Cell Proliferation Control: from Intrinsic Transcriptional Programs to Extrinsic Stromal Networks

DISSERTATION

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

Huayang Liu

Graduate Program in Molecular Genetics

The Ohio State University

2015

Dissertation Committee:

Professor Gustavo Leone, Advisor

Professor Helen Chamberlin

Professor Harold Fisk

Professor Tsonwin Hai

Huayang Liu

2015

Abstract

Robust mechanisms to control cell proliferation have evolved to maintain the integrity of organ architecture and tissue homeostasis. Disruption of these mechanisms, either intracellular circuits or extracellular inputs for cell proliferation control, inevitably occur under oncogenic conditions, rendering cells with unlimited proliferative capacity and subsequent malignant transformation. Here, we investigated how two critical intracellular proliferative pathways, Myc and E2f, are integrated to control cell cycle transitions in normal and Rb deficient cells using a murine intestinal model. We show that in contrast to the current paradigm, Myc and E2f1-3 have little impact on normal G1-

S transitions. Instead Myc and E2f1-3 synergistically control an S-G2 transcriptional program in intestinal crypts required for completing normal cell divisions and maintaining crypt-villi integrity. Surprisingly, Rb deficiency results in the Myc- dependent accumulation of E2f3a protein and the genome-wide repositioning of Myc and

E2f3 on chromatin associated with ‘super activation’ of a G1-S transcriptional program, ectopic S phase entry and rampant cell proliferation. These findings reveal that Rb deficient cells hijack and redeploy Myc and E2f3 from an S-G2 program essential for normal cell cycles to a G1-S program that re-engages ectopic cell cycles, exposing an unanticipated addiction of Rb-null ‘cancer-like cells’ on Myc. ii

We have also studied how cell non-autonomous inputs derived from stromal compartment in the tumor microenvironment impacts the proliferation of tumor compartment. We performed a genome-wide stroma-specific RNAi screen in ‘cancer sensitized’ C. elegans and discovered 39 stromal factors that suppress the proliferation of adjacent ‘Ras pathway-sensitized’ epithelial cells, with minimal impact on the proliferation of normal or ‘Wnt pathway-sensitized’ epithelial cells. These candidate genes encode histone variants and components of protein complexes known to converge on the control of chromatin dynamics, cytoplasmic polyadenylation and translation. An expression signature of the corresponding 33 human orthologs was selectively enriched in laser capture microdissected tumor stroma of breast cancer patients. Furthermore, depletion of stromal factors from human breast fibroblasts increased the proliferation of co-cultured breast cancer cells, but had no effect on normal mammary epithelial cells.

This cross-species approach identified unanticipated stromal regulatory hubs with cell non-autonomous tumor suppressive function, exposing the conserved and selective nature of stroma-tumor communication.

iii

Dedication

This document is dedicated to my family.

Acknowledgments

The time in the graduate school has been a memorable experience - it all seems like yesterday. I want to thank everyone who has been there to help me reach this point - thank you!

I would like to thank my advisor, Dr. Gustavo Leone, for his passionate and unconditioned support during the study. Rather than giving me a fish for the day, he teaches me how to fish that I can feed myself on the scientific journey for a lifetime. I would also like to thank my co-advisor, Dr. Helen Chamberlin, who has always been there to offer help on the projects and more generally, being an supportive mentor all along.

I would also like to thank my advisor committee, Dr. Harold Fisk and Dr.

Tsonwin Hai, who have been critical and suggestive on my endeavors. A few hours on the committee meeting have always driven me to push the science to the next level.

I am no doubt in debt to lab members, past or present, directly involved in the projects or just chatting over daily life. I am grateful when I reach out, you guys are always there to lend me a hand.

I am grateful to work with the undergraduate students, Amneet Bajwa, Karl Eiring,

Nicholas Fackler, Benjamin Hemmelgarn, Michael Lause, Stephan Reyes and Kyle Toth.

It has been a joyful experience, working with them as a mentor as well as a friend.

I would always treasure the support from my family and my wife, Yian Ruan. It's their love that has been accompanying me all along this journey that is full of hope and surprises. It's their love that enables me for all the achievements. It's their love that will lead me to the next station for the exploration in the future.

Thanks to everyone and best wishes!

Vita

2004 to 2008 ...... B.S. Chemistry and Biology, Tsinghua

University

2008 to 2009 ...... Graduate Teaching Associate, Department

of Molecular Genetics, The Ohio State

University

2009 to present ...... Graduate Research Associate, Department

of Molecular Genetics, The Ohio State

University

Publications

1. Liu, H.†, Sullivan, N.J.†, Bertos, N., Dowdle, J.A., Rambani, K., Mair, M., Daniel, P.,

Toth, K., Lause, M., Harrigan, M.E., Eiring, K., Sullivan, C., Chang, S.W., Kladney,

R., Tang, X., McElroy, J., Lu, Y., Tofigh, A., Fernandez, S.A., Parvin, J.D., Macrae,

E., Majumder, S., Shapiro, C.L., Yee, L.D., Hallett, M., Ostrowski, M.C., Park, M.,

Chamberlin, H.M.*, Leone, G*. Discovery of stromal regulatory networks that

suppress cancer cell proliferation. Manuscript under review. (†: co-first authors; *: co-

correspondence authors)

vii

2. Liu, H., Tang, X., Srivastava, A., Pécot, T., Daniel, P., Hemmelgarn, B., Reyes,

S., Fackler, N., Bajwa, A., Kladney, R., Koivisto, C., Chen, Z., Wang, Q., Huang,

K., Machiraju, R., Sáenz-Robles, M.T., Cantalupo, P., Pipas, J.M., Leone, G.

Redeployment of Myc and E2f1-3 drives cancer-like cell cycles. Manuscript in

revision.

3. Liu, H., Rakijas J., de Bruin, A., Leone, G. Emerging roles of E2Fs in cancer.

Commissioned review in preparation.

4. Sáenz-Robles, M.T., Chong, J.L., Koivisto, C., Trimboli, A., Liu, H., Leone, G.,

Pipas, J.M. Viral oncogene expression in the stem/progenitor cell compartment of

the mouse intestine induces adenomatous polyps. Mol Cancer Res. (2014);

12:1355-1364.

5. Coffman, V.C., Nile, A.H., Lee, I.J., Liu, H., Wu, J.Q. Roles of formin nodes and

myosin motor activity in Mid1p-dependent contractile ring assembly during

fission yeast cytokinesis. Mol. Biol. Cell (2009); 20: 5195-5210.

Fields of Study

Major Field: Molecular Genetics

viii

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgments...... v

Vita ...... vii

List of Figures ...... xiii

List of Tables ...... xv

Chapter 1 Cell Proliferation Control ...... 1

1.1 Cell proliferation, cell cycle control and cyclin-dependent kinases...... 2

1.2 Retinoblastoma (Rb) family proteins ...... 3

1.3 E2f family transcription factors...... 3

1.4 Cdk-Rb-E2f axis in cancer ...... 5

1.5 Myc transcription factor ...... 6

1.6 Intracellular circuits and extracellular inputs for proliferation control ...... 8

1.7 Proliferation signaling autonomy and inputs from tumor microenvironment in

cancer ...... 9

Chapter 2 Redeployment of Myc and E2f1-3 drives cancer-like cell cycle ...... 12

2.1 Introduction ...... 12

2.2 Results ...... 15

2.2.1 Combined ablation of Myc and E2f1-3 results in disruption of crypt-villus

integrity ...... 15

2.2.2 Combined Myc and E2f1-3 deficiency leads to S-G2 cell cycle arrest ...... 16

2.2.3 Synergistic regulation of an S-G2 transcriptional program by Myc and E2f1-3

...... 18

2.2.4 Myc and E2f3 bind to both G1-S and S-G2 target genes in crypts ...... 19

2.2.5 Myc drives ectopic proliferation of Rb deficient intestinal cells ...... 21

2.2.6 Redeployment of Myc and E2f3 in Rb deficient cells...... 22

2.2.7 Myc is required for the accumulation of E2f3 in Rb deficient villi ...... 25

2.3 Discussion ...... 27

2.4 Methods ...... 32

2.4.1 Mice ...... 32

2.4.2 β-naphthoflavone, BrdU and EdU injection ...... 32

2.4.3 Tissue preparation and histology ...... 32

2.4.4 Isolation of villi and crypts ...... 33

2.4.5 RNA and quantitative PCR ...... 34

2.4.6 Affymetrix microarray analysis ...... 34

2.4.7 Immuno-staining ...... 35

2.4.8 ChIP-PCR ...... 36

2.4.9 ChIP-exo-seq library construction and Illumina sequencing ...... 37

2.4.10 ChIP-exo-seq data analysis ...... 38

2.4.11 Flow cytometry ...... 39

2.4.12 Data deposition ...... 39

Chapter 3 Discovery of stromal regulatory networks that suppress cancer cell proliferation 78

3.1 Introduction ...... 78

3.2 Results ...... 80

3.2.1 Generation of stroma-specific RNAi and cancer-sensitized C. elegans ...... 80

3.2.2 Genome-wide screen in C. elegans identifies stromal factors that suppress

epithelial cell proliferation ...... 82

3.2.3 Spatial and signaling specificity of stromal gene functions ...... 84

3.2.4 Stromal factors suppress ectopic Ras signaling in sensitized VPCs ...... 86

3.2.5 The C. elegans-derived expression signature is associated with breast cancer

stroma ...... 86

3.2.6 Depletion of human orthologs from stromal fibroblasts enhances tumor cell

proliferation ...... 88

3.3 Discussion ...... 89 xi

3.4 Methods ...... 91

3.4.1 C. elegans strains ...... 91

3.4.2 C. elegans tissue-specific RNAi constructs ...... 92

3.4.3 C. elegans RNAi screen and microscopy ...... 93

3.4.4 Vulval Ras/MAPK signaling assay ...... 94

3.4.5 Breast tumor stroma total RNA expression profiling ...... 95

3.4.6 Human cell culture ...... 96

3.4.7 Fibroblast mRNA quantification ...... 98

3.4.8 Human lentivirus production and shRNA transduction ...... 98

3.4.9 3-dimensional co-culture proliferation assay ...... 99

3.4.10 Statistical analyses ...... 100

Chapter 4 Discussion ...... 124

4.1 E2f and Myc control normal cell proliferation ...... 124

4.2 E2f and Myc control cancer-like cell proliferation ...... 127

4.3 Conserved cell non-autonomous networks for cell proliferation control ...... 128

4.4 Genetic selectivity for cell proliferation control by stromal networks ...... 130

References ...... 132

xii

List of Figures

Figure 2.1 Disruption of the small intestine by combined loss of Myc and E2f1-3 ...... 40

Figure 2.2 S-G2 cell cycle arrest in E2f/Myc QKO progenitor cells ...... 42

Figure 2.3 Loss of E2f1-3 and/or Myc in crypts ...... 44

Figure 2.4 Synergistic regulation of an S-G2 transcriptional program by Myc and E2f1-3

...... 46

Figure 2.5 Chromatin binding of E2f3 and Myc in wild type tissues ...... 48

Figure 2.6 DNA binding of E2f3 and Myc in wild type crypts ...... 51

Figure 2.7 Rb deficient cells require Myc to drive ectopic cell cycles ...... 53

Figure 2.8 Myc and E2f1-3 regulate an overlapping G1-S transcriptional program in Rb- null cells ...... 55

Figure 2.9 Loss of either E2f1-3 or Myc corrects aberrant transcription in Rb KO villi .. 57

Figure 2.10 Rb loss redefines the chromatin binding landscape of E2f3 and Myc ...... 59

Figure 2.11 Peak summit-distance plots and tag intensity heatmaps across tissue compartments and genetic groups ...... 61

Figure 2.12 E2f3 DNA binding in wild type and Rb KO villi ...... 63

Figure 2.13 DNA binding strength and DNA binding motifs of all Myc and E2f3 peaks 65

xiii

Figure 2.14 DNA binding strength and DNA binding motifs of Myc and E2f3 peaks associated with dysregulated genes in Rb KO villi ...... 67

Figure 2.15 Myc regulates E2f3a expression in Rb deficient villi ...... 69

Figure 2.16 Expression of Wnt/β-catenin targets in Rb KO villi and auto-regulation of

E2fs ...... 71

Figure 3.1 Stroma-specific RNAi in C. elegans containing a cancer-relevant mutation 102

Figure 3.2 Stromal RNAi in C. elegans identifies conserved factors ...... 104

Figure 3.3 Genome-wide RNAi screen identifies conserved stromal factors that suppress ectopic epithelial cell division ...... 106

Figure 3.4 Stromal factors have spatial- and signaling-specific functions ...... 108

Figure 3.5 Depletion of stromal factors induces ectopic Ras signaling in VPCs ...... 110

Figure 3.6 The human ortholog signature distinguishes breast tumor stroma from adjacent normal stroma ...... 112

Figure 3.7 The 33-ortholog signature is represented in human breast cancer ...... 114

Figure 3.8 The human ortholog signature fails to distinguish breast tumor epithelium from adjacent normal epithelium ...... 116

Figure 3.9 Depletion of human orthologs in breast fibroblasts for 3D co-culture with breast cancer or normal epithelial cells ...... 118

Figure 3.10 Depletion of human orthologs in normal breast fibroblasts enhances breast cancer cell proliferation ...... 120

xiv

List of Tables

Table 2.1 Primers for mouse genotyping, RT-qPCR and ChIP-PCR ...... 73

Table 3.1 Primers for C. elegans strain construction and human gene RT-qPCR ...... 122

Chapter 1 Cell Proliferation Control

Precise control of cell proliferation is essential for organogenesis and maintenance of normal tissue architecture throughout the life time of the organism. Cell proliferation is coupled with cell differentiation and cell death (including apoptosis, autophagy and various forms of non-apoptotic programmed cell-death) to ensure a homeostasis of cell number and thus preservation of tissue functions. Under normal physiological conditions, multiple proliferation-stimulating and proliferation-inhibitory signals operate to control cellular quiescence and tissue homeostasis. To this end, normal tissues precisely control the generation and release of promoting and suppressive signals to modulate entry into and progression through the cell growth-and-division phase. These proliferation signals are received by transmembrane cell surface receptors that are capable of binding a variety of signaling molecules such as diffusible growth factors, extracellular matrix components and cell-to-cell adhesion/interaction molecules. Upon the binding to the signaling molecules, the receptors may elicit the downstream signal transduction through branched intracellular signaling circuits that ultimately converge on the control of cell cycle. Under pathological conditions such as cancer, cells evade the critical homeostatic mechanism that normally functions to guarantee an appropriate behavior of the various cell types within a tissue, establishing sustained and uncontrolled cell proliferation as the fundamental hallmark of cancer (Hanahan and Weinberg, 2000, 2011).

1.1 Cell proliferation, cell cycle control and cyclin-dependent kinases

Much of the intracellular networks employed by normal cells to respond to proliferation signals funnel through cyclin-dependent kinases (Cdks), the components governing the transit of cells through various phases of the cell cycle clock. In response to developmental and environmental mitogenic cues, nuclear levels of G1-cyclin (cyclin

D) increases, which allows the consequent activation of cyclin D-associated Cdk4/6 and phosphorylation of corresponding substrates to drive the cells out of quiescent status for the G1/S transition and DNA replication (Sherr, 1996). Subsequently, the increase of other cyclin levels (cyclin E and cyclin A) results in the formation of active cyclin E-

Cdk2 and cyclin A-Cdk2 complexes that further reinforce the cell cycle progression.

Among the key targets of cyclinA-Cdk2 complex is Cdc20-homologure 1 (Cdh1), a regulatory component of anaphase-promoting complex/cyclosome (APC/C) (Hsu et al.,

2002; Lukas et al., 1999; Visintin et al., 1997). Cdk-mediated phosphorylation of Cdh1 suppresses Cdh1's activity, which prevents APCCdh1-mediated proteolysis of S/M phase cyclins and ensures the completion of S phase and entry into mitosis. Upon entering mitosis, cyclin B translocates into the nucleus and activates Cdk1 that licenses the execution of mitotic programs. These phosphorylation/dephosphorylation events mediated by Cdks coordinate with ubiquitination and protein degradation processes mediated by APC/C and SCF complex (Skp, Cullin, F-box containing complex) for faithful DNA replication and cell division (Nakayama and Nakayama, 2006).

1.2 Retinoblastoma (Rb) family proteins

Among the critical substrates of the Cdks are the retinoblastoma (Rb) family of pocket proteins that governs the transit of the cell through quiescent G0/G1 phase into S phase. At the molecular level, phosphorylation status of the retinoblastoma protein (Rb) and its two family members, p107 and p130, integrate most and perhaps all cell cycle regulatory signals. Under quiescent status, pocket proteins are typically hypophosphorylated that enable them to physically bind to E2f transcription factors and recruit chromatin remodeling components, such as histone deacetylases (HDACs), to establish a generally repressive chromatin environment (Blais and Dynlacht, 2007).

Hyperphosphorylation of pocket proteins by Cdks liberates E2f from the complex, resulting in a switch of suppressive chromatin status into active chromatin configuration and consequent transcription of target genes necessary for the G1/S transition.

1.3 E2f family transcription factors

E2f was originally discovered as a host factor that is associated with E1A adenoviral promoters for transactivation of viral gene expression (named after E2- assoicated factors or E2f) (Kovesdi et al., 1986). E2f was subsequently shown to be a cellular target of Rb (Chellappan et al., 1991). Since discovery of the founding E2F family member, E2f1, seven additional mammalian genes have been identified to encode

E2f family members with similar DNA binding domain structure (Cartwright et al., 1998;

Christensen et al., 2005; de Bruin et al., 2003; Di Stefano et al., 2003; Ginsberg et al.,

1994; Helin et al., 1992; Itoh et al., 1995; Kaelin et al., 1992; Logan et al., 2005; Maiti et

3 al., 2005; Sardet et al., 1995; Trimarchi et al., 1998). The initial observation that overexpression of E2f1 could induce quiescent cells to enter S phase established the connection between E2f and the control of G1/S transition during cell cycle and spurred the identification of critical E2f target genes implicated for DNA replication (Johnson et al., 1993). It has been appreciated now that E2f family member may regulate the expression of a large number of genes associated with DNA repair, mitosis, apoptosis and differentiation beyond the ones involved in DNA replication (Dimova and Dyson, 2005).

One prominent feature of mammalian E2f transcription factors is the complex redundancy and antagonism within multiple members. E2fs have been traditionally classified as activators and repressors with the notions that the compensation within each subfamily provides the robustness for the whole transcription program and the opposing roles of different subfamily ensures the appropriate outputs (Trimarchi and Lees, 2002).

The transcription activator subclass includes E2f1, E2f2 and E2f3. The combined ablation of these three family members in mouse embryo fibroblasts (MEFs) results in a profound inhibition of E2f target gene expression and the failure of cells to enter the S phase (Wu et al., 2001). The transcription repressor subfamily includes E2f4-8. Ablation of E2f4 and E2f5 in MEFs results in a failure of these cells to repress E2f target genes critical for G1/S transition and to respond to proliferation inhibitory signals (Gaubatz et al., 2000). Whereas E2f4 and E2f5 recruit transcriptional machinery and chromatin remodelers through interaction with Rb pocket proteins, E2f6, E2f7 and E2f8 lack the pocket protein-interaction domain and appear to repress transcription independent of Rb family proteins.

Upon DNA binding of E2fs, the transcriptional outputs of their target genes are refined by the acquisition of activating or repressive chromatin modifications (Blais and

Dynlacht, 2007). These crucial modifications include DNA methylations, histone modifications such as methylation and acetylation, histone variant deposition, nucleosome remodeling and chromatin compaction. Under quiescent conditions, repressor E2fs, particularly E2f4, recruit HDACs, SWI/SNF components such as

BRM/BRG-1 and histone methyl transferases such as SUV39H1 to elicit inhibitory chromatin modifications (Blais and Dynlacht, 2007). Upon mitogenic stimulation, the activator E2f1-3 associated complex replaces the repressor complex and recruits distinct chromatin modifiers including histone acetyl transferases (HATs) such as p300/CBP,

P/CAF, and Tip60, which promotes histone acetylation and consequent E2f target gene expression (Blais and Dynlacht, 2007; Frolov and Dyson, 2004). The sequential waves of

E2f-dependent transcriptional activation and repression ensures the appropriate expression of crucial target genes necessary for cells to march through all four phases of the cell cycle: G1, S, G2 and M.

1.4 Cdk-Rb-E2f axis in cancer

The Cdk-Rb-E2f signaling axis, as governed by many upstream factors, can be disrupted in a variety of ways in different types of human tumors. Some tumors lose the balanced supply of proliferative and antiproliferative signals, while others lose the cell proliferation control through dysregulation of receptors or possess mutant, dysfunctional receptors that alter the signaling inputs funneling through the Cdk-Rb-E2f axis. Prevalent

5 dysregulation may also occur at the regulation of Cdks. For example, the loci encoding

Cdk inhibitors may suffer genetic or epigenetic alterations (Malumbres and Barbacid,

2009). Alternatively, the Cdks may become unresponsive to the inhibitory actions of Cdk inhibitors due to mutations within Cdks that disrupt their interactions with corresponding inhibitors (Zuo et al., 1996). Furthermore, gain of cyclin D function may be observed in certain tumors (Jiang et al., 1992). As a result, the resulting cyclin-Cdk complexes are unleashed to inactivate Rb by hyperphosphorylation. Rb may be inactivated with mutations within the gene or, alternatively, be functionally disabled as in some virus- induced cancers through sequestration by viral oncoproteins, such as the E7 oncoprotein of human papillomavirus (Dyson et al., 1989; Sherr and McCormick, 2002). E2f, the end component of this pathway, may undergo genomic amplification and/or overexpression and thus may also be tightly linked to drive the expression disorder of target genes and uncontrolled cell proliferation in certain tumors (Chen et al., 2009b).

1.5 Myc transcription factor

Cell proliferation may depend on more than a single pathway to avoid the potential deleterious consequences in the presence of severe disruption to the components within that pathway. One strategy employed by normal cells to increase the robustness of cell proliferation control directly involves the Myc oncogene, which is also adopted by cancer cells to evade the control mechanism for tissue homeostasis.

Myc was originally discovered as a homolog to the oncogene (v-myc) of avian myelocytomatosis virus (Vennstrom et al., 1982). Myc protein belongs to Myc family of

6 transcription factors, which also includes N-Myc and L-Myc genes. This family of transcription factors contains a basic helix-loop-helix/leucine zipper (bHLH/LZ) domain.

Whereas the leucine zipper domain mediates the physical association of Myc with it binding partner Max, the bHLH domain allows Myc to bind to DNA primarily with classical consensus E-box sequence (CACGTG) (Grandori et al., 2000). Upon the binding to DNA, Myc may recruit chromatin remodelers (such as HATs) to affect the transcription of critical target genes involved in cell cycle control (such as upregulation of cyclins and Cdks) (Bouchard et al., 1999; Hermeking et al., 2000).

Normal tissues may restrain cell proliferation by forcing cells to enter irreversibly into terminally differentiated states, whereas cancer cells may use various strategies including the engagement of Myc to avoid the establishment of this post-mitotic status.

The proliferation-stimulating action of Myc in association with Max may be superseded by the complex of Max with Mxd transcription factors, in which the Mxd-Max complex elicit differentiation actions (Foley and Eisenman, 1999). The presence of excessive amount of Myc, as observed in many tumors, disrupts the balance to favor Myc-Max complex over Mxd-Max complex with impaired differentiation and unrestrained proliferation. During intestinal carcinogenesis, dysregulation of the Adenomatous polyposis coli/β-catenin pathway dramatically increases Myc levels to block the appropriate differentiation of enterocytes; instead, these cells maintain the characteristics of progenitor cells with the capability of undergoing neoplasmic proliferation (Sansom et al., 2007).

Aberrant activation of Myc is frequently observed in cancers. The dysregulation of Myc may occur directly at the Myc locus through several mechanisms. Amplification events may occur at Myc locus and thus the cancer cells may harbor multiple copies of

Myc (Meyer and Penn, 2008). Myc may also be affected due to chromosomal translocation and insertional mutagenesis as observed in Burkitt lymphoma and leukemia induced by the avian myelocytomatosis retrovirus, respectively (Dalla-Favera et al., 1982;

Hayward et al., 1981; Taub et al., 1982). Mutations to the coding sequence of Myc are also found, although only in a narrow spectrum of tumors (Bhatia et al., 1993). Recent research has also revealed that Myc may also be dysregulated with one of several mechanisms that target it expression and activity. For example, accumulation of Myc may occur in colorectal cancer due to loss of Adenomatous polyposis coli, the negative regulator of canonical Wnt signaling, and/or activation of β-catenin that stimulate high level transcription of Myc (Sansom et al., 2007). Similar scenario due to abnormal Notch signaling may also ectopically activate Myc transcription (Palomero et al., 2006; Sharma et al., 2006; Weng et al., 2006). Myc may also be stabilized at the mRNA level as well as protein level, where SCFFbw7 ubiquitin ligase responsible for Myc destruction is inactivated in certain cancers (Schiavi et al., 1992; Welcker and Clurman, 2008). Overall, loss of appropriate control for Myc exists in a significant portion of all malignancies.

1.6 Intracellular circuits and extracellular inputs for proliferation control

Normal cells absolutely require stimulatory signals to exit from a quiescent state into an active proliferative state. While past research has shed light on understanding the

8 intracellullar networks that transduce these signals, it is surprising how rudimentary our understanding still is in terms of (i) exact identities of the proliferative signals operating within normal tissues; (ii) sources of such signals; (iii) mechanisms that control the synthesis and release of these signals. The challenge partly results from the fact that the proliferation of normal cells are mainly dictated by the paracrine signals from their neighboring cells or endocrine signals. The production and transduction of these signals are highly regulated in a spatio-temporal manner, which is technically difficult to be experimentally monitored. Moreover, the growth factors may be sequestered, released and activated based on the complex enzymatic activities within the extracellular matrix

(Rifkin et al., 1999), which adds another layer of complexity for dissecting the contribution from the tissue microenvironment for cell proliferation control.

1.7 Proliferation signaling autonomy and inputs from tumor microenvironment in cancer

In contrast to the normal proliferation, the mitogenic signaling supporting cancerous proliferation is better understood. While normal cells rely largely on exogenously supplied signals, cancer cells have evolved to employ several major strategies to accomplish self-sufficiency for proliferation signals. The first strategy involves alteration in the extracellular signals so that cancer cells may instruct themselves to proliferate via autocrine of growth factors and expression of cognate receptors. The second strategy involves the signaling transduction receptors. These receptors are often overexpressed in many cancers, enabling the cancer cells to be supra-responsive to the

9 growth signals (e.g. more available receptors on the surface of cell membrane that normally may be subject to the regulations such as internalization) . Alternatively, the receptors may harbor mutational changes to elicit downstream response in a ligand- independent manner. The third strategy involves alterations and constitutive activation in the components of intracellular circuits downstream the receptors.

While achievement of growth signaling autonomy appears to be critical for the uncontrolled proliferation of cancer cells, recent studies have also suggested it may not fully represent the underlying basis that fuels tumor growth. It has been increasingly appreciated that the non-cancer cells (stromal cells) in the tumor microenvironment may also play critical roles to drive tumorigenesis, along with the cell autonomous genetic and epigenetic alterations within cancer cells (Mueller and Fusenig, 2004; Tlsty and Coussens,

2006). Stromal cells may respond to the instructions from the cancer cells and then provide the cancer cells with various growth factors to support the cancer cell proliferation (Bhowmick et al., 2004b; Vermeulen et al., 2010). In some cases, the cooperating stromal cells may also undergo alterations to sustain the growth of the genetically or epigenetically deranged cancer cells (Hill et al., 2005). This has led to the notion that the emergence of tumors may result from continuous co-evolution between the tumor and the stroma compartments (Polyak et al., 2009). Some recent seminal studies start to reveal the individual genes that function in the stroma to impact various characteristics of tumor behavior and ultimately the tumor growth, although the comprehensive networks operating in the stroma still remain largely unexplored

(Bhowmick et al., 2004a; Hill et al., 2005; Hu et al., 2012; Katajisto et al., 2008; Kim et

10 al., 2006; Scherz-Shouval et al., 2014; Trimboli et al., 2009; Yang et al., 2008; Zhu et al.,

2002). Thus, the heterotypic signaling within the tumor microenvironment may play equally important roles as the intracellular programs to drive the carcinogenesis and awaits further systematic investigation for potential therapeutic interventions. In this study, we dissected how two intracellular transcriptional programs, E2f and Myc, control cell proliferation (as described in Chapter 2). Also we have identified and explored the inputs from the microenvironment that contribute to the proliferation control of epithelial cells (as described in Chapter 3).

Chapter 2 Redeployment of Myc and E2f1-3 drives cancer-like cell

cycle

2.1 Introduction

Spatial and temporal control of cell proliferation is vital for organogenesis and maintenance of tissue integrity. Normal proliferative stimuli impinge on cell surface receptors and engage multiple intracellular signaling cascades that converge on cell cycle control. Growth factor receptor activation culminates in the induction of Myc and E2f transcription factors, which are two central components believed to link external proliferative signals to early phases of the cell cycle (Bertoli et al., 2013; Meyer and

Penn, 2008).

In response to mitogenic signals such as Wnt, Notch, TGFβ and activation of many receptor tyrosine kinases (He et al., 1998; Palomero et al., 2006; Seoane et al.,

2001; Sharma et al., 2006; Weng et al., 2006), Myc in association with Max or Miz1 transcriptionally orchestrates a broad range of biological processes, including macromolecule biosynthesis, energy production, induction of cyclin-dependent kinases

(Cdks) (Arabi et al., 2005; Bouchard et al., 1999; Gomez-Roman et al., 2003; Grandori et al., 2005; Grewal et al., 2005; Hermeking et al., 2000; Leone et al., 1997; Morrish et al.,

2008; Seoane et al., 2001; Staller et al., 2001; Zhang et al., 2007), and also physically

12 interacts with the DNA pre-replication complex (Dominguez-Sola et al., 2007), which collectively prepare cells to enter S phase. Several studies have also linked Myc to the positive regulation of E2f expression and activity (Adams et al., 2000; Leone et al., 1997;

Sears et al., 1999; Sears et al., 1997). The E2f family consists of eight related members that have transcription activation and/or repression functions (Trimarchi and Lees, 2002).

Phosphorylation of Rb by Cdks liberates classical E2f repressors (E2f4-5) from Rb-E2f complexes, leading to peak levels of E2f activators (E2f1-3) late in G1, which activate a

G1-S transcriptional program that licenses the entry of cells into S phase and commitment through cell division (Dimova and Dyson, 2005). E2f6-8 are involved in the transcriptional repression of target genes in late S phase and thus contribute to their oscillatory expression as cell march though the cell cycle (Chen et al., 2009b). The fact that MYC and E2F1-3, or components that regulate them such as RB, CDKs and p16INK4A, are invariably disrupted in human cancer highlights their central role in the control of cellular proliferation (Chen et al., 2009b; Dang, 2012).

While cell culture studies suggest essential roles for Myc and E2fs in cell proliferation, knockout mouse models of Myc and E2fs reveal a surprising paucity of major cell cycle defects under physiological conditions (Bettess et al., 2005; Chen et al.,

2009a; Chong et al., 2009; Li et al., 2006; Wenzel et al., 2011). This has led to the conclusion that a high degree of redundancy exists within these two families of transcription factors. Indeed, knockin approaches demonstrated that E2f1 can replace

E2f3a and similarly, N-Myc can replace Myc to support murine development (Malynn et al., 2000; Tsai et al., 2008). Furthermore, recent work showed that mouse retinal

13 precursors can proliferate during embryonic development in the absence of N-Myc or

E2f1-3, but not in the absence of both sets of factors (Chen et al., 2009a), suggesting redundancy among the two transcription factor families as well. Whether Myc and E2fs also collaborate in other cell types and tissues and how they might do so remains to be determined.

Here we evaluated how Myc and E2f activities control cell cycle progression and cell proliferation in the small intestine of mice. The functional unit of the small intestine consists of pouch-like invaginations called crypts and finger-like projections called villi.

Pluripotent stem cells at the base of crypts continuously divide to generate transit amplifying progenitor cells, which undergo several rounds of proliferation before they exit the cell cycle and differentiate into paneth cells that migrate to the base of crypts, and into enterocytes, goblet and enteroendocrine cells that continuously repopulate villi. The regenerative capacity and spatial compartmentalization of cell proliferation makes the small intestine an experimental tractable system to study cell proliferation control in an in vivo context (van der Flier and Clevers, 2009). Using temporally regulated gene targeting approaches we show that the combined loss of Myc and E2f1-3 has little impact on G1-S transitions in the small intestine of mice. Rather, these loss of function studies show that

Myc and E2f1-3 synergistically engage an S-G2 transcriptional program required for the completion of S phase and progression through mitosis. When Rb is inactivated, however, Myc and E2fs are redeployed and engage a distinct G1-S program that promotes unregulated cell cycles. These findings distinguish how Myc and E2f control the

14 proliferation of normal versus cancer-like epithelial cells, and expose a molecular mechanism for the unexpected dependency of Rb deficient cells on Myc.

2.2 Results

2.2.1 Combined ablation of Myc and E2f1-3 results in disruption of crypt-villus integrity

To explore whether Myc and E2f activities collaborate in the control of normal cell cycles in vivo, we examined the small intestine of mice containing an inducible intestinal-specific Cre transgene (Ah-cre) and conditional alleles of Myc and/or E2f3

(Ireland et al., 2004). Because of the potential functional redundancy between E2f3 and the other two E2f activators, we also introduced null alleles of E2f1 and E2f2 into experimental animals. Ah-cre expression in intestinal crypts was induced by intraperitoneal administration of β-naphthoflavone (β-NF) and tissue histopathology was examined 7 days later by haematoxylin and eosin staining. Ablation of either E2f1-3 (Ah- cre;E2f1-/-;E2f2-/-;E2f3loxP/loxP, E2f TKO) or Myc (Ah-cre;MycloxP/loxP, Myc KO) had little effect on intestinal architecture (Figure 2.1a). The simultaneous deletion of E2f1-3 and

Myc (Ah-cre;E2f1-/-;E2f2-/-;E2f3loxP/loxP;MycloxP/loxP, E2f/Myc QKO), however, resulted in the complete collapse of crypt-villus structure. Examination at earlier time points following β-NF injection revealed that alterations in crypt architecture preceded the changes observed in associated villi (Figure 2.1b). By two days post β-NF injection,

E2f/Myc QKO crypts had fewer cells than controls, Moreover, crypt cells had reduced nuclear basophilia, enlarged nuclei and increased cellular size, whereas cells in the

15 corresponding villi appeared normal (Figure 2.1b, c). By four days post β-NF injection, the number of cells in E2f/Myc QKO crypts decreased to less than 50% of control animals leading to marked crypt atrophy and deterioration of villus integrity (Figure 2.1d). While mice became moribund within 1-2 weeks of β-NF treatment, they subsequently recovered, groomed and appeared healthy. Inspection of their small intestine showed that residual crypts escaping Cre-mediated deletion had repopulated the intestinal epithelium

(Figure 2.1e).

2.2.2 Combined Myc and E2f1-3 deficiency leads to S-G2 cell cycle arrest

We reasoned that the acute degeneration of E2f/Myc QKO crypts could be due to decreased cell proliferation. Surprisingly, DNA synthesis was unaffected in E2f/Myc

QKO progenitor cells two days following β-NF treatment, a time when Myc and E2f1-3 proteins were already depleted (Figure 2.2a, b and Figure 2.3a-c). Expression of geminin, a protein involved in blocking the re-replication of the genome late in S phase and G2

(Lee et al., 2009; McGarry and Kirschner, 1998), was also normal in E2f/Myc QKO cells

(Figure 2.2a, b and Figure 2.3a, b). However, progression through cell division was severely impaired in E2f/Myc QKO cells as indicated by the absence of mitotic figures and Serine 10-phosphorylated histone 3 (P-H3) staining (Figure 2.2a, b). Fluorescence- activated cell sorting analysis of single cell suspensions from E2f/Myc QKO crypts showed an accumulation of cells in S phase and a reduction of cells in G2-M compared to control littermates (Figure 2.2c). Together, these findings suggest that E2f/Myc QKO progenitor cells were able to enter S phase but failed to fully progress through S-G2 and

16 into mitosis. Despite the aberrant nuclear morphology and late cell cycle arrest in

E2f/Myc QKO samples, cell type-specific marker analysis revealed an appropriate number of paneth cells within crypts and goblet cells along the crypt-villus unit (Figure

2.3d), probably reflecting pre-existing non-deleted cells that persist beyond the experimental time frame analyzed here (i.e. paneth cells live for several weeks) (Ireland et al., 2005).

While E2f/Myc QKO progenitor cells entered S phase and replicated their genome, DNA synthesis was not entirely normal since widespread DNA double-strand breaks was observed in these samples, as measured by phosphorylated H2AX (P-H2AX) immunofluorescence (IF) (Figure 2.2d, e), indicating an inability of E2f/Myc QKO cells to properly repair replication-induced DNA breaks. The increase in DNA damage is a consequence from the specific ablation of E2fs since E2f TKO, but not Myc KO intestines, also displayed higher levels of P-H2AX. To determine whether cell death, possibly due to incurred DNA damage, contributed to the degeneration of E2f/Myc QKO crypts, tissue sections were processed for immunohistochemistry (IHC) using cleaved caspase-3 specific antibodies. This analysis showed that E2f TKO crypts, but not Myc KO or E2f/Myc QKO crypts, contained apoptotic cells (Figure 2.2d, e). We considered the possibility that loss of Myc might accelerate the elimination of E2f1-3 deficient apoptotic cells in E2f/Myc QKO crypts; however, this seems unlikely since a similar analysis at one and two days following β-NF injection also failed to detect apoptotic cells in these samples (data not shown). Thus, the execution of programmed cell death in E2f1-3 deficient crypts is dependent on Myc. Interestingly Myc was recently shown to be

17 required for DNA damage induced apoptosis of intestinal crypt cells (Phesse et al., 2014).

From these findings we conclude that cell cycle arrest in S-G2 and not apoptosis underlies the observed crypt atrophy caused by combined E2f-Myc deficiency.

2.2.3 Synergistic regulation of an S-G2 transcriptional program by Myc and E2f1-3

Given the established roles of Myc and E2f as transcription factors, we reasoned that changes in gene expression might underlie the observed S-G2 block in E2f/Myc QKO cells. We thus prepared crypt-enriched fractions derived from β-NF treated control, E2f

TKO, Myc KO and E2f/Myc QKO small intestines, isolated total RNA and examined global gene expression using an Affymetrix platform (mouse 430 2.0 arrays). This comparison revealed three main categories of differentially expressed genes (Figure

2.4a). One category (group I) includes genes dysregulated in all three mutant genetic groups (E2f TKO, Myc KO and E2f/Myc QKO), indicating a requirement for both transcription factor families to regulate the expression of this subset of genes. A second category (group II) includes genes dysregulated in E2f/Myc QKO samples and either E2f

TKO or Myc KO samples, suggesting that these target genes are uniquely regulated by

E2f1-3 or Myc. The third and perhaps most interesting category (group III) includes genes that are unaffected or only marginally dysregulated in E2f TKO and Myc KO cells, but are profoundly dysregulated in E2f/Myc QKO crypts, suggesting that this group of genes are synergistically regulated, albeit not necessarily directly, by E2f1-3 and Myc.

Gene ontology (GO) analysis of each group failed to detect any significant enrichment for classic G1-S regulated mRNAs. Instead, we noted that Group III genes were

18 dramatically enriched for mitotic related functions. Quantitative polymerase chain reaction with reverse transcription (RT-qPCR) assays confirmed the acute down- regulation of S-G2 related genes (Ccna2, Plk1, Cdc2a, Ccnb1, Cdc25c and Cdc20; Figure

2.4b) and the absence of expression changes in G1-S related genes (Pcna, Cdc6 and

Mcm3, Figure 2.4c) in these samples. Immunofluorescence validated the specific decrease in S-G2 related, but not G1-S related protein products in E2f/Myc QKO crypts

(Cdc2 and Ccna2 versus Pcna and Mcm3; Figure 2.4d). We conclude that Myc and E2f synergistically regulate an S-G2 expression program required for normal cell cycle progression in vivo.

2.2.4 Myc and E2f3 bind to both G1-S and S-G2 target genes in crypts

We then queried the genome-wide occupancy of E2f3 and Myc on chromatin of intestinal cells using chromatin immunoprecipitation combined with exonuclease digestion, followed by next generation sequencing (ChIP-exo-seq) (Rhee and Pugh, 2011;

Serandour et al., 2013). The ChIP-exo-seq approach incorporates combined double/single strand exonuclease treatment of immunoprecipitated DNA to provide enhanced sensitivity and specificity over traditional ChIP-seq methods, improving the detection of chromatin occupancy by non-abundant transcription factors, such as E2fs and Myc, in whole tissues where the target cell population may represent a small proportion of all cells (see Methods). To obtain sufficient material for these assays, crypt-enriched fractions from 32 wild type mice and villus-enriched fractions from 7 wild type mice were cross-linked and immunoprecipitated. Libraries were generated and sequenced on

19 an Illumina platform and the resulting data was analyzed as described in Methods.

Briefly, E2f3 and Myc DNA binding peak summits were identified using two algorithms

(MACS2 & GEM) (Guo et al., 2012; Zhang et al., 2008) with a false discovery rate

(FDR) of less than 1%. In crypts we identified 32,482 E2f3 peak summits and 158,997

Myc peak summits. Interestingly, most sites occupied by E2f3 were also co-occupied by

Myc (Figure 2.5a). In villi, we identified 7,211 E2f3 peak summits and 7,019 Myc peak summits, but the number of targets co-occupied by these two factors in villi was decreased when compared to their co-occupancy in crypts. Twenty randomly selected

E2f3- and twenty Myc-specific peak summits were validated by ChIP-PCR assays using control, E2f3-deficient and Myc-deficient tissues (Figure 2.5b). Peak summits were then mapped to genes based on the minimal distance between summits and proximal transcription start sites (Figure 2.5c). The patterns of E2f3 and Myc chromatin binding in crypts were distinct, with E2f3 occupancy biased towards promoter regions and Myc occupancy distributed more broadly across genes (Figure 2.5d). The observed decrease in

E2f3 and Myc binding to chromatin in non-proliferative villi is consistent with lower levels of E2f3a and Myc proteins in this compartment (Figure 2.5e and Figure 2.6a).

We then profiled and compared gene expression in crypts and villi using an

Affymetrix platform (mouse 430 2.0 arrays) and as expected, cell cycle related genes were preferentially expressed in crypts and differentiation related genes were preferentially expressed in villi (Figure 2.5f and Figure 2.6b). Integration of ChIP-exo- seq and gene expression data sets revealed a marked enrichment of E2f3 binding to promoter regions of differentially expressed cell cycle related genes (Figure 2.5g)

(Gauthier et al., 2008). Importantly, E2f3 and Myc chromatin co-occupancy in crypt samples was associated with cell cycle-related genes having functions in both G1-S and

S-G2 (Figure 2.5h, i, and Figure 2.6c). Notably, E2f3 and Myc targets included Cdc6,

Mcm3 and other G1-S genes expressed highly in crypts but whose expression was not impacted by the loss of E2f/Myc. Since only the expression of S-G2 related genes was disrupted in E2f/Myc QKO crypts, we suggest that the regulation of S-G2 targets requires direct binding by E2f or Myc, whereas the regulation of G1-S targets likely involves additional factors beyond E2f and Myc.

2.2.5 Myc drives ectopic proliferation of Rb deficient intestinal cells

Previous work suggested that distinct mechanisms regulate the proliferation of normal and cancer cells (Choi et al., 2012; Landis et al., 2006; Malumbres and Barbacid,

2009; Malumbres et al., 2004; Yu et al., 2001; Yu et al., 2006). Rb loss results in excessive proliferation of transit amplifying progenitor cells, even after they migrate, differentiate and begin to populate the length of the villus (Chong et al., 2009). To explore the roles of E2f and Myc in an abnormal hyper-proliferative context, we initially queried expression profiles derived from β-NF treated RbloxP/loxP (control) and Ah- cre;RbloxP/loxP (Rb KO) villus-enriched fractions. We noted that in addition to the expected upregulation of E2f target genes in Rb deficient villi, there was a striking increase in the expression of many known Myc target genes, as defined by previous gene-expression, reporter and ChIP assays. We thus tested whether Myc may be playing a role in driving the ectopic proliferation caused by Rb deficiency. Remarkably, ablation of Myc

21 suppressed the ectopic proliferation of Rb deficient intestinal villi cells (Ah- cre;RbloxP/loxP;MycloxP/loxP, Rb/Myc DKO; Figure 2.7a-c) to a similar extent as loss of

E2f1-3 (Ah-cre RbloxP/loxP;E2f1-/-;E2f2-/-;E2f3loxP/loxP, Rb/E2f QKO), without impacting the proliferation of progenitor cells in crypts.

Evaluation of a panel of cell cycle markers showed that loss of Myc or E2f1-3 in

Rb deficient villi prevented the aberrant accumulation of critical cell cycle regulators

(Ccne1, Ccne2, Ccna2, Ccnb1 and Ccnb2; Figure 2.7d, e). We also profiled global gene expression in control, RbKO, Rb/E2f QKO and Rb/Myc DKO villi. The heatmaps and waterfall plots shown in Figure 2.8a and Figure 2.8b illustrate the magnitude by which loss of Myc or E2f1-3 in Rb deficient villi restored expression programs to normal (>75% rescue) or near normal levels (25-75% rescue). Interestingly, the expression of the majority of dysregulated target genes was commonly corrected by loss of E2f1-3 or Myc

(390 from a total 701 genes; Figure 2.8c). RT-qPCR assays confirmed the reduction in the levels of a subset of these target targets (Cdc6, Mcm3 and Mcm6; Figure 2.8d and

Figure 2.9). This 390 gene-set included classic G1-S and S-G2 regulated mRNAs.

Together, these findings distinguish the roles of Myc and E2f in controlling transcription and proliferation between wild type crypt cells and Rb mutant villus cells.

2.2.6 Redeployment of Myc and E2f3 in Rb deficient cells

We then queried Myc and E2f3 chromatin occupancy in Rb deficient small intestines and compared it to wild type intestines. To this end, crypt-enriched fractions from 27 Rb KO mice and villus-enriched fractions from 7 Rb KO mice were utilized for

ChIP-exo-seq and the raw data was processed as described earlier for wild type samples.

In crypts we identified 27,200 E2f3 peak summits and 31,244 Myc peak summits. Thus, the number of E2f3 peak summits remained similar in crypts but the number of Myc summits was significantly decreased upon loss of Rb (Figure 2.10a, b). Whereas in control crypts the majority of E2f3 binding sites were co-occupied by Myc, loss of Rb resulted in more E2f3-specific peak summits. In villi, we identified 14,668 E2f3 peak summits and 3,929 Myc peak summits, representing an increase in E2f3 summits and a decrease in Myc summits upon Rb loss. Peak summits were then mapped to genes and their spatial distribution summarized in Figure 2.10b and Figure 2.10c. Three major parameters were analyzed and compared between wild type and Rb deficient crypts and villi: DNA binding location, DNA binding strength and DNA binding sequence motifs.

DNA BINDING LOCATION: We focused further analysis on the binding of E2f3 and

Myc to target genes dysregulated by Rb deficiency (701 dysregulated genes; see Figure

2.8b). E2f3 binding, but not Myc binding, was highly enriched in promoter regions

(Figure 2.10d). We then evaluated how loss of Rb influenced the location of E2f3 binding in villi. To this end, we compared E2f3 binding in the two proliferative compartments, wild type crypts and Rb deficient villi, by measuring the distance between a given summit in crypts (wild type) and its corresponding summit in villi (Rb KO); the same comparison was performed with the villi (Rb KO) as the reference compartment. While a distance of 0-100bp between corresponding summits in crypts and villi would indicate overlapping peaks and suggest that binding of E2f3 was not altered between compartments, a distance greater than 100bp between corresponding summits would

23 suggest a movement or disappearance of E2f3 occupancy from its original position in crypts. This analysis illustrated by the ‘peak summit-distance plots’ shown in Figure

2.10e revealed that approximately half of the E2f3 summits in control crypts were retained at the same position in Rb deficient villi, whereas the other half had corresponding summits in Rb deficient villi at distal locations greater than 100bp.

Consistent with this analysis, heatmaps of sequence tags showed that a portion of E2f3 summits in control crypts and Rb deficient villi were identical, whereas a large portion of summits present in control crypts were absent in the corresponding position in Rb deficient villi (Figure 2.10g). Moreover, a significant number of new E2f3 summits in Rb deficient villi were absent in control crypts. We also generated summit-distance plots and heatmaps to compare E2f3 peak summits across compartments and genotypes (Figure

2.11a, b). Examples of E2f3 binding to specific genes are shown in Supplementary Figure

2.12a-c. Together, these findings show that Rb deficiency in villi results in the recruitment of E2f3 to some of the same chromatin positions it normally occupies in control crypts, but also to new nearby and distant chromatin positions. A parallel analysis failed to show a broad redistribution of Myc binding in Rb deficient villi (Figure 2.10f, g).

However, comparison of control and Rb deficient crypts revealed a reduction in the ‘total’ number of Myc peak summits but the appearance of ‘new’ Myc-specific summits in Rb deficient crypts (Figure 2.11c, d; also see below). Thus, while E2f3 is redistributed in Rb deficient villi, Myc is redistributed in Rb deficient crypts.

DNA BINDING STRENGTH: The intensity plots shown in Figure 2.10h (and Figure

2.13a) illustrate an increase in the E2f3- and Myc-sequence tag intensity in Rb KO versus

24 control tissues, particularly in villi, indicating that E2f3 and Myc binding strength is increased upon Rb loss.

DNA BINDING SEQUENCE MOTIFS: Sequence analysis by de novo motif algorithms shows that E2f3 predominantly utilizes E2f canonical DNA binding elements irrespective of Rb deletion status or tissue compartment being analyzed (TTCCCGCC; Figure 2.10h top panels and Figure 2.13 and 2.14). In contrast, Myc utilizes a variety of non-canonical motifs in villi and control crypts but utilizes canonical E-box elements (CACGTG) in Rb

KO crypts (Figure 2.10h bottom panels and Figure 2.13 and 2.14). Thus, while Rb loss results in a marked reduction in Myc binding in crypts (Figure 2.10b), this binding is biased towards E-box elements.

From the combined analysis described above, which includes changes in the number of binding events as well as changes in DNA binding location/distribution, DNA binding strength, and DNA binding sequence motifs, we conclude that loss of Rb results in a major spatial redeployment of E2f3 binding in villi and a refocusing of Myc binding in crypts.

2.2.7 Myc is required for the accumulation of E2f3 in Rb deficient villi

We then explored the mechanistic relationship between Myc and E2fs in regulating gene expression. Previous studies established Myc as a critical downstream effector of Apc/β-catenin signaling (Sansom et al., 2007). Several nodes of crosstalk between canonical Apc/β-catenin/Myc and Cdk/Rb/E2f pathways have also been described (Cole et al., 2010; Davidson et al., 2009; Goodrich and Lee, 1992; Heinen et al.,

2002; Morris et al., 2008; Rustgi et al., 1991). Thus, we investigated the possibility that

Apc/β-catenin/Myc signaling might be increased in Rb deficient intestinal cells and contribute to their ectopic proliferation. Immunostaining of intestinal sections showed identical levels and localization of β-catenin protein in control and Rb KO small intestines (Figure 2.15a). Moreover, comparison of global mRNA profiles from control and Rb KO villi (Figure 2.8a) showed no change in the expression of a large cadre of known β-catenin target genes, including Myc itself (Figure 2.16a). Indeed, IHC of tissue sections showed that Myc protein levels were unchanged by loss of Rb (Figure 2.15b).

These results suggest that loss of Rb does not cause a general increase in Apc/β- catenin/Myc signaling in villi of the small intestine.

We then considered the possibility that Myc may regulate E2f expression as previous cell culture studies have shown (Adams et al., 2000; Leone et al., 1997; Sears et al., 1999; Sears et al., 1997). The E2f3 locus encodes two isoforms, E2f3a and E2f3b, driven by distinct promoters (Leone et al., 2000). RT-qPCR and IHC assays showed that

E2f3a (but not E2f3b) expression was markedly increased in Rb KO villi (Figure 2.15c, d).

Importantly, loss of Myc in (Rb/Myc DKO) resulted in decreased E2f3a mRNA levels without impacting E2f3b (Figure 2.15c). IHC with E2f3a-specific antibodies confirmed the reduction of E2f3a protein levels in Rb/Myc DKO samples (Figure 2.15d).

Consequently, E2f3-ChIP assays showed a moderate but significant decrease in the recruitment of E2f3 to its target promoters (Figure 2.15e).

We then evaluated Myc binding to the E2f3 locus in both crypts and villi. ChIP- exo-seq data from crypts revealed an abundance of Myc protein residing close to the TSS

26 of the E2f3a gene, which is spatially and functionally separate from the downstream

E2f3b TSS (Figure 2.15f) (Leone et al., 2000). Myc binding to the E2f3 locus was reduced in villi. We interpret these results to mean that Myc binds to the E2f3a promoter in both control and Rb deficient crypts and that this binding is diminished as progenitor cells differentiate and migrate into the villi. Binding of E2f3 to consensus E2f binding elements located just upstream of the E2f3a TSS (Figure 2.15f), which have been shown to positively regulate its expression (Adams et al., 2000), was increased in Rb KO versus control villi. Furthermore, the expression of E2f1 and E2f2, whose promoters are occupied by E2f3 and are E2f auto-regulated (Chen et al., 2009b), was increased in Rb deficient villi (Figure 2.15g and Figure 2.16b). Together, these findings suggest that Myc binding to the E2f3a promoter in Rb deficient crypts contributes to the ectopic accumulation of ‘free’ E2f3a protein in villi, presumably through epigenetic mechanisms that are maintained as crypt cells migrate into villi. E2f3a accumulation would then engage a positive feed-back loop culminating in the amplification of overall E2f1-3 transcriptional activity, the execution of an E2f gene expression program, including G1-S genes, to drive unregulated cell cycles.

2.3 Discussion

Cell proliferation is orchestrated by developmental and environmental mitogenic cues that are transmitted through signaling pathways controlling key events in the cell cycle. The convergence of multiple pathways provides cells with the necessary robustness (redundancy) to withstand perturbations in any single pathway that might

27 otherwise compromise cell cycle progression. We show here that Myc and E2f converge on the control of S-G2 in normal proliferating cells of the small intestine, but switch to the control of G1-S when Rb is inactivated (see diagram Figure 2.15h). These experiments uncover a molecular mechanism to explain how normal Myc and E2f pathways may be exploited and redirected by inactivation of the Rb tumor suppressor to foster unregulated cell proliferation, and highlight the unprecedented addiction of these

‘cancer-like’ cells on Myc.

We show an unexpected collaboration between Myc and E2f in the synergistic regulation of an S-G2 transcriptional program required for the cell cycle progression of progenitor cells. The immediate consequence of ablating Myc and E2f1-3 from progenitor cells is an arrest in S and/or G2. This arrest is visualized by enlarged nuclei with increased DNA content, expression of late S phase markers, decreased M phase markers, decreased cellularity and eventual collapse of crypt-villus architectural integrity. Whether intestinal stem cells have a similar requirement for Myc and E2f1-3 requires further rigorous evaluation. The exact mechanisms that underlie the cooperative roles of Myc and E2f in regulating S-G2 transcription remain to be determined but likely involve the co-recruitment of Myc and E2f3 to target promoters and interactions with chromatin remodeling factors as shown previously in vitro (McMahon et al., 1998; Ogawa et al.,

2002).

The observation that G1-S programs are refractory to the combined loss of Myc and E2f1-3 is surprising. Previous in vitro and cell culture studies have led to the prevailing view depicting Myc and E2f as two distinct transcriptional programs critical

28 for regulating S phase entry of G0/G1 cells (Adhikary and Eilers, 2005; Helin, 1998). This tenet is largely based on the consequences resulting from the introduction of supra- physiological levels of Myc or E2f into cells, which lead to ectopic transcription and S phase entry. However, a number of previous studies have shown that cell proliferation in developing and adult mouse tissues is resistant to the loss of either Myc or E2f1-3

(Bettess et al., 2005; Chen et al., 2009a; Chong et al., 2009; Li et al., 2006; Wenzel et al.,

2011). More recently, the combined ablation of N-Myc with E2f1-3 in mouse embryos was shown to impede DNA replication in retinal progenitor cells (Chen et al., 2009a).

The retina is a non-regenerating tissue where cell proliferation is restricted to a narrow developmental window of time late in embryogenesis. The caveat in studying non- regenerating tissues is that in the absence of a continuous replenishing pool of progenitor cells, an arrest at any cell cycle stage would exhibit reduced expression of G1-S target genes and score negative by assays that measure DNA replication (i.e. BrdU), precluding an understanding of the underlying nature of the arrest. Thus, it remains possible that as in the small intestine shown here, Myc and E2f may also converge to regulate S-G2 in the mouse retina. While G1-S transcription is insensitive to the combined loss of Myc and

E2f1-3, we show that Myc and E2f do indeed bind to G1-S related targets. Thus, it is possible and perhaps likely that Myc and E2f may contribute to G1-S transcriptional control, but only in collaboration with additional transcriptional inputs yet-to-be identified (Yang et al., 2007).

In contrast to normal cells, ectopic cell cycle progression of Rb deficient intestinal cells is acutely dependent on Myc and E2f1-3. The analysis described here provides

29 molecular insights into how loss of Rb subverts the existing Myc transcription machinery to support unregulated proliferation. While physiological levels of Myc in normal crypts utilize a variety of non-canonical DNA binding sequences (Uribesalgo et al., 2011), loss of Rb in crypts leads to the recruitment of Myc to target genes through the classical E-box sequence motif, as recently described for over-expressed or amplified Myc (Lin et al.,

2012; Nie et al., 2012; Sabo et al., 2014; Walz et al., 2014). The mechanisms for how this shift to the use of E-box DNA binding elements in crypts is established and how this might impact transcription in villi remain to be determined. It is possible that loss of Rb results in a global reconfiguration of chromatin (Narita et al., 2003; Nielsen et al., 2001;

Petrella et al., 2011; Siddiqui et al., 2007), which exposes cryptic DNA elements that enhance Myc binding and activity. Myc may contribute to the transcriptional output in Rb deficient villi in two ways that are not mutually exclusive. On one hand, Myc binding to regulatory sequences of E2f3a in crypts may facilitate its expression, which in the absence of Rb-mediated repression, would lead to the accumulation of ‘free’ E2f3a protein. This event would then engage a positive feed-forward loop to activate expression of all the three E2f activators and increase total E2f transcriptional activity. In this view,

Myc in crypts indirectly participates to amplify the output of the E2f program. On the other hand, binding of Myc to E-box elements in Rb deficient crypts may impact the local chromatin to facilitate increased binding of E2f and possibly other transcription factors to their target sites. It’s important to note that in contrast to the inactivation of Rb in crypts shown here, the targeted inactivation of Rb in villi (using distinct cre-transgenes) (Haigis et al., 2006) does not lead to either the overt induction of E2f or the ectopic proliferation

30 of cells. While the molecular basis for this discrepancy is not fully understood, it appears that the context of where Rb deletion occurs may underlie an important aspect of Rb’s tumor suppressive functions. Thus, the observation that loss of Rb in progenitor crypt cells redirects Myc binding to E-box elements may provide a potential molecular mechanism for how Rb’s function in crypts is essential to suppress ectopic proliferation in villi.

Loss of Rb also subverts the existing E2f transcription machinery to support unregulated proliferation. Like all transcription factors, E2f3 binding to chromatin is a dynamic process that is continuously in flux due to on-off binding kinetics. Additionally,

E2f3 protein levels acutely decrease in late S-G2 as cells prepare to undergo division, thus,

E2f3 must be remade and reloaded on chromatin at the next G1-S with every turn of the cell cycle. ChIP-exo-seq analysis shows that Rb deficiency results in the accumulation and recruitment of E2f3 in villi to some of the same chromatin positions it normally occupies in control crypts, but also to new chromatin locations not normally used in proliferating crypts. Thus, we conclude that Rb loss results in the redistribution of E2f3 on chromatin. This broad redeployment of E2f3 to ‘old’ and ‘new’ distant chromatin locations in villi culminates in the precocious expression of a transcriptional program that drives ectopic proliferation. In summary, we show that Myc conspires with inactivation of Rb to regulate and redeploy E2f3 across the entire genome, providing a molecular mechanism that explains how Rb deficient cells rely on or become addicted to Myc.

2.4 Methods

2.4.1 Mice

Mouse usage has been approved by Institutional Animal Care and Use Committee at the

Ohio State University. Mouse strains used in this study (Ah-cre, RbloxP/loxP, E2f1-/-, E2f2-/-,

E2f3loxP/loxP and MycloxP/loxP) have been described before(Chong et al., 2009; Trumpp et al., 2001) and maintained in a mixed background (C57BL/6 × 129 × FVB/N).

2.4.2 β-naphthoflavone, BrdU and EdU injection

To induce Ah-cre expression, 80mg kg-1 body weight of β-naphthoflavone (Sigma-

Aldrich; N3633) dissolved in corn oil (Sigma-Aldrich; C8267) was administrated into 2- month old mice with 5 injections within 30 hours. For DNA synthesis detection, 100mg kg-1 body weight of BrdU (Sigma-Aldrich; B5002) or 5mg kg-1 body weight of EdU (Life

Technologies; C10337) dissolved in sterile phosphate-buffered saline (PBS) was intraperitoneally injected 2 hours or 1 hour before the mice were sacrificed, respectively.

EdU staining was performed following manufacturer’s protocol.

2.4.3 Tissue preparation and histology

The tissue used for RNA and histology was collected from proximal 10cm intestine. For chromatin immunoprecipitation (ChIP), the intestine was divided into three parts of equal length and the first 1/3 proximal to stomach was used. Intestinal tissues were divided into

~1cm pieces and fixed with 10% pH-buffered formalin (Fisher Scientific; 23-245-685)

32 for 48h at room temperature, embedded in paraffin and cut into 4μm sections for histological staining.

2.4.4 Isolation of villi and crypts

Isolation of villi and crypts were processed similarly as previous described (Saenz-Robles et al., 2007). Immediately after euthanizing the mice, the small intestine was dissected with all mesentery and adipose tissue removed, cut longitudinally and gently washed in

PBS. The cleaned tissue was incubated in 25mL PBS with 0.5mM EDTA and 1mM dithiothreitol (DTT) for 30min at room temperature. Villus fractions were shaken off from the intestine sequentially in multiple tubes containing 10mL ice-cold PBS with

1mM DTT. The fractions were incubated on ice for 5min to allow spontaneous precipitation of intact villi by gravity. The supernatant containing broken pieces of villi was removed. After most villi were collected, the remaining intestinal tissue was incubated in 25mL PBS with 0.9mM EDTA and 1mM DTT for another 30min at room temperature. To remove residual amount of villi and collect crypts, the tissue was further shaken sequentially in multiple tubes containing 10mL ice-cold PBS with 1mM DTT until complete dissociation of epithelial cells from the mesenchyme/muscle tissues.

Crypt-enriched fractions were filtered with 70μm cell strainer (Fisher Scientific;

22363548) to minimize the potential contamination from villus fractions and increase the yields of crypts (more available crypt factions after the removal of mixed villi). Each fraction was examined microscopically to ensure the purity of enriched villi/crypts for

33 further experiments. After brief centrifuge, the cell pellet was washed in ice-cold PBS and frozen with liquid nitrogen or processed further for ChIP assays.

2.4.5 RNA and quantitative PCR

Total RNA from purified villi/crypts was isolated using TRIzol reagent (Invitrogen;

15596-018) and further cleaned up with RNeasy Mini Kit (Qiagen). First strand cDNA was synthesized using 100ng total RNA with Superscipt III Reverse Transcriptase Kit

(Invitrogen; 18080-044) following manufacturer’s protocol. Quantitative PCR (qPCR) was performed using SYBR Green master mix (Bio-Rad; 170-8884) with Bio-Rad iQ5

Multicolor Real-Time PCR Detection System or Applied Biosystems StepOnePlus Real-

Time PCR System. Samples were analyzed with triplicates and the relative expression levels of genes were calculated with 2-∆∆Ct method. Gapdh mRNA level was used for normalization.

2.4.6 Affymetrix microarray analysis

Affymetrix microarray analysis was performed at the Ohio State University Shared

Resources (http://cancer.osu.edu/research/cancerresearch/sharedresources/microarray) with Affymetrix Mouse 430.2 Array. Four or five independent samples of each genetic group were utilized for analysis. The p-value for expression level changes was calculated in R using t-test (t.test(val1,val2)$p.value under R). The Gene Ontology analysis was performed by DAVID Bioinformatics Resources (Database for Annotation, Visualization and Integrated Discovery) at National Institute of Allergy and Infectious Diseases

(NIAID), NIH. The degree of rescue (D) for mRNA level of a given gene in the waterfall plot in Figure 2.8b is defined as:

퐴 − 퐵 퐷 = ∗ 100% 퐴

A: mRNA level fold change between Rb KO and control;

B: mRNA level fold change between Rb/E2f QKO (or Rb/Myc DKO) and control.

2.4.7 Immuno-staining

After deparaffinization, antigen retrieval was performed with boiling citrate buffer

(DAKO; S169984) for 30min except that proteinase K was used for BrdU and lysozyme.

Specifically, 2N HCl was further applied to tissue sections for exposing and detecting

BrdU and then neutralized with 0.1M sodium tetraborate (pH 8.0). For immunohistochemistry staining, the endogenous peroxidase was inactivated by incubation with 3% H2O2 for 15min at room temperature. Normal horse serum was applied as the blocking reagent except that Mouse on Mouse (M.O.M.) Blocking Reagent

(Vector Labs; MKB-2213) was applied for primary antibodies raised in mice. Primary antibodies used were BrdU (Dako; M0744), geminin (Santa Cruz; sc-13015), Serine 10- phosphorylated histone 3 (Millipore; 06-570), cleaved caspase-3 (Cell Signaling; 9661), phosphorylated H2AX (Cell Signaling; 9718), Ccna2 (Santa Cruz; sc-596), Cdc2 (Santa

Cruz; sc-54), Mcm3 (Santa Cruz; sc-9850), Pcna (Santa Cruz; sc-56), lysozyme (DAKO;

A0099), Myc (Santa Cruz; sc-764), E2f3 (Millipore; 05-551) and β-catenin (AbD

Serotec; OBT1683). Signals were visualized with biotinylated goat anti-rabbit IgG antibody (Vector Labs; BA-1000) and DAB substrate kit (Vector Labs; SK-4100), 35 donkey anti-mouse Alexa Fluor 488 or Alexa Fluor 594, donkey anti-goat Alexa Fluor

594, donkey anti-rabbit Alexa Fluor 594 (Invitrogen). TSA™ Plus Fluorescein kit

(Perkin Elmer; NEL741001KT) was utilized for visualizing geminin signals. The nuclei were counterstained with haematoxylin or 4’,6-diamidino-2-phenylindole (DAPI). The images were collected using Nikon Eclipse 50i (immunohistochemistry) or Zeiss

Axioskop 40 (immunofluorescence) microscopes with a Zeiss AxioCam HRc camera and

AxioVision software.

2.4.8 ChIP-PCR

Freshly isolated crypts/villi were incubated in PBS with 1% formaldehyde at 37°C for

15min. The DNA-protein crosslink reaction was terminated by incubation with 0.125M glycine for 5min. The sample was washed in ice-cold PBS, resuspended in cytosolic lysis buffer (5mM PIPES (pH 8.0), 85mM KCl, 0.5% NP-40 and proteinase inhibitors) and nuclear lysis buffer (50mM Tris (pH 8.0), 10mM EDTA, 1% SDS and proteinase inhibitors). The chromatin was sonicated with majority of DNA fragments between 100-

300bp. The fragmented chromatin was diluted 1:10 with IP dilution buffer (16.7mM Tris

(pH 8.0), 167mM NaCl, 1.2mM EDTA, 1.1% Triton X-100 and proteinase inhibitors) and pre-cleared by incubating with Protein G Plus/Protein A beads (Calbiochem; IP05) at

4°C for 1 hour on rotator prior to adding antibodies. Antibodies used were E2f3 (Santa

Cruz; sc-878 X), Myc (Santa Cruz; sc-764 X) and normal rabbit IgG (Santa Cruz; sc-

2027). The DNA-protein-antibody-bead complex was washed twice with each following buffer in the presence of proteinase inhibitors: low-salt buffer (20mM Tris (pH 8.0),

2mM EDTA, 1% Triton X-100, 0.1% SDS and 150mM NaCl), high-salt buffer (20mM

Tris (pH 8.0), 2mM EDTA, 1% Triton X-100, 0.1% SDS and 500mM NaCl), LiCl buffer

(10mM Tris (pH 8.0), 1mM EDTA, 250mM LiCl, 1% NP-40 and 1% deoxycholic acid) and Tris-EDTA buffer (pH 8.0). Immunoprecipitated chromatin was retrieved from

Protein G Plus/Protein A beads in 1% SDS plus 0.1M NaHCO3 and reverse-crosslinked with 0.5M NaCl and 0.1mg/mL RNase A overnight at 65°C (including non- immunoprecipitated inputs). The product was further treated with 0.1mg/mL Proteinase

K and the DNA was purified with Qiagen kits (QIAquick Purification Kit (Qiagen;

28104) for DNA detected by quantitative PCR. One-tailed Student’s t-test was performed to determine whether E2f3 binding to target genes was decreased in Rb/Myc DKO villi compared to Rb KO villi in Figure 2.15e.

2.4.9 ChIP-exo-seq library construction and Illumina sequencing

Crypts from 32 control mice and 27 Rb KO mice and villi from 7 control mice and 7 Rb

KO mice were utilized for ChIP-exo-seq. Crypts collected from every 4-5 mice of the same genotype were combined and processed as one sample. The crosslink, cytosolic lysis, nuclear lysis, sonication, dilution with IP dilution buffer, pre-clean and immunoprecipitation steps were performed as described above except Pierce Protein AG

Magnetic Beads (Thermo Scientific; 88802) were used. The library construction steps including all on-bead enzymatic reactions (end polishing, P7 exo-adapter ligation, nick repair, λ-exonuclease digestion, RecJf exonuclease digestion), elution and reverse cross- linking, primer extension and P5 exo-adaptor ligation were performed as described

(Serandour et al., 2013). The resulting DNA was enriched by 13 cycles of PCR using

NEBNext High-Fidelity PCR Master Mix (NEB; M0541S) before measuring the library concentration and DNA size distribution. Cluster generation on Illumina cBot and single- end high throughput sequencing on Illumina HiSeq 2500 platform were performed at the

Ohio State University Shared Resources.

2.4.10 ChIP-exo-seq data analysis

The 50bp sequencing reads were de-multiplexed based on the barcodes in index adaptors utilized in the library construction, followed by mapping the reads to reference mouse genome (GRCm37/mm9) with Bowtie2 aligner (version 2.2.1) (Langmead and Salzberg,

2012). Only the reads uniquely mapped to the genome were utilized for further analysis.

E2f3 and Myc DNA binding peaks were identified using Genome wide Event finding and

Motif discovery algorithm GEM (version 2.4.1) (Guo et al., 2012) with a false discovery rate (FDR) less than 1% and Model-based Analysis of ChIP-Seq algorithm MACS2

(version 2.0.10) (Zhang et al., 2008) with a FDR less than 5%. The intersection of GEM and MACS2 results were used for all downstream analysis. The DNA binding peaks were associated to genes based on the minimal distance between peak summits and proximal transcription start sites. Motif discovery and matching de novo to known motifs were performed with HOMER algorithm (version 4.6) (Heinz et al., 2010) using DNA sequences around peak summits (±50bp). Integrative Genomics Viewer algorithm IGV

(Robinson et al., 2011) was used to visualize the binding events and reads coverage. The

BED files with detailed information for identified peaks have been deposited in GEO database with accession number GSE56009.

2.4.11 Flow cytometry

Isolated crypts were homogenized using a Dounce homogenizer in ice-cold suspension buffer (25mM Tris (pH 7.5), 50mM KCl, 2mM MgCl2, 1mM EDTA and 1mM phenylmethylsulfonyl fluoride (PMSF)). After brief centrifuge, the pellet was washed twice with the suspension buffer and once with PBS. Single cell suspension was then prepared in PBS containing 0.1% Triton X-100, 25μg/mL propidium iodide (Roche;

11348639001) and 20μg/mL RNase A (Invitrogen; 12091-021) and incubated 1 hour at room temperature at dark. The suspension was then filtered through 35μm cell strainer caps (BD Falcon; 352235). The DNA content measured by fluorescence intensity of propidium iodide was analyzed using BD LSR II Flow Cytometer at the Ohio State

University Shared Resources. The data was analyzed with FlowJo cytometric analytical software.

2.4.12 Data deposition

Affymetrix gene expression and ChIP-exo-seq data has been deposited in GEO database with accession number GSE56009.

Figure 2.1 Disruption of the small intestine by combined loss of Myc and E2f1-3

(a) Haematoxylin-and-eosin (H&E) stained tissue sections from control, E2f TKO, Myc

KO and E2f/Myc QKO intestines collected 7 days after induction of Ah-cre expression.

(b) Progressive degeneration of E2f/Myc QKO crypts from day 1 to day 4 after induction of Ah-cre expression. (c) H&E stained tissue sections from control and E2f/Myc QKO intestines collected 2 days after induction of Ah-cre expression. (d) Quantification of average number of crypt cells. (e) E2f/Myc QKO tissue sections stained by H&E and immunohistochemistry (IHC) of Myc to show the regeneration of intestinal epithelium by

Myc-positive cells (arrows). Data in d presented as mean ± SD. Scale bars represent 100

µm (a, e), 50 µm (b) and 25 µm (c).

Figure 2.1

Figure 2.2 S-G2 cell cycle arrest in E2f/Myc QKO progenitor cells

(a) Immunofluorescence (IF) staining of BrdU (green), geminin (green) and P-H3 (red) in crypts from intestines harvested 2 days after induction of Ah-cre expression. Nuclei were stained with DAPI (blue). (b) Quantification of BrdU, geminin and P-H3 staining. (c)

Fluorescence-activated cell sorting analysis of cell cycle status in control and E2f/Myc

QKO crypts from intestines harvested 2 days after induction of Ah-cre expression: left, representative histograms; right, quantification. (d) IF staining of P-H2AX (red) and IHC staining of cleaved caspase-3 (brown) in crypts. Intestines were harvested 2 days (P-

H2AX) or 4 days (cleaved caspase-3) after induction of Ah-cre expression. (e)

Quantification of P-H2AX and cleaved caspase-3 staining. Data in b, c and e presented as mean ± SD. Scale bars in a and d represent 50 µm.

Figure 2.2

Figure 2.3 Loss of E2f1-3 and/or Myc in crypts

(a) Dual staining of EdU (red) and geminin (green, included here for the composition of merged images) in crypts from intestines harvested 2 days after induction of Ah-cre expression. Nuclei were stained with DAPI (blue). (b) Quantification of EdU and geminin staining. (c) IHC staining of Myc in intestines harvested 2 days after induction of Ah-cre expression. (d) Alcian Blue staining for goblet cells (blue) or IF staining of lysozyme for paneth cells (red) in intestines harvested 2 days after induction of Ah-cre expression. Nuclei were counterstained with Nuclear Fast Red or DAPI. Scale bars in a, c and d represent 50μm.

Figure 2.3

Figure 2.4 Synergistic regulation of an S-G2 transcriptional program by Myc and E2f1-3

(a) Heatmap representation for clustering of differentially expressed genes between mutant genetic groups compared to control samples. Crypts were collected 2 days after induction of Ah-cre expression. (b) Quantitative polymerase chain reaction with reverse transcription (RT-qPCR) analysis for a subset of S-G2 related genes synergistically regulated by Myc and E2f1-3. Normal expression levels are illustrated as grey dotted lines and dysregulated expression levels are illustrated as red dotted lines. (c) RT-qPCR analysis for a subset G1-S related genes. Normal expression levels are illustrated as grey dotted lines. (d) IF staining of Pcna, Mcm3, Ccna2 and Cdc2. Note that degenerating

E2f/Myc QKO crypts with less dense cells have comparable protein levels of Pcna and

Mcm3, yet significantly less Ccna2 and Cdc2, compared to other genetic groups. Data in b and c presented as mean ± SD. Scale bars in d represent 50 µm.

Figure 2.4

Figure 2.5 Chromatin binding of E2f3 and Myc in wild type tissues

(a) Heatmap representation for tag intensity of all E2f3 and Myc binding locations in wild type intestines. (b) E2f3 and Myc ChIP-PCR: control (Rb KO, n=3) and E2f1-3 deficient

(Rb/E2f QKO, n=3) villi for E2f3; control (wild type, n=4) and Myc deficient (Myc KO, n=3) crypts for Myc. The signals in control groups with E2f3- or Myc-antibody immunoprecipitation are set as 1. (c) Spatial distribution of all E2f3 and Myc peak summits in wild type crypts and villi. The number of summits in each compartment is shown in parentheses. 5’ distal: 5’ region more than -50 kilobases (kb) away from transcription start sites (TSSs). 5’ proximal: 5’ region within -50kb to -5kb of TSSs.

Promoter: within -5kb to +2kb of TSSs. Gene body: from +2kb of TSSs to ends of transcripts. 3’ region: 3’ region starting from ends of transcripts. (d) Density plots across genomic regions of all E2f3 and Myc peak summits in wild type crypts and villi. Note that gene bodies for individual genes are divided into 100 bins and summits are accordingly assigned. (e) IHC staining of E2f3a and Myc in wild type intestines. (f)

Heatmap representation for differential expression of cell cycle related genes, as annotated in Cyclebase database, in wild type crypts and villi. (g) Spatial distribution of

E2f3 and Myc peak summits associated with differentially expressed cell cycle related genes in wild type crypts and villi. (h) ChIP-exo-seq track examples showing E2f3 and

Myc binding to selected G1-S related genes in wild type crypts. E2f3 and Myc peaks are highlighted in blue and red, respectively. (i) ChIP-exo-seq track examples showing E2f3

48 and Myc binding to selected S-G2 related genes in wild type crypts. Data in b presented as mean ± SD. Scale bars in e represent 50 µm. Scale bars in h and i represent 1kb.

Figure 2.5

Figure 2.6 DNA binding of E2f3 and Myc in wild type crypts

(a) IHC staining of E2f3 using the antibody (Millipore; 05-551) in control, E2f3a-/-,

E2f3b-/- and Ah-cre;E2f1-/-;E2f2-/-;E2f3loxP/loxP intestines. E2f3b-/- but not E2f3a-/- intestines show positive-stained cells, indicating the specific recognition of E2f3a isoform by this antibody. Note the non-specific staining of blood cells. (b) Heatmap representation for differential gene expression in wild type crypts and villi. (c) ChIP-exo-seq track examples showing E2f3 and Myc binding to selected G1-S and S-G2 related genes in wild type crypts. Scale bars in a represent 25 µm. Scale bars in c represent 1kb.

Figure 2.6

Figure 2.7 Rb deficient cells require Myc to drive ectopic cell cycles

(a) H&E stained tissue sections from control, Rb KO, Rb/E2f QKO and Rb/Myc DKO intestines. Note the hyperplastic feature of Rb KO villi. (b) IF staining of BrdU and P-H3.

Note the non-specific staining of blood cells in the lumen of villi. (c) Quantification of

BrdU and P-H3 staining. (d) RT-qPCR analysis for multiple cyclins in control, Rb KO,

Rb/E2f QKO and Rb/Myc DKO villi. (e) IHC staining of Ccna2. Data in c and d presented as mean ± SD. Scale bars in a, b and e represent 50 µm.

Figure 2.7

Figure 2.8 Myc and E2f1-3 regulate an overlapping G1-S transcriptional program in Rb- null cells

(a) Heatmap representation for dysregulated expression of 701 genes in Rb KO villi overlapped between two sets of independent samples (previously published data(Chong et al., 2009) in left panel included here for comparison). Group IV: genes with expression levels rescued by loss of either E2f1-3 or Myc. Group V: genes with expression levels rescued by loss of E2f1-3. Group VI: genes with expression levels rescued by loss of

Myc. (b) Waterfall plots showing the extent of gene expression rescue in Rb/E2f QKO and Rb/Myc DKO villi for genes dysregulated in RbKO villi. (c) Venn diagram showing the overlap between genes with expression levels rescued by loss of E2f1-3 or Myc

(complete and partial rescue). (d) RT-qPCR analysis for a subset of G1-S related genes.

Data in d presented as mean ± SD.

Figure 2.8

Figure 2.9 Loss of either E2f1-3 or Myc corrects aberrant transcription in Rb KO villi

Quantitative polymerase chain reaction with reverse transcription (RT-qPCR) analysis of mRNA levels for a subset of genes in control, Rb KO, Rb/E2f QKO and Rb/Myc DKO villi. The aberrant expression of these genes in Rb KO villi were normalized by loss of either E2f1-3 or Myc. The average expression level of control samples was set as 1.

Figure 2.9

Figure 2.10 Rb loss redefines the chromatin binding landscape of E2f3 and Myc

(a) Heatmap representation for tag intensity of all E2f3 and Myc binding locations in Rb

KO intestines. (b) Spatial distribution of all E2f3 and Myc peak summits in control and

Rb KO tissues (data for control tissues included here for comparison). The number of summits in each compartment is shown in parentheses. (c) Density plots across genomic regions of all E2f3 and Myc peaks in Rb KO crypts and villi. Note that gene bodies for individual genes are divided into 100 bins and summits are accordingly assigned. (d)

Spatial distribution of E2f3 and Myc peak summits associated with dysregulated genes in

Rb KO villi. The number of summits in each compartment is shown in parentheses. (e)

Peak summit-distance plots for E2f3 summits in control crypts and Rb KO villi that are associated with dysregulated genes in Rb KO villi. (f) Peak summit-distance plots for

Myc summits in control crypts and Rb KO villi that are associated with dysregulated genes in Rb KO villi. (g) Heatmap representation for tag intensity of all E2f3 and Myc binding locations in control crypts and Rb KO villi. (h) Tag intensity plots (tags per bp per peak per 100M reads) around the peak summits associated with genes dysregulated in

Rb KO villi. The canonical DNA binding motifs for E2f3 and Myc are highlighted in blue and red boxes, respectively.

Figure 2.10

Figure 2.11 Peak summit-distance plots and tag intensity heatmaps across tissue compartments and genetic groups

(a) Peak summit-distance plots for E2f3 summits that are associated with dysregulated genes in Rb KO villi and heatmap representation for tag intensity of all E2f3 binding locations in different tissue compartments. (b) Peak summit-distance plots for E2f3 summits that are associated with dysregulated genes in Rb KO villi and heatmap representation for tag intensity of all E2f3 binding locations in different genetic groups.

(c) Peak summit-distance plots for Myc summits that are associated with dysregulated genes in Rb KO villi and heatmap representation for tag intensity of all Myc binding locations in different tissue compartments. (d) Peak summit-distance plots for Myc summits that are associated with dysregulated genes in Rb KO villi and heatmap representation for tag intensity of all Myc binding locations in different genetic groups.

Figure 2.11

Figure 2.12 E2f3 DNA binding in wild type and Rb KO villi

(a-c) Three main categories of E2f3 DNA binding in wild type and Rb KO villi are illustrated: Group 1 includes binding locations with E2f3 peaks present in Rb KO villi but absent in control villi. Group 2 includes binding locations with distinct E2f3 peaks in control and Rb KO villi (either due to additional peak summits or peak position changes).

Group 3 includes binding locations with E2f3 peaks similarly positioned in control and

Rb KO villi (but possibly with different magnitude of binding strengths). ChIP-exo-seq track examples are shown for group 1 (a), group 2 (b) and group 3 (c). Scale bars represent 1kb.

Figure 2.12

Figure 2.13 DNA binding strength and DNA binding motifs of all Myc and E2f3 peaks

(a) Tag intensity plots (tags per bp per peak per 100M reads) around the peak summits of all peaks identified in various tissues. (b) Canonical DNA binding motifs (TTCCCGCC for E2f3 and CACGTG for Myc; underlined with blue and reds lines, respectively) or strongest non-canonical motifs identified from peak sequences in various genomic regions.

Figure 2.13

Figure 2.14 DNA binding strength and DNA binding motifs of Myc and E2f3 peaks associated with dysregulated genes in Rb KO villi

Canonical DNA binding motifs (TTCCCGCC for E2f3 and CACGTG for Myc; underlined with blue and reds lines, respectively) or strongest non-canonical motifs identified from peak sequences in various genomic regions.

Figure 2.14

Figure 2.15 Myc regulates E2f3a expression in Rb deficient villi

(a) IHC staining of β-catenin. (b) IHC staining of Myc. (c) RT-qPCR analysis for E2f3a and E2f3b in control, Rb KO and Rb/Myc DKO villi. (d) IHC staining of E2f3a in control,

Rb KO, Rb/Myc DKO and Rb/E2f QKO samples. Note the non-specific staining of blood cells in the lumen of the villi. (e) ChIP PCR analysis showing E2f3 loading to target genes (Pbk and Rrm1) in control, Rb KO and Rb/Myc DKO villi (n=4 per genetic group).

The 5’ region ~1kb away from TSS of Gapdh was used as the control region lacking significant E2f3 loading. *p<0.05, one-tailed t-test. (f) ChIP-exo-seq tracks showing occupancy of E2f3 locus by E2f3 itself and Myc. Distinct promoter regions for E2f3a and

E2f3b are shaded in red and yellow, respectively. (g) RT-qPCR analysis for E2f1 and

E2f2 in control and Rb KO villi. (h) Diagrams summarizing the regulation of cell cycle by Myc and E2f1-3 in wild type and Rb deficient cells. Data in c, e and g presented as mean ± SD. Scale bars represent 25 µm (a) and 50 µm (b, d). Scale bars in f represent

1kb.

Figure 2.15

Figure 2.16 Expression of Wnt/β-catenin targets in Rb KO villi and auto-regulation of

E2fs

(a) Venn diagram showing the overlap between genes dysregulated in Rb KO villi (701 genes) and Wnt/β-catenin target genes (111 genes). (b) ChIP-exo-seq tracks showing

E2f3 binding to E2f1 and E2f2 locus. The differential binding between control and Rb

KO villi is highlighted in blue.

Figure 2.16

Primers for mouse genotyping Gene Forward primer(s) (5'-3') Reverse primer(s) (5'-3') Ah-cre CCTGACTAGCATGGCGATAC ATTGCCCCTGTTTCACTATC Rb GGCGTGTGCCATCAATG CTCAAGAGCTCAGACTCATGG TACAGTCCCAAAGCCCCAGCCAA Myc CACCGCCTACATCCTGTCCATTC G AGAAGTCACGCTATGAAACCTCA AGCCACTGGATATGATTCTTGGA E2f1 C C AGTGCCAGCGGGGCTGCTAAAG

GCCCCTAACACATGCACCCATTG E2f2 CCTGAGCGAGTCGGAGGATGG G ACCAAAGAACGGAGCCGGTTGGC G AAGGGAAGGGAAAATTAAATCT E2f3 TGTGAATAATTTTTGGCATGTTTT GA CTTATTCTGAGTGTGGACATACC G

Primers for RT-qPCR Gene Forward primer (5'-3') Reverse primer (5'-3') Pcna TAGCCACATTGGAGATGCTG CAGTGGAGTGGCTTTTGTGA Cdc6 AGGGTGACTTTGAGCCAAGA ATGAAGATTCTGGGGGCTCT Mcm3 ACCCAGGACTCCCAGAAAGT ATTCCGGTTGATGGATTCTG Ccna2 ATTGGGTCACTGGGACTCTG ATGGGAGCGTTAGGACCTCT Plk1 GCCCGTCTCCCTATTACCTG CAATGGCCTCATTTGTCTCC Cdc2a AGCTCTGGGCACTCCTAACA CCATTTTGCCAGAGATTCGT Ccnb1 TGGACTACGACATGGTGCAT CAGGTGCTGCATAACAGGAA Cdc25c CTGAGGAAGCCTGTTGTTCC AATGCCGGATACTGGTTCAG Cdc20 ATTTGGAACGTCTGCTCAGG CTTGGCCATGGTTGGATACT Ccne1 ACTTTCTGCAGCGTCATCCT TGTTGTGATGCCATGTAACG Ccne2 GACTGGATGGTGCCTTTTGT GACAGCTGCCCTCCTTTTCT Ccnb2 GGCCAAGAACGTTGTGAAAG AGGATGGCAGTCCAGTGTCT E2f1 AGAGTGAGCAGCAGCTGGAT CCATCTGTTCTGCAGGGTCT E2f2 ACGGCGCAACCTACAAAGAG GTCTGCGTGTAAAGCGAAGT E2f3a GCCTCTACACCACGCCACAAG TCGCCCAGTTCCAGCCTTC E2f3b CGGAAATGCCCTTACAGC CTCAGTCACTTCTTTGGACAG

Table 2.1 Primers for mouse genotyping, RT-qPCR and ChIP-PCR

continued

Table 2.1 continued

Kif11 GGAGGCTCTGGAGCAGTGTA GGCAGTCTGCTCTTGGAGAT Smc2 TGCTCCTCTTCAAACCTGCT TTGGGTTCTTTTCCTCTGGA Pbk AAGCTGCTTCATGGAGACATAA CCCACAGAGTAAGGCCAAAA Cenpk GCTTACCTTGGGTGCTTTTT TTTGGGATGGATCTTCTGGA Tfdp1 AGGGATCCATTGGTGGAGTA TCAATCCTCCTCGTCGTTCT Hmgb2 GCGAGGAGCACAAGAAGAAG CCTCCTCATCTTCTGGTTCG Mcm6 AAGCGAGCTTGTCAACTGGT ACCAGGTAGGGGTCTTCCTC Gapdh CGGTGTGAACGGATTTGGC TTTGATGTTAGTGGGGTCTCGC

Primers for ChIP-PCR in Figure 2.15e Gene Forward primer (5'-3') Reverse primer (5'-3') Pbk ACGCGTGAGTTTTGACCCTA CCAGAAAATGCAGGAACAGC Rrm1 GGCTTTGCCCACACTCAATA CGGAATTCGAATGACGTTCT Gapdh CTGCACCTGCTACAGTGCTC GGAAGGGAGAAAAGGCATTC

Primers for ChIP-PCR in Figure 2.5b

Tmpo E2f3 Forward (5'-3') TGGCGACCTCCCTTTGTAAT ChIP F Tmpo E2f3 Reverse (5'-3') GGACAAACCCTGGTGATTCG ChIP R Cd2ap E2f3 Forward (5'-3') AAGTTTTCCCGGGTGACTCC ChIP F Cd2ap E2f3 Reverse (5'-3') GTGGCAGCAATGTCGTTTTG ChIP R Krt19 E2f3 Forward (5'-3') ACAACCCTTATCGTGGTCCA ChIP F Krt19 E2f3 Reverse (5'-3') CCTTCAAGGTTGCTGGTATCT ChIP R Vgll4 E2f3 Forward (5'-3') TGTTAGGGTGATAGGCCTGC ChIP F Vgll4 E2f3 Reverse (5'-3') AGACCTTTGTTCCCGCTCTT ChIP R Paip2b E2f3 Forward (5'-3') CCTATTGGCTGGCTTGTTGG ChIP F Paip2b E2f3 Reverse (5'-3') TCCCTCACCTCAATTCGCTT ChIP R Tgm2 E2f3 Forward (5'-3') TCCAGAGTACATGCCAGTCA ChIP F

continued

Table 2.1 continued

Armc9 E2f3 Reverse (5'-3') ACAGAGAAACCCTGTCTCAAAA ChIP R Tgm2 E2f3 Reverse (5'-3') GCTCTTTCACAGGCCAATGT ChIP R Armc9 E2f3 Forward (5'-3') TTGACCCCAGTTGGCTTACT ChIP F Prkag2 E2f3 Forward (5'-3') GTGGGTTTAAGGAGCACAGC ChIP F Prkag2 E2f3 Reverse (5'-3') GTTAGCAGCACACCCCAAAT ChIP R Fam73a E2f3 Forward (5'-3') GTTAAACAGGCAGGCTCCAG ChIP F Fam73a E2f3 Reverse (5'-3') TGCGTTCGTACTACTGACTCA ChIP R Gsr E2f3 ChIP Forward (5'-3') TTGCGCCACTTTCTGTTTGT F Gsr E2f3 ChIP Reverse (5'-3') AAACTCCCTCCTCGTTGACC R 2410004B18Ri Forward (5'-3') CGGCCTCGTTTTCTTATCTCT k E2f3 ChIP F 2410004B18Ri Reverse (5'-3') CCACACGGAAGGCAAAGAAA k E2f3 ChIP R 1700067P10Ri Forward (5'-3') CACCAGGAGTTCGTACCAGA k F 1700067P10Ri Reverse (5'-3') TTACCCCGGAAGCCTGAATC k R Sp3 E2f3 ChIP Forward (5'-3') ATTCAGACCTTTGTGCCTGG F Sp3 E2f3 ChIP Reverse (5'-3') AGAGAAGTACATGGCTCTCAGA R Zfp36l2 E2f3 Forward (5'-3') AAAATCGGACCTTCAGCAGC ChIP F Zfp36l2 E2f3 Reverse (5'-3') GTTGGCTGACTGGTTGATGG ChIP R His2h4 E2f3 Forward (5'-3') TGAGAGGATCGGACCAGATTG ChIP F His2h4 E2f3 Reverse (5'-3') CGGTAGAGGGCAGAGAAGAG ChIP R Nhlrc2 E2f3 Forward (5'-3') CCTCCGAGGGTTGCTTTAC ChIP F Nhlrc2 E2f3 Reverse (5'-3') CTGCCCATTTGTGTGTCCC ChIP R

continued 75

Table 2.1 continued

Lamc1 Myc Forward (5'-3') TGCTTTATTCTCTTCTGATGGGT ChIP F Lamc1 Myc Reverse (5'-3') CCCAGACTATTATACCCAGCCA ChIP R Rfxap Myc Forward (5'-3') TAGCCCTGGGAAATGGAAGG ChIP F Rfxap Myc Reverse (5'-3') GACGGCAACTCAGTAAGCAG ChIP R Eif2b3 Myc Forward (5'-3') GTCTTTCCTCTGTTGCCTGC ChIP F Eif2b3 Myc Reverse (5'-3') CTGGAAAGTAAGGCCTCGGA ChIP R Ryr3 Myc ChIP Forward (5'-3') TGTCCTGGGTAGTCTTTGCC F Ryr3 Myc ChIP Reverse (5'-3') GCTGACCTGTTCTTTAGCCAG R Lingo2 Myc Forward (5'-3') CACGTGTGCTTTTCCATTGT ChIP F Lingo2 Myc Reverse (5'-3') GGGACGGAAAGTGAGACCTT ChIP R Tpm3 Myc Forward (5'-3') TGGTGGAGAGAATTGGCCAT ChIP F Tpm3 Myc Reverse (5'-3') ACAGACAGACAGACACACCC ChIP R 4931428L18Rik Forward (5'-3') ACATTTTAAGAGCTCCTGTGGA Myc ChIP F 4931428L18Rik Reverse (5'-3') AGGCTAGAAGACGTGTGTTGA Myc ChIP R Hoxb8 Myc Forward (5'-3') GAGGAATTCAATGCCGTGTG ChIP F Hoxb8 Myc Reverse (5'-3') GAAAGTTTCCCCTCACTGTTTG ChIP R Sf3b2 Myc ACTTTAGCACACTGATCAAAGC Forward (5'-3') ChIP F T Sf3b2 Myc Reverse (5'-3') AGGTGTCCTTGTACCTCCTT ChIP R Wnk4 Myc Forward (5'-3') TGGGACATAGCTTGGGGAC ChIP F

continued

Table 2.1 continued

Wnk4 Myc Reverse (5'-3') TAGAGGGAGGGGCTTCAGAA ChIP R Kcmf1 Myc Forward (5'-3') TTAACCTTGCTGAGCCTTCG ChIP F Kcmf1 Myc Reverse (5'-3') GCAGTACCTTGGACCCTAGT ChIP R Ebf2 Myc ChIP Forward (5'-3') GAGTCTTTCCCCTCTCTCCG F Ebf2 Myc ChIP Reverse (5'-3') GAAAGCAGTCTCAGAGGGGT R Pou3f4 Myc Forward (5'-3') GGATTACCGTGGACTAAAGCT ChIP F Pou3f4 Myc Reverse (5'-3') GAGCTTTCAGGTGTGCTGTT ChIP R Cox7c Myc Forward (5'-3') AGGTGAATTGCATTTTCTGATGT ChIP F Cox7c Myc Reverse (5'-3') AGGTTAGTGAAATTTTGCTTGCA ChIP R Rap2a Myc Forward (5'-3') TGGTCCTGGGAAAATGGAATG ChIP F Rap2a Myc Reverse (5'-3') ACGTCTCAGGTTCAGGTTCT ChIP R Gja5 Myc ChIP Forward (5'-3') AACACACACTCAACACGCAT F Gja5 Myc ChIP AGTGAGTGTGTATGTGTGAGTC Reverse (5'-3') R T LOC620079 Forward (5'-3') TGTTTGCCCCTATGTATCCAAA Myc ChIP F LOC620079 Reverse (5'-3') GGGAGAGGAAGGTGGATCAT Myc ChIP R Zfp353 Myc Forward (5'-3') GTCTCATCACCTCCCAACCA ChIP F Zfp353 Myc Reverse (5'-3') GCTGTCTGATGTGTGCTGTG ChIP R Arx Myc ChIP Forward (5'-3') TGCAACAGCTTTAAAAGTCACT F Arx Myc ChIP Reverse (5'-3') CTCCATTACCTTCTCCTCCCC R

Chapter 3 Discovery of stromal regulatory networks that suppress

cancer cell proliferation

3.1 Introduction

Tumor stroma includes fibroblasts, adipocytes, immune and endothelial cells housed within an extra-cellular matrix rich in growth factors, cytokines and enzymatic activities. Cancer-associated fibroblasts (CAFs) represent a major stromal component that is associated with poor survival of cancer patients (reviewed in (Paulsson and Micke,

2014)). Classic xenograft experiments demonstrated that co-injection of CAFs enhanced the initiating capacity and aggressiveness of prostate tumor cells (Olumi et al., 1999).

Experiments in the last decade suggest that increased PI3K or decreased p53, TGF- and

Notch signaling in stromal fibroblasts collaborates with genetic alterations in tumor cells to incite a number of molecular, cellular and histopathological changes associated with malignant progression of cancer (Addadi et al., 2010; Bhowmick et al., 2004a; Bronisz et al., 2012; Carstens et al., 2014; Hill et al., 2005; Hu et al., 2012; Pickard et al., 2012;

Trimboli et al., 2009). Despite accumulating clinical and experimental evidence highlighting the importance of the tumor stroma in promoting cancer initiation, progression and therapeutic resistance, the identification of stromal factors that influence the critical hallmarks of tumor cell behavior remains a major challenge. This is in part

78 due to the inherent heterogeneity in cellular composition and signaling within the tumor microenvironment (Hu et al., 2005), but also due to the absence of available genome- wide systems to systematically query and identify relevant stromal signaling networks.

The mechanism by which stromal cells acquire an activated phenotype is believed to follow a multi-step process that presumably evolves in parallel with the increasing mutation burden suffered by cancer cells. Because this process may take place over months, years or decades, the stroma-tumor communication is likely to unfold within the context of a developing and/or aging organism and is thus unlikely to be fully recapitulated by standard, static co-culture systems. We exploited C. elegans vulval development and RNAi technology as an in vivo system to identify signaling pathways in mesenchymal cells (stroma) used to suppress the proliferation of adjacent, genetically sensitized epithelial cells (tumor).

During vulval organogenesis, mesenchymal tissues play an inductive role by releasing EGF ligand (LIN-3) from the gonad-derived anchor cell, which binds and activates the EGF receptor (LET-23) in adjacent epithelial vulval precursor cells (VPCs)

(Hill and Sternberg, 1992). The nearest VPC (P6.p) receives more of the EGF ligand and thus exhibits the strongest activation of Ras signaling and adopts a primary (1o) vulval cell fate. The P6.p cell transmits lateral Notch signals to the two adjacent VPCs (P5.p and

P7.p), which then adopt a secondary (2o) vulval cell fate (Berset et al., 2001). Additional

Wnt signaling from mesenchymal cells provides spatial orientation and proliferative queues that instruct subsequent cell divisions leading to the 22-cell adult vulva

(Eisenmann et al., 1998). Disruption of Ras, Notch, or Wnt signaling alters patterning of

VPCs and leads to either decreased (vulvaless) or increased (multivulva) VPC proliferation (reviewed in (Sternberg, 2005)).

To better reflect stroma-tumor interactions, two key genetic modifications were introduced into worms that allowed stromal cells to be specifically targeted by RNAi and epithelial cells to be sensitized with cancer-relevant mutations. Using this system we performed a genome-wide screen and identified 39 stromal factors in evolutionarily conserved and unanticipated pathways that can suppress the proliferation of adjacent Ras- sensitized but not Wnt-sensitized epithelial cells. An expression signature of the corresponding 33 human orthologs was specifically represented in tumor stroma of breast cancer patients. Depletion of these factors from normal stromal breast fibroblasts enhanced the proliferation of co-cultured breast cancer cells. In summary, this cross- species approach provides genetic evidence for the utilization of conserved pathways in the co-evolution of tumor and stroma in human cancer.

3.2 Results

3.2.1 Generation of stroma-specific RNAi and cancer-sensitized C. elegans

In C. elegans, RNAi precursor molecules freely mobilize across tissues (Jose and

Hunter, 2007). We thus developed strains that enabled the targeting of RNAi specifically to mesenchymal tissues without collateral impact on neighboring VPCs (Figure 3.1a). We took advantage of RNAi-resistant animals carrying a loss-of-function mutation in rde-1

(RNAi-Defective-1, rde-1(-)), a member of the conserved piwi/sting/argonaute/zwille/EIF2C gene family (Qadota et al., 2007; Tabara et al., 1999).

RNAi competency was then reinstated selectively back in the mesenchymal tissues surrounding VPCs (anchor cell, somatic gonad and muscle) by re-expressing wild-type rde-1 using well-characterized tissue-specific promoters. Anchor cell-, somatic gonad- and muscle-specific expression was confirmed by generating ACEL::∆pes-10-gfp, ddr-2- gfp and myo-3-gfp transgenic worms, respectively, and visualizing GFP expression

(Figure 3.1b). In addition, a loss-of-function mutation in rrf-3 (rrf-3(-)) was introduced into worms containing all three tissue-specific rde-1 transgenes to enhance general RNAi sensitivity (stroma-rde; Figure 3.2a) (Simmer et al., 2002).

Mesenchymal-specific RNAi sensitivity in stroma-rde worms was confirmed by assessing the consequences of RNAi against lin-3 (epidermal growth factor-like ligand) and lin-39 (homeobox transcription factor), two genes critical for vulval induction known to function in the anchor cell and VPCs, respectively. Control RNAi-resistant rde-1(-) animals were unaffected by lin-3(RNAi) and lin-39(RNAi), whereas RNAi-competent rde-

1(+) animals exhibited a vulvaless phenotype in response to both treatments (Figure

3.1c). In stroma-rde animals lin-3(RNAi) but not lin-39(RNAi) resulted in a vulvaless phenotype (Figure 3.2b). In addition, we also generated rde-1(-) worms expressing wild- type rde-1 from a VPC-specific promoter (pB253, Figure 3.1b). In this case, lin-39(RNAi) but not lin-3(RNAi) resulted in a vulvaless phenotype (Figure 3.1c). Collectively, these results demonstrated the successful deployment of stroma-specific RNAi in C. elegans.

Next, we introduced a cancer-relevant mutation that would impart VPCs with tumor cell-like behavior and ‘sensitize’ them to undergo ectopic cell divisions upon additional stromal signals. Human Ras GTPase activating protein (GAP) genes are

81 frequently lost in human cancer (NF1, DAB2IP, RASAL2) (Holzel et al., 2010;

McLaughlin et al., 2013; Min et al., 2010). In C. elegans, GAP-1 suppresses Ras activity and a loss-of-function allele (gap-1(-)) leads to ectopic Ras signaling in 2o VPCs, as measured by the egl-17p::cfp-lacZ in vivo reporter (Figure 3.1d) (Yoo et al., 2004). The level of Ras activation attained in gap-1(-) worms, however, is insufficient to promote ectopic VPC divisions and consequently, the vulva appears anatomically normal (Hajnal et al., 1997; Hopper et al., 2000). Importantly, the gap-1(-) allele can collaborate with additional oncogenic mutations to induce excessive VPC divisions and a multivulva phenotype (Hopper et al., 2000; Yoo et al., 2004). These observations prompted us to consider gap-1(-) as a cancer-relevant ‘sensitizing’ mutation. We thus intercrossed gap-

1(-) and stroma-rde worms to generate the final strain to be used in the genome-wide

RNAi screen. In this system, we hypothesized that the gap-1(-) mutation represents the

‘first hit’ and that RNAi mediated depletion of key stromal factors would represent the

‘second hit’ necessary to elicit a multivulva phenotype.

3.2.2 Genome-wide screen in C. elegans identifies stromal factors that suppress epithelial cell proliferation

The Ahringer E. coli RNAi library containing 16,757 clones, which accounts for approximately 86% of the C. elegans genome (Kamath et al., 2003), was used to systematically feed stroma-rde;gap-1(-) larvae on a gene-by-gene basis (Figure 3.3a).

Morphological vulval abnormalities that can be visualized by standard light microscopy

(ventral protrusions and/or egg-laying defects in adults) were used as a surrogate primary

82 end-point for altered VPC proliferation. RNAi resulting in overt vulval defects in at least two of three independent replicates were classified as candidates (Figure 3.3a). The testing of over 16,000 C. elegans genes identified 483 stromal candidates that when depleted by stroma-specific RNAi, led to vulval defects. As a secondary end-point, larvae fed with each of the 483 candidate RNAi were evaluated at the cellular level in the fourth larval stage by differential interference contrast (DIC) microscopy to discriminate ectopic

VPC divisions from other tissue alterations that could lead to vulval protrusions. This secondary screen identified 39 stromal products that when depleted, promoted ectopic

VPC divisions (Figure 3.3b and 3.3c).

Several conclusions may be drawn from the identity of these 39 genes. First, 31 of the 39 (79%) genes have 33 mammalian orthologs (his-72 and tlk-1 have two orthologs each), whereas only 47% of all genes in the RNAi library have a mammalian ortholog

(Figure 3.2c). This level of enrichment for mammalian orthologs was highly significant

(p<0.0001), suggesting that stromal signaling networks revealed by this screen are conserved from nematodes to humans. Second, the vast majority of the 39 candidates were not identified by previous whole body RNAi screens using non-sensitized C. elegans, highlighting the stroma-specific nature of our screen and suggesting novel aspects of cell proliferation control that may be unique to cancer cells. Third, none of the

39 genes encode secreted factors. We interpret this observation to mean that depletion of any single stromal secreted factor is insufficient to significantly increase VPC proliferation; rather, altering regulatory hubs that simultaneously control an entire cohort of factors is more likely to have a stronger impact. Fourth, there is a striking enrichment

83 for factors that control chromatin remodeling, cytoplasmic polyadenylation and translation (Figure 3.3c). Stromal factors related to chromatin remodeling include orthologs to a histone H3 variant (H3F3A, H3F3B), three additional histones

(HIST2H2BF, HIST2H2AB and HIST1H4H), tousled-like kinase that phosphorylates histone H3 (TLK1, TLK2) (Carrera et al., 2003), and key chromatin binding factors

(RBBP4, PAF1 and RUVBL2). We also identified symplekin (SYMPK) and cytoplasmic polyadenylation element binding protein (CPEB1), two components of a five-protein complex involved in cytoplasmic polyA-tail cleavage and re-polyadenylation of specific subsets of mRNAs (D'Ambrogio et al., 2013). Five stromal factors that are part of the protein translation machinery were also identified (NCBP2, EIF4E, MDN1, TSR2 and

PLRG1). The convergence of stromal hits on a few selected processes is remarkable, involving multiple components of the same protein complex or with the same general function.

3.2.3 Spatial and signaling specificity of stromal gene functions

To determine whether the function of the 39 genes is restricted to the stromal compartment, we compared the effects of depleting them in stroma-rde;gap-1(-) versus vulva-rde;gap-1(-) worm strains. Depletion of 20 gene products (51%) in the VPC compartment induced excessive cell divisions and a multivulva phenotype (Figure 3.4a), which included 2 genes (lin-31 and lin-1) with established roles in VPCs during development (Tan et al., 1998). These findings strongly suggest that the other 19 genes function exclusively in stromal cells to suppress ectopic VPC divisions. Intriguingly, the

19 stroma-specific genes are enriched for functions related to cytoplasmic polyadenylation (cpb-3 and symk-1) and translation control (ncbp-2, ife-5, F55F10.1, plrg-1), but not chromatin remodeling. This analysis stratifies factors that function exclusively in stromal cells from factors that have more ubiquitous cell autonomous and non-autonomous roles in the control of cell fate and proliferation decisions.

To discriminate the role in controlling ‘cancer-like’ versus ‘normal’ cell divisions, the 39 genes were depleted in stroma-rde animals with an intact gap-1 allele. Depletion of only 2 genes (lin-31 and lin-1; Figure 3.4a) led to ectopic VPC divisions. These results show that depletion of the vast majority of identified stromal factors leads to excessive

VPC proliferation only in worms sensitized with a cancer-relevant mutation but not in the context of normal development. We then tested whether these stromal factors collaborate with other cancer-relevant pathways beyond Ras. To this end, the 39 genes were depleted in stroma-rde worms containing an axl-1(-) mutation. The human orthologs of axl-1,

AXIN1 and AXIN2 are part of the complex that targets β-catenin for degradation and are lost in a significant number of colorectal and liver cancers (Figure 3.4b) (Liu et al., 2000;

Satoh et al., 2000). Loss of axl-1 in worms activates Wnt signaling in VPCs but is insufficient to cause ectopic VPC divisions. When combined with other oncogenic mutations, axl-1(-) leads to a multivulva phenotype (Oosterveen et al., 2007). In this

WNT-sensitized background, stromal depletion of only 4 of the 39 genes, including lin-

31and lin-1, resulted in a multivulva phenotype (Figure 3.4a). This finding demonstrates that depletion of individual stromal genes is unlikely to simply create a generally permissive microenvironment for cell proliferation. Rather, we suggest that their

85 depletion from mesenchymal tissues generates productive signals that cooperate with defined oncogenic pathways in neighboring epithelial VPCs.

3.2.4 Stromal factors suppress ectopic Ras signaling in sensitized VPCs

We then investigated whether depletion of each of the 39 stromal factors promotes excessive VPC proliferation through a common underlying mechanism. To test whether stromal depletion of each factor might augment Ras activity beyond the levels attained in gap-1(-) sensitized worms, the Ras signaling reporter (egl-17p::cfp-lacZ) was used to quantify Ras/MAPK signaling in response to RNAi. As shown in Figure 3.5a, control gap-1(+) larvae displayed reporter activity that was restricted to the 1o VPC lineage. Depletion of most stromal genes in this non-sensitized background led to ectopic

Ras activity in presumptive 2o VPC lineages (Figure 3.5b). In the gap-1(-) strain approximately 60% of larvae exhibited low levels of reporter activity in 2o VPC lineages.

Depletion of most stromal genes in this sensitized background led to a further increase in the frequency and intensity of reporter activity in presumptive 2o and 3o VPC lineages

(Figure 3.5c). These findings suggest that depletion of the 39 genes may promote ectopic division of gap-1(-) vulval cells through synergistic activation of Ras signaling.

3.2.5 The C. elegans-derived expression signature is associated with breast cancer stroma

To directly explore the relevance of our findings to human cancer, we utilized the

C. elegans-derived 33-human ortholog signature to query The Cancer Genome Atlas

(TCGA) whole tumor RNA sequencing data sets from breast invasive carcinoma

(BRCA), head and neck squamous cell carcinoma (HNSC), prostate adenocarcinoma

(PRAD), colon adenocarcinoma (COAD), uterine corpus endometrial carcinoma

(UCEC), and pancreatic adenocarcinoma (PAAD). Only the BRCA dataset demonstrated significant differential expression of the gene signature between normal and tumor tissues

(p=0.005; Figure 3.6a and Figure 3.7a). The expression of eight of the 33 orthologs was decreased in tumors. Interrogation of the Kaplan-Meier plotter database (Gyorffy et al.,

2010) with this eight gene signature showed that low expression was associated with earlier recurrence (p=2×10-6; Figure 3.7b).

Because TCGA data sets are derived from whole tumor samples, which include a variable amount of stromal and tumor tissue, we utilized the same 33-ortholog signature to query stroma-specific expression profiles derived from breast cancer patients. To this end, we laser capture microdissected (LCM) tumor-associated stroma (n=80) and adjacent normal stroma (n=11; at least 2 cm distant from tumor) from resected tumor tissue of breast cancer patients (Figure 3.6b). RNA was then isolated from laser-captured samples and expression evaluated using an Agilent platform (Finak et al., 2008; Trimboli et al., 2009). The heatmap shown in Figure 3.6c represents the expression of the 33- ortholog signature in normal stroma and tumor stroma samples ranked along the x-axis according to the rank-sum across genes (see Methods; the complete analysis of the expression data will be published elsewhere). Compared to 10,000 permutations of 33 randomly selected genes, the C. elegans-derived 33-ortholog signature could effectively distinguish human tumor stroma from adjacent normal stroma (p=0.0362; Figure 3.6c). In

87 contrast, a parallel analysis of expression profiles derived from LCM epithelium compartment from the same breast cancer patients (n=80 tumor epithelium; n=11 normal epithelium) failed to distinguish tumor epithelium from normal epithelium (p=0.4067;

Figure 3.8). In summary, the C. elegans-derived expression signature is selectively represented in human breast tumor stroma.

3.2.6 Depletion of human orthologs from stromal fibroblasts enhances tumor cell proliferation

Fibroblasts represent a major cellular component of breast tumor stroma.

Therefore, we isolated primary breast fibroblasts from normal tissue (10 cm distant from the tumor) of breast cancer patients (n=44 see Methods). Expression of the human orthologs in these ‘normal’ fibroblasts was confirmed using NanoString assays. A representative fibroblast cell line was immortalized and depleted for each ortholog individually by shRNA lentiviral transduction; expression of 22 orthologs was effectively reduced (Figure 3.9a). We then assessed the consequence of their depletion on the proliferation of normal breast epithelial cells (MCF10A) and a panel of well- characterized breast cancer cells (MDA-MB-468, MDA-MB-231, BT474, MCF-7, and

T47D) using an established 3-dimensional (3D) co-culture assay (Figure 3.9b). In these assays, shRNA-expressing breast fibroblasts were co-embedded with normal or breast cancer cells expressing DsRed within a laminin-rich basement membrane extract in a 96- well plate format (Sasser et al., 2007a; Sasser et al., 2007b; Studebaker et al., 2008). The proliferation of MDA-MB-468, T47D and MCF7 cells was significantly enhanced by

88 fibroblasts depleted for most stromal genes, when compared to fibroblasts treated with control shRNA (p<0.05, Figure 3.10a, b). The other two breast cancer cell lines (MDA-

MB-231 and BT474) were less responsive and normal MCF10A cells were completely non-responsive to co-cultured depleted fibroblasts. These results demonstrate the selective effect of gene-depleted fibroblasts on cancer cells versus normal cells.

Collectively, the data suggests that the core pathways used in stroma-epithelium communication are conserved from worms to humans and can be co-opted during development of human cancer.

3.3 Discussion

The signaling alterations in the tumor stroma remain largely unknown but are believed to fuel progression of incipient cancers (reviewed in (Hanahan and Weinberg,

2011)) and contribute to therapeutic resistance (Nakasone et al., 2012; Straussman et al.,

2012). The mechanisms for how stromal cells impact the development of cancer are undoubtedly complex and likely involve dysregulation of tumor-promoting and tumor- inhibiting activities. Here, we exploited C. elegans vulval organogenesis and developed tissue-specific RNAi tools to systematically screen and identify stromal factors of mesenchymal origin that suppress the proliferation of epithelial cells sensitized with a cancer-relevant mutation. Using this approach we identified specialized forms of transcription, mRNA processing and translation control as key stromal processes that influence the proliferation of tumor cells via cell non-autonomous mechanisms. Whereas stromal factors involved in chromatin dynamics appear to exert their role from both

89 stromal and epithelial compartments, factors involved in cytoplasmic polyadenylation and translation control act exclusively in stromal cells. Cytoplasmic polyA-tail cleavage and re-polyadenylation is a strategy used by neurons to accumulate, silence and acutely reactivate the expression of select sets of dendritic mRNAs in a temporal and spatial dependent manner to transmit signals at synapses between connecting neurons (Huang et al., 2003). We suggest that these macromolecular machineries, conserved from C. elegans to humans, represent a major axis adopted by mesenchymal cells to communicate with their epithelial neighbors and limit their proliferative capacity.

This work revealed striking specificity in how stromal factors interact with signaling pathways in epithelial cells in order to foster their proliferation. This was best highlighted by the observation that depletion of stromal factors, which were identified in worms with a gap-1(-) background, failed to stimulate the proliferation of axl-1(-)- sensitized VPCs or non-sensitized VPCs. This specificity may involve, at least in part, the re-engagement of the Ras pathway to overcome the threshold levels of activation necessary for ectopic cell divisions. We also noted a surprising representation of the C. elegans-derived 33-human ortholog signature specifically in breast cancer, which may reflect the preponderance of alterations in the gap-1 ortholog, RASAL2, or reflect a more complex mutation profile related to Ras signaling (McLaughlin et al., 2013). This was corroborated by the impact of depleted fibroblasts on the proliferation of a panel of breast cancer cells and the absence of an effect on normal breast epithelial cells.

Previous work using mouse models of prostate cancer suggested that tumor cells can drive stromal changes that benefit cancer progression, fueling the concept of stroma-

90 tumor co-evolution (Hill et al., 2005). Our findings provide genetic evidence that extends this co-evolution concept, by showing that productive stroma-tumor interactions are predicated by the specific genetic alteration present in epithelial cells. Whether this communication continues to evolve as multi-clonal cancer cell populations expand and metastasize remains to be determined. We propose that the generation of C. elegans strains with a broader panel of cancer-relevant mutations, beyond the Ras pathway alteration described here, will provide an in vivo platform that may be exploited to fully expose stromal signaling cascades that contribute to human cancer (Siddiqui et al., 2008).

While the details may differ between worms and humans, the signaling axes utilized for stroma-tumor communication are likely to be conserved. This unbiased genetic approach may open a rich source of opportunities to develop combinatorial therapies that include the targeting of novel ‘non-essential’ stromal pathways in the tumor microenvironment.

3.4 Methods

3.4.1 C. elegans strains

All C. elegans strains were maintained under standard conditions at 20ºC and listed below: N2 wild type, AH12 gap-1(ga133), KN611 axl-1(tm1095), rde-1(ne219), NL2099 rrf-3(pk1426), CM1859 unc-119(e2498); guEx1252[myo-3p::gfp; unc-119(+)], CM1866 unc-119(e2498); guEx1259[ddr-2p::gfp; unc-119(+)], CM1869 unc-119(e2498); guEx1261[ACEL-∆pes-10::gfp; unc-119(+)], CM1871 unc-119(e2498); guEx1263[pB253(lin-31p)::gfp; unc-119(+)], CM2073 rde-1(ne219); rrf-3(pk1426); guIs37[myo-3p::rde-1(+); ddr-2p::rde-1(+); ACEL-Δpes-10::rde-1(+); sur-5p::gfp; unc-

119(+)], CM2129 rde-1(ne219); rrf-3(pk1426); gap-1(ga133); guIs37[myo-3p::rde-1(+); ddr-2p::rde-1(+); ACEL-Δpes-10::rde-1(+); sur-5p::gfp; unc-119(+), GS3582 unc-

4(e120); arIs92[egl-17::NLS-cfp-lacZ; unc-4(+); ttx-3::gfp], CM2307 unc-4(e120); gap-

1(ga133); arIs92[egl-17::NLS-cfp-lacZ; unc-4(+); ttx-3::gfp], CM2423 unc-119(e2498); rde-1(ne219); rrf-3(pk1426); gap-1(ga133); guIs39[lin-31p::rde-1(+); sur-5p::gfp; unc-

119(+)].

3.4.2 C. elegans tissue-specific RNAi constructs

First-strand cDNA was synthesized using total RNA isolated from N2 wild type animals with Superscript II Reverse Transcriptase Kit (Invitrogen) following the manufacturer’s protocol. The 5´ (~1 kb) and 3´ fragments (~2 kb) of the full-length rde-1(+) cDNA were individually amplified by PCR using Pfu DNA Polymerase Kit (Stratagene). The PCR products were digested with restriction enzymes and sequentially inserted into the pBluescript II KS(+) phagemid vector (Agilent Technologies; 212207). The full-length rde-1(+) cDNA was then excised using KpnI and SacI and cloned into the pPD49.26 C. elegans expression vector from the Fire Lab Kit (Addgene). All constructs were sequence-verified to ensure no mutations were introduced into cDNA sequences. Tissue- specific promoters were chosen based on previous studies (myo-3 (Myers and

Greenwald); ACEL (Hwang and Sternberg)) or identified by examining approximately forty reporter strains with GFP expression detected in the somatic gonad (ddr-2; http://gfpweb.aecom.yu.edu/index). The regulatory sequences were amplified by PCR using Pfu DNA Polymerase Kit with N2 genomic DNA as templates. The PCR products

(myo-3 and ddr-2) were digested with restriction enzymes and cloned into pPD95.69 vector upstream of the nls::gfp reporter cassette. ACEL sequences were cloned into pPD107.94 vector upstream of the truncated ∆pes-10 minimal promoter, and the fusion

ACEL-∆pes-10 promoter was cloned into pPD95.69 vector upstream of the nls::gfp reporter cassette. All resultant promoter::gfp reporters were utilized to generate transgenic strains to verify the expression pattern of corresponding promoters. Promoter sequences were then individually cloned into pPD49.26 containing rde-1(+) cDNA to drive tissue-specific rde-1(+) cDNA expression, and vectors were micro-injected into the gonads of adult hermaphrodite animals using standard methods (Mello et al.) with unc-

119 as the selection marker. F1 progeny without uncoordinated movements were selected. Individual F2 worms were isolated to establish independent lines.

Extrachromosomal transgene arrays were integrated into chromosomes with γ-irradiation at the Ohio State University Reactor Lab as previously described

(http://wormlab.caltech.edu/protocols/). Identical methods were used to achieve rde-1(+) cDNA expression restricted to vulval cells (vulva-rde; gap-1(-) strain) by cloning rde-1 cDNA into the NotI site in pB253 vector (Myers and Greenwald; Tan et al.).

3.4.3 C. elegans RNAi screen and microscopy

The genome-wide screen utilized the Ahringer E. coli RNAi library (Source Bioscience

LifeSciences) targeting 16,757 C. elegans genes (~86% of genome) and was performed by feeding the CM2129 strain (see above) individual E. coli clones expressing double- stranded RNA. RNAi clones from 384-well library plates were inoculated into 96-well

93 deep-well plates (VWR) with Luria broth (LB) medium (Invitrogen) containing 50

µg/mL carbenicillin (Sigma-Aldrich) for 16 hours at 37ºC. E. coli were then seeded onto

12-well plates (Corning) with standard nematode growth medium (NGM) agar but containing 2× normal peptone, and containing 1 mM IPTG (Lab Scientific) and 25

µg/mL carbenicillin and incubated for 24 hours at room temperature. Synchronized embryos were seeded onto these plates and incubated for 4 days at 20ºC. For the primary

RNAi screen, approximately 50 animals were scored for abnormal vulva-associated morphologies using an Olympus SZ60 dissection microscope. Each RNAi clone was tested in three independent replicates, and genes that promoted aberrant morphologies upon stromal RNAi in at least two of three replicates were considered primary candidate

‘hits’. All primary candidate ‘hits’ were individually re-evaluated using the same 12-well plate RNAi platform, but plates were incubated at 20°C for 48 hours and L4 stage worms were selected for imaging. DIC images were captured with a Spot RT Monochrome digital camera (Diagnostic Instruments) using a Zeiss Axioskop2 microscope to identify the final 39 stromal candidate genes whose depletion induced multivulva. RNAi clones were sequence-verified to ensure the identity of each target gene.

3.4.4 Vulval Ras/MAPK signaling assay

The GS3582 strain expressing an egl-17p::LacZ-CFP fluorescent reporter transgene

(Yoo et al.) was used to determine how whole-worm RNAi-mediated depletion of the 39 identified genes affects Ras/MAPK signaling in vulval cells. The same 12-well plate

RNAi platform described above was utilized to evaluate synchronized early L3 worms at

94 one or two rounds of VPC division following RNAi. DIC images were captured with a

Hamamatsu ORCA-ER digital camera using a Zeiss Axioplan2 microscrope. The same system was used to capture corresponding CFP fluorescent images with one-second exposures.

3.4.5 Breast tumor stroma total RNA expression profiling

All tissue specimens and associated clinical data were collected at McGill University

Health Center (Montreal, Canada) between 1999 and 2013 in accordance with protocols approved by the research ethics committee. Written consent was obtained on an individual basis for all patients participating in this study. Tissue samples were collected within 30 minutes after surgery, embedded in TissueTek OCT (Somagen, Edmonton,

Alberta, Canada) and stored in liquid nitrogen until use. Frozen specimens were cryo- sectioned (10 μm thickness) and stained using the Arcturus HistoGene LCM Frozen

Section Staining Kit (Life Technologies) before LCM was performed. All tissues included in this study were re-examined by a clinical pathologist dedicated to the project.

Tissue specimens were laser capture microdissected using a PixCell IIe LCM system

(Arcturus) within three hours following tissue staining. Total RNA was extracted using the PicoPure RNA isolation kit (Life Technologies) as per the manufacturer's directions.

Subsequently, 0.5 to 1 ng of total RNA was subjected to two rounds of T7 linear amplification using the Arcturus RiboAmp HS Plus kit (Life Technologies) and labeled with Cy3 dye according to the manufacturer's procedure. Prior to microarray hybridizations, amplified products were quantified using a spectrophotometer (Nanodrop,

Wilmington, Delaware, USA) and subjected to BioAnalyzer assay for quality control

(Agilent Technologies, Santa Clara, California, USA). SurePrint G3 Human Gene

Expression 8x60K v1 microarrays (Agilent Technologies, product G4112A) were used for all experiments. RNA samples were subjected to fragmentation followed by 18 h hybridization, washing, and scanning (Agilent Technologies, model G2505B) according to the manufacturer's protocol. Samples were hybridized against Cy5-labelled Universal

Human Reference RNA (Stratagene, Jolla, California, USA).

3.4.6 Human cell culture

All breast cancer cells were maintained in a humidified incubator at 37°C and 5% CO2 in

RPMI 1640 growth medium (Life Technologies) containing 10% fetal bovine serum

(Thermo Scientific HyClone), 2 mmol/L L-glutamine, 100 units/ml penicillin, and 100

μg/ml streptomycin (Life Technologies). All DsRed-positive human breast cancer cells were previously generated (Sasser et al.; Studebaker et al.). MCF10A breast epithelial cells were maintained in standard MCF10A growth medium: DMEM/F12 (Life

Technologies) containing 5% heat-inactivated horse serum (Life Technologies), 20 ng/ml recombinant human EGF (PeproTech), 0.5 μg/ml hydrocortisone (Sigma-Aldrich), 100 ng/ml cholera toxin (Sigma-Aldrich), 10 g/ml insulin (Sigma-Aldrich), 100 units/ml penicillin, and 100 μg/ml streptomycin (Life Technologies). MCF10A cells were infected with lentivirus expressing pLVX-DsRed-Express2-C1 and DsRed-positive cells were enriched by flow cytometry based on DsRed fluorescence. Primary human breast fibroblasts were isolated from human breast tissue as previously described (Orimo et al.)

96 with minor modifications. Tissue biopsies from grossly normal human breast tissue specimens >10 cm from breast tumors were received on ice, minced into small fragments, and digested in dissociation buffer comprised of DMEM growth medium (Life

Technologies) containing 10% fetal bovine serum (Thermo Scientific HyClone), 100 units/ml penicillin, 100 μg/ml streptomycin (Life Technologies), and 2 mg/ml collagenase III (Worthington Biochemical Corp.) for sixteen to eighteen hours at 37°C.

Digested tissue was allowed to separate by gravity to remove tissue debris, and supernatants were collected to pellet fibroblasts. The fibroblast pellets were washed and cultured in DMEM containing 10% fetal bovine serum and 25 μg/ml plasmocin

(InvivoGen) for one week. Fibroblasts were subsequently maintained in DMEM growth medium containing 10% fetal bovine serum, 100 units/ml penicillin, 100 μg/ml streptomycin. For immortalization, 4×105 HEK293T cells per well were plated into 6- well plates overnight. The following day, subconfluent HEK293T cells were co- transfected with pCL-10A1 plasmid (Imgenex), which encodes HIV-1 Gag, Pol, Tet, and

Rev proteins and pBABE-hygro-hTERT plasmid (Addgene plasmid 1773) using jetPEI

DNA transfection reagent (Polyplus Transfection) according to the manufacturer’s protocol to generate VSV-G pseudotyped lentivirus. Viral supernatants were 0.45 μm filtered and used to infect primary human breast fibroblasts with hTERT for immortalization. Fibroblasts were selected with hygromycin for ten days to generate stable hTERT-positive, immortalized human breast fibroblasts.

3.4.7 Fibroblast mRNA quantification

Total RNA was isolated from primary human breast fibroblasts using TRIzol reagent

(Life Technologies) according to the manufacturer’s protocol followed by RNA clean up using a RNA clean-up and concentration kit (Norgenbiotek). Total RNA was submitted to NanoString Technologies for mRNA quantification of candidate genes by the nCounter® gene expression analysis system.

3.4.8 Human lentivirus production and shRNA transduction

Optimal shRNA-mediated target gene knockdown efficiency was initially determined by lentiviral transduction of HEK293T cells with three to six individual gene-specific GIPZ shRNA-miR plasmids (Thermo Scientific) per candidate gene. The single shRNA plasmid demonstrating the most efficient target gene knockdown was utilized for subsequent infection of human breast fibroblasts. 4×105 HEK293T cells per well were plated into 6-well plates overnight. The following day, subconfluent HEK293T cells were triple-transfected with the pCMVR8.91 packaging plasmid (encoding HIV-1 Gag, Pol,

Tet, and Rev proteins), pMD.G (encoding vesicular stomatitis virus (VSV)-G enevolpe protein; both plasmids kindly provided by Dr. Jose Silva), and gene-specific GIPZ shRNA-miR plasmid using jetPEI™ DNA transfection reagent (Polyplus Transfection) according to the manufacturer’s protocol to generate VSV-G pseudotyped lentivirus.

Fresh medium was added the following day, and 48 hours post-transfection, HEK293T cell medium was filtered and added to 1×105 human breast fibroblasts at 70% confluence in a 6-well plate two times over two consecutive days. 72 hours post-infection, fibroblasts

98 were monitored for GFP expression and treated with 0.75 μg/ml puromycin for three to five days to select for shRNA-positive cells. RNA was isolated as described above and complementary DNA (cDNA) was generated with SuperScript III reverse transcriptase kit reagents (Invitrogen), 200 ng total RNA, and oligo(dT)20. Real-time quantitative PCR was used to validate target gene knockdown following lentiviral shRNA transduction.

3.4.9 3-dimensional co-culture proliferation assay

A 3-dimensional (3D) breast fibroblast-cancer cell co-culture proliferation assay was performed as previously described (Sasser et al.; Sasser et al.; Studebaker et al.). This

DsRed fluorescence-based quantitative growth assay measures DsRed fluorescence, which correlates precisely with breast cancer cell number. Stable, DsRed-positive human breast cancer cells (triple-negative MDA-MB-231, MDA-MB-468; estrogen receptor

(ER)-positive MCF-7, T47D; HER2-positive BT474) or normal, immortalized mammary epithelial cells (MCF10A) and human breast fibroblasts were trypsinized from standard culture plates and co-cultured in 3 mg/ml Cultrex® basement membrane extract

(Trevigen) constituted in serum-free, phenol red-free RPMI 1640 growth medium in black-walled 96-well tissue culture plates. Co-cultures were set up in triplicate and included a scrambled shRNA fibroblast co-culture control each set up day. Plates were read daily on a SpectraMax M2 fluorescence microplate reader (Molecular Devices), and autofluorescence of basement membrane extract alone was subtracted from each well.

Data were normalized to fluorescence intensities of each co-culture on day one to adjust for initial cancer cell number.

3.4.10 Statistical analyses

For human ortholog enrichment, the C. elegans sequence IDs of the 39 identified genes were entered into WormMart (http://www.wormbase.org), selecting Sequence Name,

Homologous Protein, Homolog Species, and Homolog BLASTP evalue from the

Attributes list. The number with human orthologs (Homolog BLASTP evalue <=

0.00001) were tallied. The same process for C. elegans sequence IDs from all RNAi tested in the screen was performed. A Fisher’s exact test was used to determine if the frequency of C. elegans genes with a human ortholog differed between the two gene lists.

For TCGA and breast tumor stroma gene expression data analysis, data for the 33 human orthologs were extracted from the genomic mRNA expression data set. Genes were clustered by correlation into two groups (arbitrarily labeled as "up" and "down") using PAM (Partitioning Around Medoids). Each sample was ranked for each gene (in opposite orders for the up vs. down group), and the sum of the ranks across all genes was calculated for each sample. These per-sample rank sums induce a linear ordering and ranks on the samples (which were used to order the samples in the heatmaps). The sum of the ranks of the normal samples in the linear ordering was compared to the distribution of sum of ranks obtained from 10,000 sets of 33 randomly selected genes. The number of trials with sums more extreme (two-sided) than the sum obtained from the candidate gene list was tallied and divided by 10,000 to obtain a p-value. The p-value estimates the probability of getting a better clustering of the normal samples at the extremes from random sets of genes.

100

For relapse-free survival analysis, the gene expression fold changes of tumor relative to normal tissues were calculated using TCGA data. An index was calculated for each sample, which reflected the weighted mean expression level of the 8 orthologs from breast cancer datasets collected in Kaplan-Meier Plotter tool. The weights for each gene were defined as the log10 (fold change), which were calculated from TCGA data. All the tumor samples from Kaplan-Meier Plotter were equally divided into "high" and "low" groups based on the ranking of their index. Log-rank test was performed to determine whether these two groups of samples have significant differences in term of relapse-free survival.

For co-culture data analyses, the growth rate data were matched from gene- specific shRNA and non-silencing control shRNA within the replicates and by culturing time (days). Wilcoxon signed-rank test (one tail only) was performed to determine if the human ortholog-depleted fibroblasts significantly enhance the proliferation of co-cultured epithelial cells compared to control fibroblasts. A fold change cutoff of 1.1 at day 4 from gene-specific to control shRNA was applied. All analyses were performed with the R statistical programing language (www.R-project.org).

101

Figure 3.1 Stroma-specific RNAi in C. elegans containing a cancer-relevant mutation

(a) Differential interference contrast (DIC) images and corresponding schematic diagram of a normal vulva in wild-type C. elegans at mid-larval L4 stage. AC, anchor cell; SG, somatic gonad; M, muscle. (b) DIC and fluorescence images of transgenic worms expressing GFP driven by tissue-specific promoters active in the anchor cell (arrowhead), somatic gonad, muscle, and vulval cells. Dashed lines indicate worm body boundaries.

(c) DIC images of wild-type rde-1 worms (rde-1(+)), rde-1mutants (rde-1(-)), rde-1(-) mutants with stroma-specific RNAi (stroma-rde), and rde-1(-) mutants with VPC- specific RNAi (vulva-rde) following lin-3 and lin-39 RNAi. Arrows indicate normal vulva structures. (d) DIC and fluorescence images of wild-type gap-1 (gap-1(+)) and gap-1-mutant (gap-1(-)) worms expressing a LET-60/RAS-dependent transcriptional reporter. Cell lineages are designated as P6.px (presumptive 1o cell fate) and P5.px and

P7.px (presumptive 2o cell fate). Bottom panels are overexposed to better visualize CFP expression in P5.px and P7.px cells.

102

Figure 3.1

103

Figure 3.2 Stromal RNAi in C. elegans identifies conserved factors

(a) RNAi sensitivity was enhanced by introducing a rrf-3 mutation (rrf-3(-)). Introduction of the rrf-3(-) allele did not inadvertently alter RNAi resistance of rde-1(-) strains because rrf-3(-); rde-1(-) double mutants were viable upon RNAi-mediated depletion of two essential genes, pop-1 (TCF/LEF transcription factor) and pie-1 (CCCH zinc finger protein). Embryonic lethality was evaluated in two independent rrf-3(-); rde-1(-) double mutant strains, H and K. N2, wild type. (b) Summary of vulvaless phenotypes (Vul) after

RNAi-mediated depletion of lin-3 and lin-39 in wild type rde-1 worms (rde-1(+)), rde-1 mutants (rde-1(-)) and rde-1(-) mutants with stroma-specific RNAi (stroma-rde). (c)

Percentage of genes with mammalian orthologs for all C. elegans genes included in the

RNAi library or the 39 stromal genes identified by the RNAi screen. p=0.00007, Fisher’s exact test.

104

Figure 3.2

105

Figure 3.3 Genome-wide RNAi screen identifies conserved stromal factors that suppress ectopic epithelial cell division

(a) Experimental work flow of the genome-wide stroma-specific RNAi screen. (b)

Representative DIC images of multivulva in stroma-rde;gap-1(-) worms at mid-L4 stage following stromal RNAi. White and black arrows indicate normal and ectopic vulva structures, respectively. (c) The 39 genes identified by stroma-specific RNAi screen are grouped based on annotated functions. Chr., chromosome.

106

Figure 3.3

107

Figure 3.4 Stromal factors have spatial- and signaling-specific functions

(a) Summary of multivulva phenotypes after RNAi-mediated depletion of identified genes in vulva-rde;gap-1(-), stroma-rde;gap-1(+) and stroma-rde;axl-1(-) animals. n.d., not determined due to developmental deficiencies. (b) Schematic representation highlighting the inhibitory roles of GAP-1 and AXL-1 in EGF/RAS/MAPK and canonical WNT/β-CATENIN signaling, respectively. Chr., chromosome.

108

Figure 3.4

109

Figure 3.5 Depletion of stromal factors induces ectopic Ras signaling in VPCs

(a) Representative DIC and fluorescence images of non-sensitized (gap-1(+)) and sensitized (gap-1(-)) worms expressing a LET-60/RAS-dependent transcriptional reporter. Bottom panels are overexposed to better visualize CFP expression in P5.px,

P7.px and P8.px cells. (b) Percentage of gap-1(+) animals with ectopic vulval Ras activity only in presumptive 2o lineage cells following gene-specific or empty vector

(control) RNAi. n=44 (control); n>25 (each gene). (c) Percentage of gap-1(-) animals with ectopic vulval Ras activity in both 2o and 3o lineage cells (red bars) or only in presumptive 2o lineage cells (gray bars) following gene-specific or empty vector (control)

RNAi. CFP intensity in presumptive 2o lineage cells was scored as high, intermediate

(intermed.) and low. n=99 (control); n>25 (each gene).

110

Figure 3.5

111

Figure 3.6 The human ortholog signature distinguishes breast tumor stroma from adjacent normal stroma

(a) Differential gene expression of the 33 human orthologs in six carcinoma types compared to corresponding normal tissue was determined using TCGA whole tumor

RNA sequencing data. (b) Representative images showing pre- and post-laser capture microdissected (LCM) breast tumor stroma. Epi, epithelium; Str, stroma. (c) Heatmap displaying RNA expression of the 33 human orthologs in normal (n=11) and tumor stroma (n=80) from breast cancer patients. The boundaries (dashed lines) represent the

95% distribution limits for rank-sums of 1000 randomly generated samples, where patient samples outside these boundaries have more extreme rank-sums than expected by chance.

The rank-sum p-value indicates the ability of the gene signature to distinguish normal from tumor stroma in breast cancer patients.

112

Figure 3.6

113

Figure 3.7 The 33-ortholog signature is represented in human breast cancer

(a) Heatmap displaying differential gene expression of the 33-human orthologs in tumors compared to normal tissue in TCGA breast cancer RNA sequencing data (BRCA). The yellow line demarcates the tumor from normal samples. (b) Kaplan-Meier curves of relapse-free survival for breast cancer patients (n=3458) using the 8-ortholog signature.

HR, hazard ratio.

114

Figure 3.7

115

Figure 3.8 The human ortholog signature fails to distinguish breast tumor epithelium from adjacent normal epithelium

Heatmap displaying RNA expression of the 33 human orthologs in normal (n=11) and tumor epithelium (n=80) from breast cancer patients. Each sample was given a rank for each gene and ordered among the experimental samples according to the rank-sums. The boundaries (dashed lines) represent the 95% distribution limits for rank-sums of 1,000 randomly generated samples. The p-value indicates the ability of the gene signature to distinguish normal from tumor epithelium in breast cancer patients. Epi., epithelium.

116

Figure 3.8

117

Figure 3.9 Depletion of human orthologs in breast fibroblasts for 3D co-culture with breast cancer or normal epithelial cells

(a) 22 human orthologs were knocked down by lentirviral shRNA transduction as determined by real-time quantitative PCR analysis. Data are shown as mean percent residual gene expression ± standard deviation relative to non-silencing shRNA control and normalized with GAPDH expression. (b) Schematic representation of the 3D proliferation assay. DsRed-positive human breast epithelial cells (Epi.) were co-cultured with human breast fibroblasts (Fib.) at a 1:1 ratio in basement membrane extract (BME) for four days. Epithelial cell number correlates with DsRed fluorescence. DsRed fluorescence images of MCF-7 breast cancer cells are shown over four days.

118

Figure 3.9

119

Figure 3.10 Depletion of human orthologs in normal breast fibroblasts enhances breast cancer cell proliferation

(a) Heatmap summarizing the effects of fibroblast-specific depletion of each stromal factor on the proliferation of co-cultured normal epithelial cells (MCF10A) or a panel of breast cancer cell lines. Co-culture results are presented relative to a non-silencing control shRNA at day 4 of culturing (breast cancer cells, n=3 independent experiments with each experiment performed in triplicate; MCF10A cells, n=2 independent experiments with each experiment performed in triplicate). (b) Example growth curves of normal epithelial cells and breast cancer cells co-cultured for 4 days with fibroblasts expressing non-silencing (Fib./control) or stromal factor-specific shRNA (Fib./sh-RNAs).

Data are presented as average ±SEM.

120

Figure 3.10

121

C. elegans Cloning Primers

Restriction Sequences (5´-3´, restriction enzyme site underlined) Primer enzyme rde-1 5' ATCGGTACCATGTCCTCGAATTTTCCCGAATTG KpnI fragment F rde-1 5' CAAATCTAGAAATGTCAGATGTCGTTCT XbaI fragment R rde-1 3' CATCTGACATTTCTAGATTTGTGCGA XbaI fragment F rde-1 3' TGAGAGCTCTTATGCGAACGACATTCCAGG SacI fragment R GCAGGATCCCCATTGTAGGCAATCAGTCAAACC myo-3p F BamHI CAAGGATCCCATTTCTAGATGGATCTAGTGGTCGT myo-3p R BamHI G TATGCATGCGAATAAACAAGCCATTTCCCCC ddr-2p F SphI TATGGATCCCATGCAGCAACTTGATCCTGACA ddr-2p R BamHI AATGCATGCGTTTCTTTCCGTGGTTCCTTGAG ACEL F SphI CATGCTAGCGTCGTTCAGGAATCGATTTACCGAAT ACEL R NheI AT

Human RT-qPCR Primers

Annealing T Sequences (5´-3´) Gene (°C) CCAACACAGGATGATGAGGTTATG DDX47 F 57 GCGAGCGTTTCTTCTTTTCTCC DDX47 R

TCAAGATGTCCAGCCGAAG CPEB1 F 60 CAGAGCACCGACAAACACC CPEB1 R

GGCTGGAGAAGAAAGAGGTG SYMPK F 60 ACAGGTTGGTGGCTTTGATG SYMPK R

AATCTGAAACCTGTCCCGCT SCCPDH F 60 GCATACTGAACTGGTGATTCCT SCCPDH R

TACTGGAGGGGTGAAGAAAC H3F3A F 60 AGACGCTGGAAGGGAAG H3F3A R

CGCTCCACCATCACATCCC HIST2H2BF F 64 GCTTCTCCTTCTATGCCACTTTC HIST2H2BF R

CTGTCAAGCAGAGAGAAGAGTATGC TLK1 F 60 GTGGGAGGTTTGCGTTTGG TLK1 R

Table 3.1 Primers for C. elegans strain construction and human gene RT-qPCR

continued

122

Table 3.1 continued

GGCGGGGTGAAGAAGCC H3F3B F 60 CTCCCTCACCAACCTCTGG H3F3B R

CGCATTGTGGTCCGCTTCTC NCBP2 F 62 ATGGCGTTTTCCGCATAGC NCBP2 R

GCCCGAGTTTGCCGAGTG RBBP4 F 60 CGATGTGTGTGTCCCCAG RBBP4 R

CCCTACTCCTAATCCCCCGA EIF4E F 62 CGTTTGTTTTTCTCATCTTCCCAC EIF4E R

GTCTCCATCTCTTCCCCCAA ELK3 F 60 GGACTAAGGCTGCTCCAGAAA ELK3 R

CGCTACCAATCCTTCAACAACC PRIM1 F 60 GCAGAACTACAACATCTCCTCAC PRIM1 R

CCTCAGTGTCTGCCTCTCTC USP4 F 60 CCCAGGTTTCCAAGTCCACA USP4 R

GGCGTGGTGCTGGAGATG RUVBL2 F 62 GATGGCTGTGAATGGCGTG RUVBL2 R

CAGTTGTGGAAGACGGGAG TSR2 F 60 GCAGTGGCTGTGACCTTG TSR2 R

GCAACTGAAGGGAAAGGGAAAG POGK F 62 TGTGGGGAGGGGCTGG POGK R

CTTCTGCCGCTCCTCCTC BAG2 F 60 GTCCTGGCTATTTTGGATACTGTG BAG2 R

GCGTGGTGTTTCATCCATTCC POLR3H F 60 ATGTAGAGGTCGTGTGCTCC POLR3H R

CTTCCTCGGATCAGGCGTC PAF1 F 64 CAGAGTCCGGTGGGAATTGG PAF1 R

GGGCATCTTGGCTGGGTTC PLRG1 F 60 GCATTTCACTTGTTTGTCTTCTCC PLRG1 R

AGTTGCTTCTTTGTTATTTGACCCT HEATR1 F 60 TCGCTCCAAGGTTTTTGCT HEATR1 R

CGAGATCCCTCCAAAATCAA GAPDH F 60 TTCACACCCATGACGAACAT GAPDH R

123

Chapter 4 Discussion

Precise control of cell proliferation requires coordinated intrinsic and extrinsic regulatory inputs. For the intrinsic regulatory inputs, we have investigated how two critical intracellular proliferative programs, E2f and Myc, coordinate to ensure the proper progression of cell cycle clock and subsequent generation of the daughter cells. For the extrinsic regulatory inputs, we have utilized a cross-species approach to dissect the stromal networks that function in the mesenchymal tissue compartment to control the proliferation of adjacent epithelial cells. These studies have further complemented our understanding about how the normal cell proliferation is executed as well as that how these programs may be hijacked in the context of cancer for the malignant cell proliferation.

4.1 E2f and Myc control normal cell proliferation

In the classical model of mammalian cell cycle control, phosphorylation of Rb by cyclin-dependent kinases liberates the transcription factors E2f1, E2f2 and E2f3, which in turn regulates the transcription of corresponding target genes. These events were considered to be essential for S-phase entry and ultimate generation of daughter cells.

Similarly, Myc has been deemed as a critical positive regulator for cell proliferation. Myc

124 lies downstream of many cell division signal transduction pathways. In response to these signals, Myc dimerizes with the binding partner Max, recruits chromatin modifiers such as histone acetyl transferase and regulates transcription of target genes critical for cell growth and proliferation. In doing so, increased activities of Myc could drive cells out of quiescent status and into S phase.

It should be pointed out that the concept of E2f and Myc being essential for normal cell proliferation is largely based on the experimental evidence derived from in vitro cell culture experiments. On the other hand, some studies utilizing in vivo models revealed the virtual absence of major cell proliferation defects in the absence of individual E2f members or Myc (Bettess et al., 2005; Chen et al., 2009a; Chong et al.,

2009; Li et al., 2006; Wenzel et al., 2011). This has led to the notion that significant functional redundancy exists among either E2f or Myc family of transcription factors, in which the effect of single member ablation could be masked by the compensation from other family members (Malynn et al., 2000; Tsai et al., 2008). However, this notion has been challenged recently by several studies, which show that rather normal cell division could happen, without the whole E2f activator subfamily E2f1-3, in certain tissues of mice. This indicates contributions from distinct pathways compensate each other for cell cycle progression and a more broad redundancy among the intracellular cascades that control cell proliferation.

We show here the functional redundancy also exist in different families of critical cell proliferation regulators, E2f and Myc. Normal proliferation of progenitor cells in the crypts fails to withstand the combined loss of E2f1-3 and Myc, yet tolerates rather well to

125 the singular loss of either E2f1-3 or Myc. Underlying the cooperation between E2f1-3 and

Myc is their synergistic regulation of S-G2 cell cycle progression. The global chromatin occupancy profiling reveals that E2f3 and Myc co-bind to a large portion of target genes, including the ones critical for S-G2 progression. This co-recruitment of E2f and Myc to the targets may serve as the molecular basis for how the synergistic regulation and thus genetic redundancy is achieved. The exact identities of co-factors involved in this regulation remain to be determined, but likely include some chromatin remodeling factors illustrated by previous studies (McMahon et al., 1998; Ogawa et al., 2002). The observation that combined E2f/Myc-deficient cells arrest during S-G2 does not rule out that other signaling regulatory inputs also contribute to the S-G2 progression; rather, this highlights the exquisite sensitivity of normal proliferating cells to the E2f and Myc transcriptional programs. Removal of E2f1-3 and Myc together, even in the presence of other potential forces for cell cycle progression, results in the dysregulation of downstream targets and subsequent failure of fulfilling the proliferation.

In contrast to the requirement of E2f and Myc for S-G2 cell cycle progression, it is surprising that the crypt progenitor cells could undergo G1-S transition and enter S phase in the absence of both E2f1-3 and Myc. This observation underscores the complexity of cell cycle control and suggests other yet to be identified factors maintain the cell proliferation in parallel with E2f and Myc transcriptional programs. This is, somewhat, different from many previous studies that suggest the critical roles of E2f1-3 and Myc in driving the S phase entry. Two potential reasons may underlie this difference: first, in many of the previous studies, the functions of these transcription factors are interrogated

126 by introducing exogenous amounts of E2f1-3 and Myc. Thus while it is satisfying to see genetically "high" levels of E2f1-3 and Myc may be sufficient to promote the S phase entry, it is not equivalent to suggest that genetically "normal" levels of E2f1-3 and Myc would be absolutely necessary to promote the S phase entry. Second, it appears the sensitivity of various cells or tissues towards the loss of E2f1-3 and Myc is different. How the tissue-specific or genetic context-dependent changes may affect the roles of E2f1-3 and Myc in S phase entry would be discussed further below.

4.2 E2f and Myc control cancer-like cell proliferation

As described in this study, in sharp contrast to the normally proliferating crypt progenitor cells, Rb-deficient villus cells acutely depend on both E2f1-3 and Myc for the ectopic proliferation. Especially, loss of either E2f1-3 or Myc suppresses the abnormal S phase entry of Rb-null villus cells. While it is probably more intuitive that the absence of the downstream effectors, E2f1-3, could alleviate the rampant cell proliferation in the Rb- deficient villi, Myc appears to contribute to the ectopic cell proliferation partially through regulating E2f activities and partially through its parallel inputs with E2fs.

The study here highlights that the requirement for Myc and/or E2f’s inputs for G1-

S in different contexts (proliferation of crypt progenitor cells versus Rb-deficient villi) is different. It is likely that multiple inputs dictate G1-S progression and S phase entry in crypts, allowing a more robust and adaptable proliferative response that is not so sensitive to perturbation of any single pathway. In villi, the loss of Rb is a genetic perturbation that now exaggerates the activation of downstream pathways (Myc and E2f), and thus, these

127 cells become addicted on these downstream pathways, which cannot be compensated since other proliferative signal inputs (normally present in the crypts) are absent in villi.

It would be interesting to investigate the coordination between E2f and Myc for cell proliferation control in the small intestine would apply generally in other tissues for the normal proliferation of progenitor cells versus cancer-like proliferation of post- mitotic differentiated cells. Nonetheless, similar observations have been reached recently in the context of normal or oncogenic proliferation for some other critical cell cycle regulators. For example, normal cell proliferation is maintained in the murine breast tissues without cyclin D1 or Cdk4; however, the malignant proliferation of breast tissues driven by Ras or Neu oncogenes absolutely requires cyclin D1 and Cdk4 (Yu et al., 2001;

Yu et al., 2006). Thus it appears in many cases the cancer cells exaggerate the dependency on certain downstream factors or pathways for the proliferation.

Understanding the specific vulnerability of cancer cells and how they re-wire the intracellular programs to foster the uncontrolled proliferation might provide a specific window for unique therapeutic opportunities.

4.3 Conserved cell non-autonomous networks for cell proliferation control

Along with understanding the complexity of intracellular programs for cell proliferation control and how cancer cells may distort the homeostatic control mechanism , another major advancement in the field over the decades is the appreciation of tumor as an organ system whose complexity may be comparable or even exceed that of normal healthy tissues. Rather than just a homogenous population of cancer cells,

128 tumors, especially solid tumors, may contain various portions of stromal cells. It has been increasingly clear that stromal compartment significantly impacts the tumor histopathology, including the proliferative capacity of the tumor. Given the impact of stroma on tumor development, extensive efforts have been invested trying to understand the genetic and signaling networks in the stroma that impact tumor growth. However, due to the technical obstacles, the inquiry has been done largely with a candidate gene approach using classical mouse cancer models, where individual genes are genetically manipulated to interrogate their functionality. Consequently, the systematic identification of critical stromal factors that control tumor cell proliferation remains a major challenge.

In this study we have developed genetic models in C. elegans to identify and characterize stromal networks that have a direct impact on human cancer. Starting from the simple yet powerful model, we identified a cohort of stromal factors that impact the proliferation of epithelial cells with changes in the RAS-MAPK pathway. We revealed potential mechanism (Ras signaling enhancement) for this epithelial-mesenchymal communication. We translated the relevance of these C. elegans derived stromal factors to human cancer by showing that they are specifically represented in the breast cancer stroma. We have also validated the growth suppressive role of these factors in the breast fibroblasts. So these interactions and the programs utilized for the interactions are conserved across species as well as from developmental context into cancer. Further studies are still needed to provide further mechanistic insights for the functionality of these stromal factors. For example, what are the immediate changes upon the dysregulation of these factors in the stroma? How these changes are transmitted between

129 the stroma and the tumor compartment? How the cancer cells sense and respond to these extrinsic stimuli? Hopefully the understanding of these mechanisms for the cell-cell interactions may offer the basis for effective therapeutic approaches, similar with the efforts targeting angiogenesis and immuno-surveillance.

4.4 Genetic selectivity for cell proliferation control by stromal networks

In this study we also noticed the specificity in how stromal factors interact with signaling pathways in epithelial cells in order to regulate their proliferation. First, most of the genes work with the gap-1 mutant cells but not axl-1 or wild type vulval cells. Second, the specific representation of these factors in breast cancer might be related to the prevalence of alterations in the Ras GAPs in this organ. Third, depletion of these factors in the fibroblasts seems to have a selective impact on breast cancer cells but not the normal epithelial cells. In other words, different preexisting genetic alterations in the epithelial cells may request and define distinct cooperative and productive stromal changes.

With that, we propose that the methodology here may be adopted and generation of C.elegans strains with a broader panel of cancer-relevant mutations may provide a platform for further exploration of stromal signaling cascades. It is likely another set of stromal factors may be identified in the axl-1(-) background that is divergent from what has been identified here with gap-1(-) sensitization background. More importantly, the animals may be manipulated in a another directionality for the "basal" vulva proliferation status where specific strains may carry other mutations such as dominant let-23/EGFR or

130 let-60/Ras. In these cases, the animals would display multivulva phenotype at the starting point. Thus these strains could be utilized to identify factors that, when depleted, could reduce the excessive proliferation. In other words, this may expose "oncogene-like" factors in the stroma compartment. Taken all together, this may provide the basis for generation of novel mouse cancer models to be used for dissecting the stromal networks and ultimately further understanding the tumor-stroma interactions for cell proliferation control.

131

References

Adams, M.R., Sears, R., Nuckolls, F., Leone, G., and Nevins, J.R. (2000). Complex transcriptional regulatory mechanisms control expression of the E2F3 locus. Molecular and cellular biology 20, 3633-3639.

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

Adhikary, S., and Eilers, M. (2005). Transcriptional regulation and transformation by Myc proteins. Nature reviews Molecular cell biology 6, 635-645.

Arabi, A., Wu, S., Ridderstrale, K., Bierhoff, H., Shiue, C., Fatyol, K., Fahlen, S., Hydbring, P., Soderberg, O., Grummt, I., et al. (2005). c-Myc associates with ribosomal DNA and activates RNA polymerase I transcription. Nature cell biology 7, 303-310.

Berset, T., Hoier, E.F., Battu, G., Canevascini, S., and Hajnal, A. (2001). Notch inhibition of RAS signaling through MAP kinase phosphatase LIP-1 during C. elegans vulval development. Science 291, 1055-1058.

Bertoli, C., Skotheim, J.M., and de Bruin, R.A. (2013). Control of cell cycle transcription during G1 and S phases. Nature reviews Molecular cell biology 14, 518-528.

Bettess, M.D., Dubois, N., Murphy, M.J., Dubey, C., Roger, C., Robine, S., and Trumpp, A. (2005). c-Myc is required for the formation of intestinal crypts but dispensable for homeostasis of the adult intestinal epithelium. Molecular and cellular biology 25, 7868-7878.

Bhatia, K., Huppi, K., Spangler, G., Siwarski, D., Iyer, R., and Magrath, I. (1993). Point mutations in the c-Myc transactivation domain are common in Burkitt's lymphoma and mouse plasmacytomas. Nature genetics 5, 56-61.

Bhowmick, N.A., Chytil, A., Plieth, D., Gorska, A.E., Dumont, N., Shappell, S., Washington, M.K., Neilson, E.G., and Moses, H.L. (2004a). TGF-beta signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science 303, 848-851.

132

Bhowmick, N.A., Neilson, E.G., and Moses, H.L. (2004b). Stromal fibroblasts in cancer initiation and progression. Nature 432, 332-337.

Blais, A., and Dynlacht, B.D. (2007). E2F-associated chromatin modifiers and cell cycle control. Current opinion in cell biology 19, 658-662.

Bouchard, C., Thieke, K., Maier, A., Saffrich, R., Hanley-Hyde, J., Ansorge, W., Reed, S., Sicinski, P., Bartek, J., and Eilers, M. (1999). Direct induction of cyclin D2 by Myc contributes to cell cycle progression and sequestration of p27. The EMBO journal 18, 5321-5333.

Bronisz, A., Godlewski, J., Wallace, J.A., Merchant, A.S., Nowicki, M.O., Mathsyaraja, H., Srinivasan, R., Trimboli, A.J., Martin, C.K., Li, F., et al. (2012). Reprogramming of the tumour microenvironment by stromal PTEN-regulated miR-320. Nature cell biology 14, 159-167.

Carrera, P., Moshkin, Y.M., Gronke, S., Sillje, H.H., Nigg, E.A., Jackle, H., and Karch, F. (2003). Tousled-like kinase functions with the chromatin assembly pathway regulating nuclear divisions. Genes & development 17, 2578-2590.

Carstens, J.L., Shahi, P., Van Tsang, S., Smith, B., Creighton, C.J., Zhang, Y., Seamans, A., Seethammagari, M., Vedula, I., Levitt, J.M., et al. (2014). FGFR1-WNT- TGF-beta signaling in prostate cancer mouse models recapitulates human reactive stroma. Cancer research 74, 609-620.

Cartwright, P., Muller, H., Wagener, C., Holm, K., and Helin, K. (1998). E2F-6: a novel member of the E2F family is an inhibitor of E2F-dependent transcription. Oncogene 17, 611-623.

Chellappan, S.P., Hiebert, S., Mudryj, M., Horowitz, J.M., and Nevins, J.R. (1991). The E2F transcription factor is a cellular target for the RB protein. Cell 65, 1053-1061.

Chen, D., Pacal, M., Wenzel, P., Knoepfler, P.S., Leone, G., and Bremner, R. (2009a). Division and apoptosis of E2f-deficient retinal progenitors. Nature 462, 925-929.

Chen, H.Z., Tsai, S.Y., and Leone, G. (2009b). Emerging roles of E2Fs in cancer: an exit from cell cycle control. Nature reviews Cancer 9, 785-797.

Choi, Y.J., Li, X., Hydbring, P., Sanda, T., Stefano, J., Christie, A.L., Signoretti, S., Look, A.T., Kung, A.L., von Boehmer, H., et al. (2012). The requirement for cyclin D function in tumor maintenance. Cancer cell 22, 438-451.

Chong, J.L., Wenzel, P.L., Saenz-Robles, M.T., Nair, V., Ferrey, A., Hagan, J.P., Gomez, Y.M., Sharma, N., Chen, H.Z., Ouseph, M., et al. (2009). E2f1-3 switch from

133

activators in progenitor cells to repressors in differentiating cells. Nature 462, 930-934.

Christensen, J., Cloos, P., Toftegaard, U., Klinkenberg, D., Bracken, A.P., Trinh, E., Heeran, M., Di Stefano, L., and Helin, K. (2005). Characterization of E2F8, a novel E2F-like cell-cycle regulated repressor of E2F-activated transcription. Nucleic acids research 33, 5458-5470.

Cole, A.M., Myant, K., Reed, K.R., Ridgway, R.A., Athineos, D., Van den Brink, G.R., Muncan, V., Clevers, H., Clarke, A.R., Sicinski, P., et al. (2010). Cyclin D2- cyclin-dependent kinase 4/6 is required for efficient proliferation and tumorigenesis following Apc loss. Cancer research 70, 8149-8158.

D'Ambrogio, A., Nagaoka, K., and Richter, J.D. (2013). Translational control of cell growth and malignancy by the CPEBs. Nature reviews Cancer 13, 283-290.

Dalla-Favera, R., Bregni, M., Erikson, J., Patterson, D., Gallo, R.C., and Croce, C.M. (1982). Human c-myc onc gene is located on the region of chromosome 8 that is translocated in Burkitt lymphoma cells. Proceedings of the National Academy of Sciences of the United States of America 79, 7824-7827.

Dang, C.V. (2012). MYC on the path to cancer. Cell 149, 22-35.

Davidson, G., Shen, J., Huang, Y.L., Su, Y., Karaulanov, E., Bartscherer, K., Hassler, C., Stannek, P., Boutros, M., and Niehrs, C. (2009). Cell cycle control of wnt receptor activation. Developmental cell 17, 788-799. de Bruin, A., Maiti, B., Jakoi, L., Timmers, C., Buerki, R., and Leone, G. (2003). Identification and characterization of E2F7, a novel mammalian E2F family member capable of blocking cellular proliferation. The Journal of biological chemistry 278, 42041-42049.

Di Stefano, L., Jensen, M.R., and Helin, K. (2003). E2F7, a novel E2F featuring DP- independent repression of a subset of E2F-regulated genes. The EMBO journal 22, 6289-6298.

Dimova, D.K., and Dyson, N.J. (2005). The E2F transcriptional network: old acquaintances with new faces. Oncogene 24, 2810-2826.

Dominguez-Sola, D., Ying, C.Y., Grandori, C., Ruggiero, L., Chen, B., Li, M., Galloway, D.A., Gu, W., Gautier, J., and Dalla-Favera, R. (2007). Non-transcriptional control of DNA replication by c-Myc. Nature 448, 445-451.

134

Dyson, N., Howley, P.M., Munger, K., and Harlow, E. (1989). The human papilloma virus-16 E7 oncoprotein is able to bind to the retinoblastoma gene product. Science 243, 934-937.

Eisenmann, D.M., Maloof, J.N., Simske, J.S., Kenyon, C., and 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 125, 3667-3680.

Finak, G., Bertos, N., Pepin, F., Sadekova, S., Souleimanova, M., Zhao, H., Chen, H., Omeroglu, G., Meterissian, S., Omeroglu, A., et al. (2008). Stromal gene expression predicts clinical outcome in breast cancer. Nat Med 14, 518-527.

Foley, K.P., and Eisenman, R.N. (1999). Two MAD tails: what the recent knockouts of Mad1 and Mxi1 tell us about the MYC/MAX/MAD network. Biochimica et biophysica acta 1423, M37-47.

Frolov, M.V., and Dyson, N.J. (2004). Molecular mechanisms of E2F-dependent activation and pRB-mediated repression. Journal of cell science 117, 2173-2181.

Gaubatz, S., Lindeman, G.J., Ishida, S., Jakoi, L., Nevins, J.R., Livingston, D.M., and Rempel, R.E. (2000). E2F4 and E2F5 play an essential role in pocket protein- mediated G1 control. Molecular cell 6, 729-735.

Gauthier, N.P., Larsen, M.E., Wernersson, R., de Lichtenberg, U., Jensen, L.J., Brunak, S., and Jensen, T.S. (2008). Cyclebase.org--a comprehensive multi-organism online database of cell-cycle experiments. Nucleic acids research 36, D854-859.

Ginsberg, D., Vairo, G., Chittenden, T., Xiao, Z.X., Xu, G., Wydner, K.L., DeCaprio, J.A., Lawrence, J.B., and Livingston, D.M. (1994). E2F-4, a new member of the E2F transcription factor family, interacts with p107. Genes & development 8, 2665-2679.

Gomez-Roman, N., Grandori, C., Eisenman, R.N., and White, R.J. (2003). Direct activation of RNA polymerase III transcription by c-Myc. Nature 421, 290-294.

Goodrich, D.W., and Lee, W.H. (1992). Abrogation by c-myc of G1 phase arrest induced by RB protein but not by p53. Nature 360, 177-179.

Grandori, C., Cowley, S.M., James, L.P., and Eisenman, R.N. (2000). The Myc/Max/Mad network and the transcriptional control of cell behavior. Annual review of cell and developmental biology 16, 653-699.

Grandori, C., Gomez-Roman, N., Felton-Edkins, Z.A., Ngouenet, C., Galloway, D.A., Eisenman, R.N., and White, R.J. (2005). c-Myc binds to human ribosomal DNA

135

and stimulates transcription of rRNA genes by RNA polymerase I. Nature cell biology 7, 311-318.

Grewal, S.S., Li, L., Orian, A., Eisenman, R.N., and Edgar, B.A. (2005). Myc-dependent regulation of ribosomal RNA synthesis during Drosophila development. Nature cell biology 7, 295-302.

Guo, Y., Mahony, S., and Gifford, D.K. (2012). High resolution genome wide binding event finding and motif discovery reveals transcription factor spatial binding constraints. PLoS computational biology 8, e1002638.

Gyorffy, B., Lanczky, A., Eklund, A.C., Denkert, C., Budczies, J., Li, Q., and Szallasi, Z. (2010). An online survival analysis tool to rapidly assess the effect of 22,277 genes on breast cancer prognosis using microarray data of 1,809 patients. Breast cancer research and treatment 123, 725-731.

Haigis, K., Sage, J., Glickman, J., Shafer, S., and Jacks, T. (2006). The related retinoblastoma (pRb) and p130 proteins cooperate to regulate homeostasis in the intestinal epithelium. The Journal of biological chemistry 281, 638-647.

Hajnal, A., Whitfield, C.W., and Kim, S.K. (1997). Inhibition of Caenorhabditis elegans vulval induction by gap-1 and by let-23 receptor tyrosine kinase. Genes & development 11, 2715-2728.

Hanahan, D., and Weinberg, R.A. (2000). The hallmarks of cancer. Cell 100, 57-70.

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

Hayward, W.S., Neel, B.G., and Astrin, S.M. (1981). Activation of a cellular onc gene by promoter insertion in ALV-induced lymphoid leukosis. Nature 290, 475-480.

He, T.C., Sparks, A.B., Rago, C., Hermeking, H., Zawel, L., da Costa, L.T., Morin, P.J., Vogelstein, B., and Kinzler, K.W. (1998). Identification of c-MYC as a target of the APC pathway. Science 281, 1509-1512.

Heinen, C.D., Goss, K.H., Cornelius, J.R., Babcock, G.F., Knudsen, E.S., Kowalik, T., and Groden, J. (2002). The APC tumor suppressor controls entry into S-phase through its ability to regulate the cyclin D/RB pathway. Gastroenterology 123, 751-763.

Heinz, S., Benner, C., Spann, N., Bertolino, E., Lin, Y.C., Laslo, P., Cheng, J.X., Murre, C., Singh, H., and Glass, C.K. (2010). Simple combinations of lineage- determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Molecular cell 38, 576-589.

136

Helin, K. (1998). Regulation of cell proliferation by the E2F transcription factors. Current opinion in genetics & development 8, 28-35.

Helin, K., Lees, J.A., Vidal, M., Dyson, N., Harlow, E., and Fattaey, A. (1992). A cDNA encoding a pRB-binding protein with properties of the transcription factor E2F. Cell 70, 337-350.

Hermeking, H., Rago, C., Schuhmacher, M., Li, Q., Barrett, J.F., Obaya, A.J., O'Connell, B.C., Mateyak, M.K., Tam, W., Kohlhuber, F., et al. (2000). Identification of CDK4 as a target of c-MYC. Proceedings of the National Academy of Sciences of the United States of America 97, 2229-2234.

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

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

Holzel, M., Huang, S., Koster, J., Ora, I., Lakeman, A., Caron, H., Nijkamp, W., Xie, J., Callens, T., Asgharzadeh, S., et al. (2010). NF1 is a tumor suppressor in neuroblastoma that determines retinoic acid response and disease outcome. Cell 142, 218-229.

Hopper, N.A., Lee, J., and Sternberg, P.W. (2000). ARK-1 inhibits EGFR signaling in C. elegans. Molecular cell 6, 65-75.

Hsu, J.Y., Reimann, J.D., Sorensen, C.S., Lukas, J., and Jackson, P.K. (2002). E2F- dependent accumulation of hEmi1 regulates S phase entry by inhibiting APC(Cdh1). Nature cell biology 4, 358-366.

Hu, B., Castillo, E., Harewood, L., Ostano, P., Reymond, A., Dummer, R., Raffoul, W., Hoetzenecker, W., Hofbauer, G.F., and Dotto, G.P. (2012). Multifocal epithelial tumors and field cancerization from loss of mesenchymal CSL signaling. Cell 149, 1207-1220.

Hu, M., Yao, J., Cai, L., Bachman, K.E., van den Brule, F., Velculescu, V., and Polyak, K. (2005). Distinct epigenetic changes in the stromal cells of breast cancers. Nature genetics 37, 899-905.

Huang, Y.S., Carson, J.H., Barbarese, E., and Richter, J.D. (2003). Facilitation of dendritic mRNA transport by CPEB. Genes & development 17, 638-653.

137

Hwang, B.J., and Sternberg, P.W. (2004). A cell-specific enhancer that specifies lin-3 expression in the C. elegans anchor cell for vulval development. Development 131, 143-151.

Ireland, H., Houghton, C., Howard, L., and Winton, D.J. (2005). Cellular inheritance of a Cre-activated reporter gene to determine Paneth cell longevity in the murine small intestine. Developmental dynamics : an official publication of the American Association of Anatomists 233, 1332-1336.

Ireland, H., Kemp, R., Houghton, C., Howard, L., Clarke, A.R., Sansom, O.J., and Winton, D.J. (2004). Inducible Cre-mediated control of gene expression in the murine gastrointestinal tract: effect of loss of beta-catenin. Gastroenterology 126, 1236-1246.

Itoh, A., Levinson, S.F., Morita, T., Kourembanas, S., Brody, J.S., and Mitsialis, S.A. (1995). Structural characterization and specificity of expression of E2F-5: a new member of the E2F family of transcription factors. Cellular & molecular biology research 41, 147-154.

Jiang, W., Kahn, S.M., Tomita, N., Zhang, Y.J., Lu, S.H., and Weinstein, I.B. (1992). Amplification and expression of the human cyclin D gene in esophageal cancer. Cancer research 52, 2980-2983.

Johnson, D.G., Schwarz, J.K., Cress, W.D., and Nevins, J.R. (1993). Expression of transcription factor E2F1 induces quiescent cells to enter S phase. Nature 365, 349-352.

Jose, A.M., and Hunter, C.P. (2007). Transport of sequence-specific RNA interference information between cells. Annual review of genetics 41, 305-330.

Kaelin, W.G., Jr., Krek, W., Sellers, W.R., DeCaprio, J.A., Ajchenbaum, F., Fuchs, C.S., Chittenden, T., Li, Y., Farnham, P.J., Blanar, M.A., et al. (1992). Expression cloning of a cDNA encoding a retinoblastoma-binding protein with E2F-like properties. Cell 70, 351-364.

Kamath, R.S., Fraser, A.G., Dong, Y., Poulin, G., Durbin, R., Gotta, M., Kanapin, A., Le Bot, N., Moreno, S., Sohrmann, M., et al. (2003). Systematic functional analysis of the Caenorhabditis elegans genome using RNAi. Nature 421, 231-237.

Katajisto, P., Vaahtomeri, K., Ekman, N., Ventela, E., Ristimaki, A., Bardeesy, N., Feil, R., DePinho, R.A., and Makela, T.P. (2008). LKB1 signaling in mesenchymal cells required for suppression of gastrointestinal polyposis. Nature genetics 40, 455-459.

138

Kim, B.G., Li, C., Qiao, W., Mamura, M., Kasprzak, B., Anver, M., Wolfraim, L., Hong, S., Mushinski, E., Potter, M., et al. (2006). Smad4 signalling in T cells is required for suppression of gastrointestinal cancer. Nature 441, 1015-1019.

Kovesdi, I., Reichel, R., and Nevins, J.R. (1986). Identification of a cellular transcription factor involved in E1A trans-activation. Cell 45, 219-228.

Landis, M.W., Pawlyk, B.S., Li, T., Sicinski, P., and Hinds, P.W. (2006). Cyclin D1- dependent kinase activity in murine development and mammary tumorigenesis. Cancer cell 9, 13-22.

Langmead, B., and Salzberg, S.L. (2012). Fast gapped-read alignment with Bowtie 2. Nature methods 9, 357-359.

Lee, G., White, L.S., Hurov, K.E., Stappenbeck, T.S., and Piwnica-Worms, H. (2009). Response of small intestinal epithelial cells to acute disruption of cell division through CDC25 deletion. Proceedings of the National Academy of Sciences of the United States of America 106, 4701-4706.

Leone, G., DeGregori, J., Sears, R., Jakoi, L., and Nevins, J.R. (1997). Myc and Ras collaborate in inducing accumulation of active cyclin E/Cdk2 and E2F. Nature 387, 422-426.

Leone, G., Nuckolls, F., Ishida, S., Adams, M., Sears, R., Jakoi, L., Miron, A., and Nevins, J.R. (2000). Identification of a novel E2F3 product suggests a mechanism for determining specificity of repression by Rb proteins. Molecular and cellular biology 20, 3626-3632.

Li, F., Xiang, Y., Potter, J., Dinavahi, R., Dang, C.V., and Lee, L.A. (2006). Conditional deletion of c-myc does not impair liver regeneration. Cancer research 66, 5608- 5612.

Lin, C.Y., Loven, J., Rahl, P.B., Paranal, R.M., Burge, C.B., Bradner, J.E., Lee, T.I., and Young, R.A. (2012). Transcriptional amplification in tumor cells with elevated c- Myc. Cell 151, 56-67.

Liu, W., Dong, X., Mai, M., Seelan, R.S., Taniguchi, K., Krishnadath, K.K., Halling, K.C., Cunningham, J.M., Boardman, L.A., Qian, C., et al. (2000). Mutations in AXIN2 cause colorectal cancer with defective mismatch repair by activating beta- catenin/TCF signalling. Nature genetics 26, 146-147.

Logan, N., Graham, A., Zhao, X., Fisher, R., Maiti, B., Leone, G., and La Thangue, N.B. (2005). E2F-8: an E2F family member with a similar organization of DNA- binding domains to E2F-7. Oncogene 24, 5000-5004.

139

Lukas, C., Sorensen, C.S., Kramer, E., Santoni-Rugiu, E., Lindeneg, C., Peters, J.M., Bartek, J., and Lukas, J. (1999). Accumulation of cyclin B1 requires E2F and cyclin-A-dependent rearrangement of the anaphase-promoting complex. Nature 401, 815-818.

Maiti, B., Li, J., de Bruin, A., Gordon, F., Timmers, C., Opavsky, R., Patil, K., Tuttle, J., Cleghorn, W., and Leone, G. (2005). Cloning and characterization of mouse E2F8, a novel mammalian E2F family member capable of blocking cellular proliferation. The Journal of biological chemistry 280, 18211-18220.

Malumbres, M., and Barbacid, M. (2009). Cell cycle, CDKs and cancer: a changing paradigm. Nature reviews Cancer 9, 153-166.

Malumbres, M., Sotillo, R., Santamaria, D., Galan, J., Cerezo, A., Ortega, S., Dubus, P., and Barbacid, M. (2004). Mammalian cells cycle without the D-type cyclin- dependent kinases Cdk4 and Cdk6. Cell 118, 493-504.

Malynn, B.A., de Alboran, I.M., O'Hagan, R.C., Bronson, R., Davidson, L., DePinho, R.A., and Alt, F.W. (2000). N-myc can functionally replace c-myc in murine development, cellular growth, and differentiation. Genes & development 14, 1390-1399.

McGarry, T.J., and Kirschner, M.W. (1998). Geminin, an inhibitor of DNA replication, is degraded during mitosis. Cell 93, 1043-1053.

McLaughlin, S.K., Olsen, S.N., Dake, B., De Raedt, T., Lim, E., Bronson, R.T., Beroukhim, R., Polyak, K., Brown, M., Kuperwasser, C., et al. (2013). The RasGAP gene, RASAL2, is a tumor and metastasis suppressor. Cancer cell 24, 365-378.

McMahon, S.B., Van Buskirk, H.A., Dugan, K.A., Copeland, T.D., and Cole, M.D. (1998). The novel ATM-related protein TRRAP is an essential cofactor for the c- Myc and E2F oncoproteins. Cell 94, 363-374.

Mello, C.C., Kramer, J.M., Stinchcomb, D., and Ambros, V. (1991). Efficient gene transfer in C.elegans: extrachromosomal maintenance and integration of transforming sequences. The EMBO journal 10, 3959-3970.

Meyer, N., and Penn, L.Z. (2008). Reflecting on 25 years with MYC. Nature reviews Cancer 8, 976-990.

Min, J., Zaslavsky, A., Fedele, G., McLaughlin, S.K., Reczek, E.E., De Raedt, T., Guney, I., Strochlic, D.E., Macconaill, L.E., Beroukhim, R., et al. (2010). An oncogene- tumor suppressor cascade drives metastatic prostate cancer by coordinately activating Ras and nuclear factor-kappaB. Nat Med 16, 286-294. 140

Morris, E.J., Ji, J.Y., Yang, F., Di Stefano, L., Herr, A., Moon, N.S., Kwon, E.J., Haigis, K.M., Naar, A.M., and Dyson, N.J. (2008). E2F1 represses beta-catenin transcription and is antagonized by both pRB and CDK8. Nature 455, 552-556.

Morrish, F., Neretti, N., Sedivy, J.M., and Hockenbery, D.M. (2008). The oncogene c- Myc coordinates regulation of metabolic networks to enable rapid cell cycle entry. Cell Cycle 7, 1054-1066.

Mueller, M.M., and Fusenig, N.E. (2004). Friends or foes - bipolar effects of the tumour stroma in cancer. Nature reviews Cancer 4, 839-849.

Myers, T.R., and 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, 20368-20373.

Nakasone, E.S., Askautrud, H.A., Kees, T., Park, J.H., Plaks, V., Ewald, A.J., Fein, M., Rasch, M.G., Tan, Y.X., Qiu, J., et al. (2012). Imaging tumor-stroma interactions during chemotherapy reveals contributions of the microenvironment to resistance. Cancer cell 21, 488-503.

Nakayama, K.I., and Nakayama, K. (2006). Ubiquitin ligases: cell-cycle control and cancer. Nature reviews Cancer 6, 369-381.

Narita, M., Nunez, S., Heard, E., Lin, A.W., Hearn, S.A., Spector, D.L., Hannon, G.J., and Lowe, S.W. (2003). Rb-mediated heterochromatin formation and silencing of E2F target genes during cellular senescence. Cell 113, 703-716.

Nie, Z., Hu, G., Wei, G., Cui, K., Yamane, A., Resch, W., Wang, R., Green, D.R., Tessarollo, L., Casellas, R., et al. (2012). c-Myc is a universal amplifier of expressed genes in lymphocytes and embryonic stem cells. Cell 151, 68-79.

Nielsen, S.J., Schneider, R., Bauer, U.M., Bannister, A.J., Morrison, A., O'Carroll, D., Firestein, R., Cleary, M., Jenuwein, T., Herrera, R.E., et al. (2001). Rb targets histone H3 methylation and HP1 to promoters. Nature 412, 561-565.

Ogawa, H., Ishiguro, K., Gaubatz, S., Livingston, D.M., and Nakatani, Y. (2002). A complex with chromatin modifiers that occupies E2F- and Myc-responsive genes in G0 cells. Science 296, 1132-1136.

Olumi, A.F., Grossfeld, G.D., Hayward, S.W., Carroll, P.R., Tlsty, T.D., and Cunha, G.R. (1999). Carcinoma-associated fibroblasts direct tumor progression of initiated human prostatic epithelium. Cancer research 59, 5002-5011.

141

Oosterveen, T., Coudreuse, D.Y., Yang, P.T., Fraser, E., Bergsma, J., Dale, T.C., and Korswagen, H.C. (2007). Two functionally distinct Axin-like proteins regulate canonical Wnt signaling in C. elegans. Developmental biology 308, 438-448.

Orimo, A., Gupta, P.B., Sgroi, D.C., Arenzana-Seisdedos, F., Delaunay, T., Naeem, R., Carey, V.J., Richardson, A.L., and 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, 335-348.

Palomero, T., Lim, W.K., Odom, D.T., Sulis, M.L., Real, P.J., Margolin, A., Barnes, K.C., O'Neil, J., Neuberg, D., Weng, A.P., et al. (2006). NOTCH1 directly regulates c- MYC and activates a feed-forward-loop transcriptional network promoting leukemic cell growth. Proceedings of the National Academy of Sciences of the United States of America 103, 18261-18266.

Paulsson, J., and Micke, P. (2014). Prognostic relevance of cancer-associated fibroblasts in human cancer. Seminars in cancer biology 25, 61-68.

Petrella, L.N., Wang, W., Spike, C.A., Rechtsteiner, A., Reinke, V., and Strome, S. (2011). synMuv B proteins antagonize germline fate in the intestine and ensure C. elegans survival. Development 138, 1069-1079.

Phesse, T.J., Myant, K.B., Cole, A.M., Ridgway, R.A., Pearson, H., Muncan, V., van den Brink, G.R., Vousden, K.H., Sears, R., Vassilev, L.T., et al. (2014). Endogenous c-Myc is essential for p53-induced apoptosis in response to DNA damage in vivo. Cell death and differentiation 21, 956-966.

Pickard, A., Cichon, A.C., Barry, A., Kieran, D., Patel, D., Hamilton, P., Salto-Tellez, M., James, J., and McCance, D.J. (2012). Inactivation of Rb in stromal fibroblasts promotes epithelial cell invasion. The EMBO journal 31, 3092-3103.

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

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

Rhee, H.S., and Pugh, B.F. (2011). Comprehensive genome-wide protein-DNA interactions detected at single-nucleotide resolution. Cell 147, 1408-1419.

Rifkin, D.B., Mazzieri, R., Munger, J.S., Noguera, I., and Sung, J. (1999). Proteolytic control of growth factor availability. APMIS : acta pathologica, microbiologica, et immunologica Scandinavica 107, 80-85.

142

Robinson, J.T., Thorvaldsdottir, H., Winckler, W., Guttman, M., Lander, E.S., Getz, G., and Mesirov, J.P. (2011). Integrative genomics viewer. Nature biotechnology 29, 24-26.

Rustgi, A.K., Dyson, N., and Bernards, R. (1991). Amino-terminal domains of c-myc and N-myc proteins mediate binding to the retinoblastoma gene product. Nature 352, 541-544.

Sabo, A., Kress, T.R., Pelizzola, M., de Pretis, S., Gorski, M.M., Tesi, A., Morelli, M.J., Bora, P., Doni, M., Verrecchia, A., et al. (2014). Selective transcriptional regulation by Myc in cellular growth control and lymphomagenesis. Nature 511, 488-492.

Saenz-Robles, M.T., Markovics, J.A., Chong, J.L., Opavsky, R., Whitehead, R.H., Leone, G., and Pipas, J.M. (2007). Intestinal hyperplasia induced by simian virus 40 large tumor antigen requires E2F2. Journal of virology 81, 13191-13199.

Sansom, O.J., Meniel, V.S., Muncan, V., Phesse, T.J., Wilkins, J.A., Reed, K.R., Vass, J.K., Athineos, D., Clevers, H., and Clarke, A.R. (2007). Myc deletion rescues Apc deficiency in the small intestine. Nature 446, 676-679.

Sardet, C., Vidal, M., Cobrinik, D., Geng, Y., Onufryk, C., Chen, A., and Weinberg, R.A. (1995). E2F-4 and E2F-5, two members of the E2F family, are expressed in the early phases of the cell cycle. Proceedings of the National Academy of Sciences of the United States of America 92, 2403-2407.

Sasser, A.K., Mundy, B.L., Smith, K.M., Studebaker, A.W., Axel, A.E., Haidet, A.M., Fernandez, S.A., and Hall, B.M. (2007a). Human bone marrow stromal cells enhance breast cancer cell growth rates in a cell line-dependent manner when evaluated in 3D tumor environments. Cancer letters 254, 255-264.

Sasser, A.K., Sullivan, N.J., Studebaker, A.W., Hendey, L.F., Axel, A.E., and Hall, B.M. (2007b). Interleukin-6 is a potent growth factor for ER-alpha-positive human breast cancer. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 21, 3763-3770.

Satoh, S., Daigo, Y., Furukawa, Y., Kato, T., Miwa, N., Nishiwaki, T., Kawasoe, T., Ishiguro, H., Fujita, M., Tokino, T., et al. (2000). AXIN1 mutations in hepatocellular carcinomas, and growth suppression in cancer cells by virus- mediated transfer of AXIN1. Nature genetics 24, 245-250.

Scherz-Shouval, R., Santagata, S., Mendillo, M.L., Sholl, L.M., Ben-Aharon, I., Beck, A.H., Dias-Santagata, D., Koeva, M., Stemmer, S.M., Whitesell, L., et al. (2014). The reprogramming of tumor stroma by HSF1 is a potent enabler of malignancy. Cell 158, 564-578. 143

Schiavi, S.C., Belasco, J.G., and Greenberg, M.E. (1992). Regulation of proto-oncogene mRNA stability. Biochimica et biophysica acta 1114, 95-106.

Sears, R., Leone, G., DeGregori, J., and Nevins, J.R. (1999). Ras enhances Myc protein stability. Molecular cell 3, 169-179.

Sears, R., Ohtani, K., and Nevins, J.R. (1997). Identification of positively and negatively acting elements regulating expression of the E2F2 gene in response to cell growth signals. Molecular and cellular biology 17, 5227-5235.

Seoane, J., Pouponnot, C., Staller, P., Schader, M., Eilers, M., and Massague, J. (2001). TGFbeta influences Myc, Miz-1 and Smad to control the CDK inhibitor p15INK4b. Nature cell biology 3, 400-408.

Serandour, A.A., Brown, G.D., Cohen, J.D., and Carroll, J.S. (2013). Development of an Illumina-based ChIP-exonuclease method provides insight into FoxA1-DNA binding properties. Genome biology 14, R147.

Sharma, V.M., Calvo, J.A., Draheim, K.M., Cunningham, L.A., Hermance, N., Beverly, L., Krishnamoorthy, V., Bhasin, M., Capobianco, A.J., and Kelliher, M.A. (2006). Notch1 contributes to mouse T-cell leukemia by directly inducing the expression of c-myc. Molecular and cellular biology 26, 8022-8031.

Sherr, C.J. (1996). Cancer cell cycles. Science 274, 1672-1677.

Sherr, C.J., and McCormick, F. (2002). The RB and p53 pathways in cancer. Cancer cell 2, 103-112.

Siddiqui, H., Fox, S.R., Gunawardena, R.W., and Knudsen, E.S. (2007). Loss of RB compromises specific heterochromatin modifications and modulates HP1alpha dynamics. Journal of cellular physiology 211, 131-137.

Siddiqui, S.S., Loganathan, S., Krishnaswamy, S., Faoro, L., Jagadeeswaran, R., and Salgia, R. (2008). C. elegans as a model organism for in vivo screening in cancer: effects of human c-Met in lung cancer affect C. elegans vulva phenotypes. Cancer biology & therapy 7, 856-863.

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

Staller, P., Peukert, K., Kiermaier, A., Seoane, J., Lukas, J., Karsunky, H., Moroy, T., Bartek, J., Massague, J., Hanel, F., et al. (2001). Repression of p15INK4b

144

expression by Myc through association with Miz-1. Nature cell biology 3, 392- 399.

Sternberg, P.W. (2005). Vulval development. WormBook, 1-28.

Straussman, R., Morikawa, T., Shee, K., Barzily-Rokni, M., Qian, Z.R., Du, J., Davis, A., Mongare, M.M., Gould, J., Frederick, D.T., et al. (2012). Tumour microenvironment elicits innate resistance to RAF inhibitors through HGF secretion. Nature 487, 500-504.

Studebaker, A.W., Storci, G., Werbeck, J.L., Sansone, P., Sasser, A.K., Tavolari, S., Huang, T., Chan, M.W., Marini, F.C., Rosol, T.J., et al. (2008). Fibroblasts isolated from common sites of breast cancer metastasis enhance cancer cell growth rates and invasiveness in an interleukin-6-dependent manner. Cancer research 68, 9087-9095.

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

Tan, P.B., Lackner, M.R., and Kim, S.K. (1998). MAP kinase signaling specificity mediated by the LIN-1 Ets/LIN-31 WH transcription factor complex during C. elegans vulval induction. Cell 93, 569-580.

Taub, R., Kirsch, I., Morton, C., Lenoir, G., Swan, D., Tronick, S., Aaronson, S., and Leder, P. (1982). Translocation of the c-myc gene into the immunoglobulin heavy chain locus in human Burkitt lymphoma and murine plasmacytoma cells. Proceedings of the National Academy of Sciences of the United States of America 79, 7837-7841.

Tlsty, T.D., and Coussens, L.M. (2006). Tumor stroma and regulation of cancer development. Annual review of pathology 1, 119-150.

Trimarchi, J.M., Fairchild, B., Verona, R., Moberg, K., Andon, N., and Lees, J.A. (1998). E2F-6, a member of the E2F family that can behave as a transcriptional repressor. Proceedings of the National Academy of Sciences of the United States of America 95, 2850-2855.

Trimarchi, J.M., and Lees, J.A. (2002). Sibling rivalry in the E2F family. Nature reviews Molecular cell biology 3, 11-20.

Trimboli, A.J., Cantemir-Stone, C.Z., Li, F., Wallace, J.A., Merchant, A., Creasap, N., Thompson, J.C., Caserta, E., Wang, H., Chong, J.L., et al. (2009). Pten in stromal fibroblasts suppresses mammary epithelial tumours. Nature 461, 1084-1091.

145

Trumpp, A., Refaeli, Y., Oskarsson, T., Gasser, S., Murphy, M., Martin, G.R., and Bishop, J.M. (2001). c-Myc regulates mammalian body size by controlling cell number but not cell size. Nature 414, 768-773.

Tsai, S.Y., Opavsky, R., Sharma, N., Wu, L., Naidu, S., Nolan, E., Feria-Arias, E., Timmers, C., Opavska, J., de Bruin, A., et al. (2008). Mouse development with a single E2F activator. Nature 454, 1137-1141.

Uribesalgo, I., Buschbeck, M., Gutierrez, A., Teichmann, S., Demajo, S., Kuebler, B., Nomdedeu, J.F., Martin-Caballero, J., Roma, G., Benitah, S.A., et al. (2011). E- box-independent regulation of transcription and differentiation by MYC. Nature cell biology 13, 1443-1449. van der Flier, L.G., and Clevers, H. (2009). Stem cells, self-renewal, and differentiation in the intestinal epithelium. Annual review of physiology 71, 241-260.

Vennstrom, B., Sheiness, D., Zabielski, J., and Bishop, J.M. (1982). Isolation and characterization of c-myc, a cellular homolog of the oncogene (v-myc) of avian myelocytomatosis virus strain 29. Journal of virology 42, 773-779.

Vermeulen, L., De Sousa, E.M.F., van der Heijden, M., Cameron, K., de Jong, J.H., Borovski, T., Tuynman, J.B., Todaro, M., Merz, C., Rodermond, H., et al. (2010). Wnt activity defines colon cancer stem cells and is regulated by the microenvironment. Nature cell biology 12, 468-476.

Visintin, R., Prinz, S., and Amon, A. (1997). CDC20 and CDH1: a family of substrate- specific activators of APC-dependent proteolysis. Science 278, 460-463.

Walz, S., Lorenzin, F., Morton, J., Wiese, K.E., von Eyss, B., Herold, S., Rycak, L., Dumay-Odelot, H., Karim, S., Bartkuhn, M., et al. (2014). Activation and repression by oncogenic MYC shape tumour-specific gene expression profiles. Nature 511, 483-487.

Welcker, M., and Clurman, B.E. (2008). FBW7 ubiquitin ligase: a tumour suppressor at the crossroads of cell division, growth and differentiation. Nature reviews Cancer 8, 83-93.

Weng, A.P., Millholland, J.M., Yashiro-Ohtani, Y., Arcangeli, M.L., Lau, A., Wai, C., Del Bianco, C., Rodriguez, C.G., Sai, H., Tobias, J., et al. (2006). c-Myc is an important direct target of Notch1 in T-cell acute lymphoblastic leukemia/lymphoma. Genes & development 20, 2096-2109.

Wenzel, P.L., Chong, J.L., Saenz-Robles, M.T., Ferrey, A., Hagan, J.P., Gomez, Y.M., Rajmohan, R., Sharma, N., Chen, H.Z., Pipas, J.M., et al. (2011). Cell proliferation in the absence of E2F1-3. Developmental biology 351, 35-45. 146

Wu, L., Timmers, C., Maiti, B., Saavedra, H.I., Sang, L., Chong, G.T., Nuckolls, F., Giangrande, P., Wright, F.A., Field, S.J., et al. (2001). The E2F1-3 transcription factors are essential for cellular proliferation. Nature 414, 457-462.

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

Yang, Z.F., Mott, S., and Rosmarin, A.G. (2007). The Ets transcription factor GABP is required for cell-cycle progression. Nature cell biology 9, 339-346.

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

Yu, Q., Geng, Y., and Sicinski, P. (2001). Specific protection against breast cancers by cyclin D1 ablation. Nature 411, 1017-1021.

Yu, Q., Sicinska, E., Geng, Y., Ahnstrom, M., Zagozdzon, A., Kong, Y., Gardner, H., Kiyokawa, H., Harris, L.N., Stal, O., et al. (2006). Requirement for CDK4 kinase function in breast cancer. Cancer cell 9, 23-32.

Zhang, H., Gao, P., Fukuda, R., Kumar, G., Krishnamachary, B., Zeller, K.I., Dang, C.V., and Semenza, G.L. (2007). HIF-1 inhibits mitochondrial biogenesis and cellular respiration in VHL-deficient renal cell carcinoma by repression of C-MYC activity. Cancer cell 11, 407-420.

Zhang, Y., Liu, T., Meyer, C.A., Eeckhoute, J., Johnson, D.S., Bernstein, B.E., Nusbaum, C., Myers, R.M., Brown, M., Li, W., et al. (2008). Model-based analysis of ChIP- Seq (MACS). Genome biology 9, R137.

Zhu, Y., Ghosh, P., Charnay, P., Burns, D.K., and Parada, L.F. (2002). Neurofibromas in NF1: Schwann cell origin and role of tumor environment. Science 296, 920-922.

Zuo, L., Weger, J., Yang, Q., Goldstein, A.M., Tucker, M.A., Walker, G.J., Hayward, N., and Dracopoli, N.C. (1996). Germline mutations in the p16INK4a binding domain of CDK4 in familial melanoma. Nature genetics 12, 97-99.

147