenotype was calculated by reduction to 40% or less compared to the empty vector control reference. The hits that replicated the effect of RNAi-mediated suppression of Muv phenotype <40% in least 3 of the 4 reading sessions were retained. They were further validated using a high-magnification (100X) imaging assay.

Sequencing of hits to confirm right targets

The RNAi clones of the hits (after secondary screening) from the mini- RNAi library were grown overnight in 3 ml of sterile LB broth supplemented with 3µl of

50mg/ml Carbenicillin. A small aliquot (30 µl) was used for high-magnification RNAi experiments. The remaining bacterial culture was used to extract the RNAi-encoding plasmid. Standard mini-prep method using Qiagen miniprep kit was used to extract the plasmid. The purified plasmid was sequenced using M13 primer. In case of unsuccessful sequencing results, the bacterial cultures were cultured again and plasmids were extracted with mini-prep method. The successful sequence results were subject to Blast analysis to identify/confirm the targeted gene. Wormbase was used to obtain the data on common worm gene name, chromosome number, and their general function (if available), which is reported in Table 2.1.

High magnification imaging: Sample preparation and Microscopy

The validation of the screen hits was done at 100X magnification under an oil immersion lens using a Zeiss Axioscope microscope. RNAi treatment was completed as described above. At the L4 stage, RNAi fed larvae were harvested using M9 washing from the 12-well plates. Agar-slides were prepared using 3.5% Noble agar supplemented with 10 mM NaN3 to mount RNAi treated larvae for imaging. Only worms exhibiting a normal vulva phenotype are reported in the Table; over- or under-induction or morphological defects were considered abnormal.

Figure 2.1: A genome wide RNAi screen for cell non-autonomous suppressor genes of oncogenic over-proliferation

(A) Cartoon representation of the mesodermal tissue specific RNAi in C. elegans strains

(continued)

Figure 2.1: A genome wide RNAi screen for cell non-autonomous suppressor genes of oncogenic over-proliferation (continued)

(B) Baseline Muv and WT phenotype at adult and L4 stage that were used as screening assay (Muv to Normal vulva phenotype) at dissection microscope level (for primary and secondary screens) and at high (100X) magnification level (for validation screen).

(D) Schematic for identification of candidate genes in primary screen

Figure 2.2: Rescreening and validation of candidate genes

(A) Secondary screens and high magnification validation screening schematic

(continued)

Figure 2.2: Rescreening and validation of candidate genes (continued)

(B) Suppression of Muv phenotype in mesenchymal tissue specific RNAi strain with let-

60(gf) sensitization

(C) Representative images of vulva of experimental strain in various functional categories of hits. All images were taken at 100X magnification.

Table 2.1: List of final candidates genes. Column 1 represents chromosome number on which identified gene is located. Column 2 represents gene ID, Column 3 represents commonly known names of the genes in C. elegans, (continued) 49

Table 2.1: List of final candidates genes (continued)

Column 4 represents known functions of the C. elegans genes, and, Column 5 represents number of worms showing normal vulva phenotype out of total number of C. elegans that were screened.

Chapter 3

Identified Cell non-autonomous candidate genes converge on select conserved gene

families or molecular complexes

Let’s find out who all have same kind of light-sabers! - Ritav, 3.5 years old (Unpublished).

3.1 Introduction

To glean the information on the function of candidate genes from the screen, we performed bioinformatics analyses and literature search. In this chapter, we present the curated data on the human orthologs of the candidate genes, known function of human orthologs, and their functional categories. We show that majority of the candidate genes are evolutionarily conserved between C. elegans and Humans and that candidate genes converge on select gene families, biological processes or molecular complexes.

3.2 Results

Identified Cell non-autonomous proliferation regulatory genes are enriched in conserved genes between C. elegans to Humans

Having found 47 candidate genes from a genome-wide screen that markedly suppress let-60(gf) mediated oncogenic proliferation, we wished to query how many of the candidate genes are conserved in Humans and what are their known functions? 51

Bioinformatics analyses revealed that among these 47 genes, 37 have orthologs in humans while other 10 were specific to C. elegans (Figure 3.1A). This represents

78.72% of identified genes are conserved. The RNAi library clones represented 16,256 individual genes of C. elegans, of which 7,670 (47.2%) are conserved between C. elegans and Humans. The enrichment of conserved genes in our candidate list is statistically significant. (Figure 3.2B). This analysis underscored the importance of these genes in potentially conserved signaling pathways or biological processes.

Identified Cell non-autonomous candidate genes converge on select conserved gene families or molecular complexes

Several candidate genes appear to be functionally related and represent categories including histone genes, ribosomal genes, ATPase synthase genes, vesicular transport genes, and metabolic genes. Interestingly, knockdown of no gene encoding a secreted growth factors act as a suppressor of the Muv phenotype based on our threshold/cutoff criteria (Table 3.2).

3.3 Discussion

Conservation of signaling pathways and importance of C.elegans in signaling studies

Many developmental signaling pathways are conserved evolutionarily across species and often are sabotaged in pathological conditions. Over the past decades, with the advent of molecular and system biology methods and techniques for genome-scale investigations have revealed important genes, networks, regulatory hubs and signal

52 transduction pathways across the species underscoring the importance of use of model systems in our understanding of important biological process.

C. elegans vulvagenesis is a prime example of inter-cellular signaling interactions between different cell types. The EGFR-Ras-MPK pathway is highly conserved between

C. elegans and Humans (Figure 8B, chapter 1). Indeed, several of the components of this pathway were first identified in the genetic screens in the C. elegans with respect to its vulva development and later found to function in the similar way in other species and in human cell culture studies (Lowenstein et al., 1992; Clark et al., 1992). Several genome- wide or targeted screens in C. elegans vulva development have identified genes that function in the regulation of proper fate determination and proliferation of vulva precursor cells. Some of these genes act as synthetic multivulva (SynMuv) genes that are categorized in different classes. Some SynMuv genes are shown to act cell non- autonomously to regulate over-proliferation of VPCs by regulating the expression of lin-

3/EGF in the adjacent hypodermal cells. It is noteworthy that these screens were done in the whole-body fashion with tissue-specific role inferred from the mosaic studies of the transgenes.

Screens in our lab (previously performed and reported here in) are done in mesenchymal tissue-specific RNAi-mediated knockdown of genes in genetically engineered strains. This study, to our knowledge, is among the first to identify mesenchymal cell based, cell non-autonomous enhancers of oncogenic over-proliferation in vivo.

What it means to identify components of a signaling pathway, biological process or molecular complex?

The genes in each category function in some biological process and/or are part of a molecular complex. For example, the nine of the chromatin remodeling genes represent core histone proteins representing (HIST1H4G, HIST2H2AB, HIST2H2BF, HIST2H3D,

H3F3B and HDDC2) of human homologs as inferred from Wormbase, The HIST1H4G,

HIST2H2AB, HIST2H2BF, HIST2H3D are parts of core components of nucleosome.

Nucleosomes wrap and compact DNA into chromatin thereby limiting DNA accessibility to the molecular machineries that require DNA as a template. The accessibility of DNA is regulated by remodeling of nucleosomes and by complex set of post-translational modifications of histones. Thus, these histone proteins play a central role in transcription regulation, DNA repair, DNA replication and chromosomal stability. Notably,

HIST2H2BF, H3F3B are also reported to be strongly expressed in nucleus and in the extracellular region (Gonzales et al., 2009). Likewise, HDDC2 is expressed strongly in the extracellular exosomes (Gene cards). In the ribosomal proteins category, of 6 ribosomal proteins, 5 represent components of the large subunit of ribosomal protein complex while one of them represent part of the small subunit. Similarly, in the ATPases category of identified genes, importantly, both vha-3 and vha-11 (the vacuolar H

ATPases) share an operon and are prominently expressed in non-VPC cells. The three F-

ATPases are components of the molecular Fo and F1 complex of mitochondria.

Convergence of the hits in select molecular complexes or components of a biological process suggests significance of role of these biological process or molecular

54 complexes in cell-cell communication. These analyses identify an attractive research avenue and warrants further in depth studies.

Mesoderm-specific Worm RNAi Strain Human Chr Gene ID common Description of Human orthologs High Mag Ortholog gene name (Normal/total) with let-60 (gf) I C04F12.4 rpl-14 RPL14 ribosomal protein L14 87/90 I C37A2.7 C37A2.7 RPLP2 ribosomal protein, large, P2 44/50 I C04F12.1 C04F12.1 AGO4 argonaute RISC catalytic component 4 61/72 I F57B10.5 F57B10.5 TMED7 transmembrane Emp24 Protein Transport Domain Containing 7 20/50 I F36A2.8 phip-1 PHPT1 phosphohistidine phosphatase 1 40/50 I F35C12.2 ncx-4 SLC24A2 solute Carrier Family 24 (Sodium/Potassium/Calcium Exchanger), Member 2 37/50 I D1007.4 D1007.4 GEMIN6 gem (nuclear organelle) associated protein 6 47/50 I C27A12.2 C27A12.2 ZNF79 zinc finger protein 79 8/13 II F08G2.2 his-43 HIST2H2AB histone cluster 2, H2ab 21/25 II W09H1.2 his-73 HIST2H3D histone cluster 2, H3d 134/152 II W01D2.1 rpl-37 RPL37 ribosomal protein L37 82/95 II H17B01.1 fgt-1 SLC2A3 solute carrier family 2 (facilitated glucose transporter), member 3 58/86 II Y53F4B.18 Y53F4B.18 FAAH2 fatty acid amide hydrolase 2 104/105 III K11H12.2 rpl-15 RPL15 ribosomal protein L15 45/52 III W04A8.1 W04A8.1 MCPH1 microcephalin 1 43/50 IV B0035.9 his-46 HIST1H4G histone cluster 1, H4g 41/50 IV F17E9.10 his-32 HIST2H3D histone cluster 2, H3d 40/50 IV F55G1.2 his-59 HIST2H3D histone cluster 2, H3d 38/42 IV Y57G11C.16 rps-18 RPS18 ribosomal protein S18 18/22 IV Y38F2AL.4 vha-3 ATP6V0C ATPase, H+ transporting, lysosomal 16kDa, V0 subunit c 64/80 IV Y38F2AL.3 vha-11 ATP6V1C1 ATPase, H+ transporting, lysosomal 42kDa, V1 subunit C1 21/28 IV R11A8.6 iars-1 IARS isoleucyl-tRNA synthetase 38/58 IV B0035.16 mttu-1 TRMU tRNA 5-methylaminomethyl-2-thiouridylate methyltransferase 43/50 IV F40F11.2 mig-38 PRG4 proteoglycan 4 (megakaryocyte stimulating factor) 13/29 V F45F2.9 F45F2.9 HDDC2 HD domain containing 2 74/100 V T10C6.12 his-3 HIST2H2AB histone cluster 2, H2ab 46/50 V F45F2.12 his-8 HIST2H2BF histone cluster 2, H2bf 14/23 V F32D1.2 hpo-18 ATP5E ATP synthase, H+ transporting, mitochondrial F1 complex, epsilon subunit 47/50 V C06H2.1 atp-5 ATP5H ATP synthase, H+ transporting, mitochondrial Fo complex, subunit d 44/50 V T05H4.12 atp-4 ATP5J ATP synthase, H+ transporting, mitochondrial Fo complex, subunit F6 95/99 V C31B8.4 C31B8.4 GOLGA4 golgin A4 44/50 V D2023.2 pyc-1 PC pyruvate carboxylase 38/50 V F46B6.5 F46B6.5 SRRM5 Serine/Arginine Repetitive Matrix 5 40/50 X F45E1.6 his-71 H3F3B H3 histone, family 3B (H3.3B) 39/50 X F55D10.2 rpl-25.1 RPL23A ribosomal protein L23a 47/50 X C18B12.6 C18B12.6 ERGIC2 ERGIC And Golgi 2 31/38 X W04G3.8 lpr-3 TSC22D2 TSC22 Domain Family, Member 2 20/40 Nematode specific proteins I W09C5.12 W09C5.12 none none 44/72 I F17B5.5 clec-110 none none 27/50 I F55D12.8 ncRNA none none 36/50 i F55D12.1 F55D12.1 none none 39/50 II Y46G5A.36 Y46G5A.36 none none 25/50 II C14A4.12 C14A4.12 none none 41/52 V C53A5.1 ril-1 none none 46/50 X C14A6.3 C14A6.3 none none 26/50 X F15G10.3 piRNA none none 28/36 X K09E3.5 K09E3.5 none none 21/28

WormMart # Genes % Conserved p-value Total genes in RNAi library 16,256 Conserved genes in RNAi library 7,670 47.2 Identified genes/hits 47 Conserved genes/hits 37 78.72 0.0001

Figure 3.1: Candidate genes conservation analysis in humans

(A) Table of hits with human orthologs (B) Conserved genes enrichment analysis 56

Table 3.1: List of hits arranged in functional categories

Mesenchyme-specific Worm Human RNAi Strain Chr Gene ID common Description of Human orthologs Ortholog High Mag (Normal/total) gene name with let-60(wt) Histone proteins X F45E1.6 his-71 H3F3B H3 histone, family 3B (H3.3B) 50/50 V F45F2.9 F45F2.9 HDDC2 HD domain containing 2 50/50 IV B0035.9 his-46 HIST1H4G histone cluster 1, H4g 40/40 II F08G2.2 his-43 HIST2H2AB histone cluster 2, H2ab 65/65 V T10C6.12 his-3 HIST2H2AB histone cluster 2, H2ab 50/50 V F45F2.12 his-8 HIST2H2BF histone cluster 2, H2bf 50/50 IV F17E9.10 his-32 HIST2H3D histone cluster 2, H3d 38/38 IV F55G1.2 his-59 HIST2H3D histone cluster 2, H3d 50/50 II W09H1.2 his-73 HIST2H3D histone cluster 2, H3d 49/50 Ribosomal proteins I C04F12.4 rpl-14 RPL14 ribosomal protein L14 50/50 III K11H12.2 rpl-15 RPL15 ribosomal protein L15 45/50 X F55D10.2 rpl-25.1 RPL23A ribosomal protein L23a 50/50 II W01D2.1 rpl-37 RPL37 ribosomal protein L37 40/50 IV Y57G11C.16 rps-18 RPS18 ribosomal protein S18 50/50 I C37A2.7 C37A2.7 RPLP2 ribosomal protein, large, P2 32/42 ATP Synthase proteins V F32D1.2 hpo-18 ATP5E ATP synthase, H+ transporting, mitochondrial F1 complex, epsilon subunit 42/42 V C06H2.1 atp-5 ATP5H ATP synthase, H+ transporting, mitochondrial Fo complex, subunit d 50/50 V T05H4.12 atp-4 ATP5J ATP synthase, H+ transporting, mitochondrial Fo complex, subunit F6 50/50 Vesicular Transport proteins IV Y38F2AL.4 vha-3 ATP6V0C ATPase, H+ transporting, lysosomal 16kDa, V0 subunit c 50/50 IV Y38F2AL.3 vha-11 ATP6V1C1 ATPase, H+ transporting, lysosomal 42kDa, V1 subunit C1 50/50 X C18B12.6 C18B12.6 ERGIC2 ERGIC And Golgi 2 50/50 V C31B8.4 C31B8.4 GOLGA4 golgin A4 50/50 I F57B10.5 F57B10.5 TMED7 transmembrane Emp24 Protein Transport Domain Containing 7 25/25 Metabolic proteins IV R11A8.6 iars-1 IARS isoleucyl-tRNA synthetase 43/50 I F36A2.8 phip-1 PHPT1 phosphohistidine phosphatase 1 50/50 V D2023.2 pyc-1 PC pyruvate carboxylase 50/50 V F46B6.5 F46B6.5 SRRM5 Serine/Arginine Repetitive Matrix 5 50/50 II H17B01.1 fgt-1 SLC2A3 solute carrier family 2 (facilitated glucose transporter), member 3 38/39 I F35C12.2 ncx-4 SLC24A2 solute Carrier Family 24 (Sodium/Potassium/Calcium Exchanger), Member 2 50/50 IV B0035.16 mttu-1 TRMU tRNA 5-methylaminomethyl-2-thiouridylate methyltransferase 50/50 II Y53F4B.18 Y53F4B.18 FAAH2 fatty acid amide hydrolase 2 50/50 IV F40F11.2 mig-38 PRG4 proteoglycan 4 (megakaryocyte stimulating factor) 43/50 III W04A8.1 W04A8.1 MCPH1 microcephalin 1 49/50 Other proteins I D1007.4 D1007.4 GEMIN6 gem (nuclear organelle) associated protein 6 50/50 I C04F12.1 C04F12.1 AGO4 argonaute RISC catalytic component 4 32/32 X W04G3.8 lpr-3 TSC22D2 TSC22 Domain Family, Member 2 34/34 I C27A12.2 C27A12.2 ZNF79 zinc finger protein 79 35/35 Nematode specific proteins I W09C5.12 W09C5.12 none none 50/50 I F17B5.5 clec-110 none none 46/46 V C53A5.1 ril-1 none none 50/50 X C14A6.3 C14A6.3 none none 50/50 X F15G10.3 piRNA none none 38/38 X K09E3.5 K09E3.5 none none 34/34 I F55D12.8 ncRNA none none 40/40 II Y46G5A.36 Y46G5A.36 none none 43/43 i F55D12.1 F55D12.1 none none 49/49 II C14A4.12 C14A4.12 none none 48/49

Chapter 4

Candidate genes specifically suppress proliferation of let-60 mutant epithelial cells

Video-gamer 1: Irish people love potatoes. Video-gamer 2: Come on, EVERYONE loves potatoes! Ritav comments: NO, ... snakes certainly don’t!!! - Breakfast-time commentary on youtube videos (Unpublished).

4.1 Introduction

Differential effect of mesodermal genes on proliferation of adjacent normal cells or mutant cells

Activation of let-60/Ras pathway plays a crucial role in the regulation of cell proliferation. During normal vulva development, proliferative signal, Lin-3/EGF, from neighboring anchor cell activates let-60/Ras signaling cascade transiently and in a very controlled manner, leading to proliferation of VPCs to form 22 cells of vulva. In addition to lin-3/EGF signal, Wnt signals from gonad and muscle and the lateral notch signaling among the VPCs also plays role in determining their fate (primary versus secondary), subsequent proliferation and localization of the daughter cells (vulva symmetry).

Mutations in the components of let-60/Ras pathway, such as let-60(n1046)/Ras gain-of- function mutation, lead to its constitutive activation resulting in over proliferation of vulva precursor cells manifested as multiple vulva phenotype.

Mutations in Ras genes are common across wide spectrum of tumors originating from different organ sites. Numerous studies have shown tissue specific role of various 58

Ras (N-Ras, H-Ras and K-Ras) oncogenes in tumor growth. For example, over 95% of pancreatic cancer, over 50% of colorectal cancer and over 20% of lung cancer cases harbor KRas mutations (Peeters et al., 2015, Hobbs and Rossman, 2016). One hallmark of these tumors is abundance of stromal component, which is reported to reprogram in response to interactions with mutant cells and co-evolve with malignant cells (Ueno and

Talbot, 2004; Xu and Apte, 2014; Isella et al., 2015; Bremnes et al., 2011; El-Nikhely and Savai, 2012). Numerous studies have reported differential gene expression of tumor- associated fibroblasts, the predominant stromal cells in tumors, compared to the normal fibroblasts (Erez et al., 2010; Franco et al., 2010; Berdiel-Acer et al., 2014, Finak et al.,

2008). The co-culture studies of cancer associated fibroblasts (CAFs) or normal fibroblasts (NFs) with mutant cells suggest role of pathological signaling cooperation between CAFs and mutant cells (Erez et al., 2010; Cirri and Chiarugi, 2011). Are there some mesenchymal genes that exclusively play role in pathological interactions with mutant cells, but not in the interactions with normal cells to promote their proliferation?

The candidate genes from our genome-wide screen provide us unique opportunity to examine if they play differential role in influencing proliferation of adjacent mutant or normal VPCs (epithelial cells).

In this chapter, we utilized mesoderm-specific RNAi strain of C. elegans to investigate if the candidate genes have effect on the proliferation of VPCs, carrying the wild-type allele of let-60 gene as opposed to let-60(gf) allele, during the vulva development. We present the data of this targeted mini-screen to evaluate the effect of

RNAi-mediated knock down of the candidate genes in mesodermal tissues on the proliferation of the VPCs carrying wild-type allele of let-60.

4.2 Results and Discussion

The enhancers of let-60(gf)/Ras are dependent on let-60 genotype

We examined whether knockdown of the candidate genes (from the genome-wide screen) specifically inhibit let-60(gf) mediated over-proliferation, or instead have a general role in the proliferation of VPCs during vulval development. For this query, we tested each gene for its effect in CM2073, a strain genetically identical to CM2453 except the strain is homozygous wild type at let-60/Ras (Figure 4.1A) (Liu et al., submitted).

L4 worms were evaluated for whether vulval morphology is normal or fewer VPCs undergo proliferation. Our results suggest that a majority of the genes (38 of 47) have no knockdown phenotype on their own, and none has a robust effect. We interpret that the enhancer genes specifically impact the let-60(gf)/Ras mediated over-proliferation

(Figures 4.2, 4.3).

The results of this mini-screen suggest that RNAi of the majority (38 of 47) of the candidate genes did not have effect on the vulva phenotype of mesodermal tissue-specific

RNAi strain carrying let-60(wt) allele. The RNAi of a few (nine) candidate genes resulted in abnormal vulva phenotype, however, RNAi-mediated suppression of proliferation was very weak in strain with let-60(+) compared to the strong suppression of proliferation in the let-60(gf) case (Figure 4.2). Furthermore, these 9 genes do not fall in the same

60 functional category of candidate genes indicating no specific role of particular category of genes in general suppression of proliferation of VPCs.

It is noteworthy that the candidate genes do not represent genes functioning down-stream of let-60/Ras. In the previous whole body RNAi screens for suppressors of the Muv phenotype associated with let-60(gf) have identified a number of important genes (Singh and Han, 1995; Gu et al., 1998, Li et al., 2000; Kao et al., 2004). None of these genes overlaps with those from our screen. We attribute these differences to two main sources. First, the previous screens did not preferentially select for genes that function non-autonomously. Consequently, they uncovered many suppressors that function downstream of let-60/Ras, within the VPCs themselves. This feature also skewed the results against the identification of essential genes, as body-wide disruption would render the animals non-viable. Our tissue-specific RNAi based approach has allowed discovery of new roles for essential genes in influencing inappropriate vulval cell division, as majority of the genes are known or predicted to be essential. Second, we selected genes that specifically suppress excessive cell division (leading to Normal vulva phenotype), rather than vulval cell division generally (which leads to a Vulvaless, or Vul phenotype). In contrast, many previously described let-60(gf) suppressors reduce the Muv phenotype, but confer a reduced VPC division (Vul phenotype) on their own. Overall, our tissue-specific screen has enriched for new genes not previously identified to play a role in vulval development.

It is well recognized that the tumor microenvironment plays a crucial role in determining tumor growth, disease progression and metastasis in several types of cancers.

(Hanahan and Folkman, 1996; Bergers and Hanahan, 1999; Orimo et al., 2005; Kalluri and Zeisberg, 2006; Finak et al., 2008; Gabrilovich and Nagaraj, 2009; Erez and

Hanahan, 2010; Franco and Hayward, 2010; Cirri and Chiarugi, 2011; Chow and Merad,

2011; Mantovani and Jaillon, 2011; Porta et al., 2011; Nieman et al., 2011; Beck et al.,

2011; Daldrup-Link et al., 2011; Rattigan et al., 2012; Hanahan & Coussens, 2012;

Martinez-Outschoorn and Lisanti, 2014). The cancer associated fibroblasts (CAFs), the mesenchymal cells, are reported as most prominent cell types inhabiting in the stromal compartment of tumors that can determine clinical outcome in breast and colorectal cancers. (Franco et al., 2010; Cirri & Chiarugi, 2011). It is also reported that stromal cells and tumor cells co-evolve and stromal compartment increases as the tumor grows (Erkan,

2013). Differential gene expression in CAFs compared to normal fibroblasts is independent indicator of clinical outcome and disease reoccurrence. (Finak et al., 2008).

In this context, differences in the effect of genes in the activated versus normal let-60/Ras pathway may suggest the altered signaling between mesenchymal cells and mutant cells versus normal cells.

In normal development and in cancer, fibroblasts provide growth factors, cytokines and chemokines to provide growth signals and modulate immune response.

These growth-regulatory signals are under tight regulation and surveillance such that they are spatiotemporally expressed only when and where they are required and repressed in the ectopic cells. This feature is evident in C. elegans vulva development as well (Cui,

Chen, et al., 2006). Moreover, the Ras-signaling pathway is highly conserved between C. elegans and Humans (Figure1. 8B, chapter 1). Given that majority of the candidate genes

62 are conserved from C. elegans to Humans, the collaboration of mesenchymal signaling pathways with the active signaling pathway of epithelial cells may also be conserved across the species. It will be interesting to examine two important features: (a) whether expression of the candidate genes is different in the CAFs versus NFs derived from tumor versus normal tissues (b) effect of knockdown of these genes in the stromal cells on the proliferation of the co-cultured epithelial cells.

Taken together, these results provide interesting avenues to explore the specific signaling collaborations of mesodermal cells with the adjacent epithelial cells leading to regulation of proliferative potential of mutant or normal cells.

4.3 Materials and methods

Tissue-specific RNAi C. elegans models

The mesodermal tissue-specific RNAi strain (CM2073) was generated and characterized by previous lab members (Huayang Liu, et al). This strain carries a transgene that drives expression of rde-1(+) in the mesodermal tissues (anchor cell, gonad and muscle) specifically in a worm with rde-1(-) background rendering RNAi competence in the rde-1(+) rescued tissues (Figure 4.1A). In addition, rrf-3(-) mutation was introduced to enhance the RNAi effect. The baseline phenotype if this strain is normal vulva (Figure 4.1B).

Experimental design of the targeted RNAi screen

A targeted candidate-based mini-screen was designed to test effect of candidate genes (from the genome-wide screen) on the mesodermal tissue-specific RNAi strain

63 with wild-type let-60 genotype (CM2073). The RNAi-clone carrying bacteria were cultured overnight and spread on 3 wells of a 12-well NGM plates supplemented with

IPTG. The bacterial bed was allowed to grow and produce IPTG-induced RNAi at room temperature. Gravid hermaphrodites were bleached using a standard bleaching protocol to extract eggs. Each well of 12-well plate was spotted with 100-150 eggs and were transferred to 20ºC incubator to grow to L4 stage (Figure 4.1C).

Sample preparation and Microscopy

The slides for high-magnification imaging were prepared using 3.5% Noble agar supplemented with 1% 1M NaN3. The L4 stage worms were harvested from all the 3 wells of each RNAi and mounted on the agar slide. The worms at L4 stage were imaged using 100X magnification.

Cut-offs /threshold for candidate gene selection

The baseline phenotype of the experimental strain (CM2073) was normal vulva

(Figure 4.1B). Suppression of vulva proliferation from normal vulva phenotype leads to abnormal vulva structures. The vulva phenotype deviating from normal vulva phenotype was considered as having effect of RNAi of the candidate gene.

Figure 4.1: C. elegans transgenic model and targeted screen

(A) Cartoon representation of the tissue-specific RNAi strain (right top) cartoon representation of tissues in which rde-1(+) is reinstated using mesenchymal (continued)

Figure 4.1: C. elegans transgenic model and targeted screen (continued)

(anchor cell, gonad and muscle) tissue specific promoters (shown in green color). AC, anchor cell, SG, somatic gonad, M, muscle, VPC, vulva precursor cells. (adapted from

Liu et al) (Lower right panel) image of vulva phenotype of this strain at larva L4 stage at

100X magnification. The baseline vulva phenotype of this strain is normal vulva.

(B) Schematic of RNAi mini-screen. The RNAi-clone containing bacteria were cultured overnight at 37ºC incubator. Next day, bacterial cultures were spread on the NGM agar plates that were supplemented with IPTG and Carbenicillin. The bacterial bed was allowed to grow for a day. Next day, eggs harvested from the experimental strain were seeded on each well and transferred to incubator to allow them to hatch, feed on RNAi- bacteria until they reach L4 stage of larva development. At L4 stage, larvae were harvested and imaged using 100X Nomarski microscopy.

Figure 4.2: Results of targeted mini-screen of candidate genes in mesenchymal tissue-specific RNAi strain

(A) Graph representing quantification of high-mag data for each candidate RNAi treated worms. Percent of normal vulva were counted in each RNAi treatment. The baseline of this strain is normal vulva. Black bar represent control RNAi treatment representing all imaged worms had normal vulva.

(B) Representative 100X Nomarski images of vulva. (left panel) Normal vulva at L4 stage for control RNAi (black arrow), (right panel) image of vulva of worms treated with

RNAi (iars-1, C14A4.12 and rpl-37) from different functional groups. Abnormal or under-induced vulva at L4 stage (while arrows).

Figure 4.3: Comparison of effect of stromal-specific RNAi on suppression of over- proliferation of let-60(gf)-mutant versus suppression of normal let-60(wt) mediated proliferation of VPCs

(A) Graph representing quantification of high-mag data for each candidate RNAi treated worms comprising let-60(gf) mutation background. Percent of normal vulva were counted in each RNAi treatment. The baseline of this strain is multivulva (Muv) phenotype. Black bar represent control RNAi treatment.

(B) Graph representing quantification of high-mag data for each candidate RNAi treated worms comprising let-60(wt) background. Percent of normal vulva were counted in each

RNAi treatment. The baseline of this strain is normal vulva. Black bar represents control

RNAi treatment.

Mesenchyme-specific Mesenchyme-specific Worm Human RNAi Strain RNAi Strain Chr Gene ID common Description of Human orthologs Ortholog High Mag (Normal/total) High Mag (Normal/total) gene name with let-60 (gf) with let-60(wt) Histone proteins X F45E1.6 his-71 H3F3B H3 histone, family 3B (H3.3B) 39/50 50/50 V F45F2.9 F45F2.9 HDDC2 HD domain containing 2 74/100 50/50 IV B0035.9 his-46 HIST1H4G histone cluster 1, H4g 41/50 40/40 II F08G2.2 his-43 HIST2H2AB histone cluster 2, H2ab 21/25 65/65 V T10C6.12 his-3 HIST2H2AB histone cluster 2, H2ab 46/50 50/50 V F45F2.12 his-8 HIST2H2BF histone cluster 2, H2bf 14/23 50/50 IV F17E9.10 his-32 HIST2H3D histone cluster 2, H3d 40/50 38/38 IV F55G1.2 his-59 HIST2H3D histone cluster 2, H3d 38/42 50/50 II W09H1.2 his-73 HIST2H3D histone cluster 2, H3d 134/152 49/50 Ribosomal proteins I C04F12.4 rpl-14 RPL14 ribosomal protein L14 87/90 50/50 III K11H12.2 rpl-15 RPL15 ribosomal protein L15 45/52 45/50 X F55D10.2 rpl-25.1 RPL23A ribosomal protein L23a 47/50 50/50 II W01D2.1 rpl-37 RPL37 ribosomal protein L37 82/95 40/50 IV Y57G11C.16 rps-18 RPS18 ribosomal protein S18 18/22 50/50 I C37A2.7 C37A2.7 RPLP2 ribosomal protein, large, P2 44/50 32/42 ATP Synthase proteins V F32D1.2 hpo-18 ATP5E ATP synthase, H+ transporting, mitochondrial F1 complex, epsilon subunit 47/50 42/42 V C06H2.1 atp-5 ATP5H ATP synthase, H+ transporting, mitochondrial Fo complex, subunit d 44/50 50/50 V T05H4.12 atp-4 ATP5J ATP synthase, H+ transporting, mitochondrial Fo complex, subunit F6 95/99 50/50 Vesicular Transport proteins IV Y38F2AL.4 vha-3 ATP6V0C ATPase, H+ transporting, lysosomal 16kDa, V0 subunit c 64/80 50/50 IV Y38F2AL.3 vha-11 ATP6V1C1 ATPase, H+ transporting, lysosomal 42kDa, V1 subunit C1 21/28 50/50 X C18B12.6 C18B12.6 ERGIC2 ERGIC And Golgi 2 31/38 50/50 V C31B8.4 C31B8.4 GOLGA4 golgin A4 44/50 50/50 I F57B10.5 F57B10.5 TMED7 transmembrane Emp24 Protein Transport Domain Containing 7 20/50 25/25 Metabolic proteins IV R11A8.6 iars-1 IARS isoleucyl-tRNA synthetase 38/58 43/50 I F36A2.8 phip-1 PHPT1 phosphohistidine phosphatase 1 40/50 50/50 V D2023.2 pyc-1 PC pyruvate carboxylase 38/50 50/50 V F46B6.5 F46B6.5 SRRM5 Serine/Arginine Repetitive Matrix 5 40/50 50/50 II H17B01.1 fgt-1 SLC2A3 solute carrier family 2 (facilitated glucose transporter), member 3 58/86 38/39 I F35C12.2 ncx-4 SLC24A2 solute Carrier Family 24 (Sodium/Potassium/Calcium Exchanger), Member 2 37/50 50/50 IV B0035.16 mttu-1 TRMU tRNA 5-methylaminomethyl-2-thiouridylate methyltransferase 43/50 50/50 II Y53F4B.18 Y53F4B.18 FAAH2 fatty acid amide hydrolase 2 104/105 50/50 IV F40F11.2 mig-38 PRG4 proteoglycan 4 (megakaryocyte stimulating factor) 13/29 43/50 III W04A8.1 W04A8.1 MCPH1 microcephalin 1 43/50 49/50 Other proteins I D1007.4 D1007.4 GEMIN6 gem (nuclear organelle) associated protein 6 47/50 50/50 I C04F12.1 C04F12.1 AGO4 argonaute RISC catalytic component 4 61/72 32/32 X W04G3.8 lpr-3 TSC22D2 TSC22 Domain Family, Member 2 20/40 34/34 I C27A12.2 C27A12.2 ZNF79 zinc finger protein 79 8/13 35/35 Nematode specific proteins I W09C5.12 W09C5.12 none none 44/72 50/50 I F17B5.5 clec-110 none none 27/50 46/46 V C53A5.1 ril-1 none none 46/50 50/50 X C14A6.3 C14A6.3 none none 26/50 50/50 X F15G10.3 piRNA none none 28/36 38/38 X K09E3.5 K09E3.5 none none 21/28 34/34 I F55D12.8 ncRNA none none 36/50 40/40 II Y46G5A.36 Y46G5A.36 none none 25/50 43/43 i F55D12.1 F55D12.1 none none 39/50 49/49 II C14A4.12 C14A4.12 none none 41/52 48/49

Table 4.1: List of candidate genes and their effect in mesodermal-RNAi strains with and without let-60(gf) mutation.

Chapter 5

Spatial function of candidate genes in suppression of let-60(gf) mediated hyper-

proliferation

Ritav: Why can’t we see air? Mom: Because it has no color! Ritav: Then, why can we see water?? - Ritav, 3 year old (Unpublished).

5.1 Introduction

Tissue-specific gene function and intercellular signaling during neoplastic growth

Tissue and cell-type identity lie at the core of normal physiology and disease. In normal tissues, different resident specific cell types are constantly processing and responding to signals from one another and from the extracellular matrix (ECM) around them in order to coordinate their function, organization, and rates of death and division.

The spatial and temporal variation in expression of a gene carries crucial information of its function and role in tissue homeostasis (Bassett, 1999; Wilfinger and Meyer, 2013).

Though most disease-causing genes are expressed throughout the body, they typically have an effect in one tissue, and unraveling that effect can reveal the molecular basis of the disease. Understanding the functional genetic underpinnings of individual cell types in complex tissues and neoplastic growth is crucial for developing improved diagnostics and therapeutics. For example, PTEN, a gene that has causal relationship to

Cowden Symdrome, is well known for its tumor suppressor function in the epithelial cells. Recently, our laboratory identified role of PTEN in fibroblasts that enhances tumorigenic potential of adjacent ErBb2 mutant cells leading to significantly bigger tumors (Trimboli et al., 2009). Similarly, studies in other laboratories have shown role of p53 in fibroblast compartment of the liver tumors leading to their increased tumor growth

(Trachootham et al., 2013; Addadi et al., 2010). Both, PTEN and p53 are well known for their tumor suppressive functions in epithelial compartments in tumor growth.

Fibroblasts-specific depletion of these genes unraveled their novel cell non-autonomous role in the tumor growth and disease progression (Trimboli et al., 2009; Trachootham et al., 2013; Addadi et al., 2010). These reports prompted us to identify whether the candidate genes function exclusively in mesenchymal cells or in both mesenchymal and epithelial cells to suppress the Muv phenotype?

In this chapter, we utilized epithelial tissue-specific RNAi strain that also contains let-60(gf) and rrf-3(-) alleles to investigate role of candidate genes in epithelial cells with respect to suppression of Muv phenotype.

5.2 Results and Discussion

Cell type specificity of hits (mesodermal versus VPC)

We examined whether knockdown of the 47 suppressor genes specifically inhibits let-60(gf) mediated over-proliferation in a cell non-autonomous manner or can function in both mesodermal (gonad, anchor cell and muscles) as well as epithelial (VPCs) cells to result in suppression of let-60(gf) mediated Muv phenotype. For this query, we tested

71 each gene for its effect in CM2611, a strain genetically identical to CM2453 except the strain includes a transgene that targets RNAi knockdown specifically to the epidermal

VPCs (rather than mesoderm), and is RNAi-defective in the rest of the worm. RNAi- treated worms were evaluated for whether they produced a normal vulval development, compared to excessive division (Muv phenotype). A majority of the genes did not suppress the Muv phenotype in this case, indicating that they impact VPC proliferation only when targeted in non-VPCs. (Figure 5.1). Only 8 of 47 genes (his-32, rpl-15,

C37A2.7, F57B10.5, iars-1, Y53F48.18, D1007.4 and lpr-3) showed significant change in the Muv phenotype compared to the control RNAi. Interestingly, two of the seven genes significantly increased the Muv phenotype compared to the control. These results suggest that these genes (iars-1 and lrp-3) play opposing roles in the mesodermal compartment compared to the VPC compartment with respect to proliferation while six genes (his-32, rpl-15, C37A2.7, F57B10.5, D1007.4 and Y53F48.18) function as suppressors of the Muv phenotype both cell autonomously and cell non-autonomously.

In summary, we showed by utilizing tissue-specific RNAi strategies that candidate genes suppress let-60(gf)/Ras mediated oncogenic proliferation in vivo. Here we presented data to suggest that the knockdown of most candidate genes elicit their effect on robust suppression of Muv phenotype cell non-autonomously. Six genes (his-

32, rpl-15, C37A2.7, F57B10.5, D1007.4 and Y53F48.18) also function in VPCs to suppress Muv phenotype albeit to much weaker degree compared to when knocked down in the mesodermal cells. Interestingly, two genes (iars-1 and lrp-3) play opposing roles in the mesodermal compartment compared to the VPC compartment with respect to VPC

72 proliferation. Taken together, these results stratify candidate genes that specifically function cell non-autonomously from the ones that function both cell autonomously as well as cell non-autonomously in let-60(gf)/Ras mediated proliferation of VPCs.

5.3 Materials and methods

C. Nematode handling and processing

C. elegans strains were cultured and maintained under standard conditions on

NGM made with 2x (5 g/L) peptone and supplemented with 1µM IPTG and 50 µg/ml carbenicillin. All experiments were performed at 20°C unless otherwise noted.

Tissue-specific RNAi C. elegans models

The VPC tissue specific RNAi strain (CM2611) strain carries a transgene that drives expression of rde-1(+) in the vulva epithelial tissue (VPCs) specifically in a worm with rde-1(-) background rendering RNAi competence in the rde-1(+) rescued tissues

(Figure 5.1A). In addition, rrf-3(-) and let-60(n1460) alleles were introduced to enhance the RNAi effect and activation of Ras-pathway, respectively. The baseline phenotype if this strain is multi-vulva (Figure 5.1B).

VPC-specific RNAi mini-screen set-up

A targeted candidate-based mini-screen was conducted to test effect of candidate genes (from the genome-wide screen) on the epithelial tissue-specific RNAi strain with let-60(gf) genotype (CM2611). The RNAi-clone carrying bacteria were cultured overnight and spread on 3 wells of a 12-well NGM plates supplemented with IPTG. The bacterial bed was allowed to grow and produce IPTG-induced RNAi at room temperature. Gravid hermaphrodites were bleached using a standard bleaching protocol to extract eggs. Each well of 12-well plate was spotted with 100-150 eggs and were transferred to 20 ºC incubator to grow to L4 stage (Figure 5.1C).

Sample preparation and Microscopy

The slides for high-magnification imaging were prepared using 3.5% Noble agar supplemented with 1% 1M NaN3. The L4 stage worms were harvested from all the 3 wells of each RNAi and mounted on agar slide. The worms at L4 stage were imaged using 100X magnification (Figure 5.1A).

Threshold for hit calling /Cut-offs and statistics

The baseline phenotype of the experimental strain (CM2611) is multiple vulva

(figure 5.1b). This mini screen was done at 100X magnification under an oil immersion lens using a Zeiss Axioscope microscope. RNAi treatment was completed as described above. At the L4 stage, RNAi fed larvae were harvested using M9 washing from the 12- well plates. Agar-slides were prepared using 3.5% Noble agar supplemented with 10 mM

NaN3 to mount RNAi treated larvae for imaging. Only worms exhibiting a normal vulva phenotype are reported in Table 3; over- or under-induction or morphological defects were considered abnormal.

Using Z-test for proportions, statistically significant differences of RNAi of candidate genes compared to control RNAi was calculated.

Figure 5.1: C. elegans model and targeted screen in VPC-tissue specific RNAi strain

(A) Cartoon representation of the tissue-specific RNAi strain (right top) cartoon representation of tissues in which rde-1(+) is reinstated using vulva precursor cells

(VPCs) tissue specific promoter (shown in green color). AC, anchor cell, SG, somatic gonad, M, muscle, VPC, vulva precursor cells. (adapted from Liu et al) (continued)

Figure 5.1: C. elegans model and targeted screen in VPC-tissue specific RNAi strain

(continued)

(Lower right panel) image of multi vulva (Muv) phenotype of this strain at larva L4 stage at 100X magnification. The baseline vulva phenotype of this strain is Muv.

(B) Schematic of RNAi mini-screen. The RNAi-clone containing bacteria were cultured overnight at 37ºC incubator. Next day, bacterial cultures were spread on the NGM agar plates that were supplemented with IPTG and Carbenicillin. The bacterial bed was allowed to grow for a day at room temperature. Next day, eggs harvested from the experimental strain were seeded on each well and transferred to incubator to allow them to hatch, feed on RNAi-bacteria until they reach L4 stage of larva development. At L4 stage, larvae were harvested and imaged using 100X Nomarski microscopy.

Figure 5.2: Results of targeted mini-screen of candidate genes in vulva epithelium tissue-specific RNAi strain

(A) Graph of high-mag data stromal-specific compared to VPC-specific.

(B) Representative images of hits

Figure 5.3: Comparison of effect of stromal-specific RNAi to VPC-specific RNAi on suppression of over-proliferation of let-60(gf)-mutant VPCs.

(A) Quantitative representation of % of C. elegans exhibiting normal vulva at L4 stage after treatment with RNAi in mesodermal-specific RNAi strain with let-60(gf) sensitization. Black bar, control RNAi. Blue bars, candidate gene RNAi (Left). Cartoon representation and phenotype of the experimental strain used in these experiments

(Right).

(B) Quantitative representation of % of C. elegans exhibiting normal vulva at L4 stage after treatment with RNAi in VPC-specific RNAi strain with let-60(gf) sensitization. Black bar, control RNAi. Blue bars, candidate gene RNAi (Left). Cartoon representation and phenotype of the experimental strain used in these experiments (Right). * (p<0.01) compared to control RNAi treatment. 79

Mesenchyme-specific VPC-specific RNAi Worm Human RNAi Strain Strain Chr Gene ID common Description of Human orthologs Ortholog High Mag (Normal/total) High Mag (Normal/total) gene name with let-60 (gf) with let-60(gf)

Chromatin remodeller proteins X F45E1.6 his-71 H3F3B H3 histone, family 3B (H3.3B) 39/50 6/50 V F45F2.9 F45F2.9 HDDC2 HD domain containing 2 74/100 12/87 IV B0035.9 his-46 HIST1H4G histone cluster 1, H4g 41/50 5/51 II F08G2.2 his-43 HIST2H2AB histone cluster 2, H2ab 21/25 5/50 V T10C6.12 his-3 HIST2H2AB histone cluster 2, H2ab 46/50 3/50 V F45F2.12 his-8 HIST2H2BF histone cluster 2, H2bf 14/23 5/50 IV F17E9.10 his-32 HIST2H3D histone cluster 2, H3d 40/50 10/50 IV F55G1.2 his-59 HIST2H3D histone cluster 2, H3d 38/42 3/50 II W09H1.2 his-73 HIST2H3D histone cluster 2, H3d 134/152 8/50 Ribosomal proteins I C04F12.4 rpl-14 RPL14 ribosomal protein L14 87/90 10/84 III K11H12.2 rpl-15 RPL15 ribosomal protein L15 45/52 17/95 X F55D10.2 rpl-25.1 RPL23A ribosomal protein L23a 47/50 3/50 II W01D2.1 rpl-37 RPL37 ribosomal protein L37 82/95 5/50 IV Y57G11C.16 rps-18 RPS18 ribosomal protein S18 18/22 6/50 II W01D2.1 rpl-37 RPL37 ribosomal protein L37 82/95 5/50 I C37A2.7 C37A2.7 RPLP2 ribosomal protein, large, P2 44/50 10/50 ATP Synthase proteins V F32D1.2 hpo-18 ATP5E ATP synthase, H+ transporting, mitochondrial F1 complex, epsilon subunit 47/50 9/50 V C06H2.1 atp-5 ATP5H ATP synthase, H+ transporting, mitochondrial Fo complex, subunit d 44/50 5/48 V T05H4.12 atp-4 ATP5J ATP synthase, H+ transporting, mitochondrial Fo complex, subunit F6 95/99 3/40 Vesicular Transport proteins IV Y38F2AL.4 vha-3 ATP6V0C ATPase, H+ transporting, lysosomal 16kDa, V0 subunit c 64/80 5/81 IV Y38F2AL.3 vha-11 ATP6V1C1 ATPase, H+ transporting, lysosomal 42kDa, V1 subunit C1 21/28 13/90 X C18B12.6 C18B12.6 ERGIC2 ERGIC And Golgi 2 31/38 4/50 V C31B8.4 C31B8.4 GOLGA4 golgin A4 44/50 7/50 I F57B10.5 F57B10.5 TMED7 transmembrane Emp24 Protein Transport Domain Containing 7 20/50 21/52 Metabolic proteins IV R11A8.6 irs-1 IARS isoleucyl-tRNA synthetase 38/58 0/50 I F36A2.8 phip-1 PHPT1 phosphohistidine phosphatase 1 40/50 7/50 V D2023.2 pyc-1 PC pyruvate carboxylase 38/50 6/50 V F46B6.5 F46B6.5 SRRM5 Serine/Arginine Repetitive Matrix 5 40/50 1/50 II H17B01.1 fgt-1 SLC2A3 solute carrier family 2 (facilitated glucose transporter), member 3 58/86 5/48 I F35C12.2 ncx-4 SLC24A2 solute Carrier Family 24 (Sodium/Potassium/Calcium Exchanger), Member 2 37/50 5/50 IV B0035.16 mttu-1 TRMU tRNA 5-methylaminomethyl-2-thiouridylate methyltransferase 43/50 9/50 II Y53F4B.18 Y53F4B.18 FAAH2 fatty acid amide hydrolase 2 104/105 10/50 IV F40F11.2 mig-38 PRG4 proteoglycan 4 (megakaryocyte stimulating factor) 13/29 5/50 III W04A8.1 W04A8.1 MCPH1 microcephalin 1 43/50 5/50 Other proteins I D1007.4 D1007.4 GEMIN6 gem (nuclear organelle) associated protein 6 47/50 11/50 I C04F12.1 C04F12.1 AGO4 argonaute RISC catalytic component 4 61/72 8/80 X W04G3.8 lpr-3 TSC22D2 TSC22 Domain Family, Member 2 20/40 0/50 I C27A12.2 C27A12.2 ZNF79 zinc finger protein 79 8/13 9/50 Nematode specific proteins I W09C5.12 W09C5.12 none none 44/72 9/50 I F17B5.5 clec-110 none none 27/50 6/50 V C53A5.1 ril-1 none none 46/50 3/50 X C14A6.3 C14A6.3 none none 26/50 3/50 X F15G10.3 piRNA none none 28/36 8/50 X K09E3.5 K09E3.5 none none 21/28 5/50 I F55D12.8 ncRNA none none 36/50 6/50 II Y46G5A.36 Y46G5A.36 none none 25/50 5/50 i F55D12.1 F55D12.1 none none 39/50 4/50 II C14A4.12 C14A4.12 none none 41/52 1/50

Table 5.1: Most of candidate genes function in cell non-autonomous manner.

(continued from page 80)

Table showing comparison of knock down of candidate genes in mesodermal (gonad, anchor cell and muscles) tissues versus epithelial (VPCs) tissues on the oncogenic Ras mediated over-proliferation.

Chapter 6

Concluding remarks and future perspectives

Ritav : I got candies, balloons and LED lit rings! What did you get? Ritav’s Friends: We too! It was fun loot!!! - Pin˜ata party at 5th B’Day Celebration (Unpublished).

In this thesis, we have presented a genome-wide screen to identify genes in mesodermal (gonad, AC and muscles) tissues that influence proliferation of the let-

60(gf)/Ras mutant epithelial (VPCs) cells.

For this study, we developed transgenic strains of C. elegans that possess two properties: (i) mesodermal (gonad, AC, muscle) tissues-specific RNAi, and, (ii) over- proliferation of VPCs as a result of let-60(gf)/Ras mutation. This strain allowed screening for genes that function in mesodermal tissues to influence proliferation of let-60(gf)/Ras.

The genome-wide screen resulted in 47 candidate genes that show this effect and that could be functionally categorized into five categories. Importantly, the identified genes are enriched for the conserved genes between C. elegans and humans. Remarkably, genes in the same category encode proteins that are part of molecular complexes. For example, most of the histone proteins are part of nucleosome, most of the ribosomal genes are part of large subunit of ribosome, the ATP synthase proteins are part of mitochondrial F0F1 complex, and lysomal ATPases (vha-3 and vha-11) have same operon in C. elegans

81 genome. This information suggests that the screen enriched for biologically meaningful process(es) that take part in the inter-cellular communication of mesodermal and let-

60(gf)-mutated cells. This level of enrichment of select molecular complexes or functionally related genes brings confidence in the results of screen and provides initial hints on the molecular or biological processes underlying the communication of mesenchymal-mutant epithelial cells.

The targeted screen in the mesenchyme-specific RNAi strain with wild-type let-

60/Ras genotype provided insights on role of these genes in the mesodermal cells in the process of intercellular communication with normal epithelial cells. Compared to an effect on the proliferation of mutant epithelial cells (VPCs), most of the genes did not have effect on the proliferation of normal epithelial cells (VPCs). These findings suggest that these genes are specific to intercellular communication of mesenchymal and mutant epithelial cells and play role in regulating the over-proliferation rather than normal proliferation.

On the other hand, the targeted screen of candidate genes in VPC-specific RNAi strain suggested that most of the genes function in the mesenchymal compartment only and do not have effect on proliferation from the VPC compartment. Some of the genes suppressed over-proliferation of let-60(gf)-mutant VPCs but only weakly and not as robustly as their function from the mesodermal tissue side. Two of the genes enhanced the over-proliferation (Muv phenotype) rather than suppressing it, indicating their opposing roles in the mesodermal and epithelial cells on influencing the proliferation of the VPCs.

To elicit their cell non-autonomous effect on regulating the proliferation of the mutant cells, these genes potentially could use several mechanisms. The cell-cell communication could broadly be speculated in these forms: (i) by direct cell-cell contact,

(ii) by changing the secretome (ligands to EGFR/RAS/MAPK and WNT pathways), (iii) exosome/vesicle mediated transportation or (iv) micro RNAs. Given that there a few genes that are involved in vesicle transport, they might be eliciting their effect via exosomal communication between the mesenchymal and epithelial cells. Further studies to explore the above-mentioned potential intercellular communication mechanisms might help in unraveling the molecular mechanisms underlying cell-cell communication that involve genes identified from this screen.

We focused our efforts on validating (in secondary and high-magnification screening assays) genes that, when knocked down, had suppressed the over-proliferation to a great extent (<30% from baseline). The set of genes that had intermediated suppression phenotype in our primary screen may also involve bona fide cell non- autonomous modulators of proliferation. It might be informative to identify these less robust candidate genes and perform network analyses on these larger data-sets in disclosing links between the current hits to biologically relevant pathways underlying cell-cell communication.

Our current work provides a candidate list to direct future studies on cell non- autonomous regulators of oncogenic proliferation in mammalian model systems and represents and important step forward to understand the mechanisms underlying

83 mesenchymal (stromal) cell-mutant cell interactions within the microenvironment they reside.

In RNAi-based genome-wide screens, we need to be cautious about the confounding effects of RNAi, such as off-target effects or targeting of multiple genes in a gene family due to underlying sequence similarity of these genes. There is a possibility that our data contains genes that are suppressing Muv phenotype because of off-target effect. While the enrichment of components of select molecular complexes suggests enrichment of certain biologically important processes/molecular complexes and not off- target effects, it is still a good idea to validate these genes one more time. Currently, the general consensus is that the more (3 to 4) RNAi sequences targeting the same gene at different sequence sites that result in the same phenotype, the more confidence one places on that result. However, if only one out these RNAi for the same gene showing a phenotype, it is more likely the phenotype is due to an off target effect. Therefore, I would test these genes for off-target effects using 2-3 designed RNAi sequences to target these genes and test for their effect on the phenotype. One quick way to conduct this test would be to use soaking-in-RNAi solution assay rather than having to clone these newly designed targets in the expression vector and transform them in HT115 bacteria (the feeding-based RNAi assay). Additional testing will confirm the results and will set a confident path forward to investigate the cell non-autonomous role of these genes in the mammalian model systems. On the contrary, given that the RNAi clones are comprised of long genomic sequences, these RNAi clones might be knocking down multiple genes in a gene family, especially that have highly similar sequences. It might be especially true for

84 the histone genes recovered in our screen as they have highly overlapping sequences such that RNAi probe against one histone gene may be targeting other histone genes as well, thus resulting as “hits”. One way to address this problem is to utilize antibodies specific to targeted histone proteins to visualize their knockdown at the protein level by in situ immunohistochemistry (IHC) assays.

Our current experimental strain involves let-60(gf) mutation that has ~80+12%

Muv phenotype as baseline. This was a useful model to identify cell non-autonomous genes that dramatically suppress the Muv phenotype upon RNAi-mediated knockdown. A designed hypomorphic allele of let-60(gf), such that the Muv phenotype is ~40-60% would provide an ideal situation to not only identify the suppressors but also enhancers of the Muv phenotype upon RNAi-mediated knockdown to obtain a complete spectrum

(both enhancers and suppressors) of cell non-autonomous modulators of proliferation.

It is known that let-60/Ras- pathway and Wnt- pathway act in parallel in VPC and converge on lin-39 during the fate specification of the VPCs (Eisenmann, 1998).

They demonstrated that loss of function mutation in bar-1/beta-catenin can suppress the

Muv phenotype of the let-60(gf)/Ras mutant C. elegans from 95% to 22% (Eisenmann,

1998). Given that both let-60/Ras and Wnt pathways are involved in the VPC fate determination and proliferation, designing a screen using an experimental mesenchyme-

RNAi strain with background sensitization with cancer associated Wnt-pathway components such as bar-1/Beta-catenin, pry-1 or apr-1/APC will shed light on specific cell non-autonomous genetic networks that collaborate with Wnt-pathway mutant cells.

The technological advances of the past decades have transformed the scope of

85 biological questions that we can address in vivo using genetic systems. The tissue- specific RNAi system and functional screens provides a unique opportunity to “perturb and observe” each gene within the genomes in tissue specific manner, and decipher the sources of complex systems/tissue level phenotypes and underlying causes. We attempted to understand some of the underlying disease phenotypes, such as uncontrolled proliferation, and its cell extrinsic regulators, to understand potential causation of complicated diseases like cancers. Here, we have identified multiple functionally related genes affecting the over-proliferation of mutant cells in cell non-autonomous manner.

The worth of a functional genomic screen will be determined by the amount of insight brought forth from its results. Further studies to decipher the relevance of our identified genes to cancer cell proliferation are a subject of great interest moving forward. As a next step, we will test hypothesis regarding the roles of the genes identified by this screen in mammalian cancer models. Deciphering the mechanisms and clinical relevance of the candidate genes involved in stromal-cancer cell communications leading to disease progression will ultimately keep researchers busy for years to come.

References

Addadi, Y., Moskovits, N., Granot, D., Lozano, G., Carmi, Y., Apte, R. N., Oren, M. (2010). p53 status in stromal fibroblasts modulates tumor growth in an SDF1- dependent manner. Cancer Research, 70(23), 9650–9658.

Aroian, R. V., & Sternberg, P. W. (1991). Multiple functions of let-23, a Caenorhabditis elegans receptor tyrosine kinase gene required for vulval induction. Genetics, 128(2), 251–267.

Bacman, D., Merkel, S., Croner, R., Papadopoulos, T., Brueckl, W., & Dimmler, A. (2007). TGF-beta receptor 2 downregulation in tumour-associated stroma worsens prognosis and high-grade tumours show more tumour-associated macrophages and lower TGF-beta1 expression in colon carcinoma: a retrospective study. BMC Cancer, 7(1), 156.

Bassett, D. E., Eisen, M. B., & Boguski, M. S. (1999). Gene expression informatics --it’s all in your mine. Nature Genetics, 21(1 Suppl), 51–55.

Beck, A. H., Sangoi, A. R., Leung, S., Marinelli, R. J., Nielsen, T. O., van de Vijver, M. J., Koller, D. (2011). Systematic analysis of breast cancer morphology uncovers stromal features associated with survival. Science Translational Medicine, 3(108), 108ra113.

Beitel, G. J., Clark, S. G., & Horvitz, H. R. (1990). Caenorhabditis elegans ras gene let- 60 acts as a switch in the pathway of vulval induction. Nature, 348(6301), 503– 509.

Berdiel-Acer, M., Sanz-Pamplona, R., Calon, A., Cuadras, D., Berenguer, A., Sanjuan, X., Molleví, D. G. (2014). Differences between CAFs and their paired NCF from adjacent colonic mucosa reveal functional heterogeneity of CAFs, providing prognostic information. Molecular Oncology, 8(7), 1290–1305. 87

Bergers, G., Javaherian, K., Lo, K. M., Folkman, J., & Hanahan, D. (1999). Effects of angiogenesis inhibitors on multistage carcinogenesis in mice. Science (New York, N.Y.), 284(5415), 808–812.

Bhowmick, N. A., Chytil, A., Plieth, D., Gorska, A. E., Dumont, N., Shappell, S., Moses, H. L. (2004). TGF-beta signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science (New York, N.Y.), 303(5659), 848–851.

Bremnes, R. M., Dønnem, T., Al-Saad, S., Al-Shibli, K., Andersen, S., Sirera, R., Busund, L.-T. (2011). The role of tumor stroma in cancer progression and prognosis: emphasis on carcinoma-associated fibroblasts and non-small cell lung cancer. Journal of Thoracic Oncology: Official Publication of the International Association for the Study of Lung Cancer, 6(1), 209–217.

Brenner, S. (1974). The genetics of Caenorhabditis elegans. Genetics, 77(1), 71–94.

Bronisz, A., Godlewski, J., Wallace, J., Merchant, A., Nowicki, M., Mathsyaraja, H., Ostrowski, M. (2011). Reprogramming of the Tumor Microenvironment by Stromal Pten-regulated miR-320. Nature Cell Biology, 14(2), 159–167.

Burdine, R. D., Branda, C. S., & Stern, M. J. (1998). EGL-17(FGF) expression coordinates the attraction of the migrating sex myoblasts with vulval induction in C. elegans. Development (Cambridge, England), 125(6), 1083–1093.

Ceol, C. J., & Horvitz, H. R. (2004). A new class of C. elegans synMuv genes implicates a Tip60/NuA4-like HAT complex as a negative regulator of Ras signaling. Developmental Cell, 6(4), 563–576.

Chen, N., & Greenwald, I. (2004). The lateral signal for LIN-12/Notch in C. elegans vulval development comprises redundant secreted and transmembrane DSL proteins. Developmental Cell, 6(2), 183–192.

Chow, A., Brown, B. D., & Merad, M. (2011). Studying the mononuclear phagocyte system in the molecular age. Nature Reviews. Immunology, 11(11), 788–798.

Cirri, P., & Chiarugi, P. (2011). Cancer associated fibroblasts: the dark side of the coin. American Journal of Cancer Research, 1(4), 482–497.

Clark, S. G., Chisholm, A. D., & Horvitz, H. R. (1993). Control of cell fates in the central body region of C. elegans by the homeobox gene lin-39. Cell, 74(1), 43–55.

Clark, S. G., Stern, M. J., & Horvitz, H. R. (1992). C. elegans cell-signalling gene sem-5 encodes a protein with SH2 and SH3 domains. Nature, 356(6367), 340–344.

Cui, M., Chen, J., Myers, T. R., Hwang, B. J., Sternberg, P. W., Greenwald, I., & Han, M. (2006). SynMuv genes redundantly inhibit lin-3/EGF expression to prevent inappropriate vulval induction in C. elegans. Developmental Cell, 10(5), 667–672.

Cui, M., Kim, E. B., & Han, M. (2006). Diverse chromatin remodeling genes antagonize the Rb-involved SynMuv pathways in C. elegans. PLoS Genetics, 2(5), e74.

Cunha, G. R. (2008). Mesenchymal-epithelial interactions: past, present, and future. Differentiation; Research in Biological Diversity, 76(6), 578–586.

Daldrup-Link, H. E., Golovko, D., Ruffell, B., Denardo, D. G., Castaneda, R., Ansari, C., … Coussens, L. M. (2011). MRI of tumor-associated macrophages with clinically applicable iron oxide nanoparticles. Clinical Cancer Research: An Official Journal of the American Association for Cancer Research, 17(17), 5695–5704.

DeNardo, D. G., Andreu, P., & Coussens, L. M. (2010). Interactions between lymphocytes and myeloid cells regulate pro- versus anti-tumor immunity. Cancer Metastasis Reviews, 29(2), 309–316. http://doi.org/10.1007/s10555-010-9223-6

Downey, C. L., Simpkins, S. A., White, J., Holliday, D. L., Jones, J. L., Jordan, L. B., Speirs, V. (2014). The prognostic significance of tumour–stroma ratio in oestrogen receptor-positive breast cancer. British Journal of Cancer, 110(7), 1744–1747.

Dumenil, C., Vieira, T., Rouleau, E., Antoine, M., Duruisseaux, M., Poulot, V., Wislez, M. (2015). Is there a specific phenotype associated with the different subtypes of KRAS mutations in patients with advanced non-small-cell lung cancers? Lung Cancer (Amsterdam, Netherlands), 90(3), 561–567.

Egeblad, M., Ewald, A. J., Askautrud, H. A., Truitt, M. L., Welm, B. E., Bainbridge, E., … Werb, Z. (2008). Visualizing stromal cell dynamics in different tumor microenvironments by spinning disk confocal microscopy. Disease Models & Mechanisms, 1(2-3), 155–167; discussion 165.

Egeblad, M., Nakasone, E. S., & Werb, Z. (2010). Tumors as organs: complex tissues that interface with the entire organism. Developmental Cell, 18(6), 884–901.

Eisenmann, D. M., & Kim, S. K. (1994). Signal transduction and cell fate specification during Caenorhabditis elegans vulval development. Current Opinion in Genetics & Development, 4(4), 508–516.

Eisenmann, D. M., & Kim, S. K. (2000). Protruding vulva mutants identify novel loci and Wnt signaling factors that function during Caenorhabditis elegans vulva development. Genetics, 156(3), 1097–1116.

Eisenmann, D. M., Maloof, J. N., Simske, J. S., Kenyon, C., & Kim, S. K. (1998). The beta-catenin homolog BAR-1 and LET-60 Ras coordinately regulate the Hox gene lin-39 during Caenorhabditis elegans vulval development. Development (Cambridge, England), 125(18), 3667–3680.

El-Nikhely, N., Larzabal, L., Seeger, W., Calvo, A., & Savai, R. (2012). Tumor-stromal interactions in lung cancer: novel candidate targets for therapeutic intervention. Expert Opinion on Investigational Drugs, 21(8), 1107–1122.

Erez, N., Truitt, M., Olson, P., Arron, S. T., & Hanahan, D. (2010). Cancer-Associated Fibroblasts Are Activated in Incipient Neoplasia to Orchestrate Tumor-Promoting Inflammation in an NF-kappaB-Dependent Manner. Cancer Cell, 17(2), 135–147.

Erkan, M. (2013). Understanding the stroma of pancreatic cancer: co-evolution of the microenvironment with epithelial carcinogenesis. The Journal of Pathology, 231(1), 4–7.

Errico, A. (2015). Genetics: Stromal signatures drive the oncogenic phenotype of colorectal cancer. Nature Reviews Clinical Oncology, 12(5), 249–249.

Farmer, P., Bonnefoi, H., Anderle, P., Cameron, D., Wirapati, P., Becette, V., … Delorenzi, M. (2009). A stroma-related gene signature predicts resistance to neoadjuvant chemotherapy in breast cancer. Nature Medicine, 15(1), 68–74.

Fay, D. S., & Yochem, J. (2007). The SynMuv genes of Caenorhabditis elegans in vulval development and beyond. Developmental Biology, 306(1), 1–9.

Feitelson, M. A., Arzumanyan, A., Kulathinal, R. J., Blain, S. W., Holcombe, R. F., Mahajna, J., … Nowsheen, S. (2015). Sustained proliferation in cancer: Mechanisms and novel therapeutic targets. Seminars in Cancer Biology, 35, Supplement, S25–S54.

Ferguson, E. L., & Horvitz, H. R. (1985). Identification and characterization of 22 genes that affect the vulval cell lineages of the nematode Caenorhabditis elegans. Genetics, 110(1), 17–72.

Ferguson, E. L., & Horvitz, H. R. (1989). The multivulva phenotype of certain Caenorhabditis elegans mutants results from defects in two functionally redundant pathways. Genetics, 123(1), 109–121.

Ferguson, E. L., Sternberg, P. W., & Horvitz, H. R. (1987). A genetic pathway for the specification of the vulval cell lineages of Caenorhabditis elegans. Nature, 326(6110), 259–267.

Finak, G., Bertos, N., Pepin, F., Sadekova, S., Souleimanova, M., Zhao, H., … Park, M. (2008). Stromal gene expression predicts clinical outcome in breast cancer. Nature Medicine, 14(5), 518–527.

Franco, O. E., Shaw, A. K., Strand, D. W., & Hayward, S. W. (2010). Cancer associated fibroblasts in cancer pathogenesis. Seminars in Cell & Developmental Biology, 21(1), 33–39.

Fraser, A. G., Kamath, R. S., Zipperlen, P., Martinez-Campos, M., Sohrmann, M., & Ahringer, J. (2000). Functional genomic analysis of C. elegans chromosome I by systematic RNA interference. Nature, 408(6810), 325–330.

Fu, L., Zhang, C., Zhang, L.-Y., Dong, S.-S., Lu, L.-H., Chen, J., … Guan, X.-Y. (2011). Wnt2 secreted by tumour fibroblasts promotes tumour progression in oesophageal cancer by activation of the Wnt/β-catenin signalling pathway. Gut, 60(12), 1635– 1643.

Gabrilovich, D. I., & Nagaraj, S. (2009). Myeloid-derived suppressor cells as regulators of the immune system. Nature Reviews. Immunology, 9(3), 162–174.

Gache, C., Berthois, Y., Martin, P.-M., & Saez, S. (1998). Positive Regulation of Normal and Tumoral Mammary Epithelial Cell Proliferation by Fibroblasts in Coculture. In Vitro Cellular & Developmental Biology. Animal, 34(4), 347–351.

Gaggioli, C., Hooper, S., Hidalgo-Carcedo, C., Grosse, R., Marshall, J. F., Harrington, K., & Sahai, E. (2007). Fibroblast-led collective invasion of carcinoma cells with differing roles for RhoGTPases in leading and following cells. Nature Cell Biology, 9(12), 1392–1400.

Gleason, J. E., Korswagen, H. C., & Eisenmann, D. M. (2002). Activation of Wnt signaling bypasses the requirement for RTK/Ras signaling during C. elegans vulval induction. Genes & Development, 16(10), 1281–1290.

Gonzales, P. A., Pisitkun, T., Hoffert, J. D., Tchapyjnikov, D., Star, R. A., Kleta, R., … Knepper, M. A. (2009). Large-scale proteomics and phosphoproteomics of urinary exosomes. Journal of the American Society of Nephrology: JASN, 20(2), 363–379.

Green, J. L., Inoue, T., & Sternberg, P. W. (2007). The C. elegans ROR receptor tyrosine kinase, CAM-1, non-autonomously inhibits the Wnt pathway. Development (Cambridge, England), 134(22), 4053–4062.

Greenwald, I. (1997). Development of the Vulva. In D. L. Riddle, T. Blumenthal, B. J. Meyer, & J. R. Priess (Eds.), C. elegans II (2nd ed.). Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press.

Greenwald, I. S., Sternberg, P. W., & Horvitz, H. R. (1983). The lin-12 locus specifies cell fates in Caenorhabditis elegans. Cell, 34(2), 435–444.

Gu, T., Orita, S., & Han, M. (1998). Caenorhabditis elegans SUR-5, a Novel but Conserved Protein, Negatively Regulates LET-60 Ras Activity during Vulval Induction. Molecular and Cellular Biology, 18(8), 4556–4564.

Hanahan, D., & Coussens, L. M. (2012). Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell, 21(3), 309–322.

Hanahan, D., & Folkman, J. (1996). Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell, 86(3), 353–364.

Hanahan, D., & Weinberg, R. A. (2000). The hallmarks of cancer. Cell, 100(1), 57–70.

Hanahan, D., & Weinberg, R. A. (2011). Hallmarks of cancer: the next generation. Cell, 144(5), 646–674.

Han, M., Golden, A., Han, Y., & Sternberg, P. W. (1993). C. elegans lin-45 raf gene participates in let-60 ras-stimulated vulval differentiation. Nature, 363(6425), 133–140.

Han, M., & Sternberg, P. W. (1990). let-60, a gene that specifies cell fates during C. elegans vulval induction, encodes a ras protein. Cell, 63(5), 921–931.

Hill, R. J., & Sternberg, P. W. (1992). The gene lin-3 encodes an inductive signal for vulval development in C. elegans. Nature, 358(6386), 470–476.

Hill, R., Song, Y., Cardiff, R. D., & Van Dyke, T. (2005). Selective evolution of stromal mesenchyme with p53 loss in response to epithelial tumorigenesis. Cell, 123(6), 1001–1011.

Hobbs, G. A., Der, C. J., & Rossman, K. L. (2016). RAS isoforms and mutations in cancer at a glance. Journal of Cell Science.

Horvitz, H. R., & Sulston, J. E. (1980). Isolation and genetic characterization of cell- lineage mutants of the nematode Caenorhabditis elegans. Genetics, 96(2), 435– 454.

Huang, L. S., & Sternberg, P. W. (2006). Genetic dissection of developmental pathways. WormBook: The Online Review of C. Elegans Biology, 1–19.

Inoue, T., Sherwood, D. R., Aspöck, G., Butler, J. A., Gupta, B. P., Kirouac, M., … Sternberg, P. W. (2002). Gene expression markers for Caenorhabditis elegans vulval cells. Gene Expression Patterns: GEP, 2(3-4), 235–241.

Inoue, T., Wang, M., Ririe, T. O., Fernandes, J. S., & Sternberg, P. W. (2005). Transcriptional network underlying Caenorhabditis elegans vulval development. Proceedings of the National Academy of Sciences of the United States of America, 102(14), 4972–4977.

Isella, C., Terrasi, A., Bellomo, S. E., Petti, C., Galatola, G., Muratore, A., … Medico, E. (2015). Stromal contribution to the colorectal cancer transcriptome. Nature Genetics, 47(4), 312–319.

Josson, S., Matsuoka, Y., Chung, L. W. K., Zhau, H. E., & Wang, R. (2010). Tumor- stroma co-evolution in prostate cancer progression and metastasis. Seminars in Cell & Developmental Biology, 21(1), 26–32.

Kalluri, R., & Zeisberg, M. (2006). Fibroblasts in cancer. Nature Reviews. Cancer, 6(5), 392–401.

Kao, G., Tuck, S., Baillie, D., & Sundaram, M. V. (2004). C. elegans SUR-6/PR55 cooperates with LET-92/protein phosphatase 2A and promotes Raf activity independently of inhibitory Akt phosphorylation sites. Development (Cambridge, England), 131(4), 755–765.

Katajisto, P., Vaahtomeri, K., Ekman, N., Ventelä, E., Ristimäki, A., Bardeesy, N., … Mäkelä, T. P. (2008). LKB1 signaling in mesenchymal cells required for suppression of gastrointestinal polyposis. Nature Genetics, 40(4), 455–459.

Katz, W. S., Hill, R. J., Clandinin, T. R., & Sternberg, P. W. (1995). Different levels of the C. elegans growth factor LIN-3 promote distinct vulval precursor fates. Cell, 82(2), 297–307.

Katz, W. S., Lesa, G. M., Yannoukakos, D., Clandinin, T. R., Schlessinger, J., & Sternberg, P. W. (1996). A point mutation in the extracellular domain activates LET-23, the Caenorhabditis elegans epidermal growth factor receptor homolog. Molecular and Cellular Biology, 16(2), 529–537.

Kim, B.-G., Li, C., Qiao, W., Mamura, M., Kasprzak, B., Kasperczak, B., … Letterio, J. J. (2006). Smad4 signalling in T cells is required for suppression of gastrointestinal cancer. Nature, 441(7096), 1015–1019.

Kimble, J. (1981). Alterations in cell lineage following laser ablation of cells in the somatic gonad of Caenorhabditis elegans. Developmental Biology, 87(2), 286– 300.

Kornfeld, K., Guan, K. L., & Horvitz, H. R. (1995). The Caenorhabditis elegans gene mek-2 is required for vulval induction and encodes a protein similar to the protein kinase MEK. Genes & Development, 9(6), 756–768.

Kottakis, F., & Bardeesy, N. (2015). Gene signatures from pancreatic cancer tumor and stromal cells predict disease outcome. Nature Genetics, 47(10), 1102–1103.

9 5

Lackner, M. R., Kornfeld, K., Miller, L. M., Horvitz, H. R., & Kim, S. K. (1994). A MAP kinase homolog, mpk-1, is involved in ras-mediated induction of vulval cell fates in Caenorhabditis elegans. Genes & Development, 8(2), 160–173.

Lee, E. Y. H. P., & Muller, W. J. (2010). Oncogenes and Tumor Suppressor Genes. Cold Spring Harbor Perspectives in Biology, 2(10).

Li, W., Han, M., & Guan, K. L. (2000). The leucine-rich repeat protein SUR-8 enhances MAP kinase activation and forms a complex with Ras and Raf. Genes & Development, 14(8), 895–900.

Lowenstein, E. J., Daly, R. J., Batzer, A. G., Li, W., Margolis, B., Lammers, R., Schlessinger, J. (1992). The SH2 and SH3 domain-containing protein GRB2 links receptor tyrosine kinases to ras signaling. Cell, 70(3), 431–442.

Lühr, I., Friedl, A., Overath, T., Tholey, A., Kunze, T., Hilpert, F., … Bauer, M. (2012). Mammary fibroblasts regulate morphogenesis of normal and tumorigenic breast epithelial cells by mechanical and paracrine signals. Cancer Letters, 325(2), 175– 188.

Malumbres, M., & Barbacid, M. (2003). RAS oncogenes: the first 30 years. Nature Reviews Cancer, 3(6), 459–465. http://doi.org/10.1038/nrc1097

Mantovani, A., Cassatella, M. A., Costantini, C., & Jaillon, S. (2011). Neutrophils in the activation and regulation of innate and adaptive immunity. Nature Reviews. Immunology, 11(8), 519–531.

Martincorena, I., & Campbell, P. J. (2015). Somatic mutation in cancer and normal cells. Science (New York, N.Y.), 349(6255), 1483–1489.

Martinez-Outschoorn, U., Sotgia, F., & Lisanti, M. P. (2014). Tumor microenvironment and metabolic synergy in breast cancers: critical importance of mitochondrial fuels and function. Seminars in Oncology, 41(2), 195–216.

McKay, M. M., & Morrison, D. K. (2007). Integrating signals from RTKs to ERK/MAPK. Oncogene, 26(22), 3113–3121.

McMillin, D. W., Negri, J. M., & Mitsiades, C. S. (2013). The role of tumour–stromal interactions in modifying drug response: challenges and opportunities. Nature Reviews Drug Discovery, 12(3), 217–228.

Moghal, N., & Sternberg, P. W. (2003). The epidermal growth factor system in Caenorhabditis elegans. Experimental Cell Research, 284(1), 150–159.

Mohler, W. A., Shemer, G., del Campo, J. J., Valansi, C., Opoku-Serebuoh, E., Scranton, V., … Podbilewicz, B. (2002). The type I membrane protein EFF-1 is essential for developmental cell fusion. Developmental Cell, 2(3), 355–362.

Myers, T. R., & Greenwald, I. (2007). Wnt signal from multiple tissues and lin-3/EGF signal from the gonad maintain vulval precursor cell competence in Caenorhabditis elegans. Proceedings of the National Academy of Sciences of the United States of America, 104(51), 20368–20373.

Newman, A. P., & Sternberg, P. W. (1996). Coordinated morphogenesis of epithelia during development of the Caenorhabditis elegans uterine-vulval connection. Proceedings of the National Academy of Sciences of the United States of America, 93(18), 9329–9333.

Nieman, K. M., Kenny, H. A., Penicka, C. V., Ladanyi, A., Buell-Gutbrod, R., Zillhardt, M. R., … Lengyel, E. (2011). Adipocytes promote ovarian cancer metastasis and provide energy for rapid tumor growth. Nature Medicine, 17(11), 1498–1503.

Orimo, A., Gupta, P. B., Sgroi, D. C., Arenzana-Seisdedos, F., Delaunay, T., Naeem, R., … Weinberg, R. A. (2005). Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF- 1/CXCL12 secretion. Cell, 121(3), 335–348.

Payne, S. R., & Kemp, C. J. (2005). Tumor suppressor genetics. Carcinogenesis, 26(12), 2031–2045.

Peeters, M., Kafatos, G., Taylor, A., Gastanaga, V. M., Oliner, K. S., Hechmati, G., van Krieken, J. H. (2015). Prevalence of RAS mutations and individual variation patterns among patients with metastatic colorectal cancer: A pooled analysis of randomised controlled trials. European Journal of Cancer (Oxford, England: 1990), 51(13), 1704–1713.

Pietras, K., & Östman, A. (2010). Hallmarks of cancer: Interactions with the tumor stroma. Experimental Cell Research, 316(8), 1324–1331.

Polyak, K., Haviv, I., & Campbell, I. G. (2009). Co-evolution of tumor cells and their microenvironment. Trends in Genetics: TIG, 25(1), 30–38.

Pon, J. R., & Marra, M. A. (2015). Driver and passenger mutations in cancer. Annual Review of Pathology, 10, 25–50.

Porta, C., Riboldi, E., Totaro, M. G., Strauss, L., Sica, A., & Mantovani, A. (2011). Macrophages in cancer and infectious diseases: the “good” and the “bad.” Immunotherapy, 3(10), 1185–1202.

Provenzano, P. P., Eliceiri, K. W., Campbell, J. M., Inman, D. R., White, J. G., & Keely, P. J. (2006). Collagen reorganization at the tumor-stromal interface facilitates local invasion. BMC Medicine, 4(1), 38.

Puri, S., & Hebrok, M. (2010). Cellular plasticity within the pancreas--lessons learned from development. Developmental Cell, 18(3), 342–356.

Qadota, H., Inoue, M., Hikita, T., Köppen, M., Hardin, J. D., Amano, M., Kaibuchi, K. (2007). Establishment of a tissue-specific RNAi system in C. elegans. Gene, 400(1-2), 166–173.

Qian, B.-Z., & Pollard, J. W. (2010). Macrophage diversity enhances tumor progression and metastasis. Cell, 141(1), 39–51.

Rattigan, Y. I., Patel, B. B., Ackerstaff, E., Sukenick, G., Koutcher, J. A., Glod, J. W., & Banerjee, D. (2012). Lactate is a mediator of metabolic cooperation between stromal carcinoma associated fibroblasts and glycolytic tumor cells in the tumor microenvironment. Experimental Cell Research, 318(4), 326–335.

Salser, S. J., Loer, C. M., & Kenyon, C. (1993). Multiple HOM-C gene interactions specify cell fates in the nematode central nervous system. Genes & Development, 7(9), 1714–1724.

Seydoux, G., Savage, C., & Greenwald, I. (1993). Isolation and characterization of mutations causing abnormal eversion of the vulva in Caenorhabditis elegans. Developmental Biology, 157(2), 423–436.

Shaye, D. D., & Greenwald, I. (2002). Endocytosis-mediated downregulation of LIN- 12/Notch upon Ras activation in Caenorhabditis elegans. Nature, 420(6916), 686– 690.

Shemer, G., & Podbilewicz, B. (2002). LIN-39/Hox triggers cell division and represses EFF-1/fusogen-dependent vulval cell fusion. Genes & Development, 16(24), 3136–3141.

Sherman, M. H., Yu, R. T., Engle, D. D., Ding, N., Atkins, A. R., Tiriac, H., Evans, R. M. (2014). Vitamin D receptor-mediated stromal reprogramming suppresses pancreatitis and enhances pancreatic cancer therapy. Cell, 159(1), 80–93.

Sherwood, D. R., & Sternberg, P. W. (2003). Anchor cell invasion into the vulval epithelium in C. elegans. Developmental Cell, 5(1), 21–31.

Simmer, F., Tijsterman, M., Parrish, S., Koushika, S. P., Nonet, M. L., Fire, A., Plasterk, R. H. A. (2002). Loss of the putative RNA-directed RNA polymerase RRF-3 makes C. elegans hypersensitive to RNAi. Current Biology: CB, 12(15), 1317– 1319.

Simon, M. A. (1994). Signal transduction during the development of the Drosophila R7 photoreceptor. Developmental Biology, 166(2), 431–442.

Singh, N., & Han, M. (1995). sur-2, a novel gene, functions late in the let-60 ras- mediated signaling pathway during Caenorhabditis elegans vulval induction. Genes & Development, 9(18), 2251–2265.

Sternberg, P. W. (1988). Lateral inhibition during vulval induction in Caenorhabditis elegans. Nature, 335(6190), 551–554.

Sternberg, P. W., & Horvitz, H. R. (1986). Pattern formation during vulval development in C. elegans. Cell, 44(5), 761–772.

Sulston, J. E., & Horvitz, H. R. (1977). Post-embryonic cell lineages of the nematode, Caenorhabditis elegans. Developmental Biology, 56(1), 110–156.

Sulston, J. E., & White, J. G. (1980). Regulation and cell autonomy during postembryonic development of Caenorhabditis elegans. Developmental Biology, 78(2), 577–597.

Sundaram, M., & Han, M. (1995). The C. elegans ksr-1 gene encodes a novel Raf-related kinase involved in Ras-mediated signal transduction. Cell, 83(6), 889–901.

Sundaram, M., & Han, M. (1996). Control and integration of cell signaling pathways during C. elegans vulval development. BioEssays: News and Reviews in Molecular, Cellular and Developmental Biology, 18(6), 473–480.

Sundaram, M. V. (2004). Vulval development: the battle between Ras and Notch. Current Biology: CB, 14(8), R311–313.

Tabara, H., Sarkissian, M., Kelly, W. G., Fleenor, J., Grishok, A., Timmons, L., Mello, C. C. (1999). The rde-1 gene, RNA interference, and transposon silencing in C. elegans. Cell, 99(2), 123–132.

Takahashi, Y., Ishii, G., Taira, T., Fujii, S., Yanagi, S., Hishida, T., Ochiai, A. (2011). Fibrous stroma is associated with poorer prognosis in lung squamous cell

100

carcinoma patients. Journal of Thoracic Oncology: Official Publication of the International Association for the Study of Lung Cancer, 6(9), 1460–1467.

Tan, J., Buache, E., Chenard, M.-P., Dali-Youcef, N., & Rio, M.-C. (2011). Adipocyte is a non-trivial, dynamic partner of breast cancer cells. The International Journal of Developmental Biology, 55(7-9), 851–859.

Tan, P. B., & Kim, S. K. (1999). Signaling specificity: the RTK/RAS/MAP kinase pathway in metazoans. Trends in Genetics: TIG, 15(4), 145–149.

Thomas, J. H., Ceol, C. J., Schwartz, H. T., & Horvitz, H. R. (2003). New genes that interact with lin-35 Rb to negatively regulate the let-60 ras pathway in Caenorhabditis elegans. Genetics, 164(1), 135–151.

Timmons, L., Court, D. L., & Fire, A. (2001). Ingestion of bacterially expressed dsRNAs can produce specific and potent genetic interference in Caenorhabditis elegans. Gene, 263(1-2), 103–112.

Timmons, L., & Fire, A. (1998). Specific interference by ingested dsRNA. Nature, 395(6705), 854.

Trachootham, D., Chen, G., Zhang, W., Lu, W., Zhang, H., Liu, J., & Huang, P. (2013). Loss of p53 in stromal fibroblasts promotes epithelial cell invasion through redox- mediated ICAM1 signal. Free Radical Biology & Medicine, 58, 1–13.

Trimboli, A. J., Cantemir-Stone, C. Z., Li, F., Wallace, J. A., Merchant, A., Creasap, N., Leone, G. (2009). Pten in stromal fibroblasts suppresses mammary epithelial tumours. Nature, 461(7267), 1084–1091.

Troppmair, J., Bruder, J. T., Munoz, H., Lloyd, P. A., Kyriakis, J., Banerjee, P., Rapp, U. R. (1994). Mitogen-activated protein kinase/extracellular signal-regulated protein kinase activation by oncogenes, serum, and 12-O-tetradecanoylphorbol-13-acetate requires Raf and is necessary for transformation. The Journal of Biological Chemistry, 269(9), 7030–7035.

101

Ueno, H., Jones, A. M., Wilkinson, K. H., Jass, J. R., & Talbot, I. C. (2004). Histological categorisation of fibrotic cancer stroma in advanced rectal cancer. Gut, 53(4), 581–586.

Vogelstein, B., Papadopoulos, N., Velculescu, V. E., Zhou, S., Diaz, L. A., & Kinzler, K. W. (2013). Cancer genome landscapes. Science (New York, N.Y.), 339(6127), 1546–1558.

Wang, M., & Sternberg, P. W. (2000). Patterning of the C. elegans 1 degrees vulval lineage by RAS and Wnt pathways. Development (Cambridge, England), 127(23), 5047–5058.

Watson, I. R., Takahashi, K., Futreal, P. A., & Chin, L. (2013). Emerging patterns of somatic mutations in cancer. Nature Reviews. Genetics, 14(10), 703–718.

Wilfinger, A., Arkhipova, V., & Meyer, D. (2013). Cell type and tissue specific function of islet genes in zebrafish pancreas development. Developmental Biology, 378(1), 25–37.

Williams, N. G., & Roberts, T. M. (1994). Signal transduction pathways involving the Raf proto-oncogene. Cancer Metastasis Reviews, 13(1), 105–116.

Wiseman, B. S., & Werb, Z. (2002). Stromal effects on mammary gland development and breast cancer. Science (New York, N.Y.), 296(5570), 1046–1049.

Xu, Z., Pothula, S. P., Wilson, J. S., & Apte, M. V. (2014). Pancreatic cancer and its stroma: A conspiracy theory. World Journal of Gastroenterology : WJG, 20(32), 11216–11229.

Yang, F.-C., Ingram, D. A., Chen, S., Zhu, Y., Yuan, J., Li, X., Clapp, D. W. (2008). Nf1-dependent tumors require a microenvironment containing Nf1+/-- and c-kit- dependent bone marrow. Cell, 135(3), 437–448.

Yauch, R. L., Gould, S. E., Scales, S. J., Tang, T., Tian, H., Ahn, C. P., de Sauvage, F. J. (2008). A paracrine requirement for hedgehog signalling in cancer. Nature, 455(7211), 406–410. 102

Yoo, A. S., Bais, C., & Greenwald, I. (2004). Crosstalk between the EGFR and LIN- 12/Notch pathways in C. elegans vulval development. Science (New York, N.Y.), 303(5658), 663–666.

103