A Dissertation Submitted To

The role and function of the Ras-related protein TC21 in Neurofibromatosis type 1

A dissertation submitted to the Graduate School of the University of Cincinnati in partial fulfillment of the requirement for the degree of Doctor of Philosophy in the Department of Cancer and Cell Biology of the College of Medicine

Deanna M Patmore BS, Voorhees College June 2007

Dissertation Committee: Nancy Ratner, PhD (Chair) Vaughn Cleghon, PhD George Thomas, PhD Susanne Wells, PhD Yi Zheng, PhD ABSTRACT

Neurofibromatosis type 1 is a common autosomal dominant disorder affecting 1in

3500 individuals worldwide. Neurofibromin, the protein mutated in NF1 disease, is a

GTPase activating protein (GAP) for Ras proteins, inactivating the Ras proteins H-Ras,

N-Ras, K-Ras, M-Ras, R-Ras, and TC21. Missense mutations in the GAP related domain of neurofibromin cause NF1 disease, indicating that increased Ras activity is likely critical for disease pathogenesis. Loss of NF1 in Schwann cells causes formation of benign tumors known as neurofibromas. These tumors can become malignant forming malignant peripheral nerve sheath tumors (MPNSTs), which are a major source of morbidity for NF1 patients. TC21 is a member of the R-Ras family of small Ras

GTPases. TC21 is an oncogene able to transform epithelial and fibroblast cell lines, and it is also capable of inducing tumors in vivo . PI3K is the predominant effector of TC21 transforming activity, and studies support the idea that not all features of NF1 mutant cells can be ascribed to the activation of the classical Ras proteins (e.g. H-, N-, K-Ras).

We hypothesize that the effects of neurofibromin mutation that are unrelated to classical

Ras-GTP may be explained by activation of the non-classical Ras protein TC21. The signaling pathway of TC21 is unclear. We activated all Ras proteins in vivo by deletion of Nf1 and using these mice along with mice deficient for TC21 -/-, we examined the role and function of TC21 in development and tumorigenesis. Additionally, we used NF1 -/- human MPNST cell lines with an acute loss of TC21 by shRNA to examine tumor growth in xenograft mice. We found that TC21 loss delayed benign neurofibroma formation in Nf1fl/fl;DhhCre mice. Nf1 loss increased mRNA encoding the cytokine transforming growth factor-beta (TGF-beta) and rendered Schwann cell progenitors

i insensitive to TGF-beta; these phenotypes could be rescued by TC21 and were mediated through TGF-beta receptors. Conversely, growth of Nf1;Trp53 brain tumors and NF1 -/- MPNST sarcomas were accelerated by TC21 loss. MPNST from Nf1;Trp53 mice and NF1 -/- MPNST xenografts had increased levels of TGF-beta mRNA and protein and blocking TGF-beta decreased sarcoma size induced by shTC21. The results are important because TGF-beta acts as a tumor suppressor in numerous types of benign tumors and is also known for its oncogenic role in cellular transformation and tumor progression. Our data elucidates TGF-beta as a therapeutic target in NF1 malignancy.

iii

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to my mentor, Dr. Nancy Ratner, for her continuous support during the past five years of my graduate studies. Dr. Ratner’s leadership and guidance has significantly helped in molding me as a scientist. Her mentorship allowed me to equip myself with the scientific knowledge and skills necessary to establish a life-long career in basic research. During my graduate studies,

I have developed a keen appreciation for cancer research and particularly pediatric cancer research.

Additionally, I would like to thank my thesis committee members, Dr. Vaughn

Cleghon, Dr. George Thomas, Dr. Susanne Wells and Dr. Yi Zheng for their insightful ideas, comments and advice throughout the development of my thesis project, as well as their significant contribution to my scientific training.

I would also like to acknowledge and thanks the present and former members of the Ratner lab for their technical guidance and support, as well as their valued time spent in discussion of data and analytical interpretation.

Last but not least, I would like to give my deepest gratitude to my family and friends who have always believed in and supported me. Without you these past five years would have been impossible. I dedicate this thesis study to my mother, Mrs.

Yvette Lemonius. Your unconditional love and support, your determination and faith have made me the person I am today. I love you always.

TABLE OF CONTENTS

ABSTRACT……………………………………………………………….……………..i

ACKNOWLEDGEMENTS...... iv

TABLE OF CONTENTS ...... vi

LIST OF FIGURES ...... viii

LIST OF TABLES...... x

CHAPTER 1: INTRODUCTION

OVERVIEW……………………………………………………………………….1

NEUROFIBROMATOSIS TYPE 1: THE DISEASE………………………….2

NF1 MOUSE MODELS………………………………………………………….5

THE NF1 GENE……………………………………………………………..…....6

THE RAS SUPER-FAMILY PROTEINS……………...……………………...... 8

THE RAS-RELATED PROTEIN, TC21……………………………………..…9

INDIVIDUAL FUNCTIONS OF RAS PROTEINS………………………..…...10

TRANSFORMING GROWTH FACTOR BETA…………………………...…...12

CROSSTALK BETWEEN RAS AND TGF-BETA PATHWAYS………..…..14

CHAPTER 2: IN VIVO REGULATION OF TGF-BETA BY THE RAS PROTEIN TC21,

REVEALED THROUGH THE LOSS OF THE NF1 RAS-GAP

ABSTRACT……………………………………………………………………....20

BACKGROUND………………………………………………………………….21

MATERIALS AND METHODS…………………………………………………24

RESULTS………………………………………………………………………....30

DISCUSSION……………………………………………………………………..44

CHAPTER 3: DISCUSSION AND FUTURE DIRECTIONS……………………...... 69

REFERENCES……………………………………………………………………………78

vii

LIST OF FIGURES

FIGURE 1.1: RAS EFFECTOR PATHWAYS………………………………….18

FIGURE 1.2: TGF-BETA SIGNALING…………………………………………..19

FIGURE 2.1: INSERTION INTO THE MOUSE TC21 GENE CAUSES A NULL

ALLELE……………………………………………………………………………..48

FIGURE 2.2: LOSS OF TC21 EXTENDS SURVIVAL OF NEUROFIBROMA-BEARING

MICE BUT DECREASES SURVIVAL OF NPCis MICE……………………….51

FIGURE 2.3: Nf1 -/- TUMOR INITIATING CELLS ARE DECREASED WHEN TC21 IS

ABSENT WHILE RESULTING TUMORS SHOW NEUROFIBROMA HITOLOGY,

TUMOR SIZE AND TUMOR NUMBER…………………………………………..52

FIGURE 2.4: A TGF-BETA AUTOCRINE LOOP IN Nf1 -/- SCHWANN CELL

PRECURSORS……………………………………………………………………..54

FIGURE 2.5: TGF-BETA EXPRESSION IN SCIATIC NERVES AND

NEUROFIBROMAS………………………………………………………………..56

FIGURE 2.6: SURVIVAL OF NF1 MUTANT SCHWANN CELL PRECURSORS IS

DEPENDENT ON TGF-BETA AND AKT………………………………………..57

FIGURE 2.7: Nf1-/- SCHWANN CELL PRECURSORS EXPRESS HIGH LEVELS OF

PHOSPHO-AKT BUT NOT PHOSPHO-SMAD2/3……………………………..59

FIGURE 2.8: LOSS OF TC21 DECREASES THE SURVIVAL OF NPCis MICE BY

INCREASING BRAIN TUMORS………………………………………………….61

FIGURE 2.9: LOSS OF TC21 IN MPNST CELLS INCREASE SARCOMA

GROWTH…………………………………………………………………………….62

viii

FIGURE 2.10: TGF-BETA MEDIATES AGGRESSIVE GROWTH OF MPNST

XENOGRAFTS……………………………………………………………………………..64

FIGURE 2.11: TGF-BETA LIGANDS AND RECEPTOR EXPRESSION IN MPNST

CELLS……………………………………………………………………………………….66

FIGURE 2.12: MPNST CELLS EXPRESS TGF-BETA AND LOSE TGFbRII...... 67

APPENDIX A: TGFb2 mRNA EXPRESSION IN MPNST CELLS AFTER INHIBITION

OF JNK OR sFOS…………………………………………………………………………77

LIST OF TABLES

TABLE 1: LOSS OF TC21 PARTIALLY RESCUES Nf1 MUTANT EMBRYONIC

LETHALITY……………………………………….…………………………………….50

Chapter 1

Introduction

Overview

The neurofibromatoses are a set of distinct genetic disorders that cause tumors to grow along or within various types of nerves. These disorders include neurofibromatosis type 1 (NF1), neurofibromatosis type 2 (NF2) and schwannomatosis.

Together these disorders affect approximately 100,000 persons in the US. The most common of these disorders is NF1, which affects 1 in 3000 live births (1) while NF2 affects 1:25,000 births and Schwannomatosis affects an estimated 1:40,000 births (2).

These disorders each predispose affected individuals to develop Schwann cell tumors and they arise from mutations in different genes, each of which plays a key role in regulating Schwann cell function.

The NF1 gene on human chromosome 17q11, encodes an intracellular signaling molecule that functions as a GTPase activating protein for Ras proteins (3). Loss of NF1 results in activation of at least seven Ras proteins. In contrast, the NF2 gene on human chromosome 22q12 encodes a cytoskeletal-membrane linking protein (4). A candidate gene, INI1, has been recently identified in Schwannomatosis and investigating its role and function in this disorder is progressing (5).

In these studies we focus on NF1, a known inhibitor of Ras proteins. The loss of the NF1 gene in neurofibromatosis type 1 is correlated with increased activation of Ras proteins. However, not all the phenotypes of the disease can be explained by the activation of the commonly studied canonical Ras proteins (H-, N- and K-Ras). We

1 therefore investigated the role of the little studied Ras related protein, TC21 in NF1 development and tumorigenesis. We show that TC21 is an important Ras protein in neurofibromatosis type 1. Our studies examine the role of TC21 in NF1, showing that if regulates the cytokine, TGF-beta, elucidating a role for TGF-beta as a therapeutic target particularly in malignant NF1.

Neurofibromatosis Type 1: The Disease

Formally known as Von Recklinghausen disease, NF1 is an autosomal dominant disorder affecting the nervous system, brain and bones (6). The disorder has almost

100% penetrance, although clinical manifestations within a family frequently vary. 50% of NF1 cases are sporadic (7). A formal diagnostic criterion was established by the

National Institutes of Health Consensus Development Conference in 1987. The current diagnosis of NF1 is made in an individual with any of the two following clinical features: neurofibromas, cafe´ -au-lait spots; freckling; optic pathway gliomas (OPGs); lisch nodules; distinctive bony lesions; and a first-degree family relative with NF1 (8).

Additionally learning disabilities and cardiovascular defects are also seen in NF1 patients. Studies show that in 60% of NF1 cases, patients suffer from cognitive defects and academic learning difficulties, the most common neurological disability (9, 10). NF1 patients have also been shown to have hypertension and cardiovascular defects.

Studies show that NF1 patients have early cardiac morphologic and functional changes that are associated with hypertension (11).

The major feature of the disease are neurofibromas; benign Schwann cell tumors arising in peripheral nerves. These tumors are composed of Schwann cells, fibroblasts,

2 perineural cells, and mast cells. The primary neoplastic cell in neurofibromas is considered to be the NF1 -deficient Schwann cell (12). Neurofibromas can be subdivided into various groups and these differ depending on their location, pattern of growth, association with NF1 and their ability to transform into malignant tumors. Investigators and clinicians who study these lesions have varying terminologies (13). Three subsets of neurofibromas are cutaneous, subcutaneous and plexiform. The location and number of neurofibromas vary among individuals (14).

Cutaneous neurofibromas are soft fleshy tumors usually appearing in late childhood or young adulthood. These tumors rarely cause pain or neurologic deficits and do not transform into malignant tumors. Cutaneous neurofibromas may cause discomfort or extreme cosmetic disfigurement and can be removed by a surgery.

Subcutaneous neurofibromas are firm, tender nodules along the course of peripheral nerves that usually appear during adolescence or young adulthood (15).

Plexiform neurofibromas are usually inherited and are present in 30% of patients with neurofibromatosis type 1(16). They can involve long portions of one nerve or bundles of nerves. They can be deep inside the body and are only visualized with scans, such as MRIs, or they can be superficial and involve the skin, extremities, head or neck. Plexiform neurofibromas can be present at birth but may continue to appear through late adolescence and early adulthood. They tend to enlarge with age especially through the first decade of life and can become severely disfiguring (17).

Neurofibromas are divided into three categories according to their growth

(superficial, displacing, and invasive) and can be measured volumetrically with MRI

(18). Displacing and invasive plexiform neurofibromas can progress into highly

3 aggressive sarcomas known as malignant peripheral nerve sheath tumors (MPNSTs).

Therefore, plexiform neurofibromas represent a major cause of morbidity and death in

NF1 patients (19). The progression of plexiform neurofibromas into MPNSTs occurs in approximately 8–13% of all NF1 patients (20). Due to the superficial or invasive nature of plexiform neurofibromas, these tumors can be visualized by MRI. In clinical trials, the volumetric measurements resulting from MRIs allow physicians to determine if new therapeutic agents are having an effect on the growth of lesions that are either invasive or inoperable.

Studies have shown that NF1 patients with subcutaneous neurofibromas are three times more likely to have internal plexiform neurofibromas or MPNSTs than individuals without subcutaneous neurofibromas. Additionally, individuals with internal plexiform neurofibromas were 20 times more likely to have MPNSTs than individuals without internal plexiform neurofibromas (21). Therefore, NF1 patients with benign neurofibromas need to be monitored for malignancy. Since there is currently no treatment for neurofibromas (other than surgical dissection which can be dangerous based on the location of the tumor), results from the monitoring of NF1 patients with internal plexiform neurofibromas are beneficial in developing new treatment strategies.

MPNSTs are one of the most clinically aggressive cancers associated with NF1, typically arising from pre-existing plexiform neurofibromas. However, studies have shown that patients without a previous history of NF1 or plexiform neurofibromas can develop MPNSTs, sporadic MPNSTs (22). MPNST patients commonly suffer from pain and neurologic deficits. MPNSTs have an overall poor prognosis because they frequently metastasize and are resistant to therapy. Currently the standard of care for

MPNSTs consists of surgical excision with postoperative radiotherapy. This regimen does not improve the long-term survival rates of patients; however, it delays the time to local recurrence (23). Chemotherapy is a second adjuvant option for treating MPNSTs but its use remains controversial (24).

Recent studies with inhibitors in NF1 mouse models show promising targeted therapeutic treatments for MPNST. These include combinatorial trials of rapamycin

(targeting mTOR in the PI3K/AKT pathway) and heat shock protein inhibitors (25).

There are numerous studies like these that may lead to clinical trials in patients.

NF1 Mouse Models

Several genetically modified mouse models have been developed to directly test the role specific gene mutations and growth factors play in NF1-associated peripheral nerve sheath tumors pathogenesis. Additionally, these models allow for preclinical testing. The overall goal of constructing these mouse models is recapitulate the clinical features of human NF1, including the development of neurofibromas and MPNSTs.

It has been shown that NF1 is necessary for normal development through studies targeting mutation in both alleles of Nf1. Nf1 -/- mouse embryos die by embryonic day

13.5 (90% are dead by E12.5 on the C57Bl/6 genetic background) in utero due to malformations affecting the brain and heart (26-28). Further studies of mice heterozygous for Nf1 (Nf1 +/-) showed that there were viable and fertile; however, adult

Nf1 +/- mice do not spontaneously develop neurofibromas (29). This observation suggested that neurofibromas do not develop in Nf1 +/- mice because inactivation of the remaining functional Nf1 allele in murine Schwann cells occurs at a very low frequency.

Subsequent studies demonstrated neurofibromas do form when Nf1 -/- embryonic stem cells are added to wild type blastocysts (30) or in Nf1 +/- mice after wounding of peripheral nerves (31). A conditional (cre/lox) allele of Nf1 was generated, and in mouse models where Nf1 ablation was driven by three different neural crest promoters, defects in neurons were detected, but nerve hyperplasia and Schwann cell tumors were not

(32). Ablation of Nf1 in Schwann cell lineage using the Krox20 promoter in an Nf1 +/- background formation of neurofibroma-like lesions were observed (33). This suggests that pathogenesis of neurofibromas requires an Nf1 -/- Schwann cells and a haploinsufficient tumor environment. However, additionally studies using the DHH promoter to ablate Nf1 in the Schwann cell lineage show that mice develop both dermal and plexiform neurofibromas with an intact Nf1 environment (34).

In addition to modeling neurofibromas in mice, there are studies that utilize mouse models for NF1 malignancy, MPNST. Mice harboring Nf1 and p53 placed in cis on mouse chromosome 11 mice developed sarcomas with the histologic and immunohistochemical characteristics of MPNSTs (30, 35, 36). The mouse model with mutations in Nf1 and Ink4a/Arf is another MPSNT model where mice develop MPNST like sarcomas near the dorsal root ganglia and peripheral nerves (37).

The NF1 Gene

The NF1 gene, isolated in 1990, spans over 350 kb of DNA, mapping to human chromosome 17q11 (38). NF1 encodes an 11-13kb mRNA which is translated into the

2818 amino acid protein neurofibromin (39). Neurofibromin is a highly conserved protein whose central region contains a functional Ras GTPase activating protein-related

6 domain (Ras-GRD) (40). Neurofibromin is a GAP for Ras proteins (41), therefore sustained activation of each expressed Ras protein is predicted in cells that rely on neurofibromin function. Neurofibromin is a tumor suppressor that limits cell growth, and its absence or reduced expression therefore leads to increased cell growth. Inactivation of the NF1 gene, either by mutation or allelic loss, leads to loss of function and subsequent development of many different types of tumors seen in the disease (39, 42).

Mutations in both NF1 alleles have been detected in neurofibromas (43-48),

MPNSTs (49), and neurofibroma Schwann cells. However, in neurofibroma fibroblasts mutations in NF1 is not observed (50, 51). Neurofibromin contains several domains, the

Ras-GAP domain, the Sec14 domain and the PH domain (52-55). The Ras-GAP-related domain (Ras-GRD) is most extensively studied because this domain inhibits the activity of Ras GTPase proteins by catalyzing the hydrolysis of active Ras guanosine triphosphate (Ras-GTP) to inactive Ras guanosine diphosphate (Ras-GDP). Therefore, loss of neurofibromin results in constitutive activation of Ras proteins and downstream signaling; cell growth, proliferation, migration and survival (56, 57). In addition to regulating Ras proteins, neurofibromin also increase cyclic adenosine monophosphate

(cAMP) levels (58, 59). Increased levels of cAMP are associated with reduced cell growth. Both direct and indirect interaction between neurofibromin and several other proteins have been reported. These include the transmembrane proteoglycan syndecan, actin, tubulin, and intermediate filaments among others (60-62). Little is known about the function of neurofibromin with these proteins or if these interactions explain the different symptoms related to NF1(63).

The Ras Super-Family Proteins The Ras proteins are molecular switches that cycle between an inactive GDP- bound form and an active GTP-bound form. Ras proteins are activated by guanine exchange factors (GEFs) that promote them to bind GTP. GTPase activating proteins

(GAPs) facilitate the hydrolysis from an active GTP bound Ras protein to an inactive

GDP bound Ras protein (64, 65) (Figure 1.1). Ras proteins transmit signals from cell surface receptors, such as growth factor receptors, G-protein coupled receptors.

Signals are then passed from protein-to-protein along several different pathways, eventually affecting mitogenic functions. Mutation or deregulation of the Ras genes cause disruption of these signals and occur in many tumor types (66).

Members of the Ras family of proteins include the commonly studied H-Ras, K-

Ras and N-Ras and the other subgroups R-Ras, Ral and Rap. H-Ras, N-Ras, and K-

Ras proteins consist of 188-189 amino acid (p21 proteins), encoded by the three genes.

These Ras isoforms are highly homologous (67). H-Ras, K-Ras and N-Ras genes are activated by mutation in up to 50% of human cancers (68). The R-Ras subgroup, encoded by the R-Ras, R-Ras2/TC21 (subsequently TC21 ) and R-Ras3/M-Ras genes, also has oncogenic potential (69, 70). The Ras oncoproteins ( H-Ras, N-Ras, and K-

Ras ), and the related proteins R-Ras and TC21, share a high level of amino acid similarity. R-Ras and H-Ras are 55% homologous in their amino acid identity, and TC21 is 70% homologous to R-Ras (71). Ras and Ras-related proteins are often deregulated in cancers, leading to increased invasion and metastasis; and decreased apoptosis

(72).

Ras proteins couple the signals of activated growth factor receptors to downstream effectors that interact with the active GTP-bound form of Ras. Ras

8 effectors include protein kinases and lipid kinases, which transmit signals to the cell’s nucleus, recruit Ras proteins to the plasma membrane, and allow them to associate with substrates (73). Ras signaling is critical to activating a variety of downstream cascades involved in processes such as actin cytoskeletal integrity, proliferation, differentiation, cell adhesion, apoptosis and cell migration. Of these, the best characterized is the

Raf/MAPK kinase signaling pathway (74). Mitogen-activated protein kinase (MAPK) components include the ERK pathway, the SAPK/JNK pathway, and the p38 pathway.

MAPK pathways are signaling cascades that are well-conserved and are involved in the transduction of mitogen signals into cellular responses in a variety of organisms. These pathways activate various substrates in the cell and cellular responses including apoptosis, proliferation, and differentiation (75).

The Ras-related Protein, TC21

TC21 is a member of the R-Ras sub-family of Ras proteins. The TC21 gene was cloned from a human teratocarcinoma cell line (76). TC21 contains an 11-amino acid N- terminal extension, compared to other Ras proteins (K-, H- and N-Ras), making it

23kDa, as opposed to 21kDa. At the C-terminus, a TC21 CAAX box is a substrate for lipid modification. The CVIF motif makes TC21 a substrate for gerangeranylation as well as farnesylation, rendering TC21 (like K-Ras and N-Ras) insensitive to farnesyltransferase inhibitors (77). The TC21 protein is an oncogene that has been shown to drive the malignant transformation of both epithelial and fibroblast cell lines

(70) and it also induces tumors in vivo . Overexpression of TC21 has been observed in squamous cell carcinoma (78), and is sufficient to transform MCF-10A breast cancer

9 cells (79). TC21 is the only Ras-related protein that is mutated in human tumor cell lines

(80) and its mutations are associated with the growth of certain tumor types including ovarian cancer, breast cancer, oral squamous cell carcinoma (SCC), esophageal SCC and malignant skin cancer (78, 81, 82).

Thus far, very little is known about the role of TC21 in cell signaling. In 1999,

Graham et. al., showed that transfection of active TC21 into NIH3T3 cells results in increased Rac activity coupled with increased levels of phospho p38 MAPK (70).

Additional studies have shown that overexpressing active TC21 in EpH4 cells induces tumorigenicity through the phosphoinositide 3-kinase, p38 MAPK, and mTOR pathways which correlates with loss of sensitivity to the normal growth inhibitory role of TGF-beta

(83). TC21 stimulates similar downstream signaling pathways as other Ras proteins but with different potency (70, 84, 85). Studies so far show the main effector of TC21 is

PI3K (86-88). TC21-GTP binds to p100 and activates Rac in a PI3K-dependent manner

(87). Additionally, TC21-GTP forms a complex with the p85 subunit of PI3K, which correlates with increased Akt activation (86).

Individual Functions of Ras Proteins

Specific Ras proteins are mutationally activated in specific types of human cancers. Mouse model studies have shown that double mutant H-Ras -/-;N-Ras -/- mice are fertile and viable (89), both TC21 -/- and R-Ras -/- mice are fertile and viable (90, 91) but K-

Ras -/- mice die during embryogenesis (92). TC21 is most similar to drosophila (dRas2) and yeast Ras2. In the fly, overexpression of dRas1, the Ras homologous to H-, N-, and

K-Ras, causes phenotypes different from the overexpression of dRas2 even when driven by the same promoter (93-96).

Ras proteins also differ in localization to different intracellular membrane compartments. Acylation and de-acylation control transport of Ras proteins from the plasma membrane and intracellular membranes (97). Ras signaling that is both Golgi- dependent and Golgi-independent can be stimulated by growth factors. In 1994, Lisanti et. al. (98) isolated caveloae and characterized the complexes and their associated proteins by micro-sequencing. They found that TC21, but not other Ras family members, was present in these complexes. It has also been shown that neurofibromin binds to caveolin-1 to regulate Ras and AKT (99) and recruitment of H-Ras to caveolae occurs upon the activation of EGFR (100). The association of N- and K-Ras with caveloae has been shown in other cell types to regulate EGFR-directed signaling (101).

EGFR acquisition is a common feature of Nf1 mutant mouse and human cells (102), and EGFR alone can drive pre-neoplastic changes in peripheral nerve (103).

Another mechanism by which specific Ras proteins might differ in function is through different interactions with their effectors. A yeast two hybrid system approach has been used to examine the direct comparison of TC21-GTP and Ras-GTP interactions with effector proteins. It was seen that expression library screening failed to identify unique TC21 binding partners (87). This data suggests that TC21 stimulates similar downstream signaling pathways, but with different effectiveness from other Ras proteins but does not exclude the possibility of specific effectors (70, 84, 85).

Along with their individual functions, Ras proteins interact and cooperate with several other signaling cascades in tumorigenesis. One such pathway is the transforming growth factor-betas (TGF-beta) pathway that plays an essential role in both tumor initiation and tumor progression (104).

Transforming Growth Factor Beta

The transforming growth factor-betas (TGF-beta) are 25 kDa cytokines that play critical roles in processes including mammalian development and homeostasis, regulating a large number of biological processes including cell growth, migration, differentiation, ECM production, angiogenesis and immunity and tumor initiation and progression (104). TGF-beta is capable of influencing these various processes because of its ability to induce extremely variable cellular responses depending on cell type and stimulation context. Its versatility includes inhibiting epithelial cell growth and immune responses while promoting epithelial mesenchymal transition (EMT) and fibroblast activation (105). Normal epithelial and differentiated carcinoma cells are generally growth-inhibited by TGF-beta, but de-differentiated or Ras-transformed cells often grow increasingly malignant following TGF-beta treatment.

There are three TGF-beta ligands, TGF-b1, TGF-b2 and TGF-b3. TGF-b1 is the most commonly upregulated in tumor cells (106). TGF-beta signals via a heterotetrameric signaling complex comprising the type I and type II TGF-beta receptors

(TGFbRI and TGFbRII, respectively). When TGF-beta ligands bind to TGFbRII, TGFbRI is recruited and trans-phosphorylated (Figure 1.2). This leads to activation of the downstream effectors SMAD2 and SMAD3. The phosphorylated SMAD2/3 then binds with SMAD4 to translocate to the nucleus resulting in active gene transcription (107).

TGF-beta also activates many Smad-independent effectors such as MAPKs, PI3K,

AKT, mTOR, PP2A, Par6, PAK2 and c-Abl (108). Of these non-Smad effectors, studies show that PI3K plays a critical role in regulating both the inhibitory and stimulatory responses of TGF-beta. For example, the ability of TGF-beta to promote epithelial

12 growth arrest has been shown to require the SMADs collaborating with the FoxO family of transcription factors to upregulate various cyclin dependent kinase inhibitors (CDKIs)

(109, 110).

One intriguing aspect of TGF-beta proteins is the dual role they play in tumorigenesis. These proteins have the ability to act as tumor suppressors or tumor promoters depending on the points of disruption in the TGF-beta signaling pathway and the context of these disruptions (107, 111). In its role as a tumor suppressor, mutations in the genes encoding TGFbR1 and TGFbR2 or reduced expression or phosphorylation of other signaling pathway components have been found in human cancers (112). A lack of or reduction in the expression of TGF-beta receptors are correlated with poor prognosis. Additionally, mouse models with genetic deletions or reduced expression in

TGF-beta signaling have more malignant phenotypes (113). TGF-beta suppresses the initiation of tumors and early development by inhibiting cell cycle progression, induction of apoptosis, and suppression of growth factor, cytokine and chemokine expression.

As an oncogene, TGF-beta is produced in large quantities by many tumor types.

Decreased or altered TGF-beta responsiveness (mutated or lost receptors) and increased expression or activation of TGF-beta ligands is common as tumors progress.

High levels of TGF-beta ligands and SMAD activity is present in some aggressive, highly proliferative tumors and correlates with poor prognosis in patients. However, in some cancer type where the receptors are mutated, for example mutations in the

TGFbR2 in colon cancer, a favorable prognosis after adjuvant chemotherapy is observed (114, 115). Mouse models show that enrichment of TGF-beta signaling by expression of either a constitutively active TGFb1 or TGFbRI in mammary epithelial

13 cells increases pulmonary metastases, whereas systemic inhibition of TGF-beta signaling suppresses pulmonary metastases (116). The mechanisms of tumor promotion by TGF-beta include deregulation of cyclin-dependent kinase inhibitors, alteration in cytoskeletal architecture, upregulation of proteases and extracellular matrix formation, decreased immune surveillance and increased angiogenesis (111).

TGF-beta is also known for its effects on stromal cells. In addition to epithelial cells, TGF-beta signaling in stromal cells also exerts significant effects on tumor development and growth (107). TGF-beta causes cancer progression through paracrine effects on the tumor stroma. These effects of TGF-beta implicate stimulation of angiogenesis, escape from immunosurveillance and recruitment of myofibroblasts.

Crosstalk between Ras and TGF-beta Pathways

Both the Ras and TGF-beta signaling pathways are essential in tumorigenesis.

These two pathways have been shown to synergize depending on tumor stage. Ras signaling is implicated in TGF-beta’s conversion from a tumor suppressor to a tumor promoter, however the mechanism for this synergy is still not understood (117). Studies show that cancer cells can undergo an increase in malignancy and preserve a functional TGF-beta/SMAD core signaling, while tumor suppressive functions of the canonical TGF-beta/SMAD pathway are concurrently inhibited (118). Additionally, investigations show that Ras signaling mainly interferes with TGF-beta/SMAD signaling at the level of SMAD2/3 activation.

Numerous studies have shown synergy between the Ras/Raf/MAPK and TGF- beta/SMAD pathways in promoting tumor malignancy. Studies on keratinocyte,

14 mammary, prostate and hepatocyte carcinogenesis models showed a synergistic cooperation of TGF-beta and Ras to induce progression to undifferentiated, invasive tumors (119-122). Generation of metastatic adenocarcinomas has been shown in the intestinal epithelium by overexpressing K-Ras in combination with deletion of TGFRII.

Neither oncogenic K-RAS nor inactivation of TGFRII on its own was able to induce colorectal tumors (123). Cooperation of hyperactive Raf/MAPK and TGF-beta/Smad signaling has also been shown to be required for epithelial to mesenchymal transition

(EMT) and metastasis of mammary epithelial cells in vivo. However, activation of PI3K caused protection against TGF-beta-induced apoptosis and scattering of cells leading to tumorigenesis in the absence of metastatic colonization (117, 124).

The linker region of SMAD2/3 is phosphorylated at several serine/threonine sites through growth factor-mediated MAPK-ERK activation which results in cytoplasmic maintenance of SMAD2/3 and downregulation of TGF-beta signaling (125). In contrast, predominant localization of phosphorylated SMAD2/3 to the nucleus has been reported in invasive late-stage colorectal carcinoma (126, 127). Studies have also shown that mutations of the SMAD3 linker region prevent JNK-dependent phosphorylation resulting in a preserved tumor-suppressive function of TGF-beta and inhibition of tumor cell invasion (128).

Several SMAD-independent mechanisms have been reported in the literature involving Ras and TGF-beta pathways. R-RAS transformed and TC21 overexpressing

EpH4 mammary epithelial cells are insensitive to TGF-beta-mediated growth inhibition along with increased proliferation and malignancy in response to exogenous TGF-beta.

These effects of TGF-beta were mediated through Smad-independent mechanisms and

15 required the activation of TGF-beta-associated kinase 1 (TAK1) and its downstream effectors, the JNK/p38MAPK/PI3K/AKT and mTOR pathways (83, 129).

Studies have also demonstrated a mechanistic link between Ras, p53 and TGF- beta, since Ras-MAPK activity induces p53 N-terminal phosphorylation which enables the interaction of p53 with TGF-beta-activated SMADs and promotes TGF-beta- dependent cytostasis (130). Interestingly, the effects of activated Ras that lead to tumor growth are balanced by the wild type p53/SMAD cooperation maintaining TGF-beta growth control and therefore limit neoplastic transformation. A breast carcinoma study shows that Ras-activated mutant-p53 and TGF-beta cooperation counteracts the activity of p63 (131). These data provide understanding of the mechanism of the crosstalk between RAS and TGF-beta, since Ras signaling promotes mutant-p53 phosphorylation and is required for the formation of the mutant-p53/Smad complex. Therefore, Ras signaling is thought to play a significant regulatory role on the composition of co- activators and co-repressors of SMAD transcriptional complexes.

Investigations have also shown a role for other TGF-beta family members in tumor development and progression including activin A, bone morphogenic proteins

(BMPs), and nodal. As with all TGF-beta family members, the promotion or suppression of tumorigenesis depends on the stage and type of tumor. The TGF-beta family member activin A uses different receptors from TGF-beta but depends on SMAD proteins

(SMAD2/3/4) for its downstream signaling effects. Activin has shown a tumor suppressive role in breast (132), colon (133) and hepatocellular carcinoma (HCC) (134), and an oncogenic role in lung (135), prostate (136), esophageal cancer (137) and oral cancers (138).

In contrast to TGF-beta and activins, the BMP subfamily signals through the

SMAD1/5/8 of the canonical TGF-beta pathway. The Ras-Erk signaling cascade has been shown to enhance the transcriptional activity of SMAD1 in response to BMP signaling (139). SMAD-independent activities of BMPs have also been reported and include signaling through direct interaction of the cytoplasmic tail of BMP receptor II

(140, 141). Studies in colon, lung and non-small cell lung carcinoma (NSCLC) have shown a role for BMPs in tumor progression. Studies in colon cancer show frequent inactivation of the BMP signaling pathway by mutation of the receptors or SMAD proteins (142). In lung cancer, mutations in Ras proteins along with silencing of BMP expression has been demonstrated and NSCLC patients with K-Ras mutations were more likely to have epigenetically silencing of members of the BMP pathway than those with wild type K-Ras (143). The molecular mechanisms involving Ras and BMP signaling pathways is still not understood.

The Ras and TGF-beta pathways are two of the major pathways found to be deregulated in a variety of cancers. Therefore, a better understanding of the crosstalk between these two signaling cascades is needed and they remain attractive targets for cancer therapy. The success of targeting these pathways individually has been limited due to the toxicity of drugs. Targeting both pathways in cancer therapy may have promising results but the dual role of TGF-beta as both a tumor suppressor and tumor promoter needs to be further investigated. Both the deleterious and beneficial roles of

TGF-beta in tumor development and the complexity of its interaction with the Ras pathway which differs in cell type, tumor type and tumor stage complicate possible cancer therapies.

Figure 1.1: Ras effector pathways. Multiple effector pathways contribute to Ras function. Once activated, Ras proteins signal through multiple effector pathways, activating many different signal transduction pathways. These pathways are involved in proliferation, differentiation, apoptosis and senescence. Among all the pathways that are known to be activated by Ras the best characterized are the Raf/MAPK, PI3K and the Ral pathways.

Figure 1.2: TGF-beta Signaling. The TGF-beta ligands signal through TGF-b receptors

(TGFbRI and TGFbRII). Canonical signaling is through the phosphorylation of SMAD2 and SMAD3, that then combine with SMAD4 to enter the nucleus and mediate growth inhibition. TGF-beta binding to its receptors activates many non-canonical signaling pathways, including small GTPases (RhoA, PKN, and Rock), p38 kinase and PI3 kinase pathways.

Chapter 2

In vivo regulation of TGF-beta by the Ras protein TC21 revealed through the loss of the NF1 RasGAP.

Abstract

Ras proteins are involved in transforming growth factor -beta −mediated developmental pathways, and paradoxically in both tumor suppression and tumor progression. Whether specific Ras proteins integrate with TGF-beta signaling pathways in vivo is unknown. We activated all Ras proteins in vivo by absence of the Nf1 Ras-

GAP. In mice lacking both Nf1 and the Ras-related protein TC21/R-Ras2 benign neurofibroma formation was delayed. Conversely, TC21 loss in NF1 deficient models accelerated growth of brain tumors and sarcomas. This duality implicated TGF- beta signaling. We found that elevated TGF -beta expression in Nf1 Schwann cell precursors was reversed by TC21 loss. TC21 loss also blocked an Nf1/ TGFbRII/AKT dependent autocrine precursor survival loop and decreased precursor numbers, implying that delayed benign tumor formation resulted from effects on tumor initiation.

Increased NF1 -/- sarcoma size induced by TC21 loss was also TGF-beta dependent, and reversed by TGF-beta antibodies in xenografts. TGF-beta dependent effects on sarcoma were non-cell autonomous effects on endothelial cells and myofibroblasts, as

MPNST cells did not express TGFbRII. Gene expression analysis, RNA, and protein analyses confirmed increases in TGF-beta ligands and absence of TGFbRII in human malignant peripheral nerve sheath tumors. Thus, TC21 is a critical in vivo regulator of

TGF-beta.

Background

Ras proteins are molecular switches that cycle between an inactive GDP-bound form and an active GTP-bound form in which they signal through effector pathways including Ral, phosphoinositide 3-kinase (PI3K)-AKT and Raf-MEK-ERK to regulate cellular functions including proliferation, cell death and cell differentiation (144, 145).

Individual members of the Ras superfamily can have unique roles in diverse cell compartments (66, 92, 146, 147). The commonly studied Ras oncogenic proteins encoded by the H-Ras, K-Ras and N-Ras genes (Ras) are activated by mutation in up to 50% of human cancers (68). The related R-Ras family, encoded by the R-Ras,

TC21/R-Ras2 (subsequently TC21) and R-Ras3/M-Ras genes, also has oncogenic potential (69, 70). Here we focus on TC21, a poorly studied transforming oncogene mutated in human tumor cell lines (80) shown to induce lymphoma in vivo (148), in the context of the NF1 Ras GTPase activating proteins (GAP) neurofibromin.

Ras signaling is inactivated by GAPs, including the NF1 tumor suppressor protein neurofibromin. Neurofibromin is a GAP for all Ras proteins (41), so that sustained activation of each expressed Ras protein is predicted in cells that rely on neurofibromin function. Mutations in the NF1 gene result in neurofibromatosis type 1

(NF1) (149). NF1 patients develop disfiguring benign peripheral nerve tumors called neurofibromas; the plexiform neurofibroma can transform into sarcomas known as malignant peripheral nerve sheath tumors (MPNST), a leading cause of death in adults with NF1 (20). We reasoned that specific roles of individual Ras proteins would be revealed in the setting of NF1 loss in peripheral nerve cells. Consistent with this idea,

21 increased migration of Nf1 mutant Schwann cells was rescued by a dominant negative allele of TC21 (150).

Ras signaling is critical to activate a variety of downstream kinase cascades.

While mitogen-activated protein kinases (MAPKs) Erk1/2, c-Jun N-terminal kinase, and p38 SAPK MAPK act as downstream effectors of TC21 in certain cell lines (70), PI3K-AKT is currently believed to be a major TC21 effector (83, 91, 151). A recent in vivo study implicates TC21 in PI3K signaling downstream of the antigen receptor in T-cells (91). In addition to PI3K signaling, TC21 was recently linked to TGF-beta in vitro ; overexpression of an activated TC21 allele caused cells to lose responsiveness to the growth inhibitory effects of transforming growth factor beta (TGF-beta) (83).

A fascinating aspect of TGF-beta proteins is their ability to act as tumor suppressors or tumor promoters (107, 111). TGF-beta ligands 1, 2 and 3 regulate cell proliferation, cell death and cell differentiation through interaction with receptors TGFbRI

(ALK5) and TGFbRII, and the TGFbRIII co-receptor (104, 105). Downstream of TGF- beta receptors, activation of SMAD proteins and PI3K-AKT signaling are thought to control TGF-beta tumor suppression, while activation of RhoA, TAK1, and PI3K-AKT are implicated in TGF-beta oncogenesis (107). Decreased or altered TGF-beta responsiveness and increased expression or activation of TGF-beta ligands is common as tumors progress. For example, TGF-beta can promote tumorigenesis through sequestration of mutant p53 (131), and have non-autonomous effects on tumor stroma

(152).

We identify TC21 as a regulator of TGF-beta function in vivo . Previous studies examined crosstalk between the Ras and TGF-beta signaling pathways, mainly in vitro (118,

119, 124, 125, 153). Ras/MAPK activates the TGF-beta promoter (154, 155), and Ras/Erk signaling blocks SMAD translocation to the nucleus through phosphorylation of SMAD2/3

(125). In vivo, RAS/MAPK can phosphorylate p53 which then interacts with SMAD proteins

(130). TGF-beta can also activate Ras signaling by TGFbR-mediated phosphorylation of the

ShcA protein (153).

Neurofibromas and MPNSTs are believed to derive from neural crest lineage cells, more mature Schwann cell precursors, and/or from differentiated mature Schwann cells (34, 156). To study development, we crossed TC21 deficient mice (91) to Nf1 +/- mice (26). To study tumorigenesis, we crossed TC21 deficient mice to Nf1 fl/fl ;DhhCre mice that form neurofibromas, and to a sarcoma and brain tumor model, Nf1 +/-;Trp53 +/-

(NPCis), and used a xenograft model of human MPNST. We found that Nf1 mutation renders Schwann cell progenitors insensitive to TGF-beta mediated cell death and define a TGF-beta autocrine survival loop , which correlates with benign neurofibroma formation. In the absence of TC21, Nf1 mutants restore TGF-beta sensitivity and benign tumorigenesis is delayed. Conversely, loss of TC21 increases TGF-beta induced malignancy in the NPCis model and in NF1 MPNST xenografts, loss of TC21 accelerates tumor growth in a non-cell autonomous manner. These results suggest that

TC21 is a major regulator of TGF-beta production and functions on tumorigenesis in vivo .

Materials and Methods

Mice

We housed mice in a temperature- and humidity-controlled vivarium on a 12 hr dark-light cycle with free access to food and water. The animal care and use committee of Cincinnati Children’s Hospital Medical Center approved all animal use. TC21 and R-

Ras insertional mutant mice (denoted TC21 +/- and R-Ras +/-) were obtained from E.

Ruoshlati laboratory on a C57Bl/6/129 mixed background after four generations of backcross onto C57/Bl6 (90, 91). We maintained TC21 +/- and R-Ras +/- mice on a

C57BL/6 background, and bred them to homozygosity. We mated TC21 -/- and R-Ras -/- mice to Nf1 +/- mice on the C57/Bl6 background (26) to obtain TC21 -/-;Nf1 +/- and R-Ras -/-

;Nf1 +/- mice, which we intercrossed to obtain mutant embryos. TC21 -/- mice were also mated to Nf1 fl/fl ;DhhCre mice (34) to obtain TC21 +/-; Nf1 fl/+ ;DhhCre mice. F1 mice were mated to each other to obtain TC21 -/-;Nf1 fl/fl ;DhhCre mice and control littermates TC21 -/-

;Nf1 fl/+ ;DhhCre mice. We also bred TC21 -/- mice to NPCis C57BL/6 mice (36) to obtain

TC21 +/-;NPCis mice. These mice were mated with TC21 -/- to obtain TC21 -/-;NPCis mice.

TC21 +/-;NPCis littermates were bred to C57BL/6 wild type mice for parallel controls. For tumor experiments we analyzed male mice.

Cell culture and Reagents

Schwann cell precursor spheres: We dissociated dorsal root ganglia (DRG) from

E12.5 embryos with 0.25% trypsin (Mediatech, VA) for 20 min at 37°C and obtained single-cell suspensions with narrow-bore pipettes. We used trypan blue exclusion (Stem

Cell Technologies, Vancouver, BC) to identify live cells, and plated 2x10 4 live cells per well in 24-well low-binding plates (Fisher Scientific, PA) in sphere medium. Sphere medium (157): DMEM:F-12 (3:1) + 20 ng/ml rhEGF (R&D Systems, MN), 20 ng/ml rh bFGF (R&D Systems), 1% B-27 (Invitrogen, CA), and 2 mg/ml heparin (Sigma-Aldrich

Corp., MO). We maintained cultures at 37°C and 5% CO2. Five to seven days after primary spheres formed, we dissociated them in 0.05% trypsin-EDTA (Mediatech, VA) for 5 min at 37°C and plated cells at 2x10 3 cells/well in sphere medium. For experiments, we used cells at passage 2 or passage 3 and plated single cells at 500 cells per well in 24-well low-binding plates in sphere medium, and added inhibitors or antibody to cells 24 hours after plating. Spheres were counted 3 days later. For each experiment, we show a representative of at least three independent experiments. Anti-

TGF-beta antibody, rhTGFb1 and normal IgG were purchased from R&D Systems (MN) and were reconstituted as per manufacturer’s instructions. TGFbR1 inhibitor, SB

431542, MEK1 inhibitor, PD 98059 (Cayman Chemicals, MI), AKT inhibitor, MK-2206

(Selleck Chemicals, TX), p38 SAPK inhibitor, SB 203580 and the ROCK inhibitor, (Y-

27632) (Calbiochem, NJ) were dissolved in DMSO.

Mouse Schwann cells: Schwann cells were isolated from embryonic day 12.5 dorsal root ganglia using preliminary growth on neuronal axons as previously described

(158). Purified Schwann cells were cultured on poly-L-lysine-coated plates in DMEM with 10% fetal bovine serum, 10 ng/ml beta-HRG, and 2 µM forskolin (Calbiochem, NJ).

Cells were used between passages 1 and 3.

MPNST cells: The human NF1 deficient MPNST cell line S462TY was derived from the S462 cell line (159, 160). Cells were maintained in DMEM + 10% FBS and 1% penicillin/streptomycin.

Immunohistochemistry and Histology

We anaesthetized mice with Isoflurane by inhalation (Butler Animal Health

Supply). Anaesthetized neurofibroma bearing mice and mice with GEM-PNST were transcardially perfused with 0.9% saline followed by ice cold 4% paraformaldehyde

(w/v). We embedded sciatic nerves and tumor specimens in paraffin and cut 6 µm cross sections. E12.5 embryo hind limbs were fixed in 4% paraformaldehyde for 3 days and then transferred to 20% sucrose for a week before blocking in O.C.T (EM Sciences,

NJ). Frozen sections were cut at 12 µm and stored at -80°C.

For immunohistochemistry, following deparaffinization and rehydration of paraffin sections or post-fixing of frozen sections with 4% paraformaldehyde, sections were permeabilized with 0.2% TX 100 and blocked with 10% normal serum for one hour at room temperature. Primary antibodies were: Phospho-SMAD2(serine

465/467)/SMAD3(serine 423/425), phospho-AKT (serine 243), phospho 44/42 MAPK

(Cell Signaling, MA), anti-neurofilament and meca-32 (DSHB, IO). Secondary incubations used host-appropriate secondary antibodies. Neurofibroma sections were also submitted to the CCHMC Pathology Laboratory for histological characterization; hematoxylin and eosin for nuclear counts, toluidine blue for mast cells, and rabbit polyclonal anti-cow S100-beta (Dako, Carpenteria, CA).

Western analysis

We lysed cells, tissue and tumor sections in ice cold 0.1% NP40, 10% glycerol, salts, phosphatase and protease inhibitors as described (161). Tumors were lysed using a tissue ruptor (Qiagen, CA). Proteins were separated on 4 – 20% TrisHCl acrylamide gels (Biorad, CA) and transferred to PVDF membranes (Millipore, MA). Membranes were probed with antibodies for TC21 (Abnova, Taipei, Taiwan), phospho-AKT, phospho-S6K, phospho-SMAD2/3, phospho 44/42 MAPK and beta-Actin (1:1000) (Cell

Signaling). We detected signals using horseradish peroxidase-conjugated secondary antibodies (BioRad) and the ECL Plus developing system (Amersham Biosciences, NJ).

Quantitative real-time reverse transcription-PCR

Total RNA was isolated from tumors, spheres and sciatic nerves using an

RNeasy kit (Qiagen) and used as a template for cDNA synthesis (Invitrogen Superscript

III). Triplicate reactions were used to perform qRT–PCR (ABI 7500 Sequence Detection

System, CA) as described (162). Values for individual genes of interest were normalized to values for GAPDH (mouse samples) or beta-actin (human samples) and used to calculate fold change in gene expression by the ∆∆ ct method.

Lentiviral infection

For lentiviral shRNA infection, S462TY MPNST cells at 50% confluence were infected with lentiviral particles containing TRIPZ shRNAs-targeting TC21 or non-target

27 control (Open Biosystems, AL). The Cincinnati Children’s Hospital Medical Center

(CCHMC) Viral Vector Core produced virus using a 4-plasmid packaging system.

Lentiviral particles were incubated with S462TY MPNST cells in the presence of polybrene (8 µg/ml; Sigma-Aldrich) for 72 hours (three rounds) followed by selection in puromycin (Sigma-Aldrich) at a concentration of 2.5 µg/ml, killing uninfected cells within

3 days. Cells were then maintained in media containing puromycin (2.5 µg/ml).

Doxycycline (MP Biomedicals, OH) (2 µg/ml) induced shRNA expression.

Mouse xenograft

We injected 2.3x10 6 S462TY MPNST cells in a total volume of 150 µl in 30% matrigel (BD Biosciences, MD) into each flank of 5- to 6-week-old female athymic nude

(nu/nu) mice (Harlan, IN). In this model, tumors form 3 – 4 weeks after injection and mice require sacrifice at about 6 weeks after injection. Left flanks were injected with shTRIPZ non-target control and right flanks were injected with shTRIPZ TC21 cells in each animal. Mice were maintained on 1875ppm doxycycline feed (Test Diet, IN) to maintain expression of the shRNA. We measured tumors and weighed mice twice weekly once tumors began enlarging. Tumor volume was calculated according to the following formula: L * W 2 (π/6), where L is the longest diameter and W is the width. Mice were sacrificed before tumor size reached 10% body weight (~3,000 mm 3). For treatment with anti-TGF-beta, we injected mice intraperitoneally with 3mg/kg of normal rabbit IgG or anti-TGF-beta antibody (R&D Systems) every two weeks. We dissected tumors from anaesthetized mice and either flash froze them and stored them at -80°C,

28 or fixed them in 4% paraformaldehyde for histology. Fixed tumors were submitted to the

CCHMC Pathology Laboratory for paraffin embedding and sectioning.

Human model sample preparation, microarray hybridization

The human NF1 tumor data set consisted of 3 normal nerves, 13 dermal neurofibromas, 13 plexiform neurofibromas and 6 MPNSTs; all samples (except for the normal nerves) have been previously described (163). Expression profiles were generated using Affymetrix HG-U133 plus 2 and MOE430 2.0 oligonucleotide microarrays. Affymetrix Microarray Suite 5.0 was used to generate ‘CEL’ files for each sample that were normalized using the Robust Multichip Analysis (RMA) algorithm as implemented in Bioconductor/R. Affymetrix probes were remapped to RefSeq genes

(version 11.0.1). Comparisons and data visualization were performed using

GeneSpring GX v7.3.1 (Agilent Technologies).

Results

Loss of TC21 extends the survival of genetically engineered mouse (GEM) neurofibroma bearing mice but decreases survival of NPCis mice predisposed to malignancy.

To define roles for the Ras-related protein TC21, we used TC21 mutant mice

(henceforth TC21 -/-) with an insertion in the first intron of TC21 (Figure 2.1A). We examined numerous organs from adult TC21 -/- mice as well as in peripheral nerve

Schwann cells cultured from embryonic mice (Figure 2.1B, 2.1C) for the presence of

TC21 mRNA and protein. TC21 mRNA was reduced at least 10-fold and protein expression was lost, supporting the finding of Delgado et al. (91) that this TC21 mutation functions as a null allele.

To assess the potential role of TC21 in NF1 signaling, we crossed TC21 -/- mice with Nf1 +/- mice. Nf1 -/- embryos die by embryonic day 12.5 (26). In contrast, 90% of

TC21 -/-;Nf1 -/- embryos survived to embryonic day 14.5 (E14.5) and 10% of TC21 -/-;Nf1 -/- embryos survived to E16.5 (Table 1). The TC21 -/-;Nf1 -/- embryos showed heart defects common in Nf1 -/- embryos, indicating that TC21 is unlikely to be responsible for this phenotype. The partial rescue of embryo viability was specific to the loss of TC21 as it was not observed in R-Ras -/-;Nf1 -/- embryos on the same genetic background (Table 1).

Thus TC21 plays a role in Nf1 phenotypes, but other Ras proteins likely account for some aspects of embryonic development.

To determine whether TC21 is relevant to tumorigenesis, we generated

Nf1 fl/fl ;DhhCre; TC21 -/- mice and NPCis; TC21 -/- mice. Nf1 fl/fl ;DhhCre mice form benign

30 neurofibromas subsequent to Nf1 gene conditional inactivation by Cre recombinase in

Schwann cell precursors beginning at E12.5. Nf1 fl/fl ;DhhCre mice die due to neurofibroma compression of the spinal cord beginning as early as five months of age

(34). Mice with Nf1 and p53 mutations in cis on mouse chromosome 11 serve as a useful model (30, 35) of soft tissue sarcomas and brain tumors with histology of glioblastoma multiforme (36). We use the NPCis terminology for these mice.

Loss of TC21 significantly extended survival in Nf1 fl/fl ;DhhCre mice (p<0.0001;

Figure 2.2A). Thus, Nf1 fl/fl ;DhhCre; TC21 -/- (n=15) mice survived up to 20 months while littermate controls, Nf1 fl/fl ;DhhCre (n=14), were all dead by 15 months. Mice required sacrifice due to lethargy, weight loss, and/or dehydration secondary to paralysis that correlated with tumor formation and spinal cord compression in this model system. A second cohort of mice (n= 15) was analyzed and showed identical results (data not shown).

To test for effects of TC21 loss in the NPCis model, we generated NPCis; TC21 -/- mice. When NPCis; TC21 -/- mice (n=13) were aged using littermates as controls (n=7), the NPCis;TC21 -/- mice all died by 7 months while the littermate controls survived up to

13 months. The reduced life span of NPCis; TC21 -/- mice was statistically significant

(p=0.0001; Figure 2.2B). Therefore, loss of TC21 in benign tumors extends survival, while paradoxically in a model of aggressive tumors, loss of TC21 decreases survival.

As described below, NPCis; TC21 -/- mice died early due to rapid formation of aggressive brain tumors (Figure 2.8).

A role for TC21 in Nf1 tumor initiation

We evaluated paraffin sections from neurofibromas by hematoxylin and eosin

(H&E) staining for morphology. We used anti-S100-beta staining to mark Schwann cells and toluidine blue staining for mast cells to confirm grade 1 GEM neurofibroma histology per defined criteria in groups of Nf1 fl/fl ;DhhCre mice with or without TC21 (164)

(Figure 2.3A). Surprisingly, there was no significant difference in the number or size

(diameter) of the neurofibromas in these mice at the time of sacrifice (Figures 2.3B,

2.3C).

An alternative explanation for how loss of TC21 extends survival in Nf1 fl/fl ;DhhCre mice is that TC21 diminishes numbers of neurofibroma-initiating or neurofibroma- sustaining cells. Multi-potent self-renewing cells have been identified in developing dorsal root ganglia (DRG) and neurofibromas and proposed to contribute to neurofibroma initiation or growth (37, 165). To test whether loss of TC21 affects these cells, we used an in vitro model system. We quantified defined numbers of Schwann cell precursors from E12.5 DRG after plating primary cells at clonal density. Nf1 -/- DRG cells give rise to more spheres than do wild-type or Nf1+/- DRG cells, and Nf1 -/- DRG sphere cells form neurofibroma-like lesions upon xenotransplantation (165). We hypothesized that TC21 -/-;Nf1 -/- DRG cells would give rise to significantly fewer spheres than Nf1 -/- cells if TC21 affected tumor initiation. Cells from Nf1 -/- DRGs formed significantly more primary spheres than cells from wild type (WT) DRG. Importantly,

TC21 -/-;Nf1 -/- DRG cells formed WT levels of spheres (p < 0.0001; Figure 2.3D) consistent with a role of TC21 regulating numbers of tumor-initiating cells early in Nf1 tumorigenesis.

A TGF-beta autocrine loop in Nf1 -/- Schwann cell precursors prevents cell death

Given the duality of TGF-beta proteins in their ability to act as tumor suppressors or tumor promoters and our observation that TC21 acts similarly (TC21 loss enhanced survival of mice with benign neurofibromas yet led to more rapid death in the NPCis model (Figure 2.2)), and because TC21 and TGF-beta have been linked in vitro (149), we examined TGF-beta signaling in the context of Nf1 and/or TC21 null alleles. We analyzed Schwann cell precursors for effects of Nf1 loss that require TC21. Cells from secondary spheres were plated at clonal density and tested for response to TGFb1. WT

Schwann cell precursor numbers were reduced by exposure to TGFb1 in a dose dependent manner. These cells die as read out by TUNEL staining of fixed paraffin embedded sections of spheres (Figure 2.4A). In contrast, Nf1 -/- Schwann cell precursors did not die when exposed to TGFb1. The loss of TC21 restored sensitivity to TGFb1 (p

= 0.005; Figure 2.4B).

We considered the hypothesis that Nf1 -/- Schwann cell precursors become dependent on TGF-beta for their survival. To test this hypothesis, we examined the mRNA expression levels of TGFb1 in sphere cells using quantitative real time PCR. Nf1 -/- Schwann cell precursors had significantly higher levels of TGFb1 compared to WT precursors (p= 0.04;

Figure 2.4C). Importantly, upon losing TC21 in an Nf1 -/- background, TGFb1 levels decreased to levels found in WT spheres. TGFb1 protein in media conditioned by Schwann cell precursor spheres was also 20-fold higher in Nf1 -/- Schwann cell precursors compared to WT and TC21 -/-

;Nf1 -/- precursors (p = 0.01; Figure 2.4D). To test if TGF-beta produced by Nf1 -/- spheres was functional, we treated Nf1 -/- Schwann cell precursors plated at clonal density with a function- blocking anti-TGF-beta antibody (166). Nf1 -/- Schwann cell precursors treated with IgG control

33 formed healthy spheres (Figure 2.4E), but Nf1 -/- Schwann cell precursors treated with anti-

TGF-beta antibody formed many fewer spheres (Figure 2.4F, 2.4G). In contrast, treatment of

WT or TC21 -/-;Nf1 -/- Schwann cell precursors with concentrations of anti-TGF-beta antibody that diminished Nf1-/- Schwann cell precursors sphere formation did not alter the number of spheres formed (p < 0.0001; Figure 2.6A). Thus production of TGF-beta and effects of function blocking antibody support the idea that Nf1 -/- Schwann cell precursors uniquely secrete TGF- beta which enhances their survival. The data support the hypothesis that Nf1 -/- Schwann cell precursors are dependent on TGF-beta for survival.

We tested if TGFb1 mRNA is also elevated at a later stage of Schwann cell differentiation or in neurofibromas. TGFb1 mRNA expression was only slightly (2-fold) elevated in sciatic nerves from Nf1 fl/fl ;DhhCre mice compared to WT sciatic nerve

(Figure 2.5A); TGFb1 mRNA was unchanged in wild type levels compared to

Nf1 fl/fl ;DhhCre;TC21 -/- sciatic nerve. There was no significant difference in TGFb1 levels between the neurofibromas from Nf1 fl/fl ;DhhCre (n=8) and Nf1 fl/fl ;DhhCre;TC21 -/- (n=10) mice (Figure 2.5B). Therefore, the elevation of TGFb1 mRNA and the rescue in TGFb1 mRNA expression levels seen upon loss of TC21 in an Nf1 -/- background is most pronounced early in Schwann cell development.

Survival of Nf1-/- Schwann cell precursors is dependent on TGF-beta and AKT

To determine the mechanism through which TGF-beta affects survival of Nf1 -/-

Schwann cell precursors, we tested the effects of specific inhibitors. Treatment of Nf1 -/-

Schwann cell precursors with a TGF-beta receptor 1 (TGFbR1) inhibitor (SB 431542),

34 inhibited sphere formation and was without effect on the formation of WT and TC21 -/-

;Nf1 -/- spheres (p < 0.0001; Figure 2.6B). The PI3K/AKT pathway can be important in cell survival and can be a downstream effector of TGF-beta signaling. Similar to the treatment with either an anti-TGF-beta antibody or the TGFbR1 inhibitor, treatment with the AKT inhibitor (MK02206) reduced Nf1 -/- sphere formation significantly (p < 0.0001;

Figure 2.6C). In contrast, there was no effect of inhibitors of other pathways downstream of Ras or TGF-beta. Thus, neither the MEK1 inhibitor PD 98059, the

ROCK inhibitor Y-27632 nor the p38 SAPK inhibitor SB 20358 affected Nf1 -/- sphere formation (Figure 2.7A). Efficacy of inhibitors was validated by western blotting (data not shown). Confirming a role for AKT signaling in a TGF-beta autocrine loop, we treated spheres with a shRNA targeting AKT and found that compared to a non-target control

(NT), the formation of Nf1 -/- spheres was significantly reduced (p < 0.001; Figure 2.6D).

The shAKT had no discernible effects on WT or TC21 -/-;Nf1 -/- sphere formation.

Altogether, this data shows the formation and survival of Nf1 -/-spheres in vitro are dependent on TC21, TGF-beta, TGFbRI and AKT.

To test if the loss of TC21 affects either AKT signaling or TGFbR-SMAD signaling in vivo we used immunochemistry. Schwann cell precursors are present in

E12.5 DRG and peripheral nerve. These cells are the targets for Nf1 loss in the

Nf1 fl/fl ;DhhCre model. We examined the levels of AKT signaling by staining with anti- phospho-AKT in E12.5 DRG tissue sections. Nf1 -/- DRG showed elevated levels of phospho-AKT as compared to WT and TC21 -/-;Nf1 -/- embryos (Figures 2.6E, 2.6F). Nf1 -

/- spheres from E12.5 DRG grown in vitro also contained increased phospho-AKT as compared to WT and TC21 -/-;Nf1 -/- spheres (Figure 2.7B). In contrast, phospho-

SMAD2/3 staining, which marks canonical TGF-beta signaling, was similar across genotypes (Figures 2.7C, 2.7D). To test if phospho-AKT or phospho-SMAD is altered in neurofibromas we stained neurofibromas with the anti phospho-AKT or anti phospho-

SMAD2/3. Levels of phospho-AKT and phospho-SMAD2/3 expression were similar in neurofibromas from Nf1 fl/fl ;DhhCre and Nf1 fl/fl ;DhhCre;TC21 -/-mice (data not shown), consistent with the idea that TC21 affects events shortly after Nf1 loss. Taken together our data suggest that the effect of loss of TC21 in an Nf1 -/- background is pronounced in

Schwann cell precursors. The high expression of phospho-AKT in Nf1 -/- Schwann cell precursors in vivo is consistent with the finding that cultured Nf1 -/- precursors are resistant to cell death.

To test if AKT is necessary for TGFb1 mRNA production in Schwann cell precursors, we examined the levels of TGFb1 mRNA in Nf1 -/- Schwann cell precursors treated with the AKT inhibitor (MK-2206), and found that this inhibitor decreased TGFb1 mRNA levels (Figure 2.6G).

Taken together, these data support a model based on the existence of a TGF-beta autocrine survival loop involving AKT in Nf1 mutant Schwann cell precursor cells, as shown in Figure

2.6H. Presence of the autocrine loop is correlated with decreased survival of neurofibroma- bearing mice, and the autocrine loop is lost in the absence of TC21. We propose that the absence of the autocrine loop causes the observed delay in benign tumor formation.

Loss of TC21 in NPCis mice results in aggressive brain tumors

In the Nf1-driven model of GEM-sarcoma and GEM-GBM, loss of TC21 causes early lethality (Figure 2.2). Many NPCis; TC21 -/- mice showed circling behavior

36 consistent with possible brain tumors. Histological analysis of paraffin embedded brains from these mice showed that 12 of 13 NPCis; TC21 -/- mice had GEM brain tumors

(Figure 2.8A), whereas the NPCis controls did not develop brain tumors and instead died of sarcomas. GEM brain tumors in NPCis; TC21 -/- mice were grade 2 or 3 as defined by Reilly et al. (36). Thus, tumors showed nuclear pleomorphism with spindle cells, increased mitotic activity, no sign of necrosis and were S100-beta positive (Figure

2.8B).

We examined the same mice for sarcomas. NPCis; TC21 -/- mice had fewer sarcomas than NPCis controls (Figure 2.8C). We defined sarcomas in paraffin sections as GEM-PNST by staining with H&E, Ki67 and S100-beta staining (Figure 2.8D).

Histological analysis of the sarcomas of both genotypes showed that they were all grade 3 tumors, with high mitotic activity, scattered S100-beta cells and adjacent nerve bundles. We conclude that the absence of TC21 leads to rapid formation of grades 2 and 3 brain tumors and TC21 acts as a suppressor of tumorigenesis in a Nf1 +/-;p53 +/- background.

Loss of TC21 in MPNST cells increases tumor growth TGF-beta levels.

To investigate the role of TC21 in sarcomas and to provide a model for mechanistic studies, we infected the NF1 patient derived S462TY MPNST cell line

(subsequently TY) with a stably expressing doxycycline-inducible TRIPZ lentivirus encoding shRNA targeting TC21 or control TRIPZ non-target shRNA. Cells were injected subcutaneously into nu/nu mice and observed for tumor growth. Mice with

MPNST cells expressing TC21 shRNA consistently had larger tumors than control mice with MPNST cells expressing the non-target shRNA (Figures 2.9A, 2.9B). Western blot analysis confirmed that TC21 protein expression remained low in TC21 shRNA tumors at the end of the experiment (Figure 2.9C). Thus, loss of TC21 in GEM brain tumors and in a human sarcoma xenograft increases tumor growth.

In spite of the increased tumor size in MPNST xenografts expressing shTC21, histological analysis of MPNST xenograft tumors revealed no obvious difference in morphology (H&E) or cell proliferation (Ki67) compared to control tumors (data not shown). Closer examination revealed changes in tumor vasculature; we examined tissue sections for blood vessels, as TGF-beta is known to induce blood vessel formation. Counting numbers of blood vessels per area in sections stained with the epithelial marker meca-32 revealed that shTC21 tumors had significantly more vessels per mm 3 than the non-target controls (p=0.0001; Figure 2.9D). This increased tumor growth upon loss of TC21 correlates with change in tumor vasculature. Finally, consistent with altered tumor stroma upon loss of TC21, an antibody against smooth muscle actin (SMA) to detect myofibroblasts showed an increase in immunoreactivity in tumors from MPNST xenografts expressing shTC21 (Figures 2.9E, 2.9F).

Increased MPNST growth due to loss of TC21 requires TGF-beta.

To investigate whether TGF-beta plays a causal role in malignant tumor growth when TC21 is lost as it does in benign tumors, we treated mice harboring MPNST xenografts with an anti-TGF-beta antibody that blocks the function of TGFb1, 2 and 3 in

38 vivo (166). Mice were subcutaneously injected with S462TY MPNST cells expressing

TC21 shRNA or a non-targeting control. After tumors grew to an approximate size of

250 mm 3, rabbit IgG control or anti-TGF-beta antibody was administered by I.P. injection. Six weeks after the injection of mice with the IgG control antibodies, tumors from mice injected with MPNST cells expressing shTC21 remained larger than those injected with the non-target control cells (p=0.004; Figure 2.10A). Strikingly, mice injected with the anti-TGF-beta antibody had tumors of similar size whether the cells were expressing shTC21 or non-target control (Figure 2.10B). Western blotting analysis of tumors from mice xenografted with shRNA TC21 tumors injected with either IgG or anti-TGF-beta antibody showed sustained reduced levels of TC21 (data not shown).

Thus, loss of TC21 in NF1 -/- MPNST cells increases tumor growth, and this aggressive growth requires TGF-beta.

In order to delineate signaling pathways modulated by the TGF-beta antibody, we first examined the expression of phospho-SMAD, an immediate downstream effector of TGF-beta, in tumor sections. Phospho-SMAD was reduced in tumor cells and tumor stroma from mice treated with the anti-TGF-beta antibody compared with those treated with IgG (Figures 2.10C, 2.10E). Because TC21 -/-;Nf1 -/- Schwann cell precursors show reduced levels of TGF-beta and reduced phospho-AKT (Figure 2.9), we hypothesized that in the malignant setting with increased TGF-beta expression, we might also detect increased phospho-AKT expression. Indeed, phospho-AKT was elevated in the MPNST expressing shTC21, and phospho-AKT levels were subsequently reduced upon treatment with the anti-TGF-beta antibody (Figures 2.10D, 2.10E).

Tumor sections were also stained with the epithelial marker meca-32. The

39 shTC21 tumors treated with anti-TGF-beta had significantly fewer vessels than those treated with IgG (p < 0.0001; Figure 2.11A). Together this data suggests that in the NF1 -

/- setting, in transformed cells, loss of TC21 drives tumorigenesis in a TGF-beta dependent manner; TGF-beta produced by MPNST cells acts in a non-autonomous fashion on stromal cells to increase blood vessels and promotes formation of myofibroblasts.

MPNST cells express TGF-beta and lose TGFbRII

To begin to clarify whether cell autonomous and/or non-cell autonomous effects underlie changes in tumor stroma we first determined if altered TGF-beta ligands or

TGF-beta receptors are a general feature of neurofibromas or MPNST. We carried out this analysis because TGF-beta is frequently increased in expression in malignant tumors, while TGFbRs are frequently lost in malignancy (167). In this case, TGF-beta mainly has effects on tumor stroma. We analyzed the relative abundance of TGF-beta ligands and receptors in a large panel of human neurofibromas and MPNSTs compared to cultured neurofibroma Schwann cells and MPNST cell lines, cultured normal human

Schwann cells and normal human nerves by Affymetrix gene expression analysis. This data set allowed analysis of a large number of samples. The heat map (Figure 2.12A) displays fold changes between primary normal Schwann cells (NHSC, n=10), nerve

(Nerve, n=3), NF-1 derived primary benign neurofibroma Schwann cells (NFSC, n=22), benign neurofibroma (NF, n=22), malignant peripheral nerve sheath tumor cell lines

(MPNST cell lines, n=13), and MPNST tumor (MPNST, n=6) samples that have been

40 described previously (163). Differentially-expressed genes were identified by comparing

NFSC, NF, MPNST cell, MPNST groups to NHSC (fold change cutoff = 1.5, t-test p- value < 0.05 with FDR (Benjamini-Hochberg) correction).

Prominent changes included up-regulation of TGFb3 in most neurofibromas and

MPNST, with TGFb2 up-regulation specifically in 50% of neurofibroma Schwann cell samples, 50% MPNST cell lines and all human MPNST. TGFbRII was down regulated in all human MPNST, and in a subset (22/48) of neurofibromas and in neurofibroma

Schwann cell samples (Figure 2.12A). TGFbRI mRNA was slightly up-regulated but did not reach statistical significance across sample types. These data are consistent with data in many tumor types showing that during progression to malignancy, TGF-beta ligands increase and receptors decrease (107).

To confirm the increased expression of TGF-beta in MPNST and to test if TGF- beta expression is regulated by TC21 in sarcoma cells as it was in Schwann cell precursors, we examined expression of TGF-beta mRNA and protein in NF1 human

MPNST cells and in the sarcomas from NPCis mice. In benign neurofibromas and

Schwann cell precursors, loss of TC21 resulted in a reduction in TGF-beta mRNA levels in the Nf1 deficient setting. In contrast, loss of TC21 in Nf1 mutant NPCis mouse tumors resulted in an elevation of TGFb1 mRNA (p=0.01; Figure 2.11B). There was no significant difference in TGFb2 expression in these tumors (Figure 2.11C). In human

NF1 mutant MPNST xenotransplanted tumors TGFb2 mRNA was expressed, and levels were significantly elevated in tumors from MPNST cells expressing shTC21 (p=0.01;

Figure 2.12B); TGFb1 was unchanged and only a trend toward significance for increased expression of TGFb3 mRNA (Figures 2.11D, 2.11E). To confirm that the

TGFb2 protein was produced, we examined the concentration of TGFb2 in the media of

S462TY MPNST cells expressing a non-target shRNA or shRNA targeting TC21.

Knockdown of TC21 in these NF1-deficient MPNST cells caused a dramatic 80-fold increase in secreted levels of TGFb2 (p = 0.03; Figure 2.12C). The pathway leading to up-regulation of TGF-beta mRNA remains uncertain. Blockade of the MEK, AKT, p38 and ROCK kinases failed to alter TGF-beta mRNA levels in MPNST cells (data not shown).

TGFbR receptors are frequently mutated or lost in malignancy (167). If TGF-beta receptors show loss of function in NF1, then TGF-beta effects would be unlikely to be cell autonomous. While expression of TGFbRI was similar to the control nerve, levels of

TGFbRII mRNA were decreased compared to normal nerve. TGFbRII protein was undetectable by western blotting in S462TY cells compared to an immortalized human

Schwann cell line (Figure 2.12D). In 2 of 3 additional human MPNST cell lines, TGFbRII protein was also reduced (Figure 2.12F). The absence of TGFbRII in S462TY cells predicted that the cells would not respond to TGF-beta . Indeed, no effects of TGF-beta on MPNST survival or migration were detected (not shown). To further verify this finding, we examined the expression of 3 common TGF-beta target genes, p21, p15 and c-myc (152). No significant change in expression of these genes was detected after addition of TGF-beta to S462TY cells in serum-free medium at 30 minutes or 8 hours, as assessed by mRNA levels in comparison to untreated cells (not shown).

The schematic (Figure 2.12E) summarizes data using MPNST cells. When TC21 is lost in NF1 mutant cells, TGF-beta mRNA and secreted protein increase. Because

MPNST cells lack significant amounts of TGFbRII, most TGF-beta affects surrounding

TGFbRII expressing endothelial and stromal fibroblasts, resulting in increases in blood vessels and conversion of fibroblasts to myofibroblasts and thus tumor growth.

Discussion

The absence of Nf1 is predicted to activate all Ras proteins expressed in neurofibromin- dependent cells. Using Nf1 models enabled the demonstration that the Ras related protein

TC21 critically regulates TGF-beta in neurofibroma and MPNST. Loss of TC21 in the setting of

Nf1 deficiency delayed formation of benign neurofibromas, accelerated formation of aggressive brain tumors and nerve sarcomas, and uniquely regulated expression of TGF-beta. Our identification of a specific role for TC21 is consistent with recent studies showing differential localization of other Ras proteins (147), and different embryonic survival following knockout of the K-Ras mutants, H-Ras and N-Ras genes (68).

Our in vivo studies demonstrated that TC21 functions as an oncogene in benign tumors, and revealed a new role for TC21 as a tumor suppressor in nervous system malignancy.

However, TC21 was identified as up-regulated in oral and esophageal carcinomas, breast cancer cell lines (78, 79, 81), and activation of TC21 induced lymphoma in vivo (148).

Therefore, effects of TC21 activation are likely to be cell type dependent. Mechanistically, we showed that PI3K/AKT is a critical effector of TC21. In all NF1 models tested, TC21 regulated

TGF-beta , and TGF-beta acted as an oncogene.

Loss of TC21 in GEM neurofibroma extended mouse survival without affecting neurofibroma number or size, suggesting effects on early stages of tumor formation.

Consistent with a critical role for TC21 early in neurofibroma formation, we identified an autocrine survival loop specific to Nf1 -/- Schwann cell precursors, and dependent upon TGF- beta, TGFbR, and the PI3K/AKT pathway. We have not excluded additional effects on more mature Schwann cells. TGFb1 can kill mature wild type Schwann cells as well as more

44 immature cells (168, 169)(data not shown), and as in Nf1 -/- Schwann cell precursors, adult Nf1 -

/- nerves contained increased levels of TGFb1 mRNA dependent on TC21. We failed to detect elevated levels of TGFb1 mRNA in neurofibromas, but levels of TGFb1 mRNA in Schwann cells might have been obscured by expression in neurofibroma mast cells or other cells in the tumor microenvironment (170).

Additional evidence supporting a role for TC21 early in neurofibroma formation comes from clonality assays. Numbers of Nf1 -/- primary sphere-forming cells, after acute inactivation at

E12.5 using the DhhCre allele, were reduced by loss of TC21. Nf1 -/- primary sphere-forming cells were dependent on TGF-beta that they produced for their own survival. Secreted TGF- beta is predicted to enhance numbers of developing Nf1 -/- stem/progenitor cells and thus tumorigenic potential, leading to neurofibroma formation. Consistent with this idea, targeted loss of the TGF-beta type II receptor in Schwann cells suppressed early Schwann cell death and proliferation (171). We conclude that in vivo TGF-beta controlled by TC21, acts as a brake upon the growth and the number of developing Schwann cells, and thereby as a brake on neurofibroma initiation/growth.

Precursor cell effects on neurofibroma initiation and/or growth through TC21 and TGF- beta were not sufficient to prevent tumorigenesis. Other Ras proteins and/or Nf1 functions are likely necessary for neurofibroma growth. The idea that Nf1 regulates other Ras proteins— even in peripheral nerve cells— is supported by experiments in which farnesyl transferase inhibitor which predominantly blocks H-Ras blocks Nf1 mutant Schwann cell proliferation (172) and a report that N-Ras plays a major role in MPNST cells (173). Our finding that loss of TC21 delayed Nf1 -/- embryonic lethality but did not rescue embryos to birth is also consistent with roles of multiple Ras proteins in development. How each Ras protein contributes to embryonic

45 development and tumorigenesis downstream of NF1 loss of function remains to be determined. In addition, other Ras proteins may regulate TGF-beta expression or signaling in non-NF1 tumor settings. The diverse regulation of TGF-beta pathways in specific cell types may result in part because specific Ras proteins are cell type specific (174).

We found that loss of TC21 enhances tumorigenesis in malignancy. In the NPCis mouse model, loss of TC21 dramatically accelerated formation of brain tumors. In a xenograft model, NF1-deficient sarcoma cells showed accelerated tumor growth and increased levels of

TGF-beta when they expressed shTC21. Furthermore, inhibiting TGF-beta blocked the elevated tumor growth caused by loss of TC21. The action of TGF-beta as a growth promoter in malignancy correlated with increased TGF-beta ligands in neurofibroma and MPNST and decreased expression of TGFbRII in human MPNST cell lines and tumors, as shown by gene expression and confirmed by protein analysis in cell lines. Genomic mutation or loss of one or more TGF-beta receptors is common in many tumor types (111, 112). Despite absence of

TGF βRII in S462TY cells, blocking TGF β led to decreased expression of phospho-SMAD. This may result from use of a mutant p53 SMAD complex (131) and/or more complex activation of

SMAD downstream of activin receptors. One possibility is that stromal cells secrete activins or

BMPs that indirectly alter AKT and SMAD phosphorylation in tumor cells (107).

Our data are consistent with important non-cell autonomous effects of MPNST cell produced TGF-beta. This interpretation is consistent with increased blood vessels per area, and increased SMA (myo-fibroblast) expression upon down-regulation of TC21, and our finding that both phenotypes were blocked by anti-TGF-beta antibody. A non-cell autonomous effect of TGF-beta in malignancy has been documented previously in models of carcinoma, using genetic loss of TGFbRII (152).

In summary, TC21 has a dual effect: in developing Nf1 progenitor cells this Ras-related protein promotes cell survival by promoting TGF-beta production and formation of an autocrine survival loop, so that loss of TC21 results in decreased precursors and delayed tumor formation. In MPNST cells, loss of TC21 dramatically increases TGF-beta production, increasing vascularization and tumor growth. Our results linking TC21 to regulation of TGF- beta production are likely to be relevant to cancer in general, as NF1 mutations are being increasingly reported in sporadic cancers (175-177).

Figure 2.1: Insertion into the mouse TC21 gene causes a null allele. (A) Model illustrating the retroviral insertion site of the TC21 gene in intron 1. The positions of primers used for genotyping are shown. (B) Gel electrophoresis of the quantitative real time PCR across TC21 coding sequence. (C) Western blots using anti-TC21. TC21 protein was absent in tissues from TC21 -/- animals. The loading control for heart was H-

Ras rather than beta-actin. (D) Western blots of TC21 protein using Schwann cell lysates from wild type and TC21 -/- mouse tissues.

Table 1: Loss of TC21 partially rescues Nf1 mutant embryonic lethality. Mice of specified genotype and strain were mated, with the day of the copulatory plug designated E0.5. Pregnant females were sacrificed and embryos removed at embryonic day 12.5, 14.5 or 16.5 and genotyped for Nf1, TC21, and R-Ras. The TC21 -/-;Nf1 -/- embryos showed heart defects common in Nf1 -/- embryos, indicating that TC21 is unlikely to be responsible for this phenotype.

Figure 2.2: Loss of TC21 extends survival of neurofibroma-bearing mice but decreases survival of NPCis mice. (A) Kaplan-Meyer survival curves for

Nf1 fl/fl ;DhhCre mice (n=14); open circles (O) and Nf1 fl/fl ;DhhCre ;TC21 -/- mice (n=15), closed squares ( ■) (log-rank test, p < 0.0001). (B) Kaplan-Meyer survival curves for

NPCis mice (n=7); open circles (O) and NPCis; TC21 -/- mice (n=13), closed squares ( ■)

(log-rank test, p = 0.0001).

Figure 2.3: Nf1 -/- tumor initiating cells are decreased when TC21 is absent while resulting tumors show neurofibroma histology, tumor size and tumor number . (A)

Paraffin-embedded tissue sections from neurofibromas stained with hematoxylin and eosin (H&E), anti-S100-beta antibody visualized with DAB (brown) and toluidine blue staining showing metachromatic mast cell infiltration (black arrows). Scale bar: 50 µm.

(B) Quantification of the average tumor numbers per mouse for Nf1 fl/fl;DhhCre (n=9) and

TC21 -/-;Nf1 fl/fl ;DhhCre (n=10) mice. p = ns (not significant). (C) Quantification of the size

(diameter, mm) of tumors dissected from Nf1 fl/fl ;DhhCre (n=9) and TC21 -/-;Nf1 fl/fl ;DhhCre

(n=10) mice. p = ns. (D) Number of primary spheres wild type (WT), Nf1 -/- and TC21 -/-

;Nf1 -/-) from embryonic (E12.5) DRG plated at clonal density (p < 0.0001).

Figure 2.4: A TGF-beta autocrine loop in Nf1 -/- Schwann cell precursors. (A)

Quantification of number of TUNEL positive cells per high power field from paraffin embedded spheres. (B) Wild type (WT), Nf1 -/- and TC21 -/-;Nf1 -/- Schwann cell precursor spheres were treated with indicated concentrations of TGFb1 and sphere numbers counted (p = 0.005 ANOVA). (C) Quantitative real time PCR mRNA expression of

TGFb1 in embryonic spheres. Values are fold change (FC) expression relative to WT spheres (t-test between Nf1 -/- and TC21 -/-;Nf1 -/-, p = 0.04). (D) TGFb1 protein concentration in media conditioned by spheres, analyzed by ELISA (p = 0.01 ANOVA).

(E, F) Phase contrast micrographs showing Nf1 -/- spheres after treatment with rabbit IgG

(IgG) or anti-TGF-beta antibody (Anti-TGFb). (G) Quantification of Nf1 -/- spheres after treatment with IgG (IgG) or anti-TGF-beta antibody (anti-TGFb) (p = 0.005).

Figure 2.5: TGF-beta expression in sciatic nerves and neurofibromas. Quantitative real time PCR mRNA expression of TGFb1 in (A) Nf1 fl/fl ;DhhCre and

Nf1 fl/fl ;DhhCre; TC21 -/- sciatic nerves and (B) Nf1 fl/fl ;DhhCre and Nf1 fl/fl ;DhhCre; TC21 -/-;

GEM neurofibromas relative to wild type (WT) sciatic nerves.

Figure 2.6: Survival of Nf1 mutant Schwann cell precursors is dependent on

TGF −−−beta and AKT. Wild type (WT), Nf1 -/- and TC21 -/-;Nf1 -/- Schwann cell precursor spheres treated with the indicated concentrations of (A) anti-TGF-beta antibody (B)

TGFbR1 inhibitor, SB 431542 and (C) AKT inhibitor, MK-2206 or (D) shRNA targeting

AKT (shAKT) compared to a non-targeting control (NT) p values are <0.0001. Sphere numbers were counted after treatment. (E) WT, Nf1 -/- and TC21 -/-; Nf1 -/- E12.5 DRG paraffin tissue sections stained with anti-phospho-AKT. Adjacent sections were stained with anti-neurofilament (to highlight neurons in DRG). Scale bar: 50 µm. (F)

Quantification of staining intensity for phospho-AKT in DRG paraffin tissue sections (p =

0.0002). (G) Quantitative real time PCR, TGFb1 mRNA expression in Nf1 -/- Schwann cell precursor spheres treated with 1 µM AKT inhibitor, MK-2206. (H) Model of TC21 dependent TGF-beta autocrine loop in Nf1 -/- Schwann cell precursors.

Figure 2.7: Nf1 -/- Schwann cell precursors express high levels of phospho-AKT but not phospho-SMAD2/3. (A) Nf1 -/- spheres treated with indicated concentration of

MEK inhibitor, PD 98059, ROCK inhibitor, Y-27632 and P38 inhibitor, SB 203580.

Number of spheres was counted after 3 days of treatment. (B) Western blot analysis showing phospho-AKT protein expression in spheres of specified genotypes. Beta-actin serves as loading control. (C) Quantification of staining intensity for phospho-Smad2/3 in wild type (WT), Nf1 -/- and TC21 -/-;Nf1 -/- E12.5 DRG paraffin stained tissue sections.

Grading was: 0 (none); 1+ (faint); 2+ (some); 3+ (high); 4+ (saturated), defined by three observers. (D) Adjacent wild type (WT), Nf1 -/- and TC21 -/-; Nf1 -/- E12.5 DRG paraffin stained tissue sections for phospho-Smad2/3 and neurofilament (to highlight neurons in

DRGs). Scale bars: 50 µm.

Figure 2.8: Loss of TC21 decreases the survival of NPCis mice by increasing brain tumors. (A) Number of brain tumors from NPCis and NPCis; TC21 -/- mice (grade 2 or 3 GEM brain tumors). (B) Paraffin-embedded brain sections from NPCis;TC21 -/- mice stained with H&E and S100-beta. Scale bar: 50 µm. (C) Percentage of NPCis and

NPCis; TC21 -/- mice with grade 3 GEM-PNST sarcomas. (D) Paraffin-embedded brain sections from NPCis;TC21 -/- mice stained with H&E, Ki67 and S100 β. Scale bar: 50 µm.

Figure 2.9: Loss of TC21 in MPNST cells increases sarcoma growth. (A) Tumor volume quantification over time from mice injected with S462TY MPNST cells expressing non-Target control (NT) or shTC21. (B) Gross photographs of nude mice tumors. (C) Western blot showing TC21 protein expression in S462TY MPNST

62 xenograft tumors reduced subsequent to introduction of shTC21 RNA. Beta-actin serves as loading control. (D) Quantification of anti-meca staining and measurement of numbers of vessels per high power field (hpf) in non-target and shTC21 xenograft tumors. (E, F) Smooth muscle actin (SMA) immunostaining in xenograft paraffin sections. Brown labeling shows immunoreactivity. Scale bar: 50uM.

Figure 2.10: TGF-beta mediates aggressive growth of MPNST xenografts. Tumor volume quantification over time in mice injected with S462TY MPNST cells expressing non-target control or shTC21. Mice treated with (A) IgG or (B) anti-TGF-beta antibody by I.P. injection. Quantification of staining intensity for (C) phospho-SMAD2/3 and (D) phospho-AKT in non-target (NT) or shTC21 xenograft tumors. Grading was as defined in Figure 4 legend. (E) Immunohistochemistry staining of xenograft tumor sections for phospho-SMAD2/3 and phospho-AKT. Scale bar: 25 µm.

Figure 2.11: TGF-beta ligands and receptor expression in MPNST cells. (A)

Quantification of anti-meca staining for numbers of vessels per high power field (hpf) in non-target and shTC21 xenograft tumors treated with normal IgG (IgG) or anti-TGF-beta antibody (anti-TGFb) (p < 0.0001). Quantitative real time PCR (B) TGFb1 (p = 0.01) and

(WT) sciatic nerve. Quantitative real time PCR (D) TGFb1 and (E) TGFb3 mRNA expression in shTC21 S462TY xenografts relative to shnon-target control (NT). (F)

Western blot analysis 3 MPNST cell lines analyzed for TGFbRII. Actin was used as a loading control.

Figure 2.12: MPNST cells express TGF-beta and lose TGFbRII. (A) Heat map of

TGF-beta ligands and receptor that differ in human neurofibromas (NF (n=22)) and/or

MPNSTs (n=6) and/or cultured human primary neurofibroma Schwann cells (NFSC

(n=22)) and MPNST cell lines (MPNST cell (n=13)), all referenced to normal human

Schwann cells (NHSC (n=10)) and normal human peripheral nerves (Nerve (n=3)). We applied a fold change cutoff = 1.5 and a t-test p-value < 0.05 with FDR (Benjamini-

Hochberg) correction). Scale bar on left shows relative gene expression as colors. (B)

Quantitative real time PCR, TGFb2 mRNA expression in shTC21 S462TY xenografts relative to non-target control (NT) (p = 0.01). (C) TGFb2 protein concentration expression in media conditioned by MPNST cells, analyzed by ELISA (p = 0.03). (D)

Western blot analysis of immortalized human Schwann cells (iHSC) and S462TY

MPNST analyzed for TGFbRI and TGFbRII. Actin serves as loading control. (E) Model of TC21 regulating TGF-beta in MPNST cells.

Chapter 3

Discussion and Future Directions

Our studies have revealed TC21 as an important Ras protein in NF1. TC21 plays an essential role in Nf1 development by partially rescuing Nf1 -/- embryonic lethality.

TC21 has a dual effect in Nf1 tumorigenesis; loss of TC21 in neurofibroma-bearing mice significantly increases survival while loss of TC21 in GEM-PNST mice significantly decreases survival. Also, suppression of TC21 in NF1 MPNST cells increases tumor growth and tumor vasculature in xenograft mice. We have shown that these effects of

TC21 in NF1 are mediated through TGFb and that TC21 is responsible for regulating

TGFb in NF1. Our studies show that TGFb expression is necessary for Nf1 -/- Schwann cell precursor survival, which is critical in tumor progression. TC21 regulates TGFb expression in a paracrine dependent manner in NF1 malignant tumors.

NF1 patients have a risk of developing malignant peripheral nerve sheath

(MPNST) tumors, which are highly aggressive soft tissue sarcomas. Fifty percent of

MPNST cases are NF1 patients and the remaining fifty percent are sporadic cases.

These tumors are usually resected if possible, and then patients are treated with chemotherapy. However, a five year survival is seen in 21% for NF1 patients and 41% of sporadic cases (20, 24). Differences between these two MPNST groups are still not clear. Microarray studies using cell lines derived from both NF1 and sporadic MPNST tumors have not been able to identify any significant signatures (162, 178). Additionally, the median age of onset for MPNST in NF1 patients is 21 years whereas in the sporadic cases it is 62 years (20). This would suggest that genetics play an important role between these two groups. Therefore, studies leading to effective therapeutic targets

69 are of critical importance to the NF1 field. There are currently no effective treatments for

NF1. However, multiple clinical trials based on basic research studies have been initiated in the past years.

In order to develop effective therapies, we must first understand the underlying mechanisms involved in neurofibroma tumorigenesis. Loss of the NF1 gene is hypothesized to lead to elevated levels of all Ras proteins due to the fact that NF1 is a

Ras-GAP. However, studies of Ras in NF1 have been focused mainly on the canonical

Ras proteins H, N, and K-Ras. Investigating the role and functions of other Ras proteins in NF1 is important and our data verifies that TC21 is an important Ras protein in neurofibromatosis type 1.

The Role of TC21 in Neurofibromatosis type 1

Regarding its role in tumorigenesis, we have shown that TC21 plays a part in the initiation of neurofibromas. In vitro , loss of TC21 in an Nf1 -/- background rescued the increased numbers of spheres formed from embryonic day 12.5 dorsal root ganglia.

This suggests a role for TC21 in the survival/growth and/or proliferation of these

Schwann cell precursors and therefore, plays a role in the formation of neurofibromas.

This role for TC21 in survival/growth also correlates with the fact that in neurofibroma bearing mice, where loss of TC21 increased survival.

TC21 has been shown to act as an oncogene in several systems (70, 84) but its role in NF1 had not been studied in detail. We had previously shown that TC21 plays an essential role in the migration of Nf1 -/- Schwann cells (150) and, in our current studies,

70 delineate it as an important Ras protein for both the development of Nf1 -/- embryos and

NF1 tumorigenesis. The fact that loss of TC21 was only sufficient to provide a partial rescue to Nf1 -/- mice embryonic lethality shows the importance of other Ras proteins.

R-Ras -/-;Nf1 -/- embryos only survived to embryonic day 12.5, so that R-Ras is excluded for a unique role in development of Nf1 -/- embryos. Studies using the loss of multiple

Ras proteins need to be carried out to determine which proteins are essential for Nf1 embryo survival.

To confirm the role of TC21 in tumorigenesis, an essential experiment would be to investigate the capability of TC21 -/-;Nf1 -/- Schwann cell precursors to form tumors in athymic nude mice. TC21 -/-;Nf1 -/- Schwann cell precursors dissociated from E12.5

DRGs would grow as primary spheres in culture and then subcutaneously inject 1- 4.5

*10 5 cells into the flanks of athymic nude mice. We have previously shown that Nf1 -/-

Schwann cell precursors are able to form very small tumor-like nodules in nude mice whereas wild type cells cannot (165). Therefore, if TC21 is important for tumor formation, then the TC21 -/-;Nf1 -/- Schwann cell precursors should form smaller or no tumors compared to Nf1 -/- Schwann cell precursors.

TC21 and TGF-beta in NF1

Our studies of the role of TC21 in NF1 led to our finding that TGF-beta also plays an important in NF1 tumorigenesis. In vitro studies show that both TGF-beta mRNA and protein levels are elevated in Nf1 -/- Schwann cell precursors but rescued upon the loss of TC21 and occurs through an autocrine loop involving TC21, TGF-beta and AKT. Nf1 -/-

Schwann cell precursors are dependent on these factors for survival and therefore, loss of TC21 leads to a reduction in TGF-beta and AKT and results in a decrease in precursor proliferation and survival, correlating with the survival we observe in the neurofibroma bearing mice with loss of TC21. Since loss of Nf1 leads to an upregulation in TGF-beta production, an important experiment would be to treat neurofibroma bearing mice with a TGF-beta neutralizing antibody or inhibitor. If blocking TGF-beta in this context leads to increased survival or reduction in tumor number or size, this would suggest TGF-beta as a possible therapeutic target in neurofibromas. It will also be important to determine if a TGF-beta antibody or inhibitor can be used in combination with other drugs for treating neurofibromas.

Although TGF-beta is well studied in many solid tumors, its role in NF1 has not been thoroughly explored. Studies thus far have shown a role for TGF-beta in two of the cell types found in neurofibromas, mast cells and fibroblasts. It has been shown that

Nf1 +/− mast cells secrete elevated concentrations of TGF-beta and in response to secretion of TGF-beta, both murine Nf1 +/− fibroblasts and fibroblasts from human neurofibromas proliferate and synthesize excessive collagen (170).

In NF1 malignancy, TC21 acts as a tumor suppressor. In Nf1 +/-;Trp53 +/- (NPCis) mice, that form sarcomas and brain tumors, loss of TC21 decreases the survival of mice. Additionally, the mice with loss of TC21 form more aggressive brain tumors compared to littermate controls. Using a xenograft model, we showed that reducing

TC21 expression in NF1 -/- MPSNT cell lines also led to increased tumor growth.

Analysis of both sets of tumors showed elevated levels of TGF-beta in tumors lacking

TC21. In the malignant systems, TC21 acts as a tumor suppressor by causing

72 increased TGF-beta production. By treating xenograft mice with a TGF-beta neutralizing antibody, tumor growth was rescued to that of controls. Thus blocking both TC21 and

TGF-beta in MPNSTs is not sufficient to inhibit tumor growth. Essential experiments would be to use the TGF-beta antibody in combination with other drugs to examine the effects of tumor growth. Inhibiting the Raf/MAPK pathway in MPNST shows a delay in tumor growth and it is possible that a combination of blocking both TGF-beta and the

Raf/MAPK pathways would prove synergistic and lead to further reduction of MPNST tumor growth. Additionally, the effects of blocking TGF-beta in the NPCis model lacking

TC21 would be interesting. Since losing TC21 leads to decreased mouse survival and more aggressive brain tumors, blocking TGF-beta in these mice may rescue these phenotypes and further validate TGF-beta as a target for therapy.

Regulation of TGF-beta by TC21

Our studies show that TC21, a Ras protein, is a critical mediator of TGF-beta regulation in NF1 tumorigenesis. We show that in both the benign and malignant settings of NF1, loss of TC21 affects both TGF-beta mRNA and protein levels. In the benign setting, we believe that in Schwann cell precursors, a proposed initiation cells for neurofibromas, there is an autocrine loop involving TC21, TGF-beta and AKT. These factors, we show, are critical for the survival of Schwann cell precursors. We have examined other downstream effectors of both TGF-beta and Ras such as P38, ROCK and MEK and have found that inhibition of these pathways do not affect Schwann cell precursors. Additionally, we do not see a significant difference in the protein levels of

73 the canonical downstream effector of TGF-beta, SMAD2/3. However, the importance of this pathway must not be ignored. Experiments utilizing a shRNA targeting SMAD2 and

SMAD3 in Schwann cell precursors and observing effects on proliferation and survival should be carried out.

How TGF-beta is regulated by TC21 in MPNST cells is not clear. We show that loss of TC21 in these cells alters TGF-beta expression and this leads to increased tumor growth. However, we have been unsuccessful in determining the pathways leading to these phenotypes. TGFbRII is not expressed in the S462TY MPNST cells yet there is a reduction of phospho-SMAD2/3 and phospho-AKT after inhibiting TGF-beta in tumor cells. This suggests that other members of the TGF-beta family might be involved. Another possibility is that stromal cells secrete activins or BMPs that indirectly alter AKT and SMAD phosphorylation in tumor cells (107). Our data are consistent with important non-cell autonomous effects of MPNST cell produced TGF-beta. This interpretation is consistent with increased blood vessels per area and SMA (myo- fibroblast) expression increased upon down-regulation of TC21, and our finding that both phenotypes were blocked by anti-TGF-beta antibody. A non-cell autonomous effect of TGF-beta in malignancy has been documented previously in models of carcinoma, using genetic loss of TGFbRII (152).

In order to delineate a mechanism, we hypothesized that one possible pathway for TGF-beta’s regulation by TC21 is through activation of the AP1 family of transcription factors. Activation of Ras is known to lead to transcriptional activation of

AP1 family members such as, c-Jun and c-Fos (179-181). Additionally, there are known

AP1 binding sites on the TGF-beta promoter both in humans and mice (182, 183). As

74 expected, we found that inhibiting both c-Jun kinase (JNK) and c-Fos results in a reduction of TGF-beta mRNA levels in control MPNST cells. Interestingly, in cells with loss of TC21, TGF-beta transcription levels are significantly increased compared to control, but there is no significant change in TGF-beta mRNA levels when these cells are treated with either JNK or a dominant negative c-Fos (Appendices A1, A2). These data suggest the regulation of TGF-beta by TC21 is independent of AP-1.

Conclusion

Here we show that the non-canonical Ras protein TC21 plays an important role in neurofibromatosis type 1. Previous studies show that the activation of the canonical

Ras proteins is not sufficient to explain all phenotypes seen in the disease. These studies add TC21 to the list of essential Ras proteins deregulated in neurofibromatosis type 1.

We show that TC21 is involved in NF1 tumorigenesis; both the benign and malignant forms. Our studies are novel in that they examine TGF-beta in neurofibromatosis type 1. Our data suggest TGF-beta as a possible drug target in NF1 and that blocking TGF-beta may also be useful in combination therapies. Additionally, we propose that TC21 plays a role in sequestering SOS protein, an activator of Ras.

Our studies not only further our understanding of TC21 and its role in NF1 but they also help in understanding the mechanisms involved in NF1 tumorigenesis. Studies such as these are essential to identify effective therapeutic targets as therapies for treatment of

NF1.

Appendices

Appendix A: TGFb2 mRNA expression in MPNST Cells after inhibition of JNK or cFOS. Quantitative real time PCR, TGFb2 mRNA expression in shTC21 S462TY cells relative to non-target control after (A1) treatment with a JNK inhibitor (A2) infection with a dominant negative cFOS lentivirus.

REFERENCES

1. Riccardi, V. 1982. Neurofibromatosis: clinical heterogeneity. Current problems in cancer. 7:1-34.

2. Kalamarides, M., Acosta, M.T., Babovic-Vuksanovic, D., Carpen, O., Cichowski,

K., Gareth Evans, D., Giancotti, F., Oliver Hanemann, C., Ingram, D., Lloyd, A.C., et al.

2011. Neurofibromatosis 2011: a report of the Children’s Tumor Foundation Annual

Meeting. Acta Neuropathologica 123:369-380.

3. Ledbetter, D., Rich, D., O'Connell, P., Leppert, M., and Carey, J. 1989. Precise localization of NF1 to 17q11.2 by balanced translocation. American Journal of Human

Genetics 44:20-24.

4. Rouleau, G., Merel, P., Lutchman, M., Sanson, M., Zucman, J., Marineau, C.,

Hoang-Xuan, K., Demczuk, S., Desmaze, C., Plougastel, B., et al. 1993. Alteration in a new gene encoding a putative membrane-organizing protein causes neuro-fibromatosis type 2. Nature 363:515-521.

5. Hulsebos, T., Plomp, A., Wolterman, R., Robanus-Maandag, E., Baas, F., and

Wesseling, P. 2008. Germline mutation of INI1/SMARCB1 in familial schwannomatosis.

The American Journal of Human Genetics 80:805-810.

6. Hirsch, N., Murphy, A., and Radcliffe, J. 2001. Neurofibromatosis: clinical presentations and anaesthetic implications. British Journal of Anaesthesia 86:555-565.

7. Reynolds, R.M., Browning, G.G.P., Nawroz, I., and Campbell, I.W. 2003. Von

Recklinghausen's neurofibromatosis: neurofibromatosis type 1. The Lancet 361:1552-

1554.

8. 1988. National Institutes of Health Consensus Development Conference

Statement: neurofibromatosis. Neurofibromatosis 1:172-178.

9. Pride, N., Payne, J., Webster, R., Shores, E., Rae, C., and North, K. 2010.

Corpus callosum morphology and its relationship to cognitive function in neurofibromatosis type 1. Journal of Child Neurology 25:834-841.

10. Lorenzo, J., Barton, B., Acosta, M., and North, K. 2011. Mental, motor, and language development of toddlers with neurofibromatosis type 1. The Journal of

Pediatrics 158:660-665.

11. Tedesco, M., Di Salco, G., Natale, F., Graziano, L., Grassia, C., Calabro, R., and

Lama, G. 2005. Early cardiac morphologic and functional changes in neurofibromatosis type 1 hypertensives: an echocardiographic and tissue Doppler study. International

Journal of Cardiology 101:243-247.

12. Muir, D., Neubauer, D., Lim, I., Yachnis, A., and Wallace, M. 2001. Tumorigenic

Properties of Neurofibromin-Deficient Neurofibroma Schwann cells. American Journal of

Pathology 158:501-513.

13. Carroll, S.L., and Ratner, N. 2008. How does the Schwann cell lineage form tumors in NF1? Glia 56:1590-1605.

14. Scheithauer, B., Woodruff, J., and Erlandson, R. 1999. Tumors of the Peripheral

Nervous System. Armed Forces Institute of Pathology .

15. Ferner, R.E., Huson, S.M., Thomas, N., Moss, C., Willshaw, H., Evans, D.G.,

Upadhyaya, M., Towers, R., Gleeson, M., Steiger, C., et al. 2006. Guidelines for the diagnosis and management of individuals with neurofibromatosis 1. Journal of Medical

Genetics 44:81-88.

16. Ferner, R.E. 2006. Neurofibromatosis 1. European Journal of Human Genetics

15:131-138.

17. Dombi, E., Solomon, J., Gillespie, A., Fox, E., Balis, F., Patronas, N., Korf, B.,

Babovic-Vuksanovic, D., Packer, R., Belasco, J., et al. 2007. NF1 plexiform neurofibroma growth rate by volumetric MRI: relationship to age and body weight.

Neurology 68:643-647.

18. Mautner, V., Hartmann, M., Kluwe, L., Friedrich, R., and Funsterer, C. 2006. MRI growth patterns of plexiform neurofibromas in patients with neurofibromatosis type 1.

Neuroradiology 48:160-165.

19. Williams, V.C., Lucas, J., Babcock, M.A., Gutmann, D.H., Korf, B., and Maria,

B.L. 2009. Neurofibromatosis Type 1 Revisited. Pediatrics 123:124-133.

20. Evans, D., Baser, M., McGaughran, J., Sharif, S., Howard, E., and Moran, A.

2002. Malignant peripheral nerve sheath tumours in neurofibromatosis type 1 Journal of

Medical Genetics 39:311-314.

21. Tucker, T., Wolkenstein, P., Revuz, J., Zeller, J., and Friedman, J. 2005.

Association between benign and malignant peripheral nerve sheath tumors in NF1.

Neurology 65:205-211.

22. King, A., Debaun, M., Riccardi, V., and Gutmann, D. 2000. Malignant peripheral nerve sheath tumors in neurofibromatosis 1. American Journal of Medical Genetics

93:388-392.

23. Gachiani, J., Kim, D., Nelson, A., and Kline, D. 2007. Surgical management of malignant peripheral nerve sheath tumors. Neurosugical Focus 22.

24. Carli, M. 2005. Pediatric Malignant Peripheral Nerve Sheath Tumor: The Italian and German Soft Tissue Sarcoma Cooperative Group. Journal of Clinical Oncology

23:8422-8430.

25. De Raedt, T., Walton, Z., Yecies, J., Li, D., Chen, Y., Malone, C., Maertens, O.,

Jeong, S., Bronson, R., Lebleu, V., et al. 2011. Exploiting cancer cell vulnerabilities to develop a combination therapy for ras-driven tumors. Cancer Cell 20:400-413.

26. Brannan, C., Perkins, A., Vogel, K., Ratner, N., Nordlund, M., Reid, S., Buchberg,

A., Jenkins, N., Parada, L., and Copeland, N. 1994. Targeted disruption of the neurofibromatosis type-1 gene leads to developmental abnormalities in heart and various neural crest-derived tissues Genes & Development 8:1019-1029.

27. Lakkis, M., Golden, J., O'Shea, K., and Epstein, J. 1999. Neurofibromin deficiency in mice causes exencephaly and is a modifier for Splotch neural tube defects.

Developmental Biology 212:80-92.

28. Lakkis, M., and Epstein, J. 1998. Neurofibromin modulation of ras activity is required for normal endocardial-mesenchymal transformation in the developing heart.

Development 125:4359-4367.

29. Jacks, T., Shih, T., Schmitt, E., Bronson, R., Bernards, A., and Weinberg, R.

1994. Tumour predisposition in mice heterozygous for a targeted mutation in Nf1.

Nature Genetics 7:353-361.

30. Cichowski, K., Shih, T., Schmitt, E., Santiago, S., Reilly, K., McLaughlin, M.,

Bronson, R., and Jacks, T. 1999. Mouse Models of Tumor Development in

Neurofibromatosis Type 1. Science 286:2172-2176.

31. Rizvi, T., Huang, Y., Sidani, A., Atit, R., Largaespada, D., Boissy, R., and Ratner,

N. 2002. A Novel Cytokine Pathway Suppresses Glial Cell Melanogenesis after Injury to

Adult Nerve. Journal of Neuroscience 22:9831-9840.

32. Gitler, A.D., Zhu, Y., Ismat, F.A., Lu, M.M., Yamauchi, Y., Parada, L.F., and

Epstein, J.A. 2003. Nf1 has an essential role in endothelial cells. Nature Genetics

33:75-79.

33. Zhu, Y., Ghosh, P., Charnay, P., Burns, D., and Parada, L. 2002. Neurofibromas in NF1: Schwann Cell Origin and Role of Tumor Environment. Science 296:920-922.

34. Wu, J., Williams, J., Rizvi, T., Kordich, J., Witte, D., Meijer, D., Stemmer-

Rachamimov, A., Cancelas, J., and Ratner, N. 2008. Plexiform and Dermal

Neurofibromas and Pigmentation Are Caused by Nf1 Loss in Desert Hedgehog-

Expressing Cells. Cancer Cell 13:105-116.

35. Vogel, K.S., Klesse, L., Velasco-Miguel, V., Meyers, K., Rushing, E., and Parada,

L. 1999. Mouse Tumor Model for Neurofibromatosis Type 1. Science 286:2176-2179.

36. Reilly, K., Loisel, D., Bronson, R., McLaughlin, M., and Jacks, T. 2000.

NF1;Trp53 mutant mice develop glioblastoma with evidence of strain-specific effects

Nature Genetics 26:109-113.

37. Joseph, N.M., Mosher, J.T., Buchstaller, J., Snider, P., McKeever, P.E., Lim, M.,

Conway, S.J., Parada, L.F., Zhu, Y., and Morrison, S.J. 2008. The Loss of Nf1

Transiently Promotes Self-Renewal but Not Tumorigenesis by Neural Crest Stem Cells.

Cancer Cell 13:129-140.

38. Cawthon, R., Weiss, R., Xu, G., Viskochil, D., Culver, M., Stevens, J., Robertson,

M., Dunn, D., Gesteland, R., O'Connell, P., et al. 1990. A major segment of the neurofibromatosis type 1 gene: cDNA sequence, genomic structure, and point mutations. Cell 62:193-201.

39. Li, Y., O'Connell, P., Breidendach, H., Cawthon, R., Stevens, J., Xu, G., Neil, S.,

Robertson, M., White, R., and Viskochil, D. 1995. Genomic Organization of the

Neurofibromatosis 1 Gene (NF1). Genomics 25:9-18.

40. Buchberg, A., Cleveland, L., Jenkins, N., and Copeland, N. 1990. Sequence homology shared by neurofibromatosis type-1 gene and IRA-1 and IRA-2 negative regulators of the RAS cyclic AMP pathway. Nature 347:291-294.

41. Ohba, Y., Mochizuki, N., Yamashita, S., Chan, A., Schrader, J., Hattori, S.,

Nagashima, K., and Matsuda, M. 2000. Regulatory Proteins of R-Ras, TC21/R-Ras2, and M-Ras/R-Ras3. Journal of Biological Chemistry 275:20020-20026.

42. Viskochil, D., Buchberg, A., Xu, G., Cawthon, R., Stevens, J., Wolff, R., Culver,

M., Carey, J., Copeland, N., Jenkins, N., et al. 1990. Deletions and a Translocation

Interrupt a Cloned Gene at the neurofibromatosis Type 1 Locus. Cell 62:187-192.

43. Colman, S., Williams, C., and Wallace, M. 1995. Benign neurofibromas in type 1 neurofibromatosis (NF1) show somatic deletions of the NF1 gene. Nature Genetics

11:90-92.

44. Sawada, S., Florell, S., Purandare, S., Ota, M., Stephens, K., and Viskochil, D.

1996. Identification of NF1 mutations in both alleles of a dermal neurofibroma. Nature

Genetics 14:110-112.

45. Serra, E., Puig, S., Otero, D., Gaona, A., Kruyer, H., Ars, E., Estivill, X., and

Lazaro, C. 1997. Confirmation of a double-hit model for the NF1 gene in benign neurofibromas. American Journal of Human Genetics 61:512-519.

46. John, A., Ruggieri, M., Ferner, R., and Upadhyaya, M. 2000. A search for evidence of somatic mutations in the NF1 gene. Journal of Medical Genetics 37:44-49.

47. Eisenbarth, I., Beyer, K., Krone, W., and Assum, G. 2000. Toward a survey of somatic mutation of the NF1 gene in benign neurofibromas of patients with neurofibromatosis type 1. American Journal of Human Genetics 66:393-401.

48. Rasmussen, S., and Friedman, J. 2000. NF1 gene and neurofibromatosis 1.

American Journal of Epidemiology 151:33-40.

49. Legius, E., Marchuk, D., Collins, F., and Glover, T. 1993. Somatic deletion of the neurofibromatosis type 1 gene in a neurofibrosarcoma supports a tumour suppressor gene hypothesis. Nature Genetics 3:122-126.

50. Kluwe, L., Friedrich, R., and Mautner, V. 1999. Loss of NF1 allele in Schwann cells but not in fibroblasts derived from an NF1-associated neurofibroma. Genes

Chromosomes and Cancer 24:283-285.

51. Serra, E., Rosenbaum, T., Nadal, M., Winner, U., Ars, E., Estivill, X., and Lazaro,

C. 2001. Mitotic recombination effects homozygosity for NF1 germline mutations in neurofibromas. Nature Genetics 28:294-296.

52. Martin, G., Viskochil, D., Bollag, G., McCabe, P., Crosier, W., Haubruck, H.,

Conroy, L., Clark, R., O'Connell, P., Cawthon, R., et al. 1990. The GAP-Related Domain of the Neurofibromatosis Type 1 Gene Product Interacts with Ras p21. Cell 63:843-849.

53. Xu, G., Lin, B., Tanaka, K., Dunn, D., Wood, D., Gesteland, R., White, R., Weiss,

R., and Tamanoi, F. 1990. The Catalytic Domain of the Neurofibromatosis Type 1 Gene

Product Stimulates Ras GTPase and Compliments IRA Mutants in S. cerevisiae. Cell

63:835-841.

54. Ballester, R., Marchuk, D., Boguski, M., Saulino, A., Letcher, R., Wigler, M., and

Collins, F. 1990. The NF1 Locus Encodes a Protein Functionally Related to Mammalian

GAP and Yeast IRA Proteins. Cell 63:851-859.

55. Bonneau, F., D'Angelo, I., Welti, S., Stier, G., Ylanne, J., and Scheffzek, K. 2004.

Expression, purification and preliminary crystallographic characterization of a novel segment from the neurofibromatosis type 1 protein Acta Crystallographica. Section D,

Biological crystallography 60:2364-2367.

56. Viskochil, D. 2002. Genetics of neurofibromatosis 1 and the NF1 gene. Journal of

Child Neurology 17:562-570.

57. Yunoue, S., Tokuo, H., Fukunaga, K., Feng, L., Ozawa, T., Nishi, T., Kikuchi, A.,

Hattori, S., Kuratsu, J., Saya, H., et al. 2003. Neurofibromatosis Type I Tumor

Suppressor Neurofibromin Regulates Neuronal Differentiation via Its GTPase-activating

Protein Function toward Ras. Journal of Biological Chemistry 278:26958-26969.

58. Dasgupta, B., Dugan, L., and Gutmann, D. 2003. The Neurofibromatosis 1 Gene

Product Neurofibromin Regulates Pituitary Adenylate Cyclase-Activating Polypeptide-

Mediated signaling in Astrocytes. Journal of Neuroscience 23:8949-8954.

59. Guo, H., Tong, L., Hannan, F., Luo, L., and Zhong, Y. 2000. A neurofibromatosis-

1-regulated pathway is required for learning in Drosophila. Nature 403:895-898.

60. Malminen, M., Peltonen, S., Koivunen, J., and Peltonen, J. 2002. Functional expression of NF1 tumor suppressor protein: association with keratin intermediate filaments during the early development of human epidermis. BMC Dermatology 2.

61. Li, C., Cheng, Y., Gutmann, D., and Mangoura, D. 2001. Differential localization of the neurofibromatosis 1 (NF1) gene product, neurofibromin, with the F-actin or

86 microtubule cytoskeleton during differentiation of telencephalic neurons. Brain research.

Developmental brain research. 130:231-248.

62. Volta, M., Calza, S., Roberts, A., and Roberts, R. 2010. Characterisation of the interaction between syndecan-2, neurofibromin and CASK: dependence of interaction on syndecan dimerization. Biochemical and Biophysical research communications.

391:1216-1221.

63. Jouhilahti, E.-M., Peltonen, S., Heape, A.M., and Peltonen, J. 2011. The

Pathoetiology of Neurofibromatosis 1. The American Journal of Pathology 178:1932-

1939.

64. Bernards, A., and Settleman, J. 2004. GAP control: regulating the regulators of small GTPases. Trends in Cell Biology 14:377-385.

65. Schmidt, A., and Hall, A. 2002. Guanine nucleotide exchange factors for Rho

GTPases: turning on the switch. Genes & Development 16:1587-1609.

66. Karnoub, A., and Weinberg, R. 2008. Ras oncogenes: split personalities. Nature

Reviews Molecular Cell Biology 9:517-531.

67. Khosravi-Far, R., and Der, C. 1994. The Ras signal transduction pathway.

Cancer Metastasis Review 13:67-89.

68. Fernandez-Medarde, A., and Santos, E. 2011. Ras in Cancer and Developmental

Diseases. Genes & Cancer 2:344-358.

69. Cox, A., Brtva, T., Lowe, D., and Der, C. 1994. R-Ras induces malignant, but not morphologic, transformation of NIH3T3 cells. Oncogene 9:3281-3288.

70. Graham, S., Oldham, S., Martin, C., Drugan, J., Zohn, I., Campbell, S., and Der,

C. 1999. TC21 and Ras share indistinguishable transforming and differentiating activities. Oncogene 18:2107-2116.

71. Lowe, D., and Goeddel, D. 1987. Heterologous Expression and Characterization of the Human R-Ras Gene Product. Molecular and Cellular Biology 7:2845-2856.

72. Goodsell, D. 1999. The Ras oncogene. The Oncologist 4:263-264.

73. Rebollo, A., and Martinez-A, C. 1999. Ras proteins: recent advances and new functions. Blood 94:2971-2980.

74. Elion, E. 1998. Signal Transduction: Routing MAP Kinase Cascades. Science

281:1625-1626.

75. Groves, J., and Kuriyan, J. 2010. Molecular mechanisms in signal transduction at the membrane. Nature Structural Molecular Biology 17:659-665.

76. Drivas, G., Shih, A., Coutavas, E., Rush, M., and D'eustachio, P. 1990.

Characterization of Four Novel ras-Like Genes Expressed in a Human Teratocarcinoma

Cell Line. Molecular and Cellular Biology 10:1793-1798.

77. Carboni, J., Yan, N., Cox, A., Bustelo, X., Graham, S., Lynch, M., Weinmann, R.,

Seizinger, B., Der, C., and Barbacid, M. 1995. Farnesyltransferase inhibitors are inhibitors of Ras but not R-Ras2/TC21, transformation. Oncogene 10:1905-1913.

78. Arora, S., Matta, A., Shukla, N., Deo, S., and Ralhan, R. 2005. Identification of differentially expressed genes in oral squamous cell carcinoma. Molecular

Carcinogenesis 42:97-108.

79. Clark, G., Kinch, M., Gilmer, T., Burridge, K., and Der, C. 1996. Overexpression of the Ras-related TC21/R-Ras2 protein may contribute to the development of human breast cancers. Oncogene 12:169-176.

80. Chan, A., Miki, T., Meyers, K., and Aaronson, S. 1994. A human oncogene of the

RAS superfamily unmasked by expression cDNA cloning. Proceedings of the National

Academy of Sciences 91:7558-7562.

81. Sharma, R., Sud, N., Chattopadhyay, T., and Ralhan, R. 2005. TC21/R-Ras2 upregulation in esophageal tumorigenesis: potential diagnostic implications. Oncology

69:10-18.

82. Lee, J., Pyon, J., Lee, S., Lee, Y., Kang, S., Kim, C., Kim, D., Nam, H., Park, Y.,

Jeong, D., et al. 2011. Greater expression of TC21/R-ras2 in highly aggressive malignant skin cancer. International Journal of Dermatology 50:956-960.

83. Erdogan, M., Pozzi, A., Bhowmick, N., Moses, H.L., and Zent, R. 2007. Signaling

Pathways Regulating TC21-induced Tumorigenesis. Journal of Biological Chemistry

282:27713-27720.

84. Graham, S., Cox, A., Drivas, G., Rush, M., D'eustachio, P., and Der, C. 1994.

Aberrant Function of the Ras-Related Protein TC21/R-Ras2 Triggers Malignant

Transformation. Molecular and Cellular Biology 14:4108-4115.

85. Graham, S., Rogers-Graham, K., Figueroa, C., Der, C., and Vojtek, A. 2001.

Analyses of TC21/R-Ras2 signaling and biological activity. Methods in Enzymology

333:203-216.

86. Rong, R., He, Q., Liu, Y., Sheikh, M., and Huang, Y. 2002. TC21 mediates transformation and cell survival via activation of phosphatidylinositol 3-kinase/Akt and

NF-kB signaling pathway. Oncogene 21:1062-1070.

87. Murphy, G., Graham, S., Morita, S., Reks, S., Rogers-Graham, K., Vojtek, A.,

Kelley, G., and Der, C. 2002. Involvement of Phosphatidylinositol 3-Kinase, but Not

RalGDS, in TC21/R-Ras2-mediated Transformation. Journal of Biological Chemistry

277:9966-9975.

88. Ehrhardt, A., Ehrhardt, G., Guo, X., and Schrader, J. 2002. Ras and relatives— job sharing and networking keep an old family together. Experimental Hematology

30:1089-1106.

89. Esteban, L.M., Vicario-Abejon, C., Fernandez-Salguero, P., Fernandez-Medarde,

A., Swaminathan, N., Yienger, K., Lopez, E., Malumbres, M., McKay, R., Ward, J.M., et al. 2001. Targeted Genomic Disruption of H-ras and N-ras, Individually or in

Combination, Reveals the Dispensability of Both Loci for Mouse Growth and

Development. Molecular and Cellular Biology 21:1444-1452.

90. Komatsu, M., and Ruoslahti, E. 2005. R-Ras is a global regulator of vascular regeneration that suppresses intimal hyperplasia and tumor angiogenesis. Nature

Medicine 11:1346-1350.

91. Delgado, P., Cubelos, B., Calleja, E., Martínez-Martín, N., Ciprés, A., Mérida, I.,

Bellas, C., Bustelo, X., and Alarcón, B. 2009. Essential function for the GTPase TC21 in homeostatic antigen receptor signaling. Nature Immunology 10:880-888.

92. Johnson, L., Greenbaum, D., Cichowski, K., Mercer, K., Murphy, E., Schmitt, E.,

Bronson, R., Umanoff, H., Edelmann, W., Kucherlapati, R., et al. 1997. K-Ras is an essential gene in the mouse with partial functional overlap with N-Ras. Genes &

Development 11:2468-2481.

93. Fortini, M., Simon, M., and Rubin, G. 1992. Signalling by the sevenless protein tyrosine kinase is mimicked by Ras1 activation. Nature 355:559-561.

94. Brand, A., and Perrimon, N. 1993. Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118:401-415.

95. Prober, D., and Edgar, B. 2002. Interactions between Ras1, dMyc, and dPI3K signaling in the developing Drosophila wing. Genes & Development 16:2286-2299.

96. Bishop, J.G., and Corces, V.G. 1988. Expression of an activated ras gene causes developmental abnormalities in transgenic Drosophila melanogaster. Genes &

Development 2:567-577.

97. Rocks, O., Peyker, A., Kahms, M., Verveer, P., Koerner, C., Lumbierres, M.,

Kuhlmann, J., Waldmann, H., Wittinghofer, A., and Bastianens, P. 2005. An acylation cycle regulates localization and activity of palmitoylated Ras isoforms. Science

307:1746-1752.

98. Lisanti, M., Scherer, P., Vidugiriene, J., Tang, Z., Hermanowski-Vosatka, A., Tu,

Y., Cook, R., and Sargiacomo, M. 1994. Characterization of Caveolin-rich Membrane

Domains Isolated from an Endothelial-rich Source: Implications for Human Disease. The

Journal of Cell Biology 126:111-126.

99. Boyanapalli, M., Lahoud, O., Messiaen, L., Kim, B., Anderie de Sylor, M.,

Duckett, S., Somara, S., and Mikol, D. 2006. Neurofibromin binds to caveolin-1 and regulates ras, FAK, and Akt. Biochemical and Biophysical research communications.

340:1200-1208.

100. Mineo, C., James, G., Smart, E., and Anderson, R. 1996. Localization of

Epidermal Growth Factor-stimulated Ras/Raf-1 Interaction to Caveolae Membrane.

Journal of Biological Chemistry 271:11930-11935.

101. Kranenburg, O., Verlaan, I., and Moolenaar, W. 2001. Regulating c-Ras function: cholesterol depletion affects caveolin association, GTP loading, and signaling. Current

Biology 11:1880-1884.

102. DeClue, J., Heffelfinger, S., Benvenuto, G., Ling, B., Li, S., Rui, W., Vass, W.,

Viskochil, D., and Ratner, N. 2000. Epidermal growth factor receptor expression in neurofibromatosis type 1-related tumors and NF1 animal models. Journal of Clinical

Investigation 105:1233-1241.

103. Ling, B., Wu, J., Miller, S., Monk, K., Shamekh, R., Rizvi, T., DeCourten-Myers,

G., Vogel, K., DeClue, J., and Ratner, N. 2005. Role for the epidermal growth factor

92 receptor in neurofibromatosis-related peripheral nerve tumorigenesis. Cancer Cell 7:65-

75.

104. Bierie, B., and Moses, H. 2006. TGF-beta and cancer. Cytokine Growth Factor

Reviews 17:29-40.

105. Rahimi, R., and Leof, E. 2007. TGF-beta signaling: a tale of two responses.

Journal of Cellular Biochemistry 102:593-608.

106. Derynck, R., and Zhang, Y. 2004. Smad-dependent and Smad-independent pathways in TGF-beta family signalling. Nature 425:577-584.

107. Yang, L., Pang, Y., and Moses, H.L. 2010. TGF-β and immune cells: an important regulatory axis in the tumor microenvironment and progression. Trends in

Immunology 31:220-227.

108. Moustakas, A., and Heldin, C. 2005. Non-Smad TGF- signals. Journal of Cell

Science 118:3573-3584.

109. Seoane, J., Le, H., Shen, L., Anderson, S., and Massagué, J. 2004. Integration of

Smad and forkhead pathways in the control of neuroepithelial and glioblastoma cell proliferation. Cell 117:211-223.

110. Gomis, R., Alarcón, C., He, W., Qiongqing, W., Seoane, J., Lash, A., and

Massagué, J. 2006. A FoxO-Smad synexpression group in human keratinocytes.

Proceedings of the National Academy of Sciences 103:12747-12752.

111. Massagué, J. 2008. TGF β in Cancer. Cell 134:215-230.

112. Akhurst, R., and Derynck, R. 2001. TGF-beta signaling in cancer-a double-edged sword. Trends in Cell Biology 11:S44-51.

113. Yang, L., and Moses, H.L. 2008. Transforming Growth Factor : Tumor

Suppressor or Promoter? Are Host Immune Cells the Answer? Cancer Research

68:9107-9111.

114. Watanabe, T., Wu, T., Catalano, P., Ueki, T., Satriano, R., Haller, D., Benson, A., and Hamilton, S. 2001. Molecular predictors of survival after adjuvant chemotherapy for colon cancer. The New Englang Journal of Medicine 344:1196-1206.

115. Bruna, A., Darken, R.S., Rojo, F., Ocaña, A., Peñuelas, S., Arias, A., Paris, R.,

Tortosa, A., Mora, J., Baselga, J., et al. 2007. High TGF β-Smad Activity Confers Poor

Prognosis in Glioma Patients and Promotes Cell Proliferation Depending on the

Methylation of the PDGF-B Gene. Cancer Cell 11:147-160.

116. Arteaga, C. 2006. Inhibition of TGFbeta signaling in cancer therapy. Current

Opinion in Genetics & Development 16:30-37.

117. Grunert, S., Jechlinger, M., and Beug, H. 2003. Diverse cellular and molecular mechanisms contribute to epithelial plasticity and metastasis. Nature Reviews Molecular

Cell Biology 4:657-665.

118. Grusch, M., Petz, M., Metzner, T., Ozturk, D., Schneller, D., and Mikulits, W.

2010. The Crosstalk of RAS with the TGF-β Family During Carcinoma Progression and its implications for Targeted Cancer Therapy Current Cancer Drug Targets 10:849-857.

119. Oft, M., Peli, J., Rudaz, C., Schwarz, H., Beug, H., and Reichmann, E. 1996.

TGF-beta1 and Ha-Ras collaborate in modulating the phenotypic plasticity and invasiveness of epithelial tumor cells. Genes & Development 10:2462-2477.

120. Oft, M., Heider, K., and Beug, H. 1998. TGFb signaling is necessary for carcinoma cell invasiveness and metastasis. Current Biology 8:1243-1252.

121. Iglesias, M., Frontelo, P., Gamallo, C., and Quintanilla, M. 2000. Blockade of

Smad4 in transformed keratinocytes containing a Ras oncogene leads to hyperactivation of the Ras-dependent Erk signalling pathway associated with progression to undifferentiated carcinomas. Oncogene 19:4134-4145.

122. Park, B., Park, J., Byun, D., Park, J., and Chi, S. 2000. Mitogenic conversion of transforming growth factor-beta1 effect by oncogenic Ha-Ras-induced activation of the mitogen-activated protein kinase signaling pathway in human prostate cancer. Cancer

Research 60:3031-3036.

123. Trobridge, P., Knoblaugh, S., Washington, M.K., Munoz, N.M., Tsuchiya, K.D.,

Rojas, A., Song, X., Ulrich, C.M., Sasazuki, T., Shirasawa, S., et al. 2009. TGF-β

Receptor Inactivation and Mutant Kras Induce Intestinal Neoplasms in Mice via a β-

Catenin-Independent Pathway. Gastroenterology 136:1680-1688.e1687.

124. Janda, E. 2002. Ras and TGFbeta cooperatively regulate epithelial cell plasticity and metastasis: dissection of Ras signaling pathways. The Journal of Cell Biology

156:299-314.

125. Kretzschmar, M., Doody, J., Timokhina, I., and Massagué, J. 1999. A mechanism of repression of TGFb/Smad signaling by oncogenic Ras. Genes & Development

13:804-816.

126. Oft, M., Akhurst, R., and Balmain, A. 2002. Metastasis is driven by sequential elevation of H-ras and Smad2 levels. Nature Cell Biology 4:487-494.

127. Sekimoto, G., Matsuzaki, K., Yoshida, K., Mori, S., Murata, M., Seki, T., Matsui,

H., Fujisawa, J.i., and Okazaki, K. 2007. Reversible Smad-Dependent Signaling between Tumor Suppression and Oncogenesis. Cancer Research 67:5090-5096.

128. Mori, S., Matsuzaki, K., Yoshida, K., Furukawa, F., Tahashi, Y., Yamagata, H.,

Sekimoto, G., Seki, T., Matsui, H., Nishizawa, M., et al. 2004. TGF-β and HGF transmit the signals through JNK-dependent Smad2/3 phosphorylation at the linker regions.

Oncogene 23:7416-7429.

129. Erdogan, M., Pozzi, A., Bhowmick, N., Moses, H.L., and Zent, R. 2008.

Transforming Growth Factor- (TGF- ) and TGF- -Associated Kinase 1 Are Required for

R-Ras-Mediated Transformation of Mammary Epithelial Cells. Cancer Research

68:6224-6231.

130. Cordenonsi, M., Montagner, M., Adorno, M., Zacchigna, L., Martello, G., Mamidi,

A., Soligo, S., Dupont, S., and Piccolo, S. 2007. Integration of TGF-beta and Ras/MAPK signaling through p53 phosphorylation. Science 315:840-843.

131. Adorno, M., Cordenonsi, M., Montagner, M., Dupont, S., Wong, C., Hann, B.,

Solari, A., Bobisse, S., Rondina, M.B., Guzzardo, V., et al. 2009. A Mutant-p53/Smad

Complex Opposes p63 to Empower TGF β-Induced Metastasis. Cell 137:87-98.

132. Razanajaona, D., Joguet, S., Ay, A.S., Treilleux, I., Goddard-Leon, S., Bartholin,

L., and Rimokh, R. 2007. Silencing of FLRG, an Antagonist of Activin, Inhibits Human

Breast Tumor Cell Growth. Cancer Research 67:7223-7229.

133. Jung, B., Doctolero, R., Tajima, A., Nguyen, A., Keku, T., Sandler, R., and

Carethers, J. 2004. Loss of activin receptor type 2 protein expression in microsatellite unstable colon cancers. Gastroenterology 126:654-659.

134. Grusch, M., Drucker, C., Peter-Vorosmarty, B., Erlach, N., Lackner, A., Losert,

A., Macheiner, D., Schneider, W., Hermann, M., Groome, N., et al. 2006. Deregulation of the activin/follistatin system in hepatocarcinogenesis. Journal of Hepatology 45:673-

680.

135. Seder, C., Hartojo, W., Lin, L., Silvers, A., Wang, Z., Thomas, D., Giordano, T.,

Chen, G., Chang, A., Orringer, M., et al. 2009. Upregulated INHBA expression may promote cell proliferation and is associated with poor survival in lung adenocarcinoma.

Neoplasia 11:388-396.

136. Kang, H.-Y., Huang, H.-Y., Hsieh, C.-Y., Li, C.-F., Shyr, C.-R., Tsai, M.-Y.,

Chang, C., Chuang, Y.-C., and Huang, K.-E. 2009. Activin A Enhances Prostate Cancer

Cell Migration Through Activation of Androgen Receptor and Is Overexpressed in

Metastatic Prostate Cancer. Journal of Bone and Mineral Research 24:1180-1193.

137. Yoshinaga, K., Yamashita, K., Mimori, K., Tanaka, F., Inoue, H., and Mori, M.

2008. Activin a causes cancer cell aggressiveness in esophageal squamous cell carcinoma cells. Annals of Surgical Oncology 15:96-103.

138. Chang, K., Kao, H., Liang, Y., Cheng, M., Chang, Y., Liu, Y., Ko, T., Lee, Y.,

Tsai, C., Wang, T., et al. 2010. Overexpression of activin A in oral squamous cell carcinoma: association with poor prognosis and tumor progression. Annals of Surgical

Oncology 17:1945-1956.

139. Suzawa, M., Tamura, Y., Fukumoto, S., Miyazono, K., Fujita, T., Kato, S., and

Takeuchi, Y. 2002. Stimulation of Smad1 transcriptional activity by Ras-extracellular signal-regulated kinase pathway: a possible mechanism for collagen-dependent osteoblastic differentiation. Journal of Bone and Mineral Research 17:240-248.

140. Beck, S., and Carethers, J. 2007. BMP suppresses PTEN expression via

RAS/ERK signaling. Cancer Biology and Therapy 6:1313-1317.

141. Foletta, V., Lim, M., Soosairajah, J., Kelly, A., Stanley, E., Shannon, M., He, W.,

Das, S., Massagué, J., and Bernard, O. 2003. Direct signaling by the BMP type II receptor via the cytoskeletal regulator LIMK1. The Journal of Cell Biology 162:1089-

1098.

142. Kodach, L.L., Bleuming, S.A., Musler, A.R., Peppelenbosch, M.P., Hommes,

D.W., van den Brink, G.R., van Noesel, C.J.M., Offerhaus, G.J.A., and Hardwick, J.C.H.

2008. The bone morphogenetic protein pathway is active in human colon adenomas and inactivated in colorectal cancer. Cancer 112:300-306.

143. Kraunz, K.S., Nelson, H.H., Liu, M., Wiencke, J.K., and Kelsey, K.T. 2005.

Interaction between the bone morphogenetic proteins and Ras/MAP-kinase signalling pathways in lung cancer. British Journal of Cancer 93:949-952.

144. Mitin, N., Rossman, K.L., and Der, C.J. 2005. Signaling Interplay in Ras

Superfamily Function. Current Biology 15:R563-R574.

145. Castellano, E., and Downward, J. 2010. Role of RAS in the regulation of PI 3- kinase. Current Topics in Microbiology and Immunology 346:143-169.

146. Reuther, G., and Der, C. 2000. The Ras branch of small GTPases: Ras family members don’t fall from the tree. Current Opinion in Cell Biology 12:157-165.

147. Quinlan, M., Quatela, S., Philips, M., and Settleman, J. 2008. Activated Kras, but

Not Hras or Nras, May Initiate Tumors of Endodermal Origin via Stem Cell Expansion.

Molecular and Cellular Biology 28:2659-2674.

148. Kim, R., Trubetskoy, A., Suzuki, T., Jenkins, N.A., Copeland, N.G., and Lenz, J.

2003. Genome-Based Identification of Cancer Genes by Proviral Tagging in Mouse

Retrovirus-Induced T-Cell Lymphomas. Journal of Virology 77:2056-2062.

149. Boyd, K., Korf, B., and Theos, A. 2009. Neurofibromatosis type 1. Journal of the

American Academy of Dermatology 61:1-14.

150. Huang, Y., Rangwala, F., Fulkerson, P.C., Ling, B., Reed, E., Cox, A.D.,

Kamholz, J., and Ratner, N. 2004. Role of TC21/R-Ras2 in enhanced migration of neurofibromin-deficient Schwann cells. Oncogene 23:368-378.

151. Rosario, M., Paterson, H.F., and Marshall, C.J. 2001. Activation of the Ral and

Phosphatidylinositol 3' Kinase Signaling Pathways by the Ras-Related Protein TC21.

Molecular and Cellular Biology 21:3750-3762.

152. Lu, S., Herrington, H., Reh, D., Weber, S., Bornstein, S., Wang, D., Li, A., Tang,

C., Siddigui, Y., Nord, J., et al. 2006. Loss of transforming growth factor-beta type II receptor promotes metastatic head-and-neck squamous cell carcinoma. Genes &

Development 20:1331-1342.

153. Lee, M., Pardoux, C., Hall, M., Lee, P., Warburton, D., Qing, J., Smith, S., and

Derynck, R. 2007. TGF-b activates Erk MAP kinase signalling through direct phosphorylation of ShcA. The EMBO Journal 26:3957-3967.

154. Geiser, A., Kim, S., Roberts, A., and Sporn, M. 1991. Characterization of the mouse transforming growth factor-beta 1 promoter and activation by the Ha-ras oncogene. Molecular and Cellular Biology 11:84-92.

155. Yue, J., and Mulder, K. 2000. Requirement of Ras/MAPK Pathway Activation by

Transforming Growth Factor beta for Transforming Growth Factor beta 1 Production in a

Smad-dependent Pathway. Journal of Biological Chemistry 275:30765-30773.

156. Zheng, H., Chang, L., Patel, N., Yang, J., Lowe, L., Burns, D.K., and Zhu, Y.

2008. Induction of Abnormal Proliferation by Nonmyelinating Schwann Cells

Triggers Neurofibroma Formation. Cancer Cell 13:117-128.

157. Fernandes, K.J.L., McKenzie, I.A., Mill, P., Smith, K.M., Akhavan, M., Barnabé-

Heider, F., Biernaskie, J., Junek, A., Kobayashi, N.R., Toma, J.G., et al. 2004. A dermal

100 niche for multipotent adult skin-derived precursor cells. Nature Cell Biology 6:1082-

1093.

158. Kim, H., Rosenbaum, T., Marchionni, M., Ratner, N., and DeClue, J. 1995.

Schwann cells from neurofibromin deficient mice exhibit activation of p21ras, inhibition of cell proliferation and morphological changes. Oncogene 11:325-335.

159. Frahm, S., Mautner, V., Brems, H., Legius, E., Debiec-Rychter, M., Friedrich, R.,

Knofel, W., Peiper, M., and Kluwe, L. 2004. Genetic and phenotypic characterization of tumor cells derived from malignant peripheral nerve sheath tumors of neurofibromatosis type 1 patients. Neurobiology of Disease 16:85-91.

160. Mahller, Y.Y., Vaikunth, S.S., Currier, M.A., Miller, S.J., Ripberger, M.C., Hsu, Y.-

H., Mehrian-Shai, R., Collins, M.H., Crombleholme, T.M., Ratner, N., et al. 2007.

Oncolytic HSV and Erlotinib Inhibit Tumor Growth and Angiogenesis in a Novel

Malignant Peripheral Nerve Sheath Tumor Xenograft Model. Molecular Therapy 15:279-

286.

161. De Windt, L., Lim, H., Haq, S., Force, T., and Molkentin, J. 2000. Calcineurin promotes protein kinase C and c-Jun NH2-terminal kinase activation in the heart. Cross- talk between cardiac hypertrophic signaling pathways. The Journal of Biological

Chemistry 275:13571-13579.

162. Miller, S.J., Rangwala, F., Williams, J.P., Ackerman, P., Kong, S., Jegga, A.,

Kaiser, S., Aronow, B., Frahm, S., Kluwe, L., et al. 2006. Large-Scale Molecular

101

Comparison of Human Schwann Cells to Malignant Peripheral Nerve Sheath Tumor

Cell Lines and Tissues. Cancer Research 66:2584-2591.

163. Miller, S.J., Jessen, W.J., Mehta, T., Hardiman, A., Sites, E., Kaiser, S., Jegga,

A.G., Li, H., Upadhyaya, M., Giovannini, M., et al. 2009. Integrative genomic analyses of neurofibromatosis tumours identify SOX9 as a biomarker and survival gene. EMBO

Molecular Medicine 1:236-248.

164. Stemmer-Rachamimov, A., Louis, D., Nielsen, G., Antonescu, C., Borowsky, A.,

Bronson, R., Burns, D., Cervera, P., McLaughlin, M., Reifenberger, G., et al. 2004.

Comparative Pathology of Nerve Sheath Tumors in Mouse Models and Humans.

Cancer Research 64:3718-3724.

165. Williams, J., Wu, J., Johansson, G., Rizvi, T., Miller, S., Geiger, H., Malik, P., Li,

W., Mukouyama, Y., Cancelas, J., et al. 2008. Nf1 Mutation Expands an EGFR-

Dependent Peripheral Nerve Progenitor that Confers Neurofibroma Tumorigenic

Potential. Cell Stem Cell 3:658-669.

166. Habashi, J.P. 2006. Losartan, an AT1 Antagonist, Prevents Aortic Aneurysm in a

Mouse Model of Marfan Syndrome. Science 312:117-121.

167. Lloyd, R., Vidal, S., Horvath, E., Kovacs, K., and Scheithauer, B. 2003.

Angiogenesis in normal and neoplastic pituitary tissues. Microscopy Research and

Technique 60:244-250.

102

168. Hagedorn, L., Suter, U., and Sommer, L. 1999. P0 and PMP22 mark a multipotent neural crest-derived cell type that displays community effects in response to

TGF-beta family factors. Development 126:3781-3794.

169. Parkinson, D., Dong, Z., Bunting, H., Whitfield, J., Meier, C., Marie, H., Mirsky,

R., and Jessen, K. 2001. Transforming Growth Factor b (TGFb) Mediates Schwann Cell

Death in vitro and in vivo: Examination of c-Jun Activation, Interactions with Survival

Signals, and the Relationship of TGFb-mediated death to Schwann cell differentiation.

The Journal of Neuroscience 21:8572-8585.

170. Yang, F.C. 2006. Nf1+/- mast cells induce neurofibroma like phenotypes through secreted TGF- signaling. Human Molecular Genetics 15:2421-2437.

171. D'Antonio, M. 2006. TGFbeta Type II Receptor Signaling Controls Schwann Cell

Death and Proliferation in Developing Nerves. Journal of Neuroscience 26:8417-8427.

172. Kim, H., Ling, B., and Ratner, N. 1997. Nf1-deficient mouse Schwann cells are angiogenic and invasive and can be induced to hyperproliferate: reversion of some phenotypes by an inhibitor of farnesyl protein transferase. Molecular and Cellular

Biology 17:862-872.

173. Kraniak, J., Sun, D., Mattingly, R., Reiners, J.J., and Tainsky, M. 2010. The role of neurofibromin in N-Ras mediated AP-1 regulation in malignant peripheral nerve sheath tumors. Molecular and Cellular Biochemistry 344:267-276.

174. Rahimi, R., and Leof, E. 2009. TGFbeta versatility: PI3K as a critical mediator of distinct cell type and context specific responses. Cell Cycle 8:1813-1817.

103

175. Ding, L., Getz, G., Wheeler, D.A., Mardis, E.R., McLellan, M.D., Cibulskis, K.,

Sougnez, C., Greulich, H., Muzny, D.M., Morgan, M.B., et al. 2008. Somatic mutations affect key pathways in lung adenocarcinoma. Nature 455:1069-1075.

176. Network, C.G.A. 2008. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 455:1061-1068.

177. Hölzel, M., Huang, S., Koster, J., Øra, 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.

178. Watson, M., Perry, A., Tihan, T., Prayson, R., Guha, A., Bridge, J., Ferner, R., and Gutmann, D. 2004. Gene expression profiling reveals unique molecular subtypes of

Neurofibromatosis Type I-associated and sporadic malignant peripheral nerve sheath tumors. Brain Pathology 14:297-303.

179. Cho, H., Kang, J., Kwak, J., Lee, T., Lee, I., Park, N., Nakajima, H., Magae, J., and Chang, Y. 2007. Ascofuranone suppresses PMA-mediated matrix metalloproteinase-9 gene activation through the Ras/Raf/MEK/ERK- and Ap1- dependent mechanisms. Carcinogenesis 28:1104-1110.

180. Karin, M. 1995. The Regulation of AP-1 Activity by Mitogen-activated Protein

Kinases. Journal of Biological Chemistry 270:16483-16486.

181. Mechta, F., Lallemand, D., Pfarr, C., and Yaniv, M. 1997. Transformation by ras modifies AP1 composition and activity. Oncogene 14:847-847.

104

182. Weigert, C., Brodbeck, K., Klopfer, K., Haring, H., and Schleicher, E. 2002.

Angiotensin II induces human TGF-β1 promoter activation: similarity to hyperglycaemia. Diabetologia 45:890-898.

183. Ogawa, K., Chen, F., Kuang, C., and Chen, Y. 2004. Suppression of matrix metalloproteinase-9 transcription by transforming growth factor-b is mediated by a nuclear factor-kB site. Biochemical Journal 381:413-422.

105

106