USF Libraries
USF Digital Collections

The role of RalA and RalB in cancer

MISSING IMAGE

Material Information

Title:
The role of RalA and RalB in cancer
Physical Description:
v, 187 leaves : ill. ; 29 cm.
Language:
English
Creator:
Falsetti, Samuel C
Publication Date:

Subjects

Subjects / Keywords:
Intracellular Signaling Peptides and Proteins   ( mesh )
Neoplasm Proteins   ( mesh )
Ovarian Neoplasms   ( mesh )
Receptors, Cell Surface   ( mesh )
Proteomics   ( mesh )
Genes, ras   ( mesh )
Alkyl and Aryl Transferases   ( mesh )
Protein Prenylation   ( mesh )
Leucine -- analogs & derivatives   ( mesh )
ral GTP-Binding Proteins   ( mesh )
Ras
Geranylgeranyltransferase I inhibitors
RACK1
Ovarian cancer
Proteomics
Dissertations, Academic -- Biochemistry and Molecular Biology -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Ras genes are frequently mutated in human cancers and present compelling targets for therapeutic intervention. While previous attempts to directly inhibit oncogenic Ras function have largely been unsuccessful use of targeted agents to inhibit the three primary oncogenic pathways activated by mutated Ras: RalGEF-Ral, PI3K-Akt and Raf- MEK-Erk, is an area of intense investigation. Here, we describe the ability of a novel pharmacological inhibitor of geranylgeranyltransferase I, GGTI-2417, to inhibit Ral prenylation and localization. We further used a Ral rescue system to selectively preserve RalA and RalB function and localization during GGTI-2417 treatment and determine the precise roles for inhibition of Ral prenylation in the GGTI anti-cancer response. Specifically, we determined inhibition of RalA is required for GGTI-attenuation of anchorage independent growth whereas inhibition of RalB is required for inhibition of proliferation, induction of apoptosis, suppression of survivin and induction of p27Kip1. We next determined the role of RalGEF-Ral signaling as well as PI3K-Akt and Raf-MEKErk signal transduction pathways in an in vitro model of human ovarian surface epithelial (T80 HOSE) cell Ras-dependent transformation. Using both small interfering RNA (siRNA) and pharmacological inhibitors of Ral, PI3K and MEK we determined that Ras signaling via Ral and PI3K but not MEK is required for ovarian oncogenesis. Furthermore, stable expression of Ras mutants unable to activate Raf-MEK-Erk signaling were able to robustly transform T80 cells. Since we had confirmed the importance of Ral proteins to human epithelial malignancies we next sought to explore the molecular interactions governing Ral transformation using a proteomics approach to rapidly identify proposed Ral interacting partners. Using immunoprecipition of transiently overexpressed FLAG-tagged RalA and RalB followed by 1D-gel separation and tandem MS/MS analysis we determined a database of proposed Ral interacting proteins. One of these, RACK1, is a validated RalA and RalB interacting protein which is at least partially required for Ras and Ral transformation. These results provide both a strong impetus and a solid basis for future studies into the mechanisms of RalA- and RalB- dependent transformation.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2008.
Bibliography:
Includes bibliographical references.
Statement of Responsibility:
by Samuel C. Falsetti.
General Note:
Includes vita.
General Note:
Also available online.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002058224
oclc - 503001781
usfldc doi - E14-SFE0002307
usfldc handle - e14.2307
System ID:
SFS0026625:00001


This item is only available as the following downloads:


Full Text

PAGE 1

The Role of RalA and RalB in Cancer By Samuel C. Falsetti A dissertation submitted in partial fulfillment Of the requirements for the degree of Doctor of Philosophy Department of Molecular Medicine College of Medicine University of South Florida Major Professor: Sa d M. Sebti, Ph.D. Larry P. Solomonson, Ph.D. Gloria C. Ferreira, Ph.D. Srikumar Chellapan, Ph.D. Gary Reuther, Ph.D. Douglas Cress, Ph.D. Date of Approval: April 7th, 2008 Keywords: Ras, RACK1, Geranylgeranyltransferase I inhibitors, ovarian cancer, proteomics Copyright 2008, Samuel C. Falsetti

PAGE 2

Dedication This thesis is dedicated to my great est supporter, my wife. Without her loving advice and patience none of this would be possible.

PAGE 3

Acknowledgments I would like to extend my sincere gratitude to my wife, Nicole. She is the most inspirational person in my life and I am honored to be with her; in truth, this degree ought to come with two names printed on it. Thanks to my parents and grandparents for inspiring a love of learning and nurturing creativity in me from an early age. I wish to extend my thanks to the rest of my family: my sister Adrianna; my in-laws Bob, Gail, Michael, and Robert; and of course my adopted sisters Courtney and Andrea. Also thanks to Audrey Shor for lots of afternoon coffee sessions and invaluable career advice, Adam Carie for being a great friend as well as an excellent source of advice, Jim Hawker, Michelle Blaskovich, Cindy Boo-shay, Ryan Floyd, Kazi Aslamuzzaman, De-an Wang, Maria Balasis, Laura Francis, Lisa White, and Barbara Roberto. Additionally, I would like to thank the many memb ers of the Sebti Lab who have also contributed to my career development: Kara Forinash, Norb ert Berndt, Kun Jiang, Kuichin Zhu, Kristine Sedey and Iain Duffy. A special thanks go es to my committee members whose patience and advice has been invaluable and without whom this work would not be possible: Larry Solomonson, Gloria Ferreira, Gary Reuther, Sr ikumar Chellapan, Doug Cress and my wonderful outside advisor, Channing Der. Last, but most certainly not least, I wish to extend my most sincere thanks to my mentor Dr. Sad Sebti who has been the most invaluable source of scientific advice and inspiration over the last five years and without him none of the following work would be possible.

PAGE 4

i Table of Contents List of Tables iii List of Figures iv Abstract 1 List of Abbreviations 3 Chapter 1: Introduction 5 The Ras Superfamily 6 The Role of Pre nylation in Ras Superfamily Function 13 The Tissueand Species-specific Mechanisms of Ras Transformation 15 Use of Spontaneously Immortalized Cell Systems to Study Ras Transformation 16 Use of Genetically Defined Cell systems to Study Ras Transformation 19 Use of an In Vitro Model System of Human Ovarian Cancer to Study Mechanisms of Ra s Transformation: Implications for RalA and RalB 22 Ral Proteins as Central Me diators of Oncogenesis 26 Ral Effectors and Control of Cellular Processes 28 References 31 Chapter 2: Geranylgeranyltransferase Inhibitors Target RalB to Inhi bit Anchorage-dependent Growth and Induce Apoptosis and RalA to Inhibit Anchorage-independent Growth 47 Abstract 4 8 Introduction 49 Materials and Methods 53 Results 61 Discussion 86 Acknowledgements 92 References 93 Chapter 3: Ras Transformation in a Gene tically Defined Human Ovarian Cancer Model Requires Akt and Ral but not Raf 102 Abstract 10 3

PAGE 5

ii Introduction 104 Materials and Methods 108 Results 112 Discussion 133 Acknowledgements 140 References 141 Chapter 4: Discovery of a Proposed Data base of Ral Interacting Proteins: RACK1 Binds Ral and is Required for H and K-Ras Mediated Transformation 150 Abstract 151 Introduction 152 Materials and Methods 157 Results 161 Discussion 173 Acknowledgements 177 References 178 Chapter 5: Summary and Implications 184 About the Author End page

PAGE 6

iii List of Tables Table 1 The Ras superfamily 7 Table 2 Proposed RalA in teracting proteins 164 Table 3 Proposed RalB in teracting proteins 166

PAGE 7

iv List of Figures Figure 1. The Ras GTP-GDP cycle 9 Figure 2. The GTPase domain and point mutations 10 Figure 3. Ras signaling and transformation 12 Figure 4. Ral signaling 30 Figure 5. RalA-CCIL and RalB -CCIL are geranylgeranylated whereas the mutants RalA-CCIS and RalBCCLS are farnesylated 65 Figure 6. Geranylgeranylated a nd farnesylated RalA and Ra lB localize similarly and require prenylati on for correct localization 68 Figure 7. Farnesylated and gerany lgeranylated RalA and RalB are equivalent in mediating activation of NFB promoter activity 71 Figure 8. Ectopic expression of farnesylated RalB, but not RalA, renders cells less sensitive to GGTI-2417 inhibition of su rvival and proliferation and induction of apoptosis 74 Figure 9. Stable expression of farnesylated RalB, but not RalA, promotes resistance to the anti-prolif erative and pro-apoptotic e ffects of GGTI-2417 in MiaPaCa2 cells 78 Figure 10. Stable expression of farnesylated RalA, but not RalB, induces resistance to inhibition of anchorage independent growth by GGTI-2417 in MiaPaCa2 cells 82 Figure 11. RalB-F, but not RalA-F, inhibits th e ability of GGTI-2417 to increase p27Kip1 and decrease su rvivin protein levels 85 Figure 12. Differential transformi ng activity and signaling activation by Hand KRas12V in human ovarian surface epithelial (HOSE) cells 114

PAGE 8

v Figure 13. Ras transformation of T 80 HOSE cells requires expression of RalA and RalB 119 Figure 14. Ras transformation of HO SE cells requires expression of Akt1/2 but not Raf1or MEK1/2 121 Figure 15. K-Ras mutants which pref erentially activate RalGDS and/or PI3K but not Raf1 are capable of transforming T 80 HOSE cells to a similar or greater extant as compared to fully active K-Ras 126 Figure 16. Pharmacological inhibitors of Ral and PI3K, but not MEK, inhibit Ras transformation of T80 HOSE cells 130 Figure 17. Expression and imm unoprecipitation of FLAG-Ral72L 162 Figure 18. RACK1/ GBLP interacts with ectopically expressed Ral72L 168 Figure 19. RACK1/ GBLP interacts with endogenous RalA protein 170 Figure 20. RACK1/ GBLP depleti on phenocopies the effects of Ral depletion in Ras-transformed ovarian epithelial cells 172

PAGE 9

The Role of RalA and RalB in Cancer Samuel C. Falsetti ABSTRACT Ras genes are frequently mutated in human cancers and present compelling targets for therapeutic intervention. While previ ous attempts to directly inhibit oncogenic Ras function have largely been unsuccessful us e of targeted agents to inhibit the three primary oncogenic pathways activated by muta ted Ras: RalGEF-Ral, PI3K-Akt and RafMEK-Erk, is an area of intense investigati on. Here, we describe the ability of a novel pharmacological inhibitor of geranylgeranyltransferase I GGTI-2417, to inhibit Ral prenylation and localization. We further used a Ral rescue system to selectively preserve RalA and RalB function and localization during GGTI-2417 treatment and determine the precise roles for inhibition of Ral pre nylation in the GGTI anti-cancer response. Specifically, we determined inhibition of RalA is required for GGTI-attenuation of anchorage independent growth whereas inhibition of RalB is required for inhibition of proliferation, induction of apoptosis, suppression of survivin and induction of p27Kip1. We next determined the role of RalGEF-Ral signaling as well as PI3K-Akt and Raf-MEKErk signal transduction pathways in an in vitro model of human ovarian surface epithelial (T80 HOSE) cell Ras-dependent transforma tion. Using both small interfering RNA (siRNA) and pharmacological inhibitors of Ra l, PI3K and MEK we determined that Ras signaling via Ral and PI3K but not MEK is required for ovarian oncogenesis.

PAGE 10

Furthermore, stable expression of Ras muta nts unable to activate Raf-MEK-Erk signaling were able to robustly transform T80 cells. Si nce we had confirmed the importance of Ral proteins to human epithelial malignancies we next sought to explore the molecular interactions governing Ral transformation using a proteomics approach to rapidly identify proposed Ral interacting partners. Using i mmunoprecipition of tran siently overexpressed FLAG-tagged RalA and RalB followed by 1D-gel separation and tandem MS/MS analysis we determined a database of propos ed Ral interacting proteins. One of these, RACK1, is a validated RalA and RalB intera cting protein which is at least partially required for Ras and Ral transformation. Thes e results provide both a strong impetus and a solid basis for future studies into the mechanisms of RalAand RalBdependent transformation.

PAGE 11

List of Abbreviations AIG: Anchorage independent growth EGFR: Epidermal growth factor receptor FTase: Farnesyltransferase FTIs: Farnesyltransferase inhibitors GAP: GTPase activating protein GBLP1: guanine nucleotide binding-like protein-1 GEF: Guanine nucleotide exchange factor GGTase I: Geranylgeranyltransferase I GGTI: Geranylgeranyltransferase I inhibitors HDJ2: Human DNAJ-2 HEK: Human embryonic kidney cells HMECS: Human mammary epithelial cells HOSE: Human ovarian surface epithelial cells hTERT: human telomerase catalytic subunit IEC-6: Rat intestinal epithelial cells-6 PI3K: phosphotydil inositol 3,4,5-triphosphate kinase PLC-E: Phospholipase C epsilon PLD-1: Phospholipase D-1 PP2A: Protein phosphotase-2A RACK1: Receptor for activated protein kinase C-1 RalA: Ras-like A RalB: Ras-like B

PAGE 12

RalBP1: Ral binding protein-1 RalGDS: Ral-guanine nucleoti de dissociation stimulator RalGEF: Ral-guanine nucleotide exchange factor RasGAP: Ras GTPase activating protein RIE: Rat intestinal epithelial cells ROSE: Rat ovarian surface epithelial cells RTK: Receptor tyrosine kinase SCID: Severely compromised immuno-deficient mouse siRNA: Small interfering RNA shRNA: Small hairpin RNA SOS: Son of sevenless SV40T: simian virus40 large T antigen S40t : simian virus-40 small t antigen ZONAB : ZO-1 binding protein

PAGE 13

Chapter 1 Introduction By Samuel C. Falsetti 1,2, Sad M. Sebti 1,2,* 1Drug Discovery Program, The H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL 2Departments of Interdisciplinary Oncology a nd Molecular Medicine, The University of South Florida, Tampa, FL *Corresponding Author: 12902 Magnolia Drive, Tampa, FL 33612; Tel (813) 745-6734; Fax (813) 745-6748; email: said.sebti@moffitt.org

PAGE 14

The Ras Superfamily In order to understand the role of Ral in cancer it is necessa ry to understand the role of Ras genes in oncogenesis. Members of the Ras and Rho branches of the Ras superfamily of small GTPases are critically involved in the regulation of many biological events critical to the regulation of cellular homeostasis such as cell cycle control, cell survival, death, differentiation, development a nd growth (11, 93). The aberrant activation or inactivation of Ras family proteins is be lieved to be important in the induction of oncogenesis. In addition to the three Ras prot eins (H-, Nand K-Ras), other Ras family proteins with validated roles in oncogene sis include R-Ras, Ral, Rheb, Di-Ras and Noey2/ARHI small GTPases. Rho family GTPases (e.g., RhoB, RhoC, Rac1b, DBC2) are also implicated in oncogenesis (7 3). There are 156 known genes in the Ras superfamily and at least 166 known Ras supe rfamily isoforms (see Table 1 on following page).

PAGE 15

2 2 Ral 10 9 Other 1 1 Ran 30 27 Sar1/ Arf 63 61 Rab 22 20 Rho 39 36 Ras Known isoforms Human genes Family Total 156 166 Adapted from WennerbergK, RossmanKL, DerCJ. Journal of Cellular Science .2005; 118(5); 843-846. Table 1: The RasSuperfamily

PAGE 16

Members of the Ras superfamily are also referred to as molecular switches which cycle from the on (GTP-bound) to off position (GDP-bound) through an intrinsic GTPase activity (Figure 1, following page). Two classes of proteins regulate the catalytic rate of Ras superfamily GTPase activity: GTPase activating proteins (GAPs) and guanine-nucleotide exchange f actors (GEFs). GAP pr oteins activate the intrinsic GTPase activity of Ras family members causing Ras to remain in a preferentially GDP-bound state whereas GEFs catalyze the exchange GDP for GTP causing Ras to be in a preferentially GTP-bound state (2, 16, 18, 46, 102) Mutational activation of Ras proteins typically involves point mutation of key re sidues, for example amino acids 12 and 61 of Ras (see Figure 2, following page), essential for GAP binding; this gradually reduces the GTPase activity of Ras and keeps it in a persistently activated GTP-bound state (17, 63, 79, 87, 95, 96). The nucleotide bound state dictates the orientation of the effector loop regions of Ras family proteins; when GT P is bound the effector loop regions become accessible to the Ras-binding regions of cogna te effector proteins activating multiple downstream effector pathways (60).

PAGE 17

GEFsON Ras GDP RasGAP Sos1/2 Grb2 RTKs Ras GTP OFF Figure 1: The RasGTP-GDP Cycle Abbreviations: RTKs(receptor tyrosine ki nases), SOS (son of sevenless), Ras-GTP (GTP-bound Ras), Ras-GDP (GDP-bound Ras) GEFs(guanine nucleotide exchange factors), RasGAP(RasGTPaseactivating protein)

PAGE 18

Adapted from PaiEF, KrnngelU, PetskoGA, et al .EMBO Journal. 1990; 9(8); 2351-9. Figure 2: RasGTPaseDomains and Point Mutations H-Ras Switch I Switch II *** Amino acid 12 Amino Acid 61 CAAX

PAGE 19

Ras proteins engage over twenty known eff ectors, which in turn activate a variety of other proteins. Three of the Ras effector pathways, phos photidyl inositol tri-phosphate kinase (PI3K)/Akt, Raf/MEK/Erk and Ral gua nine nucleotide disso ciation stimulator (RalGDS)/Ral) (7, 21, 35, 81) ar e at least partially required for constitutively activated Ras to initiate oncogenesis in certain cell types (26, 31, 41, 45, 64, 65, 71) [Figure 3, see following page]. Ras proteins, as the archet ypical oncogenic small GTPases, are the most intensively studied members of the Ras supe rfamily. However, the vast majority of known small GTPases, despite being similar to Ras, are far less ex tensively studied and play undefined roles in oncogenesis. Furthe rmore, these gene products have, at least potentially, as many or more effectors as Ras. Some of these proteins, such as members of the Rac and Ral subfamilies, are known to interact directly with effectors which are distinct from those of Ras. Taken as a w hole these observations suggest that the true range of cellular pathways governed by sma ll GTPases remain to be determined.

PAGE 20

Ras-GTP RalGEF RalAGTP RAF MEK Erk1/2 PI3K Akt RalBGTP TRANSFORMATION Cell cycle progression Survival, AIG, Motility Invasion, Transcription Tiam1 NORE1 Rin1 AF6 PLC-E Figure 3: RasSignaling and Transformation Abbreviations: Ras-GTP (GTP-bound Ra s), Ral-GEF (Ralguanine nucleotide exchange factor), RalA/B-GTP (G TP-bound Ral), PLC-E (phospholipaseC epsilon), AIG (anchorage independent growth)

PAGE 21

The Role of Prenylation in Ras Superfamily Function The oncogenic functions of the Ras and Rho proteins require posttranslational processing by prenyltransferase enzymes ( 37, 39, 101). The two enzymes responsible for prenylation of Ras family proteins are farnesyltransferase (FTase) and geranylgeranyltransferase I (GGTase I) ( 6, 8, 9, 55, 74, 75, 100, 101), which covalently attach the 15-carbon farnesyl and 20-carbon ge ranylgeranyl lipids, respectively, to the cysteine of proteins with the carboxy te rminal tetrapeptide cons ensus sequence CAAX, (C is cysteine, A is any aliphatic amino acid and X is any carboxyl-terminal amino acid), see figure 2. In general, FTase farnesylates pr oteins in which X is methionine or serine (70), whereas GGTase I geranylgeranylates protei ns in which X is leucine or isoleucine (24). Ras proteins that ar e mutationally rendered unprenylatable lose their oncogenic activity and fail to properly lo calize within the cell (37, 39). Similarly, prenylation of other proteins in the Ras and Rho families is essential to their activities (1, 37, 49) The fact that prenylati on is required for the oncogen ic activity of small GTPases prompted us and others to design FTase a nd GGTase I inhibitors (FTIs and GGTIs) as potential anticancer drugs (28, 50, 68, 105). While numerous studies have shown that FTIs suppress oncogenic and tumor survival pathways, the actual mechanism by which FTIs inhibit tumor growth is not known (67, 76). Thus, while designed originally as antiRas inhibitors, the Ras isof orms most commonly mutated in human cancers (Nand KRas) escape FTI inhibition by undergoing alternative pre nylation by GGTase I (72, 85, 94). Therefore, the critical fa rnesylated proteins that FTIs target to induce these effects are not known (67, 76). Similarly, the GGTase I substrates important for the anti-tumor

PAGE 22

activity of GGTIs are not cl early understood yet. While Rac1 and Rac3 Rho family proteins have been implicated as candida te targets for GGTIs (38), other important targets remain to be identified. Previous st udies have demonstrated that GGTIs, at least partially require, the inhibition of Akt serine /threonine kinase and expression of survivin (15, 83), however it remains unknown how inhibi tion of geranylgeranylated proteins results in downregulation of these critical targets. Furthermore, GGTIs also induce p21waf, inhibit CDK activity, phosphorylation of Rb and lead to G0/G1 cell cycle accumulation (61, 68, 83-85, 90). In animal models, GGTIs both inhibit tumor growth in nude mouse xenografts and induce tumor regr ession in transgenic mice (82, 83). The GGTase I substrates targeted by GGTIs to i nduce their anti-neoplastic effects are not known. Logical candidates include other Ras and Rho family proteins with roles in oncogenesis, such as the Ra sl ike RalA and RalB small GTPases. The increasing evidence for Ral GTPases in oncogenesis prompted our interest in evaluating Ral GTPases as important targets for GGTI anti-tumor activity. Both RalA and RalB C-termini contain a CAAX sequence that predicts prenylation and RalA has been shown to be geranylgeranylated (43). Furthermore, RalA and RalB are involved in many oncogenic steps that are inhibited by GGT Is. Therefore, in chapter two of this thesis we investigated whether some of the anti-neoplastic effects of GGTIs are mediated by inhibition of the geranyl geranylation and function of RalA and/or RalB.

PAGE 23

Tissueand SpeciesSpecific Mech anisms of Ras Transformation Understanding the mechanisms of Ras transformation has become a critical component of developing a new generation of targeted therapeutics. Over 30% of human cancers have mutationally activated Ras (32). As we review in this section, over twenty years of research, using a variety of in vitro model systems, has defined a series of species and tissue-specific mechanisms of Ras-transformation. These in vitro cell systems fall into two general categorie s: spontaneously immortalized and genetically definedmodel systems; both of these are minimally transformed systems which provide a platform to assess the effects of stable expression of oncogenic Ras (30, 32). Furthermore, these systems provide an eas ily evaluable means of determining the mechanism by which Ras drives speciesand tissuespecific transformation. In this section we first review the mechanisms of Ras-transformation in spontaneously immortalized cell systems followed by the more recently used genetically defined models of human cancer.

PAGE 24

Use of Spontaneously Immortalized Ce ll Systems to Study Ras Transformation Three of the Ras signaling pathways, Ra lGEF/Ral, PI3K/AKT and Raf/Mek/Erk have been primarily implicated in the ability of oncogenic Ras to initiate and maintain the transformed phenotype (41, 45, 51, 75, 78). The Raf/Mek/Erk pathway has been widely assumed to be the primary pathway of Ras transformation in all cell types. This inaccurate assessment is mostly derived from initial reports which, correctly, determined that the transformative capabilities of Ras oncogenes relied upon activity of the Raf/Mek/Erk pathway in spontaneously immo rtalized mouse fibroblast NIH-3T3 cells (41, 45). However, a variety of other spontaneo usly immortalized cell systems have been used to describe an array of speciesand tissuespecific requirements for Ras transformation. Spontaneously immortalized ra t intestinal epithelial (RIE) cells have been used as a model system to assess the involvement of the Raf, Akt and Ral signaling pathways in Ras mediated transformation. Specificall y, Sheng and colleagues (77) found that pharmacologic blockade of PI3K/Akt inhib ited Ras mediated transformation however, myristolated-Akt was not sufficient for tr ansformation but was transforming when coexpressed with a constitutively activated C-Raf mutant (Delta-Raf-22W). Further work by Der and colleagues (64) compared transformation of NIH-3T3 side by side with RIE cells. While both Ras and a constitutively activated C-Raf (Raf-CAAX) could transform NIH-3T3 only Ras could independently transf orm RIE cells. Gangarosa and colleagues (27) compared the ability of K-Ras12V, H-Ras12V and Raf-CAAX to transform RIE (and the similar intestinal epithelial cell line6, IEC-6 cells) by measuring the ability of the

PAGE 25

oncogenes to elicit an external epidermal grow th factor receptor (EGFR) autocrine loop. In these cells oncogenic Ras, but not Raf, could elicit this autocrine loop and transform RIE and IEC6 cells. Further work by this gr oup in collaboration with Der and colleagues (65) demonstrated that while Ras transfor mation in RIE cells could be blocked by the MEK inhibitor PD98059, Ras mutants defective for Raf activation could still transform RIE cells, albeit at a reduced potency. These re sults argue that while both Raf and PI3K signaling are partially required for Ras transfor mation of rat intestin al epithelial cells neither is sufficient to indepe ndently initiate oncogenesis. Immortalized huma n mammary epithelial cells (MCF-10A) derived from dysplastic mammary tissue were used as a model system to assess the involvement of various Ras downstream signaling molecules in Ras mediated transformation. Der and colleagues (64) compared transformation of NIH-3T3 side-by-side with MCF-10A cells. Both oncogenic Ras and RafCAAX could transform NIH-3T 3 but only oncogenic Ras was sufficient to transform MCF-10A cells. The Salomon group (57) further determined that oncogenic Ras induced an EGFR-autocri ne loop in MCF-10A cells and determined that both EGFR activation a nd MEK activity were required for full transformation. These results argue for a more complex mechanism of transformation in MCF-10A cells than in NIH-3T3 cells. Spontaneously immortalized rat ovarian su rface epithelial (ROSE) cells have been used to determine the requirements for Ras mediated transformation. In these cells the Der group (88) determined that stable expression of H-Ras12V activated Raf/MEK/ERK signaling but did not activate PI3K signali ng. Here, constitutively active Raf could partially reconstitute both Ras-transforma tion and associated morphology changes. The

PAGE 26

Erickson group further defined this system (12) and determined that all three effector pathways (Raf, PI3K, and RalGDS) were required for full transformation and the production of cathepsin L.

PAGE 27

Use of Genetically Defined Cell Systems to Study Ras Transformation While spontaneously immortalized cell syst ems have proven to be useful models for the study of transformation requirements it is important to note that spontaneous immortalization proceeds through poorly defined processes. Conversely, genetically tractable models of human epithelial immortalization, while experimentally defined, offer an alternative means of investigating mechanisms of specific oncogene induced transformation in cells with a defined genetic background which closely mimics the genetic abnormalities present in the early and late stages of human carcinogenesis (51). Human embryonic kidney cells sequentially immortalized by th e catalytic subunit of human telomerase (hTERT) and the SV40-large T and small t antigens (which inactivate the key tumor suppr essors p53, Rb and protein phosphotase-2A), or HEK-HT cells, are derived from the embryonic mes oderm. HEKs are a mixed lineage cell line known to express both epithelial and mesenchymal markers that become fully transformed by stable expression of oncogeni c Ras (30). The Counter group determined that H-Ras12V37G was, by itself, sufficient for transformation in this cell line (31). This concept was further supported by the demonstration that constitutively activated Rlf or activated RalA, but not the related RalB, was capable of transforming these cells (51). In contrast to these results the Weinberg gr oup further defined the requirements for Rasmediated transformation of this system (69). Rangarajaran and colleagues found that only paired expression of H-Ras12V37G (only activating RalGDS) and H-Ras12V40C (only activating PI3K), but not H-Ras12V35S (only activating Raf) and H-Ras12V40C nor HRas12V37G and H-Ras12V35S, could fully transform HEK-HT cells. This concept was further

PAGE 28

supported by evidence that stable expression of constitutively activated Akt and Rlf (a Ral-GEF) could induce transf ormation in soft agar. In contrast, human mammary epithelial cells (HMECs) immortalized by hTERT and SV-40-large T antigen, HMECs, become fully transformed by stable expression of oncogenic Ras (19) but differ in the Rasdownstream signaling requirements from HEKHTs. The Weinberg group has investigat ed the requirements for Ras-mediated transformation of this system (69) and determ ined that only co-activation of all three primary Ras effectors (Raf, RalGDS, and PI3K) can transform HMEC cells. A variety of other human-derived genetic ally defined models of transformation have been used to address mechanisms of Ras transformation. Human fibroblasts immortalized by hTERT and SV4040-large T antigen, BJ fibroblasts, become fully transformed by stable expression of oncoge nic Ras (30). The Weinberg group further defined the requirements for Ras-mediated transformation of this system (69). Rangarajaran and colleagues found that only paired expression of H-Ras12V35S (only activating Raf-1) and H-Ras12V37G (only activating RalGDS), but not H-Ras12V35S and HRas12V40C not H-Ras12V37G and H-Ras12V40C, could fully transform BJ fibroblasts. Human esophageal epithelial cells immortalized by hTERT and SV40, EPC2 cells, can be fully transformed by either H-Ras12V, c-Myc or activated AKT (42). A variety of other tissue types following the same se quence of genetic events (hTERT activation, inactivation of p53 and R b, and transduction of oncogenic Ras), but using non-viral approaches, have been used to create other genetically defined models of human cancer. Counter and colleagues (40) report the creatio n of an alternative human mammary epithelial cell system, the first human myoblast cell system, and an alternative

PAGE 29

human embryonic kidney cell system. Linard ic and colleagues re port the creation of rhabdomyosarcoma model systems from huma n skeletal muscle cell precursors and committed human skeletal muscle myoblasts (5 3). Also, human bronchi al epithelial cells immortalized by hTERT, CDK4 and siRNA to p53, HBECs, can be partially transformed by stable transduction of K-Ras12V or mutant EGFR (E746-A750 del or LL859R). The Minna group determined that either Ras or EGFR could induce transformation and invasion in a 3-dimensional organotypic culture (103). Additionally, while not gene tically defined nor able to be fully transformed human thyroid epithelial prim ary cells can be induced to a hyper-proliferative state by expression of oncogenic Ras which closely re sembles the phenotype of an early stage thyroid tumor, follicular adenoma. The Wynford-Thomas group (4) used these cells to investigate the contributions of the Ras-eff ectors Raf, PI3K and Ra lGDS to induction of thyrocyte hyperproliferation. Co-expression of the three Ras effector mutants which only bind RalGDS, Raf, and PI3K could fully recapit ulate the hyperproliferative state induced by fully oncogenic Ras while combinations of any of these two mutants was not sufficient. Importantly for our studies, human ovarian surface epithelial cells immortalized by stable expression of hTERT and SV40-larg e T antigen, T80 cells, have been recently established as models for human ovarian onc ogenesis. These cells were created by the Bast group (54) and are fully transformed following stable expression of either H-Ras12V or K-Ras12V. A similar model system, created by Kusukari and colleagues (48), could also be transformed by stable expression of either H-Ras12V or c-erbB-2.

PAGE 30

Use of an In Vitro Model of Human Ovarian Cancer to Study the Mechanism of Ras Transformation: Implications for RalA and RalB Ovarian cancer is the primary cause of death among gynecological malignancies and is the 5th leading cause of cancer deaths among women in the United States (29). Despite the high incidence of ovarian cancer among women and the rela tive lethality of the disease little is known about the precise mechanism of transformation of human ovarian surface epithelial (HOSE) cells. Ind eed, multiple mechanisms of ovarian surface epithelial cell transformation have been proposed and are believe d to be sub-type specific (3). The two most common sub-types of ovari an cancer, high-grade serous and mucinous ovarian carcinoma (29), are comm only diagnosed at late-stage and are frequently lethal due to a high-propensity for both metastasis and drug-resistant recu rrence (3, 29). It is thus of critical importance to define the mechanism of ovarian epithelial cell transformation both for a more precise unders tanding of the underlying genetic risk factors as well as for the development of targeted therapies Epithelial neoplasias, such as ovarian can cers, exhibit multiple genetic aberrations in key tumor suppressor genes, so-called cellular gatekeepers such as p53 and Rb (32). Indeed, over 50% of ovarian cancers exhi bit missense mutations in P53 (47, 56). Furthermore, over 80% of human ovarian can cers examined have been shown to have mutational defects in one or more of the ge nes involved in the Rb tumor suppressive pathway, such as CDK4, cyclin D1, and Rb ( 33). Activation of human telomerase activity is another well-recognized defect of human epithelial neoplasms (32) with ovarian cancer being no exception to this rule (13). In addition to p53, Rb and hTERT genetic

PAGE 31

abnormalities activation of one or more proto-oncogenes is a pivotal fourth event in the development of malignant disease. One of the most well studied mechanis ms of oncogenesis in human cancer is mutational activation of members of the Ras ge ne family. Mutations in Ras gene family members (H-, K-, and NRas) are known to occur in approximately 30% of human cancers (78). Numerous studies have demonstrated that mutational activation of K-Ras, through missense mutations at codons 12 and 13, occurs with high frequency (27-69%) in certain subtypes of ovarian cancers (14, 25, 36, 62, 89). Similar mutations in the H-Ras proto-oncogene have been determined to occu r in ovarian cancer subtypes though there is considerable disagreement over the fre quency (20, 89, 104). For example, while one study has found that H-Ras activations occur with variable frequency depending on subtype (33% of borderline ovarian tumors, 13% of mucinous adenocarcinomas, and 7% of serous adenocarcinomas and in 6% of primary invasive ovar ian carcinomas) (89) others have found H-Ras mutations occu r infrequently (0-1 2.5%) (20, 104) A role for Ras mutation in ovarian ca ncer progression is especially well recognized in mucinous sub-type ovarian tu mors where the incidence of K-Ras protooncogene mutation increases profoundly duri ng the progression of disease (62). For example, Mok and colleagues found a high, pr ogression-dependent incidence of K-Ras activation at codons 12 and 13 in 13% of mucinous adenomas, 33% of mucinous tumors at the borderline, and 46% of mucinous carcinomas (62). Sim ilar results, in which K-Ras mutation incidence increases dramatically throughout disease progression, have been reported by many other groups (14, 20, 36, 62, 89). These studies have primarily examined mutations at codons 12 and 13 how ever, activating Ras mutations are also

PAGE 32

known to occur in some cancers at codon 61; thus it is possible that the actual incidence of Ras mutations in mucinous ovarian cancer has been underestimated. K-Ras mutation have also b een detected, albeit to a le sser degree (0-12%), in high-grade serous sub-type ovarian can cer, the most common form of ovarian malignancy (62). It is interesting to note that a much higher incidence of Ras mutation exists in low-grade serous malignancy. Ind eed, over 68% of low grade serous tumors have either a K-Ras or B-Ra f mutation (59). Furthermore, these mutations occur in a mutually exclusive (59) manner suggesting that the primary mechanism of K-Ras mutation in low-grade serous is through BRaf activation. Additionally, the presence of K-Ras proto-oncogene mutations in up to 48% of borderlin e ovarian epithelial tumors (62) indicates a role for K-Ras mutation in th e early events of low-grade serous ovarian neoplasia. Given the lack of B-Raf mutations in high-grade serous carcinoma the precise mechanism of Ras mediated transformation in this tumor sub-type remains a mystery, as does the etiology of disease progression. While genetic analysis of patient tumors of multiple ovarian subtypes suggest Ras mutation is a frequent occurrence analysis of downstream Ras effectors, such as Raf, has not yielded a consensus view of the mechanism for Ras transformation. However, other methods are available for determining the necessary components of ovari an cell-type specific Ras transformation. Recently, Der and colleagues described th e mechanism of Ras transformation in spontaneously immortalized rat ovarian surface epithelial (ROSE) cells (88). They assessed the transformative ability of retr ovirally transduced H-Ras and constitutively active mutants of the Ras downstream effectors Raf, RalGDS and the catalytic subunit of PI3K (p110) to determine, for the first time, which Ras effectors are capable of

PAGE 33

independent ovarian cell transformation. In this system Raf, but not p110 or Ral-GDS, was capable of independently transf orming in a manner similar to H-Ras12V. Interestingly, H-Ras12V was unable to activate PI3K catalytic activity or stimulate activation of Akt in this cell lin e. These results sugge st that Ras effector utilization varies in a tissue specific manner. While this system has been useful in suggesting which signaling cascades are required for Ras transf ormation in ovarian ep ithelial cells, the ROSE system is neither genetically defi ned nor of human origin and thus, while suggestive of a role for Raf in ovari an transformation, has unknown prognostic significance for human ov arian oncogenesis. Since that seminal discovery by Der a nd colleagues a genetically defined human model of ovarian malignancy has been devel oped which recapitulates the critical events in neoplasia (54). The T80 cell line is de rived from primary human ovarian surface epithelial cells engineered to express SV40T and hTERT. Upon the addition of retrovirally transduced H-Ras12V these cells, referred to as T80H, acquire anchorage independent growth capability in soft agar, a reliable estimate of transformation, and are capable of forming xenograft tumors in a SCID mouse model. This transformation requires continued expression of H-Ras (98) and, as such, is a bona fide model of Rasdependent, ovarian cell-type-speci fic transformation. In chapter 3 of this thesis we have used these cells as well as our own stable K-Ras12V expressing variant (T80K) to investigate the required downstream signaling components of Ras transformation in human ovarian surface epithelial cells.

PAGE 34

Ral Proteins are Central Mediators of Oncogenesis In further confirmation of a role for the Ral small GTPases in oncogenesis both RalA and RalB, which are over 85% similar, ha ve been identified as critical divergent mediators of multiple tumorigenic processe s, including metastasis, invasion, anchorage independent growth, survival and cell motility (10, 22, 66, 91, 92, 97). Specifically, depletion of RalA by siRNA has been s hown to inhibit anchorage independent proliferation of multiple human cancer cell line s, such as the HeLa cervical and SW680 prostate cancer cell lines (10) as well as multiple human pancreatic cancer cell lines (52). Also, RalA expression is required for anchor age independent growth of sequentially Ras transformed human ovarian (see Chapter 3) and kidney cells (51). Similarly, stable depletion of RalA, but not RalB, by shRNA ha s been shown to inhibit tumor formation and metastasis of multiple human pancreatic cancer cell lines in athymic nude mice (52). Further confirming a role for RalA in metast asis, stable overexpression of constitutively activated RalA has been show n to promote both standard an d experimental metastasis in vivo (86, 91) and in human prostate cancer cell s, stable expression of activated RalA promotes bone, but not brain, metastasis (99) Even in non-epithelia l cancers there is emerging evidence that RalA is involved in tumorigenicity; for example, in human HT1080 fibrosarcoma cells stable overexpression of activated RalA promotes anchorage independent growth (97). In agreement with these findings two cellu lar processes thought to promote metastasis and anchorage independent growth: cell motility and invasion, are inhibited by siRNA-mediated de pletion of RalA in multiple human renal cancer cell lines (66).

PAGE 35

While RalA has been well validated to play an essential role in anchorage independent processes in multiple tissue types the role of RalB has been determined to be far more tissue specific. For example, while de pletion of either RalA or RalB inhibits anchorage independent growth in human ovarian epithelial cells transformed by either Hor K-Ras12V (Chapter 3), only depl etion of RalA, but not RalB, inhibits H-Ras12V transformation of HEK-HT cells (51). We, as we ll as others have previously determined a role for RalB, but not RalA in the survival of multiple human cancer cell lines (10, 22). Specifically, Whites group has found that depl etion of RalB by siRNA induces apoptosis in both HeLa and SW680 human cancer cell lines (10). Similarly, our results in chapter two have established, using a chemical biol ogy approach, that inhi bition of a RalB, but not RalA, survival pathway underlies the apopt otic response to geranylgeranyltransferase I inhibitors (GGTIs) (22). Specifically, Ra l proteins have been shown to require geranylgeranylation for localization and inte raction with other proteins (34, 44, 58, 80). Importantly, stable expression of GGTI-resistant RalB in the human pancreatic carcinoma cell line MiaPaCa2 renders cells resistant to GGTI-induc ed apoptosis (22). Taken as a whole, these studies strongly suggest a role for Ra lB tumor survival in a wide array of tissue types as well as a tissue speci fic role in anchorage independent growth.

PAGE 36

Ral Effectors and Contro l of Cellular Processes Despite the wealth of information regard ing the role of Ral small GTPases in transformation very little is known about th e pathways through which Ral exerts these effects. For example, while RalA and RalB bear a high homology to Ras and over 20 effectors are known for the Ras isoforms (5, 78),only 6 Ral effectors have been described: phospholipase D1 (PLD1), the ex ocyst components Sec5/ Exo84, Filamin A, ZO-1 N-terminally associated binding pr otein (ZONAB) and Ra l binding protein-1 (RalBP1/RLIP) (23). Precise roles for thes e proteins in the various transformation specific processes that are regulated by RalA and RalB remain poorly defined. Despite the fact that RalA and RalB are commonly t hought of as Ras effector proteins, both RalA and RalB are found in the hypera ctivated state inde pendently of Ras in human pancreatic tumors (51, 52). Thus, novel means of Ral activ ation and inactivation may constitute an important and undescribed mechanism of transformation. However, no Ral GTPase activating proteins (RalGAPs), which would negatively regulate RalA and RalB, have been described to date. Also, RalA and RalB are known to regulate diverse physiological processes such as signaling via STAT3, NF -KB, JNK and AFX (23). However, the intermediate proteins in these signal tr ansduction pathways through which RalA and RalB exert these functions remain unknown. In an effort to more fully understand the protein interactions that govern the biological activities of RalA and RalB we have used proteomic analysis, in chapter 4 of this thesis, to describe a proposed database of Ral interacting proteins for the first time. We have uncovered a wealth of potential RalA and RalB interacting partners and used a

PAGE 37

systems biology approach to analyze the broa d themes that emerge from this proposed database. We have further demonstrated that one of these proteins, receptor for activated C-kinase (also known as GBLP1 or RACK1), is a potential RalA and RalB interacting protein that is required for both Ras and Ra l mediated transformation of human ovarian surface epithelial (T80) cells. In addition to uncovering a critical role for RACK1 expression in both Rasand Ra lmediated transformation our proteomics study is the first large-scale analysis of the Ral inter actome. This database will provide the emerging field of Ral study with a wealth of potential RalA and RalB biological effectors and potential regulators of Ral cellular activities.

PAGE 38

Figure 4: RalSignaling Abbreviations: Ral-GTP (GTP-bound Ral), PLD1 (phospholipaseD1), RalBP1 (RalBinding Protein 1), CD C-GDP (GDP-bound CDC42), Rac-GDP (GDP-bound Rac) Exocysttargeting Actin regulation Receptor endocytosis Mitotic spindle polarization Ral-GTP Filamin-A Exo84 Sec5 RalBP1 PLD1 Rac-GDP Cdc42-GDP ZONAB

PAGE 39

References 1. Adamson, P., C. J. Marshall, A. Hall, and P. A. Tilbrook. 1992. Posttranslational modifications of p21rho proteins. J Biol Chem 267: 20033-8. 2. Adari, H., D. R. Lowy, B. M. Willumsen, C. J. Der, and F. McCormick. 1988. Guanosine triphosphatase activating prot ein (GAP) interacts with the p21 ras effector binding domain. Science 240: 518-21. 3. Borresen, A. L. 1992. Oncogenesis in ovarian ca ncer. Acta Obstet Gynecol Scand Suppl 155: 25-30. 4. Bounacer, A., A. McGregor, J. Skinner, J. Bond, Z. Poghosyan, and D. Wynford-Thomas. 2004. Mutant ras-induced pro liferation of human thyroid epithelial cells requires three effector pathways. Oncogene 23: 7839-45. 5. Campbell, S. L., R. Khosravi-Far, K. L. Rossman, G. J. Clark, and C. J. Der. 1998. Increasing complexity of Ras signaling. Oncogene 17: 1395-413. 6. Casey, P. J., P. A. Solski, C. J. Der, and J. E. Buss. 1989. p21ras is modified by a farnesyl isoprenoid. Proc Natl Acad Sci U S A 86: 8323-7. 7. Chen, J. M., S. Manolatos, P. W. Brandt-Rauf, R. B. Murphy, R. Monaco, and M. R. Pincus. 1996. Computed three-dimensional structures for the rasbinding domain of the raf-p74 protein co mplexed with ras-p21 and with its suppressor protein, rap1A. J Protein Chem 15: 511-8. 8. Chen, W. J., D. A. Andres, J. L. Goldstein, and M. S. Brown. 1991. Cloning and expression of a cDNA encoding th e alpha subunit of rat p21ras protein farnesyltransferase. Proc Natl Acad Sci U S A 88: 11368-72.

PAGE 40

9. Chen, W. J., D. A. Andres, J. L. Goldst ein, D. W. Russell, and M. S. Brown. 1991. cDNA cloning and expression of the peptide-binding be ta subunit of rat p21ras farnesyltransferase, the c ounterpart of yeast DPR1/RAM1. Cell 66: 327-34. 10. Chien, Y., and M. A. White. 2003. RAL GTPases are linchpin modulators of human tumour-cell proliferation and survival. EMBO Rep 4: 800-6. 11. Colicelli, J. 2004. Human RAS superfamily protei ns and related GTPases. Sci STKE 2004: RE13. 12. Collette, J., A. S. Ulku, C. J. Der, A. Jones, and A. H. Erickson. 2004. Enhanced cathepsin L expression is mediat ed by different Ras effector pathways in fibroblasts and epitheli al cells. Int J Cancer 112: 190-9. 13. Counter, C. M., H. W. Hirte, S. Bacchetti, and C. B. Harley. 1994. Telomerase activity in human ovarian carcinom a. Proc Natl Acad Sci U S A 91: 2900-4. 14. Cuatrecasas, M., A. Villanueva, X. Matias-Guiu, and J. Prat. 1997. K-ras mutations in mucinous ovarian tumors: a clinicopathologic a nd molecular study of 95 cases. Cancer 79: 1581-6. 15. Dan, H. C., K. Jiang, D. Coppola, A. Hamilton, S. V. Nicosia, S. M. Sebti, and J. Q. Cheng. 2004. Phosphatidylinositol-3-OH kinase/AKT and survivin pathways as critical targets for gerany lgeranyltransferase I inhibitor-induced apoptosis. Oncogene 23: 706-15. 16. de Vos, A. M., L. Tong, M. V. Mil burn, P. M. Matias, J. Jancarik, S. Noguchi, S. Nishimura, K. Miur a, E. Ohtsuka, and S. H. Kim. 1988. Threedimensional structure of an oncogene prot ein: catalytic domain of human c-H-ras p21. Science 239: 888-93.

PAGE 41

17. Der, C. J., T. G. Krontiris, and G. M. Cooper. 1982. Transforming genes of human bladder and lung carcinoma cell lines are homologous to the ras genes of Harvey and Kirsten sarcoma viruses. Proc Natl Acad Sci U S A 79: 3637-40. 18. Downward, J., R. Riehl, L. Wu, and R. A. Weinberg. 1990. Identification of a nucleotide exchange-promo ting activity for p21ras. Proc Natl Acad Sci U S A 87: 5998-6002. 19. Elenbaas, B., L. Spirio, F. Koerner, M. D. Fleming, D. B. Zimonjic, J. L. Donaher, N. C. Popescu, W. C. Hahn, and R. A. Weinberg. 2001. Human breast cancer cells generated by oncogenic transformation of primary mammary epithelial cells. Genes Dev 15: 50-65. 20. Enomoto, T., C. M. Weghorst, M. Inoue, O. Tanizawa, and J. M. Rice. 1991. K-ras activation occurs frequently in mucinous adenocarcinomas and rarely in other common epithelial tumors of the human ovary. Am J Pathol 139: 777-85. 21. Esser, D., B. Bauer, R. M. Wolthuis, A. Wittinghofer, R. H. Cool, and P. Bayer. 1998. Structure determination of the Ras-binding domain of the Ralspecific guanine nucleotide exchan ge factor Rlf. Biochemistry 37: 13453-62. 22. Falsetti, S. C., D. A. Wang, H. Peng, D. Carrico, A. D. Cox, C. J. Der, A. D. Hamilton, and S. M. Sebti. 2007. Geranylgeranyltransferas e I inhibitors target RalB to inhibit anchorage-dependent gr owth and induce apoptosis and RalA to inhibit anchorage-independent growth. Mol Cell Biol 27: 8003-14. 23. Feig, L. A. 2003. Ral-GTPases: approaching th eir 15 minutes of fame. Trends Cell Biol 13: 419-25.

PAGE 42

24. Finegold, A. A., D. I. Johnson, C. C. Farnsworth, M. H. Gelb, S. R. Judd, J. A. Glomset, and F. Tamanoi. 1991. Protein geranylge ranyltransferase of Saccharomyces cerevisiae is specific for Cys-Xaa-Xaa-Leu motif proteins and requires the CDC43 gene product but not the DPR1 gene product. Proc Natl Acad Sci U S A 88: 4448-52. 25. Fujita, M., T. Enomoto, M. Inoue, O. Ta nizawa, M. Ozaki, J. M. Rice, and T. Nomura. 1994. Alteration of the p53 tumor suppressor gene occurs independently of K-ras activation and more frequently in serous adenocarcinomas than in other common epithelial tumors of the human ovary. Jpn J Cancer Res 85: 1247-56. 26. Fukui, Y., A. R. Saltiel, and H. Hanafusa. 1991. Phosphatidylinositol-3 kinase is activated in v-src, v-yes, and v-fps transformed chicken embryo fibroblasts. Oncogene 6: 407-11. 27. Gangarosa, L. M., N. Sizemore, R. Graves-Deal, S. M. Oldham, C. J. Der, and R. J. Coffey. 1997. A raf-independent epider mal growth factor receptor autocrine loop is necessary for Ras transformation of rat intestinal epithelial cells. J Biol Chem 272: 18926-31. 28. Garcia, A. M., C. Rowell, K. Ackermann, J. J. Kowalczyk, and M. D. Lewis. 1993. Peptidomimetic inhibitors of Ras farnesylation and function in whole cells. J Biol Chem 268: 18415-8. 29. Goodman, M. T., and H. L. Howe. 2003. Descriptive epidemiology of ovarian cancer in the United States, 1992-1997. Cancer 97: 2615-30.

PAGE 43

30. Hahn, W. C., C. M. Counter, A. S. Lundberg, R. L. Beijersbergen, M. W. Brooks, and R. A. Weinberg. 1999. Creation of human tumour cells with defined genetic elements. Nature 400: 464-8. 31. Hamad, N. M., J. H. Elconin, A. E. Karnoub, W. Bai, J. N. Rich, R. T. Abraham, C. J. Der, and C. M. Counter. 2002. Distinct requirements for Ras oncogenesis in human versus mouse cells. Genes Dev 16: 2045-57. 32. Hanahan, D., and R. A. Weinberg. 2000. The hallmarks of cancer. Cell 100: 5770. 33. Hashiguchi, Y., H. Tsuda, K. Yamamoto, T. Inoue, O. Ishiko, and S. Ogita. 2001. Combined analysis of p53 and RB path ways in epithelial ovarian cancer. Hum Pathol 32: 988-96. 34. Hinoi, T., S. Kishida, S. Koyama, M. Ikeda, Y. Matsuura, and A. Kikuchi. 1996. Post-translational modifications of Ra s and Ral are import ant for the action of Ral GDP dissociation stimulator. J Biol Chem 271: 19710-6. 35. Huang, L., X. Weng, F. Hofer, G. S. Martin, and S. H. Kim. 1997. Threedimensional structure of the Ras-interact ing domain of RalGDS. Nat Struct Biol 4: 609-15. 36. Ichikawa, Y., M. Nishida, H. Suzuki, S. Yoshida, H. Tsunoda, T. Kubo, K. Uchida, and M. Miwa. 1994. Mutation of K-ras prot ooncogene is associated with histological subtypes in human mucinous ovarian tumors. Cancer Res 54: 335.

PAGE 44

37. Jackson, J. H., C. G. Cochrane, J. R. Bourne, P. A. Solski, J. E. Buss, and C. J. Der. 1990. Farnesol modification of Kirstenras exon 4B protein is essential for transformation. Proc Natl Acad Sci U S A 87: 3042-6. 38. Joyce, P. L., and A. D. Cox. 2003. Rac1 and Rac3 are targets for geranylgeranyltransferase I inhibito r-mediated inhibition of signaling, transformation, and membrane ruffling. Cancer Res 63: 7959-67. 39. Kato, K., A. D. Cox, M. M. Hisaka, S. M. Graham, J. E. Buss, and C. J. Der. 1992. Isoprenoid addition to Ras protein is the critical modification for its membrane association and transformi ng activity. Proc Natl Acad Sci U S A 89: 6403-7. 40. Kendall, S. D., C. M. Linardic, S. J. Adam, and C. M. Counter. 2005. A network of genetic events sufficient to convert normal human cells to a tumorigenic state. Cancer Res 65: 9824-8. 41. Khosravi-Far, R., M. A. White, J. K. Westwick, P. A. Solski, M. Chrzanowska-Wodnicka, L. Van Aelst, M. H. Wigler, and C. J. Der. 1996. Oncogenic Ras activation of Raf/mitogen-activated protein kinase-independent pathways is sufficient to cause tumo rigenic transformation. Mol Cell Biol 16: 3923-33. 42. Kim, S. H., H. Nakagawa, A. Navaraj, Y. Naomoto, A. J. Klein-Szanto, A. K. Rustgi, and W. S. El-Deiry. 2006. Tumorigenic convers ion of primary human esophageal epithelial cells using oncogene combinations in the absence of exogenous Ras. Cancer Res 66: 10415-24.

PAGE 45

43. Kinsella, B. T., R. A. Erdman, and W. A. Maltese. 1991. Carboxyl-terminal isoprenylation of ras-related GTP-bindi ng proteins encoded by rac1, rac2, and ralA. J Biol Chem 266: 9786-94. 44. Kishida, S., S. Koyama, K. Matsubara, M. Kishida, Y. Matsuura, and A. Kikuchi. 1997. Colocalization of Ras and Ral on the membrane is required for Ras-dependent Ral activati on through Ral GDP dissociat ion stimulator. Oncogene 15: 2899-907. 45. Kolch, W., G. Heidecker, P. Lloyd, and U. R. Rapp. 1991. Raf-1 protein kinase is required for growth of induced NIH/3T3 cells. Nature 349: 426-8. 46. Konstantinopoulos, P. A., M. V. Karamouzis, and A. G. Papavassiliou. 2007. Post-translational modifications and regulation of the RAS superfamily of GTPases as anticancer targets. Nat Rev Drug Discov 6: 541-55. 47. Kupryjanczyk, J., A. D. Thor, R. Beauch amp, V. Merritt, S. M. Edgerton, D. A. Bell, and D. W. Yandell. 1993. p53 gene mutations and protein accumulation in human ovarian cancer. Proc Natl Acad Sci U S A 90: 4961-5. 48. Kusakari, T., M. Kariya, M. Mandai, Y. Tsuruta, A. A. Hamid, K. Fukuhara, K. Nanbu, K. Takakura, and S. Fujii. 2003. C-erbB-2 or mutant Ha-ras induced malignant transformation of immortalized human ovarian surface epithelial cells in vitro. Br J Cancer 89: 2293-8. 49. Lebowitz, P. F., W. Du, and G. C. Prendergast. 1997. Prenylation of RhoB is required for its cell transforming function but not its ability to activate serum response element-dependent transcription. J Biol Chem 272: 16093-5.

PAGE 46

50. Leftheris, K., T. Kline, G. D. Vite, Y. H. Cho, R. S. Bhide, D. V. Patel, M. M. Patel, R. J. Schmidt, H. N. Weller, M. L. Andahazy, J. M. Carboni, J. L. Gullo-Brown, F. Y. Lee, C. Ricca, W. C. Rose, N. Yan, M. Barbacid, J. T. Hunt, C. A. Meyers, B. R. Seizinger, R. Zahler, and V. Manne. 1996. Development of highly potent inhibitors of Ras farnesyltransferase possessing cellular and in vivo activity. J Med Chem 39: 224-36. 51. Lim, K. H., A. T. Baines, J. J. Fiordalisi, M. Shipitsin, L. A. Feig, A. D. Cox, C. J. Der, and C. M. Counter. 2005. Activation of RalA is critical for Rasinduced tumorigenesis of human cells. Cancer Cell 7: 533-45. 52. Lim, K. H., K. O'Hayer, S. J. Adam, S. D. Kendall, P. M. Campbell, C. J. Der, and C. M. Counter. 2006. Divergent roles for RalA and RalB in malignant growth of human pancreatic carcinoma cells. Curr Biol 16: 2385-94. 53. Linardic, C. M., D. L. Downie, S. Qualman, R. C. Bentley, and C. M. Counter. 2005. Genetic modeling of human rhabdomyosarcoma. Cancer Res 65: 4490-5. 54. Liu, J., G. Yang, J. A. Thompson-Lan za, A. Glassman, K. Hayes, A. Patterson, R. T. Marquez, N. Auersperg, Y. Yu, W. C. Hahn, G. B. Mills, and R. C. Bast, Jr. 2004. A genetically defined mode l for human ovarian cancer. Cancer Res 64: 1655-63. 55. Manne, V., D. Roberts, A. Tobin, E. O'Rourke, M. De Virgilio, C. Meyers, N. Ahmed, B. Kurz, M. Resh, H. F. Kung, and et al. 1990. Identification and preliminary characterization of protein-cy steine farnesyltransferase. Proc Natl Acad Sci U S A 87: 7541-5.

PAGE 47

56. Marks, J. R., A. M. Davidoff, B. J. Kern s, P. A. Humphrey, J. C. Pence, R. K. Dodge, D. L. Clarke-Pearson, J. D. Iglehart, R. C. Bast, Jr., and A. Berchuck. 1991. Overexpression and mutation of p53 in epithelial ovarian cancer. Cancer Res 51: 2979-84. 57. Martinez-Lacaci, I., S. Kannan, M. De Santis, C. Bianco, N. Kim, B. WallaceJones, A. D. Ebert, C. Wechselberger, and D. S. Salomon. 2000. RAS transformation causes sustained activation of epidermal growth factor receptor and elevation of mitogen-act ivated protein kinase in human mammary epithelial cells. Int J Cancer 88: 44-52. 58. Matsubara, K., S. Kishida, Y. Matsuura, H. Kitayama, M. Noda, and A. Kikuchi. 1999. Plasma membrane recruitment of RalGDS is critical for Rasdependent Ral activation. Oncogene 18: 1303-12. 59. Mayr, D., A. Hirschmann, U. Lohrs, and J. Diebold. 2006. KRAS and BRAF mutations in ovarian tumors: a compre hensive study of invasive carcinomas, borderline tumors and extraova rian implants. Gynecol Oncol 103: 883-7. 60. Milburn, M. V., L. Tong, A. M. deVos, A. Brunger, Z. Yamaizumi, S. Nishimura, and S. H. Kim. 1990. Molecular switch for signal transduction: structural differences between active and inactive forms of protooncogenic ras proteins. Science 247: 939-45. 61. Miquel, K., A. Pradines, J. Sun, Y. Qian, A. D. Hamilton, S. M. Sebti, and G. Favre. 1997. GGTI-298 induces G0-G1 block and apoptosis whereas FTI-277 causes G2-M enrichment in A549 cells. Cancer Res 57: 1846-50.

PAGE 48

62. Mok, S. C., D. A. Bell, R. C. Knapp, P. M. Fishbaugh, W. R. Welch, M. G. Muto, R. S. Berkowitz, and S. W. Tsao. 1993. Mutation of K-ras protooncogene in human ovarian epithelial tumors of borderline malignancy. Cancer Res 53: 1489-92. 63. Nakano, H., F. Yamamoto, C. Neville, D. Evans, T. Mizuno, and M. Perucho. 1984. Isolation of transforming sequen ces of two human lung carcinomas: structural and functional analysis of th e activated c-K-ras on cogenes. Proc Natl Acad Sci U S A 81: 71-5. 64. Oldham, S. M., G. J. Clark, L. M. Ga ngarosa, R. J. Coffey, Jr., and C. J. Der. 1996. Activation of the Raf-1/MAP kinase cascade is not sufficient for Ras transformation of RIE-1 epithelial cells. Proc Natl Acad Sci U S A 93: 6924-8. 65. Oldham, S. M., A. D. Cox, E. R. Reyn olds, N. S. Sizemore, R. J. Coffey, Jr., and C. J. Der. 1998. Ras, but not Src, transformation of RIE-1 epithelial cells is dependent on activation of the mitoge n-activated protein kinase cascade. Oncogene 16: 2565-73. 66. Oxford, G., C. R. Owens, B. J. Titus, T. L. Foreman, M. C. Herlevsen, S. C. Smith, and D. Theodorescu. 2005. RalA and RalB: antagonistic relatives in cancer cell migration. Cancer Res 65: 7111-20. 67. Prendergast, G. C. 2001. Actin' up: RhoB in cancer and apoptosis. Nat Rev Cancer 1: 162-8. 68. Qian, Y., A. Vogt, A. Vasudevan, S. M. Sebti, and A. D. Hamilton. 1998. Selective inhibition of type-I geranylgeranyltransferase in vitro and in whole cells by CAAL peptidomimetics. Bioorg Med Chem 6: 293-9.

PAGE 49

69. Rangarajan, A., S. J. Hong, A. Gifford, and R. A. Weinberg. 2004. Speciesand cell type-specific requirements for cellular transformation. Cancer Cell 6: 17183. 70. Reiss, Y., S. J. Stradley, L. M. Gierasch, M. S. Brown, and J. L. Goldstein. 1991. Sequence requirement for peptide r ecognition by rat brain p21ras protein farnesyltransferase. Proc Natl Acad Sci U S A 88: 732-6. 71. Rodriguez-Viciana, P., P. H. Warne, A. Khwaja, B. M. Marte, D. Pappin, P. Das, M. D. Waterfield, A. Ridley, and J. Downward. 1997. Role of phosphoinositide 3-OH kinase in cell transf ormation and control of the actin cytoskeleton by Ras. Cell 89: 457-67. 72. Rowell, C. A., J. J. Kowalczyk, M. D. Lewis, and A. M. Garcia. 1997. Direct demonstration of geranylgeranylation and farnesylation of Ki-Ras in vivo. J Biol Chem 272: 14093-7. 73. Sahai, E., and C. J. Marshall. 2002. RHO-GTPases and cancer. Nat Rev Cancer 2: 133-42. 74. Schafer, W. R., C. E. Trueblood, C. C. Yang, M. P. Mayer, S. Rosenberg, C. D. Poulter, S. H. Kim, and J. Rine. 1990. Enzymatic coupling of cholesterol intermediates to a mating pheromone precu rsor and to the ras protein. Science 249: 1133-9. 75. Seabra, M. C., Y. Reiss, P. J. Casey, M. S. Brown, and J. L. Goldstein. 1991. Protein farnesyltransferase and geranylgeranyltransferase share a common alpha subunit. Cell 65: 429-34.

PAGE 50

76. Sebti, S. M., and C. J. Der. 2003. Opinion: Searching fo r the elusive targets of farnesyltransferase inhibitors. Nat Rev Cancer 3: 945-51. 77. Sheng, H., J. Shao, and R. N. DuBois. 2001. Akt/PKB activity is required for Ha-Ras-mediated transformation of intest inal epithelial ce lls. J Biol Chem 276: 14498-504. 78. Shields, J. M., K. Pruitt, A. McFall, A. Shaub, and C. J. Der. 2000. Understanding Ras: 'it ain't over 'til it's over'. Trends Cell Biol 10: 147-54. 79. Shimizu, K., M. Goldfarb, Y. Suard, M. Perucho, Y. Li, T. Kamata, J. Feramisco, E. Stavnezer, J. Fogh, and M. H. Wigler. 1983. Three human transforming genes are related to the vira l ras oncogenes. Proc Natl Acad Sci U S A 80: 2112-6. 80. Sidhu, R. S., S. M. Elsaraj, O. Grujic, and R. P. Bhullar. 2005. Calmodulin binding to the small GTPase Ral requir es isoprenylated Ral. Biochem Biophys Res Commun 336: 105-9. 81. Sjolander, A., K. Yamamoto, B. E. Huber, and E. G. Lapetina. 1991. Association of p21ras with phosphatidylinos itol 3-kinase. Proc Natl Acad Sci U S A 88: 7908-12. 82. Sun, J., M. A. Blaskovich, D. Knowles, Y. Qian, J. Ohkanda, R. D. Bailey, A. D. Hamilton, and S. M. Sebti. 1999. Antitumor efficacy of a novel class of nonthiol-containing peptidomimetic inhi bitors of farnesyltransferase and geranylgeranyltransferase I: combinati on therapy with the cytotoxic agents cisplatin, Taxol, and gemcitabine. Cancer Res 59: 4919-26.

PAGE 51

83. Sun, J., J. Ohkanda, D. Coppola, H. Yin, M. Kothare, B. Busciglio, A. D. Hamilton, and S. M. Sebti. 2003. Geranylgeranyltransf erase I inhibitor GGTI2154 induces breast carcinoma apoptosis and tumor regression in H-Ras transgenic mice. Cancer Res 63: 8922-9. 84. Sun, J., Y. Qian, Z. Chen, J. Marfurt, A. D. Hamilton, and S. M. Sebti. 1999. The geranylgeranyltransferase I inhi bitor GGTI-298 induces hypophosphorylation of retinoblastoma and partner switching of cyclin-depende nt kinase inhibitors. A potential mechanism for GGTI-298 antitumor activity. J Biol Chem 274: 6930-4. 85. Sun, J., Y. Qian, A. D. Hamilton, and S. M. Sebti. 1998. Both farnesyltransferase and geranylgeranyltran sferase I inhibitors are required for inhibition of oncogenic K-Ras prenylation but each alone is sufficient to suppress human tumor growth in nude mouse xenografts. Oncogene 16: 1467-73. 86. Tchevkina, E., L. Agapova, N. Dyakova, A. Martinjuk, A. Komelkov, and A. Tatosyan. 2005. The small G-protein RalA stimul ates metastasis of transformed cells. Oncogene 24: 329-35. 87. Tsuchida, N., E. Ohtsubo, and T. Ryder. 1982. Nucleotide sequence of the oncogene encoding the p21 transforming protein of Kirsten murine sarcoma virus. Science 217: 937-9. 88. Ulku, A. S., R. Schafer, and C. J. Der. 2003. Essential role of Raf in Ras transformation and deregulation of matrix metalloproteinase expression in ovarian epithelial cells. Mol Cancer Res 1: 1077-88. 89. Varras, M. N., G. Sourvinos, E. Diakomanolis, E. Koumantakis, G. A. Flouris, J. Lekka-Katsouli, S. Michalas, and D. A. Spandidos. 1999. Detection

PAGE 52

and clinical correlations of ras gene mutations in human ovarian tumors. Oncology 56: 89-96. 90. Vogt, A., J. Sun, Y. Qian, A. D. Hamilton, and S. M. Sebti. 1997. The geranylgeranyltransferase -I inhibitor GGTI-298 arrests human tumor cells in G0/G1 and induces p21(WAF1/CIP1/SDI1) in a p53-independent manner. J Biol Chem 272: 27224-9. 91. Ward, Y., W. Wang, E. Woodhouse, I. Linnoila, L. Liotta, and K. Kelly. 2001. Signal pathways which promote inva sion and metastasis : critical and distinct contributions of extracellular signal-regulated kinase and Ral-specific guanine exchange factor pathways. Mol Cell Biol 21: 5958-69. 92. Webb, C. P., L. Van Aelst, M. H. Wigler, and G. F. Woude. 1998. Signaling pathways in Ras-mediated tumorigenicity and metastasis. Proc Natl Acad Sci U S A 95: 8773-8. 93. Wennerberg, K., K. L. Rossman, and C. J. Der. 2005. The Ras superfamily at a glance. J Cell Sci 118: 843-6. 94. Whyte, D. B., P. Kirschmeier, T. N. Hockenberry, I. Nunez-Oliva, L. James, J. J. Catino, W. R. Bishop, and J. K. Pai. 1997. Kand N-Ras are geranylgeranylated in cells treated with fa rnesyl protein transferase inhibitors. J Biol Chem 272: 14459-64. 95. Willumsen, B. M., R. W. Ellis, E. M. Scolnick, and D. R. Lowy. 1984. Further genetic localization of th e transforming sequences of the p21 v-ras gene of Harvey murine sarcoma virus. J Virol 49: 601-3.

PAGE 53

96. Wittinghofer, F. 1992. Three-dimensional stru cture of p21H-ras and its implications. Semin Cancer Biol 3: 189-98. 97. Yamazaki, Y., Y. Kaziro, and H. Koide. 2001. Ral promotes anchorageindependent growth of a human fi brosarcoma, HT1080. Biochem Biophys Res Commun 280: 868-73. 98. Yang, G., J. A. Thompson, B. Fang, and J. Liu. 2003. Silencing of H-ras gene expression by retrovirus-mediated siRNA decreases transformation efficiency and tumorgrowth in a model of hu man ovarian cancer. Oncogene 22: 5694-701. 99. Yin, J., C. Pollock, K. Tracy, M. Chock, P. Martin, M. Oberst, and K. Kelly. 2007. Activation of the RalGEF/Ral pathway promotes prostate cancer metastasis to bone. Mol Cell Biol 27: 7538-50. 100. Yoshida, Y., M. Kawata, M. Katayama, H. Horiuchi, Y. Kita, and Y. Takai. 1991. A geranylgeranyltransferase fo r rhoA p21 distinct from the farnesyltransferase for ras p21S Biochem Biophys Res Commun 175: 720-8. 101. Zhang, F. L., and P. J. Casey. 1996. Protein prenylation: molecular mechanisms and functional consequences. Annu Rev Biochem 65: 241-69. 102. Zhang, K., J. E. DeClue, W. C. Vass, A. G. Papageorge, F. McCormick, and D. R. Lowy. 1990. Suppression of c-ras transf ormation by GTPase-activating protein. Nature 346: 754-6. 103. Zhao, W., R. A. Ahokas, K. T. Weber, and Y. Sun. 2006. ANG II-induced cardiac molecular and cellular events: role of aldosterone. Am J Physiol Heart Circ Physiol 291: H336-43.

PAGE 54

104. Zhou, D. J., N. Gonzalez-Cadavid, H. Ahuja, H. Battifora, G. E. Moore, and M. J. Cline. 1988. A unique pattern of proto-oncogene abnormalities in ovarian adenocarcinomas. Cancer 62: 1573-6. 105. Zhu, K., A. D. Hamilton, and S. M. Sebti. 2003. Farnesyltransf erase inhibitors as anticancer agents: current status. Curr Opin Investig Drugs 4: 1428-35.

PAGE 55

47 Chapter 2: Geranylgeranyltransferase I Inhibitors Targ et RalB to Inhibit Anchorage-dependent Growth and Induce Apoptosis, and RalA to Inhi bit Anchorage-independent Growth By Samuel C. Falsetti 1,2, De-an Wang1,2, Hairuo Peng 4, Dora Carrico4, Adrienne D. Cox3, Channing J. Der 3, Andrew D. Hamilton4, Sad M. Sebti1,2,* All the work in this chapter was performed by Samuel C. Falsetti except for the cloning of the various constructs that was done in collaboration with De-an Wang and the chemical synthesis that was performe d by Hairuo Peng and Dora Carrico 1Drug Discovery Program, The H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL 2Departments of Interdisciplinary Oncology a nd Molecular Medicine, The University of South Florida, Tampa, FL 3Lineberger Comprehensive Can cer Center, University of North Carolina at Chapel Hill, NC 4Yale University, Department of Chemistry, New Haven, CT

PAGE 56

Abstract Geranylgeranyltransferase I inhibitors (GGTIs) are pres ently undergoing advanced preclinical studies and have been shown to disrupt oncogenic and tumor survival pathways, to inhibit anchorage-dependent and independent growth and to induce apoptosis. However, the geranylgeranylated proteins that are targeted by GGTIs to induce these effects are not known. Here we provide evidence that the Ras-like small GTPases RalA and RalB are exclusively gera nylgeranylated and that inhibition of their geranylgeranylation mediates, at least in part, the effects of GGTIs on anchoragedependent and independent growth and tumor apoptosis. To this end, we have created the corresponding carboxyl terminal mutants th at are exclusively farnesylated, verified that they retain the subcellular localizati on and signaling activities of the wild type geranylgeranylated proteins and that Ral GT Pases do not undergo alternative prenylation in response to GGTI treatment. By expre ssing farnesylated, GGTI-resistant RalA and RalB in Cos7 cells and the human pancreatic MiaPaCa2 cancer cells followed by GGTI2417 treatment we demonstrated that farnesylated RalB, but not RalA, confers resistance to the pro-apoptotic and anti-anchoragedependent growth effects of GGTI-2417. Conversely, farnesylated RalA but not RalB expression renders MiaPaCa2 cells less sensitive to inhibition of anchorage-independent growth. Furthermore, farnesylated RalB, but not RalA inhibits the ability of GGT I-2417 to suppress survivin and induce p27Kip1 protein levels. We conclude that RalA a nd RalB are important, functionally distinct targets for GGTI-mediated tumor apoptosis and growth inhibition.

PAGE 57

Introduction Members of the Ras and Rho branches of the Ras superfamily of small GTPases are critically involved in the regulation of many biologica l events critical to the regulation of cellular homeostasis such as cell cycle control, cell survival, death, differentiation, development and growth (9, 53). The aberrant activation or inactivation of Ras family proteins is believed to be im portant in the induction of oncogenesis. In addition to the three Ras proteins (H-, Nand K-Ras), other Ras family proteins with validated roles in oncogenesis include R-Ras, Ral, Rheb, Di-Ras and Noey2/ARHI small GTPases. Rho family GTPases (e.g., RhoB, RhoC, Rac1b, DBC2) are also implicated in oncogenesis (41). The oncogenic functions of the Ras and Rho proteins require posttranslational processing by prenyltransferase enzymes ( 20, 23, 56). The two enzymes responsible for prenylation of Ras family proteins are farnesyltransferase (FTase) and geranylgeranyltransferase I ( GGTase I) (3-5, 29, 42, 43, 55, 56), which covalently attach the 15-carbon farnesyl and 20-carbon geranylgeranyl lipids, respectively, to the cysteine of proteins with the carboxy terminal tetr apeptide consensus seque nce CAAX, (C is cysteine, A is any aliphatic amino acid a nd X is any carboxyl-terminal amino acid). In general, FTase farnesylates proteins in whic h X is methionine or serine (37), whereas GGTase I geranylgeranylates proteins in whic h X is leucine or isoleucine (12). Ras proteins that are mutationally rendered unpre nylatable lose their oncogenic activity and fail to properly locali ze within the cell (20, 23). Simila rly, prenylation of other proteins in the Ras and Rho families is essent ial to their activities (1, 20, 25)

PAGE 58

The fact that prenylati on is required for the oncogen ic activity of small GTPases prompted us and others to design FTase a nd GGTase I inhibitors (FTIs and GGTIs) as potential anticancer drugs (14, 26, 35, 57). While numerous studies have shown that FTIs suppress oncogenic and tumor survival pathways, the actual mechanism by which FTIs inhibit tumor growth is not known (34, 44). Thus, while designed originally as antiRas inhibitors, the Ras isof orms most commonly mutated in human cancers (Nand KRas) escape FTI inhibition by undergoing alternative pre nylation by GGTase I (40, 51, 54). Therefore, the critical fa rnesylated proteins that FTIs target to induce these effects are not known (34, 44). Similarly, the GGTase I substrates important for the anti-tumor activity of GGTIs are not clea rly understood yet. While we have implicated Rac1 and Rac3 Rho family proteins as candidate targ ets for GGTIs (22), other important targets remain to be identified. Clues to what these targets may be are suggested by our previous studies demonstrating that GGTIs inhibit the activation of the Akt serine/threonine kinase and e xpression of survivin (10, 49). Furthermore, GGTIs also induce p21waf, inhibit CDK activity, phosphorylat ion of Rb and lead to G0/G1 cell cycle accumulation (30, 35, 49-52). In animal models, GGTIs both inhibit tumor growth in nude mouse xenografts and induce tumor regr ession in transgenic mice (48, 49). The GGTase I substrates targeted by GGTIs to i nduce their anti-neoplastic effects are not known. Logical candidates include other Ras and Rho family proteins with roles in oncogenesis. Recently, the Ra sl ike RalA and RalB small GTPases have been shown to play critical roles in Ras-mediated growth tr ansformation of human cells (16, 27). Ral GTPases are activated by Ral guanine nucleotide exchange factors (RalGEFs; e.g.,

PAGE 59

RalGDS), and RalGEFs function as key downs tream effectors of activated Ras (38). Critical evidence for the important role of the RalGEF-Ral effector pathway in Rasmediated oncogenesis is provide d by studies in cell culture and mouse model systems. We showed previously that this effector pathway, and not the Raf-MEK-ERK mitogenactivated protein kinase eff ector pathway, is sufficient and necessary to promote Rasmediated tumorigenic growth transformation (16). Similar observations were made by Weinberg and colleagues (36), who found cell type differences in the importance of RalGEF-Ral signaling in Ras transformation. Marshall and colleagues found that mice deficient in one RalGEF (RalGDS) are deve lopmentally normal, but showed impaired skin tumor growth caused by carcinogen-induc ed Ras activation (15). An unexpected outcome of the study of the role of Ral GT Pases in oncogenesis has been the distinct functions of the highly relate d RalA and RalB isoforms. Although RalA and RalB share strong sequence (85% identity) similarities, White and colleagues found that RalA is important for tumor cell anchorage-independent proliferation, whereas RalB promotes tumor cell survival (8). We recently determin ed that RalA, but not RalB is critical for anchorage-independent and tumorigenic growth of pancreatic car cinoma cells, whereas RalB and to a lesser degree RalA is critic al for pancreatic carcinoma invasion and metastasis (28). The distinct functions of RalA and RalB may be due, in part, to distinct downstream effector utilization (7, 45). The increasing evidence for Ral GTPases in oncogenesis prompted our interest in evaluating Ral GTPases as important targets for GGTI anti-tumor activity. Both RalA and RalB C-termini contain a CAAX sequence that predicts prenylation, and RalA has been shown to be geranylgeranylated (24). Furthermore, RalA and RalB are involved in

PAGE 60

many oncogenic steps that are inhibited by GGT Is. Therefore, in this manuscript we investigated whether some of the anti-ne oplastic effects of GGTIs are mediated by inhibition of the geranylgeranylation and function of RalA and/or RalB. To this end, we have demonstrated that RalA and RalB ar e exclusively geranylge ranylated, generated farnesylated variants of Ral and verified th at they are GGTI-insensitive, and used these variants to rescue cancer cells from the an ti-neoplastic effects of GGTIs. Our data suggest that inhibition of RalB mediates the effects of GGTIs on survivin, p27Kip1, apoptosis and anchorage-depende nt growth, whereas inhibition of RalA mediates, at least in part, the effects of GGTIs on anchorage-independent growth.

PAGE 61

Materials and Methods Synthesis of CAAX Peptidomimetic The GGTase I-specific peptidomimetics GGTI-2417 and GGTI-2418 were synthesized as describe d previously (33). The FTase-specific peptidomimetics FTI-2148 and FTI-2153 were s ynthesized as described previously (48). Cloning of Ral A and Ral B mutantsWe used our previously described RalA and RalB pBABE expression constructs (16) as templa te DNA for site directed mutagenesis PCR driven by Platinum Taq polymerase (Invitrogen, Carlsbad, CA) according to manufacturers instructions with the follo wing primers (Qiagen, Valencia, CA): FLAGRalA-CCIL was generated using (F-5GCCGGATCCATG GATTACAAGGATGACGACGATAAGATCGTCGACTACCTAGCAAATAAGCCC3, R-5GCCGGATCCTTATAAAATGCAG CATCTTTCTCTGATTC-3), FLAG-RalACCIS was generated using (F -5GCCGGATCCATGGATTACAA GGATGACGACGATAAGATCGTCGACTACCTAGCAAATAAGCCC-3, R-5G CCGGATCCTTATGAAATGCAGCATCTTTCTCTGATTC-3), FLAG-RalA-SCIL was generated using (F-5GCCG GATCCATGGATTACAAGGATGACGACGA TAAGATCGTCGACTACCTAG CAAATAAG CCC-3, R-5GCCGGATCCTTATA AAATGCAGGATCTTTCTCTGATTC-3); FLAG-Ra lB-CCLL was generated using (F5GCCGGATCCATGGATTACAAGGATGACGACGATAAGATCGCT GCCAACAAGAGTAAGGGCCAG-3, R5GCCGGATCCTCATAGTAAGCAA CATCTTTCTTTAAAACT-3), FLAG-RalB -CCLS was generated using (F5GCCGGATCCATGGATTACAAGGAT GACGACGATAAGATCGCTGCCAA CAAGAGTAAGGGCCAG-3, R-5GC CGGATCCTCATGATAAGCAACATCT TTCTTTAAAACT-3), FLAG-RalB-SCLL wa s generated using (F-5GCCGGAT

PAGE 62

CCATGGATTACAAGGATGACGACGATAAGATCGCTGCCAACAAGAGTA AGGGCCAG-3, R-5G CCGGATCCTCATAGTAAGC AAGATCTTTCTTTAAAA CT-3). These PCR fragments were subclone d into pBABE molonymurine retroviral plasmid by single enzyme digest with BamH I (New England Biolabs, Ipswich, MA) using standard restriction enzyme conditi ons. Plasmid sequences were verified by standard Sanger sequencing reaction. In vitro transcription/translation/prenylation assay Plasmid DNA was amplified in a PCR reaction using AccuPrime Taq polymer ase (Invitrogen, Carlsbad, CA) using 1 g plasmid DNA by forward primer, T7-FLAG (RalA: 5GCCGGATCCTAATACGACTCACTATAGGGTCGACTACCTAGCAAATAAGCC C, RalB: GCCGGATCCTAATAC GACTCACTATAGGGGCTGCCAAC AAGAGTAAGGGCCAG) and gene specific reverse primers, the same primers as mentioned previously in the cloning prot ocol. Subsequent cDNA was isolated using QIAquick PCR purification column (Qiagen, Valencia, CA) and 500ng was used for T7in vitro transcription-translati on (TnT coupled rabbit reti culocyte lysate system, Promega, Madison, WI). Briefly, reaction co mponents were assembled on ice according to manufacturers protocol along with either 5 Ci [H3] farnesyl pyrophosphate, 5 Ci [H3]Geranylgeranyl pyrophosphate or 10 Ci [S35] methionine in the presence or absence of GGTI-2418 or FTI-2148. Th e reaction was incubated at 30C for 120 min and the reaction was stopped by additi on of an equal volume of 2x SDS-PAGE sample buffer. The samples were then loaded onto 12 % acr ylamide SDS-PAGE gel and separated at 50 V. Gels were then fixed in methanol/acetic acid (50% methanol, 10% glacial acetic acid, 40 % ddH2O) for 30 min while gently shaking. Gels were then rinsed and incubated for

PAGE 63

30 min with Amplify reagent (Amersham Bios ciences, Piscataway, NJ) for 30 min while gently shaking. Gels were then transferred to Whatman paper at 80C for 2 h on a BioRad Model 583 gel drier using a BioR ad HydroTech Vacuum Pump (BioRad, Hercules, CA). Gels were visualized by au toradiography using Kodak BioMax FX film (Kodak, Rochester, NY) at C. Cells and Culture Human tumor cell line MiaPaCa2 (pancreatic carcinoma), monkey embryonic kidney cell Cos7 and murine NIH3T3 fibroblasts were purchased from ATCC (Manassas, VA) and grown in Dulbecco s modified minimal essential media (Invitrogen, Carlsbad, CA) at 37C in a humidified incubator at 5% CO2. Transfection procedure Cells were grown to 50-70% c onfluence and transfected with Transit-LT1 (Mirus, Madison, WI) according to the manufacturers instructions. Briefly, 3 l of Transit LT1 reagent was suspended pe r 1 ml of OPTI-MEM media (Invitrogen, Carlsbad, CA) and allowed to equilibrate at 24-27C for 15 min. Approximately 1 g of plasmid per ml of media was suspended and allowed to complex with the liposomes for 15 min at 24-27C. Cells were briefly washed with OPTI-MEM and 2 ml containing 2 g media and 6 l Transit LT1 was plated on top of the cells and incubated at 37C for 6 h. Two ml of DMEM containing 10% FBS, without penicillin -streptomycin, was added and the cells were further inc ubated at 37C overnight. Immunofluorescence Cos7 cells were seeded at a 50% cell density in 6-well plates containing sterilized glass coverslips and a llowed to attach overnight. Cells were transfected overnight using TransitLT1 in Opti-MEM media following the manufacturers instructions. Media was repl aced with complete growth media containing either DMSO, 25 M FTI-2153 or 25 M GGTI-2417 and incubated at 37C, 5% CO2

PAGE 64

for 48 h. Cells were then aspirated and washed twice with sterile PBS (pH 7.4) and fixed in 4% paraformaldehyde. Cells were then ri nsed twice with sterile PBS (pH 7.4) and permeabilised, on ice, with 0.1% Triton X-100. Cells were then blocked for one hour with 1% BSA, rinsed twice and incubated overnight with 1:100 anti-FLAG M2 monoclonal antibody (Sigma-Aldrich, St. Loui s, MO) or 1:100 anti-Hemaglutin (HA) monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Cells were rinsed three times in sterile PBS (pH 7.4) and th en incubated for one hour with secondary antibody, FITC-conjugated rabbit anti-mouse (Santa Cruz Biotechnology, Santa Cruz, CA). Cells were then washed twice in ster ile PBS, mounted using Vectashield mounting reagent containing DAPI for nuclear visualiz ation and analyzed using a Leica DMIL florescent microscope (Leitz, Wetzlar, Germany) at 525 nm (FITC) and 365/420 nm DAPI. Luciferase assay NIH-3T3 cells were plated (106 cells well) in 6-well plates and transfected using standard calcium phosphate-mediated transfection with 4 g of either pBABE, HA-H-Ras12V, FLAG-RalA72L-F, FLAG-RalA72L-GG, FLAG-RalA72L-S, FLAG-RalB72L-F, FLAGRalB72L-GG, or FL AG-RalB72L-S; 1 g of NF-KB-pTAL firefly luciferase plasmid (BD Biosciences, Franklin Lakes, NJ); and 0.2 g Renilla luciferase plasmid for 5 h (both from BD Biosciences, Franklin Lakes, NJ). Cells were then incubated in full growth media overnig ht followed by serum starvation for 24 h in DMEM supplemented with 0.5% fetal calf serum. Cells were then lysed, and luciferase activity was determined using ProMega dual luciferase assay system according to the manufacturers instruction.

PAGE 65

Western blotting Cells were treated with GGTI-2417 fo r 48 h, harvested, and lysed in HEPES lysis buffer as described previously (51). Proteins were then resolved by 12.5% SDS-PAGE and immunoblotted with anti bodies against unprenylated Rap1A/Krev1(121), p21WAF1 (C-190), Akt 1-2 (N19), and RhoB (119), RhoA (26C4), antiHemaglutin (HA) monoclonal antibody and survivin (FL142) (all from Santa Cruz Biotechnology, Santa Cruz, CA), p27KIP 1 (G173), (from Pharmingen, San Diego, CA), phospho-serine 473 Akt (C ell signaling, Danvers, MA) and -actin (AC15) and anti-FLAG M2 monoclonal antibody (both fr om Sigma-Aldrich, St. Louis, MO). The ECL blotting system (NEN Life Science Products, Boston, MA) was used for detection of positive antibody reactions. Trypan blue dye exclusion assay Adherent cells were harves ted using trypsinisation and pooled with suspension cells from media s upernatant by pelleting at 300g for 5 min at 4C. The cells were then aspirated and resusp ended in an appropriate volume of media by pipetting gently up and down. Two 20 l aliquo ts were removed and combined with an equal volume of trypan blue and allowed to mix for 2 min. A 10 l volume was loaded onto a hemacytometer and cells were scored as live or dead based on trypan blue dye exclusion. TUNEL analysis Cells were seeded into 60-mm-diam eter dishes and grown in DMEM supplemented with 5% FCS for 24 h and then treated with GGTI-2417 for 48 h, which we determined to be the optimal timepoint for induction of apoptosis. Apoptosis was determined by terminal deoxynucleotidyltransf erase-mediated dUTP nick end labeling (TUNEL) by using an in situ cell death detection kit (Roche Indianapolis, IN). The cells were trypsinized, and cytospin preparati ons were obtained by centrifugation at 1,500xg

PAGE 66

(using a Cytospin-3 centrifuge, Therma-Shandon) Cells were fixed with freshly prepared paraformaldehyde (4% in phosphate-buffered sali ne [PBS], pH 7.4). Slides were rinsed with PBS, incubated in permeabilization solution, and cross-reacted with TUNEL reaction mixture for 60 min at 37C in a humid ified chamber. The slides were rinsed, mounted and analyzed under a light microscope. Creation of retrovirus Retrovirus was created by transi ent transfection of HEK-293T human embryonic kidney cells with pVPACK-Ampho, pVPACK-gag-pol and pBABE retroviral plasmids according to manufacturers protocol (Invitrogen Carlsbad, CA). Briefly 293T cells were seeded at 2.5x106 cells in a 60 mm2 dish. 3 g of each plasmid was combined and brought to a volume of 225 l in Rnase/Dnase free water following which 25 l 2.5 M CaCl2 was added dropwise while gently vortexing. This solution was then added to 250 l 2x Hepes-buffered sa line dropwise and incubated at room temperature for 5 min. The total volume of 500 l was then added to the 293T cells and incubated for 8 h at 37 C The medium was then removed and cells were incubated for 48 h. The supernatant was then passed through a 0.45 M nylon low protein-binding filter (Fisherbrand, Houston, TX) to remove cells and cell debris. Retroviral titers were determined by serial dilution and infection of NIH-3T3 cells followed by selection and colony formation in puromycin. Cellular fractionation Membrane and cytosolic fracti ons were isolated using the MEMPER membrane extraction kit (Pro mega, Madison, WI). Briefly, 1x106 Cos7 cells were transfected with the appropria te plasmid and collected seve nty-two hours later. Cells were lysed and membrane and cytosolic fractions were isolated according to manufacturers protocol. Membrane and cyto solic fractions were diluted five fold and

PAGE 67

diluted further in 2x Laemmli sample buffer (Biorad, Carlsbad, CA). Proteins were separated by SDS-PAGE and visualized by Western blotting. Creation of stable cell lines MiaPaCa2 cells were seeded into 6-well plates at a 40% confluency and incubated with 8 g/ml polybr ene (Millipore/Specialty Media, Billirika, MA) and 5x105 viral particles for 8 h. Medium wa s changed to complete growth media and cells were incubated at 37C, 5% CO2 for 24 h. Media was then replaced with complete growth media containing 4 g/ml puromycin and incubated until colonies formed. All colonies were pooled and ta ken as a single polyclonal population and cultured in complete media containing 2 g /ml puromycin. MTT metabolism assay Cells were seeded in a 96-well plate at a density of 1,500 cells per well and allowed to attach overnight. Cells were then incubated for 96 h with varying concentrations of GGTI-2417 or appropriate DMSO control. Media was aspirated after 96 h, the optimal time for study of prolifera tion using this assay, and replaced with complete medium containing 1 mg/ml MTT and incubated for 3 h at 37 C in 5% CO2 humidified incubator. Medium was then aspirated and DMSO was added. Cells were incubated 5 min at room temperature while shaking following which absorbance was determined at 495 nm. Soft agar clonogenicity assay For soft agar growth assays, the cell lines were seeded at a cell density of 1,500/well in triplicate in 12-we ll culture dishes in 0.3% agar over a 0.6% bottom agar layer. Various concentrati ons of GGTI-2417 or vehicle (DMSO) were included in the 0.3% agar layer of cells. Cu ltures were fed and treated with drug or vehicle once weekly until colonies grew to a suitable size for observation (colony growth rates were 10-14 days for RalA stable Mi aPaCa2 cells and 3 weeks for RalB stable

PAGE 68

MiaPaCa2 cells). Colonies were photographed after overnight incubation with 1 mg/ml MTT in the respective cell growth media. The growth of colonies in the presence of inhibitor was compared with the cont rol colonies treate d with vehicle. Statistical analysis Statistical analysis was performed using standard students T-Test via either Microsoft Excel (Microsoft, Re dmond, WA) or GraphPad Software (San Diego, CA) statistical analysis tools.

PAGE 69

Results Generation of CAAX box mutants that are fa rnesylated or unprenylated versions of both RalA and RalB It has been demonstrated previously that RalA is geranylgeranyl ated and that both RalA and RalB require the prenyl-accepting cysteine for proper localization (18, 24). However, direct evidence that RalA and RalB are exclusively geranylgeranylated, and thus validated targets of GGTIs, is lacking. Furthermore, whether the nature of the prenyl group influences the subcellular lo cation and function of Ral proteins is not known. To this end, we used site-direct ed mutagenesis to generate two CAAX box mutant forms of wild type or GTPase-deficient /constitutively activated (Q72L) RalA and RalB proteins. The first missense mutation re placed the carboxyl-terminal leucine with a serine residue to switch th e prenyltransferase specificit y from GGTase I (RalA-CCIL and RalBCCLL) to a site preferred by the re lated enzyme FTase (RalACCIS and RalB CCLS). The second mutation replaced the pren ylated cysteine with a serine residue to prevent prenylation (RalA SCIL and RalB SCLL). Both GTP-locked (RalA72L, RalB72L) and wild type RalA and RalB were used to generate the CAAX box mutants. Plasmid constructs containing these Ra l cDNA sequences were then used in in vitro transcription-translation-prenylation assays in rabbit reticulocyte lysates with radiolabeled [35S]-labeled-methionine, [3H]-labeled farnesyl pyrophosphate (FPP) and [3H]-labeled geranylgeranyl pyrophosphate (GG PP) to determine the relative strength of translation, farnesylation and geranylger anylation, respectively. Figure 5A (see following pages) shows that RalACCIL and RalB-CCLL were geranylgeranylated while replacement of the carboxyl-terminal amino acid with a serine resulted in incorporation

PAGE 70

of a farnesyl instead of a geranylgeranyl moiety. Replacement of the prenylation site cysteine with a serine residue (CCIL and CCLL to SCIL and SCLL) rendered both RalA and RalB unprenylated (Figure 5A, following pages).

PAGE 71

Geranylgeranylated and farnesylated RalA and RalB are not alternatively prenylated in the presence of GGTIs and FTIs, respectively. It has been well documente d that certain Ras family members are capable of being alternatively prenylated when FTase is inhibited (40, 51, 54). We reasoned that it is possible that RalA and/or RalB (geranylgera nylated or farnesylated forms) could be alternatively prenylated when GGTase I or FTase is inhibited. In order to determine whether the wild type RalACCIL and RalB-CCLL were indeed targets only of GGTase I and could not be prenylated by FTase upon GGTase I inhibition, we added GGTI-2418, a potent and selective competitive inhibitor of GGTase I (33), to the reticulocyte lysate transcription-translation-prenylation mixtur e (Figure 5B, following pages). Wild type RalA-CCIL and RalB-CCLL were geranylge ranylated and were not alternatively farnesylated when GGTase I was inhibite d by GGTI-2418, and their geranylgeranylation was little affected by inhibition of FTase by our previously characterized FTase-specific inhibitor, FTI-2148 (48). Furthermore, RalA-CCIS and RalB-CCLS remained farnesylated when GGTase I was inhibited a nd did not become geranylgeranylated when FTase I was inhibited (Figure 5B, following pa ge). Minor variations in the apparent incorporation efficiency of th e radiolabeled prenyl groups in the absence or presence of inhibitors, was seen regularly but without consistency. These results suggest that RalACCIL and RalB-CCLL are targets of GGTIs but not FTIs, and conversely, that the RalACCIS and RalB-CCLS CAAX box mutants ar e targets of FTIs, but not GGTIs. Furthermore, neither the geranylgeranylated or farnesylated Ral proteins are alternatively prenylated under the pressure of GGTI or FTI treatment, respectively. Therefore, for the rest of the manuscript we will refer to RalA-CCIL and RalB-CCLL as RalA-GG and

PAGE 72

RalB-GG, respectively. Similarly, we will re fer to RalA-CCIS and RalB-CCLS as RalAF and RalB-F, respectively.

PAGE 73

B RalA-CCIS RalA-CCILRalB-CCLS RalB-CCLL S35-Met H3-F H3-GG FTI-2148 GGTI-2418 -+ --+ --+ --+ --+--+--+--+ RalA72L S35-Met H3-F H3-GG CCIS CCIL SCIL RalA RalB72LRalB CCIS CCIL SCILCCLS CCLL SCLL CCLS CCLL SCLLA Figure 5. RalA-CCIL and RalB-CCLL are geranylgeranylated whereas the mutants RalA-CCIS and RalB-CCLS are farnesylated. A) Ral DNAswere used in transcription-translation-prenylation assays using either radiolabeled [35S]-methionine, [3H]-FPP or [3H]-GGPP then run on SDS-PAGE and visualized by autoradiography as described in Materials and Methods. Results are representative of two independent e xperiments. B) RalA-CCIS, RalA-CCIL, RalB-CCLS or RalB-CCLL DNAswere transcribed, translated and prenylated using either [35S]-Methionine, [3H]-FPP or [3H]-GGPP in the presence of either vehicle (DMSO), 250 nMFTI-2148 or 250 nMGGTI-2418 then run on SDSPAGE and visualized by autoradiography. Results are representative of two independent experiments. Figure 5. RalA-CCIL and RalB-CCLL are geranylgeranylated whereas the mutants RalA-CCIS and RalB-CCLS are farnesylated

PAGE 74

Prenylation is required for proper subcellu lar localization of RalA and RalB and both RalA and RalB are mislocalized in response to prenyltransferase inhibition. The results of Figure 1 demonstrated, in a reconstituted cell reticulocyte system, that the wild type RalA-GG, RalB-GG, and CAAX box mutant RalA-F and RalB-F variants are not alternatively prenylated in vitro. We next determined whether prenylation is required for proper localiza tion in intact cells and whether GGTI-2417 treatment is sufficient to disrupt localizati on of wild type RalA-GG and RalB-GG. We also wanted to determine if the RalA-F and RalB-F mutants display a localization similar to RalA-GG and RalB-GG. To this end, we ectopically expresse d FLAG epitope-tagged RalA-GG, RalA-F, RalA-S, RalB-GG, RalB-F or RalB-S in Cos-7 cells followed by treatment with the indicated prenyltransferase inhibitors. We found that both the geranylgeranylated and farnesyl ated forms of RalA and RalB demonstrated a similar localization to the plasma membrane (Figur e 6, following pages). This was confirmed by membrane/cytosol cellular fractionations wh ere RalA-GG and RalB-GG as well as RalAF and RalB-F localized to the membrane fract ions (Figure 7A, following pages). The fact that RalA and RalB localize similarly is interesting but different than results by Feig et al (45). The reason for this difference at pres ent is not known. The unprenylated versions of RalA (RalA-S) and RalB (RalB-S) we re diffused in the cytoplasmic and the perinuclear regions indicating that prenylation is required for proper plasma membrane localization of RalA and RalB (Figure 6, following pages). Furthermore, in response to GGTase I inhibition RalA-GG and RalB-GG we re unable to localize to the plasma membrane while no change in membrane localization was observed when FTase activity was inhibited. As expected, H-Ras was predom inantly localized to the plasma membrane,

PAGE 75

and this was only affected by FTI but not GGTI treatment. Similarly, in response to FTase inhibition, RalA-F and RalB-F were una ble to localize to the plasma membrane while no change in membrane association was observed when treated with GGTase I inhibition (Figure 6, following pages). These data demonstrate that RalA-GG and RalBGG are targets of GGTase I inhibitors and that the farnesylated Ral variants will be useful reagents to determine if inhibition of Ral GT Pases contribute to the biological effects of GGTI-2417 treatment.

PAGE 76

H-Ras RalAGG RalA-F RalA-S DMSO FTI-2153 GGTI-2417 RalBGG RalB-F RalB-S Figure 6. Geranylgeranylated (GG) and farnesylated (F) RalA and RalB localize similarly and require p renylation for correct localization

PAGE 77

Figure 6. Geranylgeranylated (GG) and farnesylated (F) RalA and RalB localize similarly and require prenylation for correct localization. Cos7 cells grown on coverslipswere transiently-transfected with plasmids expressing the indicated proteins and treated with the indicatedinhibitors (25 ), and were then analyzed for FLAG-Ral or HA-H-Ras distribution using immunoflorescenceas described under Materials and Methods.Nuclei were visualized using DAPI stain. Re sults are representative of three independent experiments.

PAGE 78

Farnesylated and geranylgeranylated RalA and RalB are equivalent in ability to activate NF-B responsive promoter elements We next wanted to verify that Ral function was retained when modified by farnesylation. It has been reported previous ly that the constitutively activated RalB72L mutant protein is capable of activating NFB-dependent transcription (17). We therefore determined if farnesylation and ge ranylgeranylation of Ral are equivalent in their ability to mediate Ral activation of an NFB -dependent promoter. In order to assess the contribution of th e prenyl moiety to NFB-driven transcription we transiently transfected NIH-3T3 cells with either the pBabe-puro empty vector, or encoding constitutively activated H-Ras (H-Ras 12V), RalB72L-GG, RalB72L-F, RalB72L-S, RalA72L-GG, RalA72L-F or RalA72L-S. As expected, H-Ras12V stimulated an approximately four-fold increase in NFB luciferase activity (Figure 7, following page). As previously reported, RalB72L-GG activated NFB-dependent promoter activity. Importantly, RalB72L-F activated NFB luciferase activity to a similar extent as the geranylgeranylated version (approximately twofold). Similar to the results with RalB, ectopic expression of RalA72L-GG and RalA72L -F both stimulated four-fold activation of NFB. Importantly, neither nonprenylated RalB -S nor RalA-S stimulated activation of NFB (Figure 7, following page). These data demonstrate that prenylation is required for RalA and RalB activation of NFB and that farnesyl and geranylgeranyl moieties are equivalent in supporting Ral activation of NFB.

PAGE 79

pBABE H-Ras12V RalB72L-F RalB72L-GG RalB72L-S RalA72L-F RalA72L-GG RalA72L-SFLAG HA -Actin 0 0 0 0 0 0 01 2 3 4 5 Fold Change NFB B FLAG-Ral HSP90 pBABE RalA-GG RalA-F RalB-GG RalB-F M CM C M CM CM C A Figure 7. Farnesylated and geranylgeranylated RalA and RalB are equivalent in mediating activation of NFB promoter activity. A) Cos7 cells were transiently transfected with the indicated plasmids and membrane (M) and cytosolic(C) fractions were isolated and probed by Western blotting; HSP90 was used as a cytosolicfraction marker. Results are representative of 2 independent e xperiments. B) NIH-3T3 cells were serum-starved and transiently co-trans fected with plasmid expressing the indicated proteins and a NFB reporter plasmid. NFB promoter activity was detected following transfection as described under Materials and Methods. Expression was analyzed by western blot analysis. Results are representative of three independent experiments. Figure 7. Farnesylated and geranylgeranylated RalA and RalB are equivalent in mediating activation of NFB promoter activity.

PAGE 80

Ectopic expression of RalB-F renders Co s7 cells resistant to inhibition of proliferation, induction of cell death and apoptosis by GGTI-2417 Since the farnesylated and geranylgeranyl ated forms of Ral proteins displayed similar subcellular localization and signaling pr operties, we reasoned that, if Ral GTPases are important functional targets of GGTIs, ect opic expression of farnesylated, but not geranylgeranylated, RalA and/or RalB would rescue cells from the effects of GGTase I inhibition. Figure 8A (see following pages) show s that ectopic expressi on of farnesylated RalB, but not RalA, rendered cells less sensitive to the induction of cell death by GGTI2417. There were no statistically significant di fferences between a ll transfected groups (pBabe empty vector control, RalA-F, RalA -GG, RalB-GG) in cell death except in the RalB-F expressing Cos7 cel ls (P<0.05 at 50, 100 and 150 M). For example, GGTI-2417 induced 39% cell death at 50 M in RalB-GG transfected Cos7 cells but when RalB-F was expressed GGTI-2417 induced only 7 % ce ll death (Figure 8A, following pages). Similarly, RalB-F expressing Cos7 cells were 4.51-fold more resist ant to inhibition of proliferation by GGTI-2417 than RalB -GG expressing cells (IC50 of 131 M and 29 M, respectively) (Figure 8B, following pages). We next determined whether RalA-F and/or RalB-F could rescue cells from GGTI-2417 treatment-induced programmed cell death (apoptosis) by TUNEL assay. Figure 8D (see following pages) shows that only ectopic expression of farnesylated RalB, but not farnesylated RalA, demonstrated a significantly protective effect from GGTI2417-induced apoptosis. GGTI-2417 increased th e fraction of apoptotic cells from about 18 to 32% in RalB-GG but only from 17 to 20% in RalB-F (Figure 8D, following pages). In contrast, GGTI-2417 potently induced apoptosis in Cos7 cells expressing either RalA-

PAGE 81

GG or RalA-F, with GGTI-2417 treatment increa sing the fraction of apoptotic cells from approximately 15% to 30% in both cell popula tions (Figure 8D, following pages). These results suggest that inhibition of RalB and not RalA geranylgeranylation is an important target for GGTI-mediated apoptosis. All cel l lines displayed similar accumulation of non-geranylgeranylated Rap1A following GGTI-2417 treatment (Figure 8C, following pages). No changes were observed in gel mobility of the human homologue of DNAJ-2 (HDJ2), an exclusively farnesylated prot ein, indicating that FTase activity was not blocked by this treatment (Figure 8C, following pages).

PAGE 82

DpBABERalA-GGRalA-F 0.00 10.00 20.00 30.00 40.00 %apoptosis -+-+-+ 0.00 10.00 20.00 30.00 40.00 %apoptosis -+ -+ -+ pBABERalB-GGRalB-F GGTI 0.00 25.00 50.00 75.00 100.00 050100150% dead * 0.0 25.0 50.0 75.0 100.0 125.0 0 50 100 15 0 GGT217M] *% proliferation pBABE RalA-GG RalA-F RalB-F RalB-GG A BGGTI 2417 [M] GGTI 2417 [M] 0 5 50 100 0 5 50 100 0 5 50 100 0 5 50 100 0 5 50 100 pBABE RalA-F RalA-GG RalB-F RalB-GG FLAG U-Rap1A HDJ2 CGGTI [M] FLAG U-Rap1A HDJ2 GGTI [M]Figure 8. Ectopic expression of farn esylated RalB, but not RalA,renders cells less sensitive to GGTI-2417 inhibition of survival and proliferation and induction of apoptosis in Cos7 cells

PAGE 83

Figure 8. Ectopic expression of farn esylated RalB, but not RalA,renders cells less sensitive to GGTI-2417 inhibition of survival and proliferation and induction of apoptosis in Cos7 cells. A-B) Inhibition of RalB prenylation is required for GGTI induction of cell death and inhibition of proliferation. Cos7 cells were transfec ted with the indicated plasmids then treated with GGTI-2417 or vehicle (DMSO) control. Cell viability was determined by the trypanblue dye exclusion assay. Data shown are the average of three independent experiments. C) Cell lysatesfrom A-B were analyzed for expression by western blot analysis as described inMaterials and Methods. D) Inhibition of RalB prenylation is required for GGTI induction of apoptosis. Cos7 cells were transiently-transfected with plasmids expressing the indicated proteins and then treated for 72 h with50 M GGTI2417 or DMSO control, collected and analyzed for apoptosis via TUNEL assay as described in Materials and Methods. Data shown are the average of three independent experiments. (*p<0.05)

PAGE 84

MiaPaCa2 human pancreatic cancer cells stab ly expressing farnesylated RalB are resistant to the anti-prolif erative and pro-apoptotic ef fects of GGTI-2417 compared to geranylgeranylated RalB-expressing cells Since Cos7 are immortalized non-transfo rmed cells of non-human primate origin we endeavored to create a more relevant cell system to characterize the growth inhibitory and pro-apoptotic effects of GGTI-2417. We fi rst determined by western blotting that both RalA and RalB are efficiently targeted by GGTI-2417 in MiaPaCa2 cells. Figure 9A (see following pages) shows that GGT I-2417 concentrations as low as 1 inhibited RalA and RalB geranylgeranylation. We next established populations of MiaPaCa2 cells stably expressing RalA and RalB prenyl isoforms through retrovi ral transduction and selection (Figure 9B, following pages). We then used these cells to examine the effects of ectopic WT or farnesylated RalA a nd RalB expression on GGTI-2417-mediated inhibition of proliferation of MiaPaCa2 cells by both MTT metabolism viability (Figure 9C) and trypan blue dye excl usion assays (Figure 9D, following pages). Similar to our observations in Cos7 cells, MiaPaCa2 cells st ably expressing farnesylated as opposed to geranylgeranylated RalB were less sensitive to the anchorage-dependent anti-proliferative effects of GGTI-2417. Specifically, using the MTT assay RalB-GG expressing cells were inhibited by 50% at 1.6+/0.3 M GGTI-2417 whereas RalB-F expressing cells (IC50 of 8.3 1.4 M, p value = 0.0015) were more than 5-fold resistant. There were no statistically significant differences betw een RalB-GG and empty vector cells. Furthermore, no statistically significant differe nces in proliferation inhibition sensitivity were observed for empty vector, RalA-F or RalA-GG expressing cells (Figure 9C, following pages). Similar results were obtai ned using total cell counting by trypan blue

PAGE 85

dye exclusion assay demonstrating no diffe rence in sensitivity between RalA-F and RalA-GG expressing cells and a greater than 6-fold statistically significant difference (p<0.0001) in IC50 values between RalB-GG and RalB-F expressing cells (IC50 values of 4.8 1.6 M and greater than 30 M, respectively) (Figure 9D, following pages). There were no statistically significant differences between empty vector (5.6 1.7 ) and RalB-GG (4.8 1.6 M) cells. These data both confirmed and extended our observations in Cos7 cells and further indi cated that inhibition of RalB, not RalA, prenylation is an important target for GGTI inhibition of anchorage-dependent cell proliferation. To further examine the role of Ral prot eins in GGTI-2417 induction of apoptosis, we determined if ectopic expression of RalA -GG, RalA-F, RalB-GG or RalB-F altered MiaPaca2 sensitivity. Specifically, we obs erved that GGTI-2417 treatment induced a greater than three-fold increase in apoptos is in all cell lines except those expressing RalB-F (Figure 9F, following pages). Of primary importance, GGTI-2417 induced a 3.6fold induction of apoptosis (1.6 to 6.07%) in RalB-GG expressing cells as compared to no statistically significant induction of cell death in RalB -F expressing cells (1.42 to 1.56%). This indicates that, consistent with the Cos7 results, RalB and not RalA is a critical target for the apoptotic cell d eath caused by GGTI-2417 trea tment of MiaPaCa2 cells.

PAGE 86

FLAG B-actin pBABERalB-GG RalB-F RalA-G RalA-FB GGTI-2417 [M]% proliferation(MTT) 0 1 3 10 30 0 1 3 10 30 % proliferation (Trypan) = RalBGG = RalB-F = RalA-GG= RalA-F C D0 1 3 10 30 0 1 3 10 30 0 25 50 75 100 0 25 50 75 10 0 0 25 50 75 0 25 50 75 10 0 100 = pBABE = pBABE ** ** ** ** ** ** A U-Rap1a -Actin 0 1 3 10 30 [GGTI-2417] M RalA U P RalB U PFigure 9. Stable expression of farnes ylated RalB, but no t RalA, promotes resistance to the anti-proliferative and pro-apoptotic effects of GGTI-2417 in MiaPaCa2 cells.

PAGE 87

GGTI 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 pBABE RalA-GG RalA-F RalB-GG RalB-F -+ -+ -+ -+ -+ % apoptosis F E 0 1 3 10 30 0 1 3 10 30 RalB-FRalB-GG FLAG U-Rap1A Vector 0 30 0 1 3 10 300 1 3 10 30 RalA-FRalA-GG Vector 0 30 GGTI [ M] FLAG U-Rap1A GGTI [ M]Figure 9. Stable expression of farnesylated RalB, but not RalA, promotes resistance to the anti-proliferative and pro-apoptotic effects of GGTI-2417 in MiaPaCa2 cells.

PAGE 88

Figure 9. Stable expression of farnes ylated RalB, but no t RalA, promotes resistance to the anti-proliferative and pro-apoptotic effects of GGTI2417 in MiaPaCa2 cells. A) Dose response to determine the efficacy of GGTI-2417 in inhibiting RalA and RalB prenylation in MiaPaCa2 cells. MiaPaCa2 cells were treated with various concentrations of GGTI-2417 for 48h and processed for Western blotting as described under Materials and Methods. B) Stable MiaPaCa2 cell lines were created using retroviral infection of puromycin resistance mark er along with the indicated transgenes as described under Materials and Methods. Cells were lysed and expression was assessed by western blot analysis. C) Stably expressing MiaPaCa2 cells were treated with the indicated concentrations of GGTI-2417 for 96 h. Proliferation was measured by the MTT viability assay as described in Materials and Methods. Data shown are the average of at least three independent experiments. D) MiaPaCa2 cells stably expressing the indicated Ral proteins were treated with the in dicated concentrations of GGTI-2417 for 72 h. Proliferation was assessed by trypanblue dye exclusion assay as described in Materials and Methods. Data shown are the average of at least three independent experiments. E) MiaPaCa2 cells stably expressing the indicated Ral proteins were treated with GGTI-2417 (30), lysed and expression was assessed by western blot analysis. Data shown are representative of three independent expe riments. F) MiaPaCa2 cells stably expressing the indicated Ral proteins were treated with GGTI-2417 for 48 h and apoptosis was assesse d by TUNEL assay as described in Materials and Methods. Data shown are the average of three independent experiments. (*, p<0.05; **, p<0.01, ***,p<0.001)

PAGE 89

Stable expression of farnes ylated RalA, but not RalB, induces partial resistance to inhibition of anchorage-independent grow th by GGTI-2417 in MiaPaCa2 cells We next determined whether GGTI-2417 inhibition of growth in soft agar might be due, in part, to inhibition of RalA and/or RalB function. To this end, we used a soft agar clonogenicity assay to measure the eff ects of GGTI-2417 on MiaPaCa2 cells stably expressing geranylgeranylated or farnesylated RalA or RalB. We found that MiaPaCa2 cells stably expressing either RalB-GG or RalB-F did not differ in sensitivity to GGTI2417 treatment. However, MiaPaCa2 cells stab ly expressing RalA-F were less sensitive to inhibition of soft agar growth by GGTI2417 than those cells stably expressing ectopic RalA-GG (Figure 10, following page). For example, at 30 GGTI-2417, RalA-GG and RalA-F inhibited colony formation by 41.3 5.9 and 17.8 8.0, respectively (p<0.001). Taken as a whole these data implicate RalA, but not RalB, inhibition of prenylation as a potential mechanism for GGTI reversion of the transformed growth phenotype of MiaPaCa2 cells as measured by anchor age-independent soft agar growth.

PAGE 90

0 3 30 100 0 3 30 100 *** ***% colony formation GGTI-2417 [ M]= RalA-GG= RalA-F= RalB-GG= RalB-F 0 20 40 60 80 100 120 0 20 40 60 80 100 120 = pBABE Figure 10.Stable expression of farnesylated RalA, but not RalB, induces resistance to inhibition of anchorage independent growthby GGTI-2417 in MiaPaCa2 cells. MiaPaCa2 cells stably-expressing the indicated Ral proteins were seeded into 12-well plates in 0.3% soft agar and treated with the indicated concentrations of GGTI-2417 for 10 days as described in Material and Methods. Data shown are the average oftwo independent experiments repeated in triplicate. (*, p<0.05; ***,p<0.001 ) Figure 10. Stable expression of farn esylated RalA, but not RalB,induces resistance to inhibition of anchor age independent growth by GGTI-2417 in MiaPaCa2 cells.

PAGE 91

Stable expression of farnes ylated RalB but not RalA in MiaPaCa2 cells inhibits the ability of GGTI-2417 to increase p27Kip1 and decrease survivin protein levels. Recent data from our lab showed that GGTIs treatment affects several signal transduction pathways critical to cell cycle division and tumor cell survival (30, 35, 4952)1. In order to examine the contribution of inhibition of the ge ranylgeranylation of RalA and/or RalB to the effects of GGTI-2417 on signaling pathways, we treated MiaPaCa2 cells stably expressing RalA-GG, RalA-F, RalB-GG and RalB-F with GGTI2417. In control, empty vector-transfect ed MiaPaCa2 cells, GGTI-2417 treatment resulted in inhibition of RalA and Rap1 gerany lgeranylation, increased the protein levels of p27Kip1, RhoA and RhoB and decreased the le vels of two anti-apoptotic proteins, activated phosphorylated Akt (P-Akt) and su rvivin (Figure 11, following page). Expression of RalA-F, RalA-GG, or RalB-GG did not affect the ab ility of GGTI-2417 to induce its effects on these signaling molecules. However, expression of RalB-F inhibited the ability of GGTI-2417 to increase p27Kip1 and to decrease survivin protein levels. For example, while p27Kip1 protein levels we re increased in cells expressing RalB-GG (3.12 0.88 fold) cells expressing RalB-F showed a 50% (1.63 0.64 fold) attenuated increase of p27Kip1 protein levels (p =0.0073). Furthermore, GGTI-2417 treatment reduced survivin levels by 46 11% in empty vector a nd RalB-GG expressing cells. However, MiaPaCa2 cells expressing RalB-F we re less sensitive to depletion of survivin levels, with only a 12 3% reduction of su rvivin levels following GGT-2417 treatment (p=0.0006). By contrast expres sion of either RalA-F or RalA-GG had no effect on GGTI 1 and Kazi A, A Carie, MA Blaskovich, C Bucher, V Thai, S Moulder, H Peng, D Carrico, E Pusateri, AD Hamilton and SM Sebti, Inhibition of Protein Geranylgeranylation Requires p27Kip1 to Induce Human Breast Cancer Cell Death: Implications for breast cancer therapy, submitted for publication.

PAGE 92

reduction of survivin or induction of p27Ki p1. To determine if knockdown of RalA or RalB affects the levels of survivin or p27K ip1, we used siRNA to specifically inhibit RalA and RalB expression in parental Mi aPaCa2 cells. We found that knockdown of RalB, but not RalA, inhibited the survivin pr otein levels by 70.2%, whereas neither RalA nor RalB knockdown affected the leve ls of p27Kip1 (data not shown).

PAGE 93

p27kip1 U-Rap1A -Actin -+ pBABE -+ RalB-F -+ RalB-GG -+ RalA-F -+ RalA-GG GGTI-2417 RalA FLAG U pAkt Akt1/2 RhoB RhoA Survivin P Figure 11. RalB-F, not RalA-F, inhibit the ability of GGTI-2417 to increase p27Kip1 and decrease survivin protein levels. Stably expressing MiaPaCa2 cells were seeded into 6-well plates and treated with GGTI-2417 (30 ) for 48 h. Cells were lysed and expr ession was assessed by western blot analysis. Data shown are representativ e of three independent experiments. Figure 11. RalB-F, not RalA-F, inhibit the ability of GGTI-2417 to increase p27Kip1 and decrease survivin protein levels.

PAGE 94

Discussion Inhibition of protein geranylgeranylation strongly attenuates multiple tumorigenic pathways such as anchorage-dependent a nd independent tumor growth and protection from apoptosis (30, 35, 49-52). However, many proteins are substrates for GGTase Imediated protein geranylgeranylation, and it is not known which of these are critical targets for the anti-neoplastic effects of GGTIs (56). Since recent studies demonstrated the important roles of the RalA and RalB small GTPases in human oncogenesis (7, 8, 15, 16, 28, 36, 38, 45), we evaluated the possibility that these GGTase I substrates are important targets for GGTIs. In this study we have demonstrated that the GTPases RalA and RalB are geranylgeranylated by GGTase I, require prenylation for proper localization and are downstream targets of pharmacological inhibitors of GGTase I. We designed farnesylated, GGTI-insensitive, variants of Ral GTPases and found that farnesylated RalB, but not RalA, confers resistance to th e pro-apoptotic and anti -anchorage-dependent growth effects of GGTI-2417 on Cos7 and Mi aPaCa2 pancreatic carcinoma cells. Conversely, farnesylated RalA, but not RalB expression renders MiaPaCa2 cells less sensitive to inhibition of anchorage-indepe ndent growth. Finally, we determined that farnesylated RalB, but not RalA inhibited the ability of GGTI-2417 to suppress survivin and induce p27Kip1. We conclude that Ra lA and RalB are important, functionally distinct targets for GGTI-mediated anti-neopl astic effects. Additionally, our studies extend recent observations showing that the hi ghly related RalA and RalB proteins serve distinct functions in oncogenesis. The determination of the GGTase I substrates that are critical for the anti-tumor activity of GGTIs is complicated by the f act that over 60 CAAX-terminating proteins

PAGE 95

may be substrates for this prenyltransferase (56). We previously established and validated a substitutive chemical biology approach to define the contribution of inhibition of specific GGTase I substrates to GGTI inhibition of cell surviv al, proliferation, transformation and cell signaling (22). In this study, we applied this approach and we generated farnesylated mutants of both RalA and RalB and showed that they localized similarly to authentic geranylgeranylated wild-type versions and similarly activated NFB and were resistant to inhibition by GGTI treatment. Thus, this substitutive chemical biology approach is a valid means of unc oupling RalA and RalB from GGTase I dependency while preserving their wild t ype subcellular loca lization and signaling activity. Our results demonstrate a partial requiremen t for inhibition of RalA prenylation in the inhibition of anchorage-independen t growth by GGTI-2417. Since expression of farnesylated RalB did not attenuate the ability of GGTI-2417 to inhibit anchorageindependent growth we reason that the ability of RalA to regulate clonogenecity is a divergent function from that of the 85% identical protein RalB. The ability of farnesylated RalA but not RalB to regulate anchorage-independent growth is similar to results reported by Chien et al (15) wher e depletion of RalA but not RalB by siRNA inhibited anchorage independent growth in HeLa and SW480 cancer cell lines. However, other processes commonly associated with an chorage-independent pr oliferation such as cell migration and chemotaxis bear a require ment for RalB as opposed to RalA; such results were first reported by Oxford et al (32) where depletion of RalB but not RalA by siRNA inhibited cell migration in a panel of renal carcinoma cell lines. In particular, we

PAGE 96

found that RNAi-mediated suppression of RalA, but not RalB greatly impaired the soft agar growth of MiaPaCa2 and other pa ncreatic carcinoma cells (27, 28). Interestingly, we found further divergent functions for the Ral family of small GTPases. Specifically, inhibi tion of RalB, but not RalA, prenylation was required for GGTI-mediated inhibition of proliferation a nd induction of apoptosis. In both the Cos7 and the MiaPaCa2 cell lines expression of RalB-F but not RalA-F rendered cells less sensitive to GGTI-mediated inhibition of proliferation and inducti on of apoptosis. In MiaPaCa2 cells this rescue was concurrent to abrogation of GGTI effects on increasing p27Kip1 and decreasing survivin protein levels, but not on inhi biting Akt activation levels and inducing RhoA and RhoB levels, bot h of which we have shown previously to be induced by GGTIs (11). These results suggest that, at least in th e MiaPaCa2 pancreatic carcinoma cells, the ability of RalB-F to abrogate the GGTI anti-proliferative and proapoptotic mechanism of action is associat ed with abrogation of p27Kip1 induction and suppression of survivin levels. In further suppo rt of this we used siRNA to specifically inhibit RalA and RalB expression and f ound that knockdown of RalB, but not RalA, strongly attenuated survivin e xpression but did not affect p27K ip1 levels. These results concur with previous data th at some tumor cells have at least a partial requirement for RalB in suppressing apoptosis (8) and suggest that RalB maintenance of survivin expression could be an important mechanism by which RalB promotes tumor cell survival. These results agree with our previous finding that ectopic expression of survivin partially abrogates GGTI induction of programme d cell death (10). Taken as a whole the results of this study raise interesting ques tions about the mechan ism of GGTI-mediated anti-neoplastic activity. Both RalA and RalB are important regulators of many cellular

PAGE 97

processes that have not prev iously been implicated in the GGTI mechanism of action. One of the most critical of these in oncogenesi s is cellular trafficking both at the receptor level through endocytosis and through forma tion and delivery of the exocyst complex (2, 6, 13, 21, 47), a process reported to be require d for the transforming ability of RalA and RalB (27). While both RalA and RalB uti lize the exocyst complex RalA seems to primarily utilize the exocyst for transforma tion whereas RalB preferentially utilizes components of the exocyst for exocyst-independe nt functions in regulating cell mobility (7, 39). In fact, RalA and RalB are the only known binding partners for two competitive regulatory elements of the exocyst; Exo84 a nd Sec5. Therefore, the inhibition of RalA and RalB by GGTI-2417 represents a novel mech anism to inhibit exocyst function in transformed cells. Beyond the exocytic and endoc ytic trafficking pathways RalA and RalB may regulate important interactions with the actin cytoskeleton through th eir association with RalBP1, Rho GTPase activating protein, and filamin, an actin-binding partner (31). Indeed, while many small GTPases, such as Cdc42, Rac, and Rho, bind filamin, only RalA and RalB bind filamin in a GTP-depe ndent manner. Therefor e, the effects of GGTIs on cytoskeletal organization may be me diated by inhibition of RalA and/or RalB geranylgeranylation. In the cour se of our study we observed that RalB-F, but not RalA-F, inhibited GGTI-induced cell rounding in MiaPaC a2 cells concurrent with inhibition of actin fiber formation (data not shown). This is indicative of a disruption in the cell-cell junctions and the underlying actin driven membrane formations. This raises the provocative hypothesis that at l east part of GGTI effects could be due to inhibition of normal cell-cell patterning. This is consistent with the fact that RalA is required for

PAGE 98

basolateral sorting and delivery of E-cadhe rin (45), an important component of cell junctions and sheet patterning. Furthermore, it has been previous ly reported that cadherins mediate growth suppr ession by potently inducing p27Kip1 levels in a variety of cell lines (46). It is thus an intriguing possibility that aber rant activation of Ral proteins could mediate further aberrati ons in vesicle sorting of cadherins and contribute to suppression of p27Kip1 levels. Di sruption of Ral-mediated ve sicle sorting of cadherins by GGTIs could constitute a mechanism for the anti-neoplastic activity of GGTase I inhibition. Further, cadherin clus tering and expression levels can regulate survivin levels (19). This suggests that sor ting of basolateral and apical membrane proteins, such as cadherins, could be a Ral-dependent pathway that is required for tumor cell proliferation and survival and is sensitive to perturbation by GGTIs. The anti-neoplastic activity of GGTIs is likely to be a conseq uence of inhibiting the function of multiple, functionally dis tinct GGTase I substrates. However, our chemical biology approach has clearly de lineated non-overlapping roles for RalA and RalB in anchorage-independent and -dependent growth and demonstr ates that inhibition of RalA and RalB geranylgeranylation is an important step in the mechanism of action of GGTIs. Specifically, we have iden tified three novel pathways that are associated with the GGTIs response: first, a RalB-dependent induction of the p27Kip1 pathway; second, a RalB maintenance of the survivin pathwa y and third, a RalA anchorage-independent proliferation pathway. The first two pathways are associated with th e ability of GGTI to inhibit anchorage-dependent proliferation and survival. Further characterizing these pathways should prove to be an interesting and important contributi on to the study of Ral biology. We feel these results should prompt a thorough examination of GGTI

PAGE 99

mechanism of action with particular attenti on focused on the Ral family in both future clinical trials and in preclinical models.

PAGE 100

Acknowledgments We would like to thank the H. Lee Moffitt Cancer Center Microscopy, and Molecular Biology core facilities. This work was suppor ted in part by a national cooperative drug discovery grant (NCDDG U19 CA67771) award from the National Cancer Institute. We would also like to thank Norbert Berndt for critical review of the manuscript.

PAGE 101

References 1. Adamson, P., C. J. Marshall, A. Hall, and P. A. Tilbrook. 1992. Posttranslational modifications of p21rho proteins. J Biol Chem 267: 20033-8. 2. Brymora, A., V. A. Valova, M. R. Larsen, B. D. Roufogalis, and P. J. Robinson. 2001. The brain exocyst complex interacts with RalA in a GTPdependent manner: identification of a novel mammalian Sec3 gene and a second Sec15 gene. J Biol Chem 276: 29792-7. 3. Casey, P. J., P. A. Solski, C. J. Der, and J. E. Buss. 1989. p21ras is modified by a farnesyl isoprenoid. Proc Natl Acad Sci U S A 86: 8323-7. 4. Chen, W. J., D. A. Andres, J. L. Goldstein, and M. S. Brown. 1991. Cloning and expression of a cDNA encoding th e alpha subunit of rat p21ras protein farnesyltransferase. Proc Natl Acad Sci U S A 88: 11368-72. 5. Chen, W. J., D. A. Andres, J. L. Goldst ein, D. W. Russell, and M. S. Brown. 1991. cDNA cloning and expression of the peptide-binding be ta subunit of rat p21ras farnesyltransferase, the c ounterpart of yeast DPR1/RAM1. Cell 66: 327-34. 6. Chen, X. W., M. Inoue, S. C. Hsu, and A. R. Saltiel. 2006. RalA-exocyst dependent recycling endosome trafficki ng is required for the completion of cytokinesis. J Biol Chem. 7. Chien, Y., S. Kim, R. Bumeister, Y. M. Loo, S. W. Kwon, C. L. Johnson, M. G. Balakireva, Y. Romeo, L. Kopelovich, M. Gale, Jr., C. Yeaman, J. H. Camonis, Y. Zhao, and M. A. White. 2006. RalB GTPase-mediated activation of the IkappaB family kinase TBK1 c ouples innate immune signaling to tumor cell survival. Cell 127: 157-70.

PAGE 102

8. Chien, Y., and M. A. White. 2003. RAL GTPases are linchpin modulators of human tumour-cell proliferation and survival. EMBO Rep 4: 800-6. 9. Colicelli, J. 2004. Human RAS superfamily protei ns and related GTPases. Sci STKE 2004: RE13. 10. Dan, H. C., K. Jiang, D. Coppola, A. Hamilton, S. V. Nicosia, S. M. Sebti, and J. Q. Cheng. 2004. Phosphatidylinositol-3-OH kinase/AKT and survivin pathways as critical targets for gerany lgeranyltransferase I inhibitor-induced apoptosis. Oncogene 23: 706-15. 11. Delarue, F. L., J. Adnane, B. Joshi, M. A. Blaskovich, D. A. Wang, J. Hawker, F. Bizouarn, J. Ohkanda, K. Z hu, A. D. Hamilton, S. Chellappan, and S. M. Sebti. 2007. Farnesyltransferase and geranylgeranyltransferase I inhibitors upregulate RhoB expression by HDAC1 dissociation, HAT association and histone acetylation of th e RhoB promoter. Oncogene 26: 633-40. 12. Finegold, A. A., D. I. Johnson, C. C. Farnsworth, M. H. Gelb, S. R. Judd, J. A. Glomset, and F. Tamanoi. 1991. Protein geranylge ranyltransferase of Saccharomyces cerevisiae is specific for Cys-Xaa-Xaa-Leu motif proteins and requires the CDC43 gene product but not the DPR1 gene product. Proc Natl Acad Sci U S A 88: 4448-52. 13. Fukai, S., H. T. Matern, J. R. Jagath, R. H. Scheller, and A. T. Brunger. 2003. Structural basis of the interaction between RalA and Sec5, a subunit of the sec6/8 complex. Embo J 22: 3267-78.

PAGE 103

14. Garcia, A. M., C. Rowell, K. Ackermann, J. J. Kowalczyk, and M. D. Lewis. 1993. Peptidomimetic inhibitors of Ras farnesylation and function in whole cells. J Biol Chem 268: 18415-8. 15. Gonzalez-Garcia, A., C. A. Pritchard, H. F. Paterson, G. Mavria, G. Stamp, and C. J. Marshall. 2005. RalGDS is required for tumor formation in a model of skin carcinogenesis. Cancer Cell 7: 219-26. 16. Hamad, N. M., J. H. Elconin, A. E. Karnoub, W. Bai, J. N. Rich, R. T. Abraham, C. J. Der, and C. M. Counter. 2002. Distinct requirements for Ras oncogenesis in human versus mouse cells. Genes Dev 16: 2045-57. 17. Henry, D. O., S. A. Moskalenko, K. J. Kaur, M. Fu, R. G. Pestell, J. H. Camonis, and M. A. White. 2000. Ral GTPases contribute to regulation of cyclin D1 through activation of NF-kappaB. Mol Cell Biol 20: 8084-92. 18. Hinoi, T., S. Kishida, S. Koyama, M. Ikeda, Y. Matsuura, and A. Kikuchi. 1996. Post-translational modifications of Ra s and Ral are import ant for the action of Ral GDP dissociation stimulator. J Biol Chem 271: 19710-6. 19. Iurlaro, M., F. Demontis, M. Corada, L. Zanetta, C. Drake, M. Gariboldi, S. Peiro, A. Cano, P. Navarro, A. Cattelino, S. Tognin, P. C. Marchisio, and E. Dejana. 2004. VE-cadherin expression and clus tering maintain low levels of survivin in endothelial cells. Am J Pathol 165: 181-9. 20. Jackson, J. H., C. G. Cochrane, J. R. Bourne, P. A. Solski, J. E. Buss, and C. J. Der. 1990. Farnesol modification of Kirstenras exon 4B protein is essential for transformation. Proc Natl Acad Sci U S A 87: 3042-6.

PAGE 104

21. Jin, R., J. R. Junutula, H. T. Matern, K. E. Ervin, R. H. Scheller, and A. T. Brunger. 2005. Exo84 and Sec5 are competitive regulatory Sec6/8 effectors to the RalA GTPase. Embo J 24: 2064-74. 22. Joyce, P. L., and A. D. Cox. 2003. Rac1 and Rac3 are targets for geranylgeranyltransferase I inhibito r-mediated inhibition of signaling, transformation, and membrane ruffling. Cancer Res 63: 7959-67. 23. Kato, K., A. D. Cox, M. M. Hisaka, S. M. Graham, J. E. Buss, and C. J. Der. 1992. Isoprenoid addition to Ras protein is the critical modification for its membrane association and transformi ng activity. Proc Natl Acad Sci U S A 89: 6403-7. 24. Kinsella, B. T., R. A. Erdman, and W. A. Maltese. 1991. Carboxyl-terminal isoprenylation of ras-related GTP-bindi ng proteins encoded by rac1, rac2, and ralA. J Biol Chem 266: 9786-94. 25. Lebowitz, P. F., W. Du, and G. C. Prendergast. 1997. Prenylation of RhoB is required for its cell transforming function but not its ability to activate serum response element-dependent transcription. J Biol Chem 272: 16093-5. 26. Leftheris, K., T. Kline, G. D. Vite, Y. H. Cho, R. S. Bhide, D. V. Patel, M. M. Patel, R. J. Schmidt, H. N. Weller, M. L. Andahazy, J. M. Carboni, J. L. Gullo-Brown, F. Y. Lee, C. Ricca, W. C. Rose, N. Yan, M. Barbacid, J. T. Hunt, C. A. Meyers, B. R. Seizinger, R. Zahler, and V. Manne. 1996. Development of highly potent inhibitors of Ras farnesyltransferase possessing cellular and in vivo activity. J Med Chem 39: 224-36.

PAGE 105

27. Lim, K. H., A. T. Baines, J. J. Fiordalisi, M. Shipitsin, L. A. Feig, A. D. Cox, C. J. Der, and C. M. Counter. 2005. Activation of RalA is critical for Rasinduced tumorigenesis of human cells. Cancer Cell 7: 533-45. 28. Lim, K. H., K. O'Hayer, S. J. Adam, S. D. Kendall, P. M. Campbell, C. J. Der, and C. M. Counter. 2006. Divergent roles for RalA and RalB in malignant growth of human pancreatic carcinoma cells. Curr Biol 16: 2385-94. 29. Manne, V., D. Roberts, A. Tobin, E. O'Rourke, M. De Virgilio, C. Meyers, N. Ahmed, B. Kurz, M. Resh, H. F. Kung, and et al. 1990. Identification and preliminary characterization of protein-cy steine farnesyltransferase. Proc Natl Acad Sci U S A 87: 7541-5. 30. Miquel, K., A. Pradines, J. Sun, Y. Qian, A. D. Hamilton, S. M. Sebti, and G. Favre. 1997. GGTI-298 induces G0-G1 block and apoptosis whereas FTI-277 causes G2-M enrichment in A549 cells. Cancer Res 57: 1846-50. 31. Ohta, Y., N. Suzuki, S. Nakamura, J. H. Hartwig, and T. P. Stossel. 1999. The small GTPase RalA targets filamin to i nduce filopodia. Proc Natl Acad Sci U S A 96: 2122-8. 32. Oxford, G., C. R. Owens, B. J. Titus, T. L. Foreman, M. C. Herlevsen, S. C. Smith, and D. Theodorescu. 2005. RalA and RalB: antagonistic relatives in cancer cell migration. Cancer Res 65: 7111-20. 33. Peng, H., D. Carrico, V. Thai, M. Blaskovi ch, C. Bucher, E. E. Pusateri, S. M. Sebti, and A. D. Hamilton. 2006. Synthesis and evalua tion of potent, highlyselective, 3-aryl-piperazinone inhibitors of protein geranylgeranyltransferase-I. Org Biomol Chem 4: 1768-84.

PAGE 106

34. Prendergast, G. C. 2001. Actin' up: RhoB in cancer and apoptosis. Nat Rev Cancer 1: 162-8. 35. Qian, Y., A. Vogt, A. Vasudevan, S. M. Sebti, and A. D. Hamilton. 1998. Selective inhibition of type-I geranylgeranyltransferase in vitro and in whole cells by CAAL peptidomimetics. Bioorg Med Chem 6: 293-9. 36. Rangarajan, A., S. J. Hong, A. Gifford, and R. A. Weinberg. 2004. Speciesand cell type-specific requirements for cellular transformation. Cancer Cell 6: 17183. 37. Reiss, Y., S. J. Stradley, L. M. Gierasch, M. S. Brown, and J. L. Goldstein. 1991. Sequence requirement for peptide r ecognition by rat brain p21ras protein farnesyltransferase. Proc Natl Acad Sci U S A 88: 732-6. 38. Repasky, G. A., E. J. Chenette, and C. J. Der. 2004. Renewing the conspiracy theory debate: does Raf function alone to mediate Ras oncogenesis? Trends Cell Biol 14: 639-47. 39. Rosse, C., A. Hatzoglou, M. C. Parrini M. A. White, P. Chavrier, and J. Camonis. 2006. RalB mobilizes the exocyst to drive cell migration. Mol Cell Biol 26: 727-34. 40. Rowell, C. A., J. J. Kowalczyk, M. D. Lewis, and A. M. Garcia. 1997. Direct demonstration of geranylgeranylation and farnesylation of Ki-Ras in vivo. J Biol Chem 272: 14093-7. 41. Sahai, E., and C. J. Marshall. 2002. RHO-GTPases and cancer. Nat Rev Cancer 2: 133-42.

PAGE 107

42. Schafer, W. R., C. E. Trueblood, C. C. Yang, M. P. Mayer, S. Rosenberg, C. D. Poulter, S. H. Kim, and J. Rine. 1990. Enzymatic coupling of cholesterol intermediates to a mating pheromone precu rsor and to the ras protein. Science 249: 1133-9. 43. Seabra, M. C., Y. Reiss, P. J. Casey, M. S. Brown, and J. L. Goldstein. 1991. Protein farnesyltransferase and geranylgeranyltransferase share a common alpha subunit. Cell 65: 429-34. 44. Sebti, S. M., and C. J. Der. 2003. Opinion: Searching fo r the elusive targets of farnesyltransferase inhibitors. Nat Rev Cancer 3: 945-51. 45. Shipitsin, M., and L. A. Feig. 2004. RalA but not RalB enhances polarized delivery of membrane proteins to the ba solateral surface of ep ithelial cells. Mol Cell Biol 24: 5746-56. 46. St Croix, B., C. Sheehan, J. W. Rak, V. A. Florenes, J. M. Slingerland, and R. S. Kerbel. 1998. E-Cadherin-dependent growth suppression is mediated by the cyclin-dependent kinase inhi bitor p27(KIP1). J Cell Biol 142: 557-71. 47. Sugihara, K., S. Asano, K. Tanaka, A. Iwamatsu, K. Okawa, and Y. Ohta. 2002. The exocyst complex binds the sma ll GTPase RalA to mediate filopodia formation. Nat Cell Biol 4: 73-8. 48. Sun, J., M. A. Blaskovich, D. Knowles, Y. Qian, J. Ohkanda, R. D. Bailey, A. D. Hamilton, and S. M. Sebti. 1999. Antitumor efficacy of a novel class of nonthiol-containing peptidomimetic inhi bitors of farnesyltransferase and geranylgeranyltransferase I: combinati on therapy with the cytotoxic agents cisplatin, Taxol, and gemcitabine. Cancer Res 59: 4919-26.

PAGE 108

49. Sun, J., J. Ohkanda, D. Coppola, H. Yin, M. Kothare, B. Busciglio, A. D. Hamilton, and S. M. Sebti. 2003. Geranylgeranyltransf erase I inhibitor GGTI2154 induces breast carcinoma apoptosis and tumor regression in H-Ras transgenic mice. Cancer Res 63: 8922-9. 50. Sun, J., Y. Qian, Z. Chen, J. Marfurt, A. D. Hamilton, and S. M. Sebti. 1999. The geranylgeranyltransferase I inhi bitor GGTI-298 induces hypophosphorylation of retinoblastoma and partner switching of cyclin-depende nt kinase inhibitors. A potential mechanism for GGTI-298 antitumor activity. J Biol Chem 274: 6930-4. 51. Sun, J., Y. Qian, A. D. Hamilton, and S. M. Sebti. 1998. Both farnesyltransferase and geranylgeranyltran sferase I inhibitors are required for inhibition of oncogenic K-Ras prenylation but each alone is sufficient to suppress human tumor growth in nude mouse xenografts. Oncogene 16: 1467-73. 52. Vogt, A., J. Sun, Y. Qian, A. D. Hamilton, and S. M. Sebti. 1997. The geranylgeranyltransferase -I inhibitor GGTI-298 arrests human tumor cells in G0/G1 and induces p21(WAF1/CIP1/SDI1) in a p53-independent manner. J Biol Chem 272: 27224-9. 53. Wennerberg, K., K. L. Rossman, and C. J. Der. 2005. The Ras superfamily at a glance. J Cell Sci 118: 843-6. 54. Whyte, D. B., P. Kirschmeier, T. N. Hockenberry, I. Nunez-Oliva, L. James, J. J. Catino, W. R. Bishop, and J. K. Pai. 1997. Kand N-Ras are geranylgeranylated in cells treated with fa rnesyl protein transferase inhibitors. J Biol Chem 272: 14459-64.

PAGE 109

55. Yoshida, Y., M. Kawata, M. Katayama, H. Horiuchi, Y. Kita, and Y. Takai. 1991. A geranylgeranyltransferase fo r rhoA p21 distinct from the farnesyltransferase for ras p21S Biochem Biophys Res Commun 175: 720-8. 56. Zhang, F. L., and P. J. Casey. 1996. Protein prenylation: molecular mechanisms and functional consequences. Annu Rev Biochem 65: 241-69. 57. Zhu, K., A. D. Hamilton, and S. M. Sebti. 2003. Farnesyltransf erase inhibitors as anticancer agents: current status. Curr Opin Investig Drugs 4: 1428-35.

PAGE 110

Chapter 3 Ras Transformation in a Genetically Defined Human Ovarian Ca ncer Model Requires Akt and Ral but not Raf By Samuel C. Falsetti 1,2, Sad M. Sebti 1,2,* 1Drug Discovery Program, The H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL 2Departments of Interdisciplinary Oncology a nd Molecular Medicine, The University of South Florida, Tampa, FL *Corresponding Author: 12902 Magnolia Drive, Tampa, FL 33612; Tel (813) 745-6734; Fax (813) 745-6748; email: said.sebti@moffitt.org

PAGE 111

Abstract In addition to the high prevalence of in activating mutations of p53 and Rb tumor suppressors, oncogenic mutations of the Ras genes are also common in human ovarian cancer; however, the downstream effectors requi red for Ras transformation in this disease are not known. Using a genetically defi ned human ovarian epithelial cell model in which stable expressi on of Hor K-Ras12V drives transformation of ovarian surface epithelial cells (T80) immortalized by SV 40 Large T antigen (SV40T) and activated human telomerase reverse transcriptase cata lytic subunit (hTERT), we have delineated a requirement for Ral A/B and Akt1/2 but not Raf-1 or Mek1/2 expres sion and activation in Ras transformation.. Knockdown of expression of the RalGDS effectors RalA and RalB, and the PI3K effectors Akt1/2, but not Raf-1 and Mek1/2 inhibited Hand K-Ras transformation. Furthermore, stable expression of K-Ras12V effector loop mutants revealed that activation of either RalGDS with Ras12V37G or PI3K withRas12V40C, but not Raf-1 with Ras12V35S, was sufficient to transform T80 cells. In further support of Ral and Akt as the primary transforming component s of the Hand K-Ras signaling cascade perturbation of Ral or Akt but not Raf-1 si gnaling pathways by pharmacologic inhibitors attenuated Ras transformation. These results sharply contrast with pr evious studies which have reduced the requirements for Ras tran sformation in spontaneously immortalized rat ovarian surface epithelial cells to Ras/Raf signaling.

PAGE 112

Introduction Ovarian cancer is the primary cause of death among gynecological malignancies and is the 5th leading cause of cancer deaths among women in the United States (12). Despite the high incidence of ovarian cancer among women and the rela tive lethality of the disease little is known about the precise mechanism of transformation of human ovarian surface epithelial (HOSE) cells. Ind eed, multiple mechanisms of ovarian surface epithelial cell transformation have been proposed and are believe d to be sub-type specific (3). The two most common sub-types of ovari an cancer, high-grade serous and mucinous ovarian carcinoma (12), are comm only diagnosed at late-stage and are frequently lethal due to a high-propensity for both metastasis and drug-resistant recu rrence (3, 12). It is thus of critical importance to define the mechanism of ovarian epithelial cell transformation both for a more precise unders tanding of the underlying genetic risk factors as well as for the development of targeted therapies for treatment. Epithelial neoplasias, such as ovarian can cers, exhibit multiple genetic aberrations in key tumor suppressor genes, so-called cel lular gatekeepers such as p53 and Rb(16). Indeed, over 50% of ovarian cancers exhibit missense mutations resu lting inactivation of p53 (25, 30). Furthermore, over 80% of huma n ovarian cancers examined have been shown to have mutational defects in one or more of the genes involved in the Rb tumor suppressive pathway, such as CDK4, cyclin D1, and Rb (17). Activation of human telomerase (hTERT) activity is another we ll-recognized defect of human epithelial neoplasms including ovarian cancer [3,7]. In addition to p53, Rb and hTERT genetic abnormalities, activation of one or more proto-oncogenes is a pivotal fourth event in the development of human epithelial malignant disease.

PAGE 113

Mutations in Ras proto-oncogene family members (H-, K-, and NRas) are known to occur in approximately 30% of hu man cancers (40). Numerous studies have demonstrated that mutational activation of K-Ras, through missense mutations at codons 12 and 13, occurs with high frequency (27-69%) in certain subtypes of ovarian cancers (7, 11, 20, 33, 44). Similar mutations in the H-Ra s proto-oncogene have been determined to occur in ovarian cancer s ubtypes though there is consider able disagreement over the frequency (9, 44, 53).A role for Ras mutation in ovarian cancer progression is especially well recognized in mucinous sub-type ovarian tumors where the incidence of K-Ras proto-oncogene mutation increases profoundly dur ing the progression of disease (33). For example, Mok and colleagues found a high, pr ogression-dependent incidence of K-Ras activation at codons 12 and 13 in 13% of mucinous adenomas, 33% of mucinous tumors at the borderline, and 46% of mucinous carcinomas (33). Similar results have been reported by many other groups (7, 9, 20, 33, 44). K-Ras mutation have also been detected, albeit to a lesser degr ee (up to12% only), in high-gr ade serous sub-type ovarian cancer, the most common form of ovarian maligna ncy (33). It is intere sting to note that a much higher incidence of Ras mutation exists in low-grade serous malignancy. Indeed, over 68% of low grade serous tumors have either a Ras or Raf mutation (31), and Ras muations have been found in up to 48% of borderline ovarian epithelial tumors (33) indicating a role for Ras mutations in the early events of lowgrade serous ovarian neoplasia. Recently, genetically defined models of human epithelial transformation have been developed which have enabled, for th e first time, the description of oncogene stimulated transformation in cells of human epithelial origin (13-15, 21, 27, 29, 37).

PAGE 114

Sequential inactivation of the tumor suppre ssors p53 and retinoblastoma (Rb) protein by ectopic expression of SV40-T-antigen (TAg) or polyoma middle T antigen (PoMT), coupled with stable expression of human telomerase catalytic subunit (hTERT) in human epithelial and fibroblastic cell lin es recapitulates the critical events in the development of human neoplastic disease and creates a minimally transformed model for the study of oncogenic signaling pathways (13, 14). Creat ion of these models has prompted a reexamination of the essential downstream signaling components of the Ras oncogene. Previous reports in the spont aneously immortalized NIH3T3 murine fibroblast model identified the Raf-MEK-Erk pathway, but not phosphatidylinositol 3 kinase (PI3K)-Akt or Ral guaunine nucleotide dissociation stim ulator (RLGDS)-Ral pathways, as being necessary and sufficient to promote Ras-mediated growth and morphologic transformation (38). However, recent work ha s demonstrated tissueand species-specific requirements for Ras downstream effectors, other than Raf, in malignant transformation (15, 27, 37).While in one study [19], the only Ras effector pathway to show independent transforming activity in the human embr yonic kidney cell system (HEK-HT) was RalGEF-RalA small GTPase pathway, whereas in another study [23] both Ras/RalGEF and Ras/PI3K pathways were required for Ras transformation in the same HEK-HT model.. In contrast, in human mammary epith elial cells (HMECs), which are minimally transformed by hTERT and TAg, Ras transforma tion requires concurrent activation of Raf, RalGEF, and PI3K. Yet in another study [26], spontaneously immortalized rat ovarian surface epithelial (ROSE) cells, Raf but not PI3K or RalG EF, was independently transforming. Therefore, the requirements for Ras transformation appear to be highly dependent on the tissue and species of origin.

PAGE 115

For ovarian cancer, while the ROSE cell model has been useful in suggesting which signaling cascades are required for Ras transformation in ovarian epithelial cells, it is neither genetically defined nor of human origin. Thus, while suggestive of a role for Raf in ovarian transformation, the ROSE model has unknown prognostic significance for human ovarian oncogenesis. Recently, a gene tically defined human model of ovarian malignancy has been developed which recapitulate s the critical events in neoplasia (29). The T80 cell line is derived from primary human ovarian surface epithelial cells engineered to express TAg and hTERT. Upon the addition of retrov irally transduced HRas12V, these cells, referred to as T80H, acquire anchorage independent growth capability in soft agar and are capable of forming xenograft tumors in a SCID mouse model. This transformation requires continued expression of H-Ras (49) and, as such, is a bona fide model of Ras-dependent, ovarian cell-type-specific transforma tion. Using this system, as well as our own stable K-Ras12V expressing variant (T80K), we selectively attenuated Ras signaling to the three primary Ras signal transduction pathways (RalGEF-Ral, PI3K-Akt, and Raf-MEK-Erk) using a combination of genetic and pharmacologi cal approaches, as well as ectopic expression of Ras effector loop mutants to determine which of the downstream signaling components are required for human ovarian su rface epithe lial cell transformation. In contrast to the ROSE cell model we have determined that RalGEF-Ral and PI3K-Akt but not Raf-MEK-Erk are the pr imary pathways of Ras transformation of human ovarian epithelial cells.

PAGE 116

Materials and Methods Cells and culture Human ovarian surface epithelial cells T80 and T80H were cultured as described previously (29). The T80K, T80K12V35S, T80K12V37G, T80K12V40C, stably expressing cell lines were maintained in sim ilar culture conditions as the T80 and T80H cells [22]. Human HEK-293T was a kind gi ft from Dr. Gary Reuther and were maintained in Dulbeccos modified minimum essential medium (DMEM) 10% fetal bovine serum and 1% penicillin/ streptomycin at 10% CO2 and 37C. Plasmids Retroviral plasmids pBAB E-puro containing the K-Ras12V, KRas12V35S, KRas12V37G and KRas12V40C transgenes were a kind gift from Dr. Gary Reuther (15). Creation of retrovirus Retrovirus was created by transi ent transfection of HEK-293T human embryonic kidney cells with pVPACK-Ampho, pVPACK-gag-pol and pBABE retroviral plasmids as previously described (10). Briefly, 293T cells were seeded at 2.5x106 cells in a 60mm2 dish. 3 g of each plasmid was combined and brought to a volume of 225 l in RNase/DNase free water, following which 25 l 2.5M CaCl2 was added dropwise while gently vortexing. Th is solution was then added to 250 l 2x HEPES Buffered Saline dropwise and incubated at room temperature for 5 min. The total volume of 500 l was then added to the 293T cells and incubated for 8 hs at 37 C The media was then removed and cells were incubated for 48 h. The supernatant was then passed through a 0.45 m nylon low protein-binding filter to remove cells and cell debris. Creation of stable cell lines T80 cells were seeded in to 6 well plates at a 40% confluency and incubated with 10 g/mL polybrene (Millipore/ Specialty Media, Billirika, MA) and a 1:3 dilution of retrovirus containing conditioned media for 72 h. Media was then replaced with complete growth medium containing 1.0 g/mL

PAGE 117

puromycin and incubated until colonies forme d. All colonies were pooled and taken as a single polyclonal population and cult ured in complete medium. Small interfering RNA (siRNA) sequences SiRNA sequences targeting RalA and RalB were purchased from Dharmacon using previ ously described descri bed sequences (6). Specifically, RalA: 5'-GACAGGUUU CUGUAGAAGAdTdT-3', RalB: 5'GGUGAUCAUGGUUGGCAGCdTdT-3' Pre-designed chemically synthesized siRNA targeting Akt1/2 (Cat #6211, Cell Signaling Technology Inc., Danvers, MA), Raf1 (Cat #M-003601-00, Dharmacon, Lafayette, CO) MEK1 (Cat #6420, Cell Signaling Technology Inc., Danvers, MA) and MEK2 (Cat #6431, Cell Signaling Technology Inc., Danvers, MA) were used according to manufacturers recommendation. Small interfering RNA (siRNA) transfection procedure Cells were grown to 50% confluence and transfected with Oligofectamine transf ection reagent (Invitrogen, Carlsbad, CA) according to manufacturers inst ructions. Briefly, 5 l of oligofectamine reagent was suspended per 1 mL of OPT I-MEM media (Invitrogen, Carlsbad, CA) and allowed to equilibrate at 24-27C. A 100 nM final concentration of siRNA was suspended and allowed to complex with the liposomes. Cells were briefly washed with OPTI-MEM and the transfection mix was plat ed on top of the cells and incubated at 37C. Following 8 h incubation 2 mL of complete growth media, without penicillinstreptomycin, was added and the cells were further incubated at 37C overnight. The media was replaced after 24 h with complete growth medium for 24 h. Experiments were performed in triplicate. Pharmacological inhibitors Cells were treated for 48 h with either DMSO vehicle (Sigma-Aldrich, St. Louis, MO), 20 M LY29004 (Calbiochem, San Diego, CA), 10 M

PAGE 118

U0126 (Sigma-Aldrich, St. Louis, MO), or 10 M GGTI-2417 (36). Experiments were performed in triplicate. Western blotting Cells were harvested, and lysed in HEPES lysis buffer as described previously (10). Proteins were then resolved by 11.5% SDS-PAGE gel and immunoblotted with antibodies against phosphorylated-Ser473 Akt1/2 (9217, Cell Signaling Technologies Inc., Danvers, MA), phosphorylated-Thr202/Tyr 204 Erk1/2 (9101, Cell Signaling Technologies Inc., Danve rs, MA), Akt1/2 (N-19, Santa Cruz Biotechnology, Santa Cruz, CA), Erk1/2 (p44/p42 MAP Kinase, 9102, Cell Signaling Technologies Inc., Danvers, MA), RalA (61022, BD Biosciences Pharmingen, Franklin Lakes, NJ), H-Ras(C-20, Sa nta Cruz Biotechnology, Sant a Cruz, CA), K-Ras (F234, Santa Cruz Biotechnology, Sant a Cruz, CA), RalB (04037, Mill ipore, Billerika, MA), actin (AC15, Sigma-Aldrich, St. Louis, MO ) and anti-FLAG M2 monoclonal antibody (Sigma-Aldrich, St. Louis, MO). The ECL blotting system (NEN Life Science Products, Boston, MA) was used for detection of pos itive antibody reactions. Experiments were performed in triplicate. Soft agar clonogenicity assay For soft agar growth assays, the cell lines were seeded at a cell density of 5000/well in tr iplicate in 12-well culture dishes in 0.3% agar over a 0.6% bottom agar layer as previously described (10). Cultures were fed once weekly until colonies grew to a suitable size for observ ation (approximately 14 days). Colonies were photographed after overnight incubation with 1 mg/ml MTT in cell growth media.

PAGE 119

Colony number was visually determined and quantified. Experiments were performed in triplicate and 3 independent e xperiments were performed.

PAGE 120

Results Differential transforming activity and si gnaling activation by Hand KRas12V isoforms in the T80 human ovarian su rface epithelial (HOSE) cell line. The incidence of K-Ras mutation in ovarian malignancy is far greater than that of H-Ras (3, 7, 9, 20), suggesting a preferential role for K-Ras in the development of ovarian neoplasia. Additionally, while Ras isoforms are commonly assumed to activate the same three primary transformative path ways; Raf/MEK/Erk, PI3K/Akt and RalGDS/ Ral, the degree to which Hand KRas isof orms activate these pathways vary according to both physical association cons tants (8, 48) as well as cell -type specific differences (15, 26, 37, 43). Therefore, we first wanted to ex amine differences in both the transformative capability and the signaling pathway utiliz ation of oncogenic, mutated Hand KRas12Vin the HOSE T80 cell model where hTERT is overexpressed and p53 and Rb are inactivated by Tag as described in Liu et al (29).To this end, we created a T80 cell line stably expressing oncogenic K-Ras12V (T80-K) as describe d under Materials and Methods. The T80 cell line stab ly expressing oncogenic H-Ras12V (T80H) was obtained from Dr. Bast (29). We first compared the downstream effector utiliz ation of the two Ras isoforms by SDS-PAGE and we stern blot analysis as we ll as Ral-GTP binding assay (figures 12A and 12B, see followi ng pages). Both Hand K-Ras12V potently stimulated GTP loading of RalA and RalB to a si milar extent. However, while both H-Ras12V and KRas12V induced phosphorylation of Akt1/2 and Erk1/2, K-Ras12V was relatively less potent; indicating a differential specificity of the Ras isoforms for activation of Akt and Erk in the human ovarian cancer cell model. Finally, we used anchorage-independent growth on soft agar to show that ec topic expression of both mutant H-Ras12V and K-

PAGE 121

Ras12V promotes soft agar clonogenicity in the T-80 model. Figure 12C shows that whereas T80 grew only 42 5colonies, T 80-H and T80-K grew 183 28 (p<0.05) and 114 13 (p<0.01) colonies, respectively.

PAGE 122

Figure 12. Differential transforming activity and signaling activation by H-and K-Ras12Vin human ovarian surface epithelial (HOSE) cells RalA RalB RalA-GTP RalB-GTP T80 T80H T80K A. T80 T80H T80K K-Ras Erk1/2 p-Erk1/2 Akt1/2 p-Akt1/2 Actin H-Ras B.

PAGE 123

Figure 12. Differential transforming ac tivity and signaling activation by H-and K-Ras12Vin human ovarian surface epithelial (HOSE) cells # ColoniesT80T80HT80K ** = P<0.05, ** = P<0.01 compared to T80 50 100 150 200 250 C. 0 T80 T80H T80K

PAGE 124

Figure 12. Differential transforming ac tivity and signaling activation by H-and K-Ras12Visoforms in the T80 human ovarian surface epithelial (HOSE) cell line. T80 cells stably transduced with H-Ras12V(T80H) or KRas12V(T80K) were collected and lysed. Proteins were separated by SDSPAGE and used for Ral-GTP quantitationusing GST-RalBPpull down assay (A) or western blot analysis (B); results are representative of three experiments. T80, T80H and T80K cells were seeded into 0.3% softagar for two weeks, colonies were manually scored (C); T80 cells formed 43+/-5 colonies per well, T80H formed 183 +/-28 colonies per well and T80K formed 114+/-13 colonies per well. Results are representative of three experiments each performed in triplicate.

PAGE 125

Ras transformation of human ovarian surfac e epithelial cells requires expression of RalA/B and Akt1/2 but not Raf1 or Mek1/2. We next determined the requirements for expression of the Ras effectors RalA, RalB, Akt1/2, and Raf1 and Mek1/2 in both Hand KRas12V mediated transformation of T80 cells. Fo r these analyses, we used siRNA to RalA, RalB, Akt1/2 and Raf1 and Mek1/2 to specifically deplet e each gene product and systematically evaluate the effects of this knockdown on Ras transforming capability as described under Materials and Methods. Figure 13A (see following pages) shows that in both T80H and T80K RalA siRNA knocked down RalA, but not RalB, expression whereas RalB siRNA knocked down RalB, but not RalA, expression ; these results demonstrated, by western blot analysis, that each siRNA was specific. Figure 13B (see following pages) shows that knockdown of expression of RalA potently inhibited H-Ras12V and K-Ras12V transformation of T80 cells by 56%% and 54%%, respectively (all p values were less than 0.01). Similarly, knockdown of expression of RalB potently inhibited H-Ras12V and K-Ras12V transformation of T80 cells by 54% 11% and 66%%, respectively (all p values were less than 0.01). Knockdown of Akt1/2 expres sion in both T80H and T80K cells (Figure 14A, see following pages) inhibited transformation by 46%% and 42%%, respectively (Figure 14C, all p values were less than 0.05). In contrast knockdown of Raf1 expression (Figure 3A, see following pages) did not have any statistically significant effect on the clonogenicity of T80K and T80H cells (Figure 14C, see following pages). Since inhibition of Ra f1 expression did not result in reduction of Ras transformation we aimed to further evaluate the role of Ra f/MEK/Erk signaling by knocking down the expression of Mek1/2. While inhibition of Mek1/2 expression

PAGE 126

resulted in the attenuation of Erk1/2 phosphor ylation (figure 14B, see following pages), it did not inhibit either H-Ras12V or K-Ras12V transformation (figure 14C, see following pages). These results suggest that RalA, Ra lB and Akt1/2, but not Raf1 or Mek1/2, are essential components of Ras mediated human ovarian epith elial cell tran sformation.

PAGE 127

Figure 13. Ras transformation of HOSE cells requires expression of RalA/B. 0.0 25.0 50.0 75.0 100.0 125.0 % control N.C. RalA RalB *** *** ** ** = T80 H-Ras12V = T80 K-Ras12V B. A. Actin RalB RalA T80H N.C. RalARalB T80K N.C. RalA RalB T80 H-Ras12V T80 K-Ras12V siRNA: N.C. RalA RalB

PAGE 128

Figure 13. Ras transformation of HOSE cells requires expression of RalA and RalB. Small interfering RNA (siRNA) specific to Ra lA and RalB or a negative control siRNA were transfected for 48 hours into T80, T80H and T80K cells. Cells were collected, counted using trypan blue dye exclusion assay and collected for western blot analysis (A) and an equal number of cells were pl ated for soft agar assay(B) and scored manually at the end of two weeks; for T 80-H N.C. cells 153+/-11 cells = 100%, for T80K N.C. cells 115+/-17 cells = 100%; results ar e representative of three experimentseach performed in triplicate.

PAGE 129

Figure 14. Ras transformation of HOSE cells requires expression of Akt1/2 but not Raf1 or MEK1/2. T80H N.C. Akt1/2 Raf1 P-Erk Erk1/2 Akt1/2 Raf1 Actin T80K N.C. Akt1/2 Raf1 A. B. Mek pErk Erk Actin N.C. Mek1/2 T80H N.C. Mek1/2 T80K

PAGE 130

Figure 14. Ras transformation of HOSE cells requires expression of Akt1/2 but not Raf1 or MEK1/2. C.% Control N.C. Akt1/2 Raf1 MEK1/2 = T80 H-Ras12V= T80 K-Ras12V 50 100 25 0 75 125 *** T80 H-Ras12V T80 K-Ras12V siRNA: N.C. Akt1/2 Raf1 MEK1/2

PAGE 131

Figure 14. Ras transformation of HOSE cells requires expression of Akt1/2 but not Raf1 or MEK1/2. Small interfering RNA (siRNA) specific to Akt1/2 and Raf1 or a ne gative control siRNA were transfected for 48 hours into T80, T80H and T80K cells. Cells were collected, counted using trypan blue dye exclusion assa y and collected for western blot analysis (A) and an equal number of ce lls were plated for soft agar assay (B) and scored manually at the end of two weeks; for T80-H N.C. cells 153+/-11 cells = 100%, for T80-K N.C. cells 115+/-17 cells = 100%; results are representative of three expe riments each performed in triplicate.

PAGE 132

K-Ras effector loop mutants which preferen tially activate RalGDS and/or PI3K, but not Raf1, are capable of transforming T80 HOSE cells to a similar, or greater, extent as compared to fully active K-Ras12V. To further evaluate the contributions of the Ras effector pathways RalGDS/Ral, PI3K/Akt, and Raf/MEK/Erk to Ras transfor mation in human ovarian surface epithelial cells we used retroviruses to stably tranduce oncogenically active K-Ras12V with effector loop amino acid substitutions (35S, 37G, and 40C) which are well characterized to restrict the interaction of Ras to, resp ectively, Raf-1, RalGDSand PI3K (22, 23, 32, 39, 45, 46). Figures 15A and 15B (see fo llowing pages) show that K-Ras12V stimulated phosphorylation of P-Erk1/2 and Akt as well as induced RalA and Ral GTP loading. KRas12V35S (which activates Raf but not PI3K or RalGDS) activated Erk1/2 phosphorylation but did not stimulate RalA or RalB GTP loading and did not activate Akt1/2 phosphorylation (figure 15A and 15B, see following pages). Similarly, KRas12V37G (which binds RalGDS but is defective in binding to Raf and PI3K) stimulated GTP-loading of RalA and RalB but did not stimulate phosphorylation of Akt1/2 and Erk1/2 (figure 15A and 15B, see following pages). K-Ras12V40C (which activates PI3K but not RalGDS or Raf) activated Akt but not Erk1/2 phosphorylation and did not stimulate GTP-loading of RalA or RalB (figure 15A and 15B, see following pages). Further evaluation of the transforming potential of these mutational variants relative to fully oncogenic K-Ras revealed that Ras e ffector loop mutants which preferentially activated RalGDS/Ral (K-Ras12V37G) and PI3K/Akt (K-Ras12V40C), but not Raf/MEK/Erk (K-Ras12V35S), were capable of transforming T80 HOSE cells to a similar or greater extent than K-Ras12V (figure 3C, see following pages). For example, KRas12V and K-

PAGE 133

Ras12V40C increased the number of colonies fr om 62 16 in pBABE (T80 cells) to 195 25 and 201 13 respectively. K-Ras12V37G was more potent and resulted in 282 20 colonies. In contrast, K-Ras12V35S was dramatically reduced in transforming potential, 50% of K-Ras12V or 95.3 colonies (p value < 0.005). These findings both confirmed and extended our results by further demonstr ating the central impor tance of PI3K/Akt and RalGDS/Ral, but not Raf-1, in ova rian epithelial cell transformation.

PAGE 134

A. Figure 15. K-Ras mutants which preferentially activate RalGDS and/or PI3K, but not Raf1, are capable of transforming T80 HOSE cells to a similar, or greater, extant as compared to fully active K-Ras12VGTP-RalB RalB GTP-RalA RalA K12V K12V35S K12V37G K12V40C pBABE B. K-Ras P-Erk Erk1/2 P-Akt Akt1/2 K12V K12V37G K12V40C pBABE K12V35S

PAGE 135

Figure 15. K-Ras mutants which preferentially activate RalGDS and/or PI3K, but not Raf1, are capable of transforming T80 HOSE cells to a similar, or greater, extant as compared to fully active K-Ras12V# ColoniesK12V K12V35S K12V37G K12V40C 100 200 300 C. 0 pBABEXX ** K12VK12V35SK12V37GK12V40CpBABE

PAGE 136

Figure 15. K-Ras mutants which preferentially activate RalGDS and/or PI3K, but not Raf1, are capable of transforming T80 HOSE cells to a similar, or greater, extent as compared to fully active K-Ras12V. T80 cells stably transduced with empty vector, KRas12V(T80K), K-Ras12V35S(T80K35S), K-Ras12V37G(T80K37G) or K-Ras12V40C(T80K40C) were collected and lysed. Cell lysateswere separated by SDS-PAGE and used fo r western blot analysis (A) or quantitation of Ral-GTP levels using GST-RalBP pull down assay (B); results are representative of three experiments. T80, T80K and T80K35S, T80K37G and T80K40C cells were seeded into soft agar for two weeks, colonies were manually scored (C); results are representative of three experiment s each performed in triplicate. **=p<0.01 (greater than T80K), xx=p<.01(less than T80K)

PAGE 137

Pharmacological inhibitors of PI3K or Ral, but not MEK, inhibit Ras transformation of HOSE cells. While we have demonstrated that expr ession and activation of RalGDS/Ral and PI3K/Akt, but not Raf-1/ Mek1/ 2 are required for Ras transfor mation we also wanted to determine the effects that pharmacological pe rturbation of these pathways would have on Ras transformation. We used well-characterized inhibitors of MEK (U0126) and PI3K (LY294006) to selectively attenuate these pathways. We also used GGTI-2417, a small molecule competitive inhibitor of geranylgeranyltransferase-I (36), to inhibit the prenylation of RalA and RalB. Both RalA and RalB require gera nylgeranylation for both localization and biological activity (19) and are necessary components of GGTI antineoplastic activities (10). As expected, pha rmacological inhibition of PI3K by LY294006 attenuated phosphorylation of Akt (Figure 16A, see following pages). LY294006 also inhibited P-Erk1/2 in both cell lines. GGTI-2417, similar to our previous reports, inhibited the prenylation of bot h RalA and RalB (see slight mobility shift in Figure 16A, see following pages). The MEK1/2 inhi bitor UO1266 partially inhibited Erk1/2 phosphorylation, without any effect on Akt phosphorylation (Figure 16A, see following pages). GGTI-2417 inhibited soft agar cl onogenicity by 85.5% (T80K, p value <0.005) and by 54.5% (T80H, p value <0.05) (Figure 16B, see following pages). LY294006 inhibited soft agar clonogenicity by 89.6% (T 80H, p value <0.01) and by 90.1% (T80K, p value <0.01). In contrast, UO126 did not attenuate either T80H or T80K soft agar clonogenicity.

PAGE 138

A. P-Erk Erk1/2 P-Akt Akt1/2 Actin DMSO LY UO GGTIDMSO LY UO GGTIT80-H T80-K RalA RalB B. DMSO LY UO GGTIDMSO LY UO GGTIT80-H T80-K Figure 16. Pharmacological inhibitors of PI3K or Ral, but not MEK, inhbitRas transformation of HOSE cells.

PAGE 139

C.% Control DMSOLY UOGGTI 50 75 100 125 ** ** *** 25 = H-Ras12V= K-Ras12V0 DMSO LY UO GGTI T80-K T80-H Figure 16. Pharmacological inhibitors of PI3K or Ral, but not MEK, inhbit Ras transformation of HOSE cells.

PAGE 140

Figure 16. Pharmacological inhibitors of PI3K or Ral, but not MEK, inhbitRas transformation of HOSE cells. T80, T80H and T80K cells were exposed to full growth media containing either 0.01% DMSO, 20M LY29004, 10M U0126, or 10M GGTI-2417 for 48-hours. Cells were then collected, counted using trypan blue dye exclusion assay and collected for western blot analysis (A) and an equal number of cells were plated for soft agar assay (B ) and scored manually at the end of two weeks; results are representative of three experiments each performed in triplicate.

PAGE 141

Discussion Several studies (7, 11, 20, 41, 44) ha ve implicated mutant K-Ras in the development of ovarian cancer, especially in mucinous sub-type malignancies, the second most common form of ovarian neoplasia (7, 33, 44). In this study we examined the roles of the three most thoroughly characterized Ra s effector pathways RalGEF-Ral, PI3K-Akt and Raf-MEK-Erk in mediating Ras anchorag e independent growth of a genetically defined model of human ovarian cancer. We determined that Ras anchorage independent growth in this tissue type proceeds prim arily through the RalGEF-Ral and PI3K-Akt pathways but not through Raf-MEK-Erk, in sh arp contrast to prev ious reports that spontaneously immortalized rat ovarian surf ace epithelial cells coul d be transformed by H-Ras12V in a Raf1-dependent manner (43). It is important to note that anchorage independent growth is only one measure of transformation and other aspects of transformation, such as invasion and metastas is may require Raf1 signal transduction. We first determined that the transforming activity of both Hand K-Ras12V required the expression and activity of RalA, RalB and Akt1/2 but nor Raf1 or Mek1/2. We next determined that K-Ras mutants whic h preferentially activat ed RalGEF, or PI3K, but not Raf, were capable of transformation to a similar or greater extent than K-Ras capable of activating all three pathways. Impor tantly, we discerned that K-Ras capable of only activating the Raf-MEK-Erk pathway wa s incapable of transforming T80 cells. These results suggest that RalGEF or PI3K activation alone is sufficient to promote Rasmediated anchorage-independent growth. Howe ver, it should be emphasized that these effector mutants do retain the ability to bind other Ras effectors that may regulate transformation. For example the 37G mutant retains the ability to activate phospholipase

PAGE 142

C epsilon. We have also used pharmacological probes to provide further support for the role of Ral and Akt but not Raf pathways in Ras mediated ovarian cell transformation. One of our initial observations was that differences exist in both transforming potency and effector utilization among Ras isoforms. Indeed, K-Ras12V was less potent than H-Ras12V in the ability to both transform and activate Akt and Erk. Interestingly, another independently derived T80K cell line has also exhibited similarly diminished transformative capability in soft agar assays ( 29), making it unlikely that this is an artifact of viral transduction. It is inte resting to speculate on the cause s of this differential activity as these results stand in marked cont rast to previous studies of H-Ras12V in rat ovarian surface epithelial (ROSE) cells in which stable expression of H-Ras12V did not stimulate Akt activation in ROSE cells (43) However, we find that H-Ras12V potently stimulates Akt activation whereas K-Ras12V was less potent. Similarly, while stable expression of either Ras oncogene stimulate d Erk hyper-phosphorylation, K-Ras12V was less potent. However, HRas12V and KRas12V were equipotent in their ability to stimulate GTPloading of RalA and RalB. While these results suggest a species-specific difference in Ras isoform effector utilizati on, at least when compared to the previous studies in ROSE cells (43), there has been no effort to systematically determine Ras isoform tissuespecific differences in effector utilization. Howe ver, previous studies in murine fibroblast NIH-3T3 cells indicate that H-Ras pref erentially activates Erk whereas K-Ras preferentially activates Akt ( 48). We next determined which limbs of the Ras pathways are required for its ability to transform HOSE cells. We found that either specific depletion of RalA or RalB by si RNA or inhibition of Ral prenylation by GGTI-2417 was sufficient to strongly attenuate transformation by

PAGE 143

Ras12V. Others have previously demonstrated a central requirement for RalGDS (27), and specifically RalA, in the tran sformative activity of H-Ras12V in HEK-HT cells. Furthermore, studies in NI H-3T3 fibroblasts have shown that ectopic expression of dominant negative RalA can partially reverse Ras-dependent transformation (1); and inhibition of RalA by either siRNA mediated depletion or overexpression of dominant negative RalA has been shown to inhibit an chorage independent pr oliferation, invasion and metastasis of human tumor cell lines of non-ovarian origin (6, 45, 51). While RalB expression was previously descri bed as dispensable for H-Ras12V mediated transformation of human embr yonic epithelial kidney cells (2 7) we have found a central requirement for RalB in both Hand K-Ras12V transformation of ovarian epithelial cells. Indeed, the effects of depleti on of RalB by siRNA were equiva lent to depletion of RalA in attenuating clonogenicity of T80H and T80K cells. However, we have also previously reported in the human pancreatic cancer cell line MiaPaCa2 biological differences between RalA and RalB, which suggested a ro le for RalA but not RalB in clonogenicity (10). Using a chemical biology approach we determined that GGTI-insensitive mutants of RalA rescued from GGTI-inh ibition of anchorage-indepe ndent growth while similar mutants of RalB rescued from divergent pr ocesses such as inhibition of anchoragedependent proliferation, induction of apoptosis, in crease in p27Kip1 and decrease in survivin in human pancreatic cancer cells ( 10). Others, such as White, Theoderescu, Der and Counter have used siRNA to RalA and RalB to similarly demonstrate divergence between the Ral isoforms in multiple human pancreatic and renal cancer cell lines, as well as certain human cervical and prostate cancer cell lines, respectively HeLa and SW480 (6, 28, 35). Given the presumed biologically divergent roles of the Ral family it is

PAGE 144

interesting to consider whether the ability of RalA and RalB expression to govern Ras transformation is through redunda nt or divergent pathways. In support of a role for RalGEF/Ral signaling in the progression of ovarian cancer, expression of K-Ras12V37G, which is deficient for activation of Raf-1 and PI3K, activated GTP-loading of RalA and Ra lB to a similar extent as K-Ras12V and was able to transform to a similar or greater extent than K-Ras12V in T80 cells. This is similar to the results in the HEK-HT system reporte d by the Counter lab in which H-Ras12V37G had independent transformative capability (27) but st ands in marked contrast to the results of the Weinberg lab where H-Ras12V37G was only transforming when combined with activation of PI3K (15). We have also identified the PI3K/Akt pathway as being central for Ras-induced ovarian oncogenesis. Indeed, previous studies on the PI3K /Akt signaling pathway in ovarian cancer provide ample evidence that ab errations in this pathway could play an important role in the progression of ovarian neoplasia. Specif ically, Akt2 kinase activity is elevated in ovarian carci noma patient biopsies (52), Akt2 gene amplification occurs in 12-18% of ovarian carcinoma samples analyzed (2, 5) and mutation of PI3K p110 subunit occurs in 20% of serous carcinomas (47). Ho wever, none of these observations have been correlated to Ras mutational status. Our resu lts demonstrate that inhibition of Akt1/2 expression by siRNA and i nhibition of Akt activation by the PI3K inhibitor LY294006 strongly attenuated both T80H and T80K transformation. These results are in concurrence with previous reports in ROSE cells in which LY294006, an inhibitor of PI3K, reversed H-Ras12V transformation (43). However, in this system since H-Ras12V expression did not result in hyperphosphorylation of Akt1, or Akt2, no role could be

PAGE 145

assigned to Akt signaling. In contrast in our present study we found both H-Ras12V and K-Ras12V could activate Akt1/2. This is consistent with previous studies in spontaneously immortalized NIH-3T3 murine fibroblasts and rat intestinal epithelial (RIE) cells that have correlated Ras isoform transformation potential to the activa tion of PI3K (26). However, it is important to note, despite the lower level of Akt activation by K-Ras12V as opposed to H-Ras12V both T80H and T80K cell lines were equally sensitive to inhibition of Akt activation by LY and expression knockdown by Akt siRNA. Ascribing a causal role for PI3K and Akt in Ras transformation of epithelial cells is particularly important given the wealth of small molecules current ly under clinical investigation as antineoplastic agents which target PI3K and Akt (18). Indeed, one of these compounds, triciribine monophosphate (TCNP) (50), is curre ntly in phase I trials in patients where tumors (including ovarian) contain persis tently activated hyper-phosphorylated Akt (Clinical trial #NCT00363454). Our results sugge st inhibition of e ither PI3K or Akt could be beneficial to pati ents harboring Ras mutation positive ovarian tumors. In further support of a role for PI3K/A kt signaling in the ovarian neoplastic processes independent of Ras we have found KRas12V40C, activates Akt to a similar extent as K-Ras12V and is capable of simila r transforming activity. However, since this effector domain mutant retains the ability to bind to other Ras effectors, more definitive demonstration of this will require a more thorough examination of the independent transforming activity of constitutively activat ed Akt and PI3K isoforms. Previous cell line systems, such as HEK-HT and HMEC have not ascribed an independent role for PI3K/Akt signaling in the induction of epithelial oncoge nesis however PI3K-Akt signaling was required for maintenance of Ra s transformation in HEK-HT cells (6, 15).

PAGE 146

Furthermore, given both frequency of PTEN deletions in ovarian cancer (24) and the presence of activating point mutations in PI 3K p110 gene (47) as well as Akt1 and Akt2 (4, 34, 42) these results are c onsistent with epid emiologic observations. To date most studies carried out have examined the overlapping incidence of BRAF and RAS mutations. While some limited studies ha ve focused on understanding the overlapping frequency of RAS and PI3K mutations in ova rian cancer these results suggest further analysis of the mutational status of the various Ras effectors, such as mutational aberrations in PI3K/Akt signaling and the various RalGEFS, relative to Ras mutation among the various sub-types of ovarian cancer is warranted. Perhaps most interestingly our results s uggest that Raf/Mek/Erk stimulation does not play a substantial role in either Hor K-Ras12V transformation of human ovarian surface epithelial cells. Stab le expression of K-Ras12V35S, deficient for activation of PI3K and RalGDS, activated Erk phosphorylation to a similar extent as K-Ras12V but lacked transformative capacity in T80 cells. Furtherm ore, specific depletion of either Raf1 or MEK1/2 by siRNA or pharmacologic inhibi tion of MEK by U0126 did not affect Ras mediated transformation. Taken together, these results suggest that ovarian cancer patients with ovarian subtypes typically devo id of BRAF mutations, and whose tumors harbor activating Ras mutations, might not bene fit from Raf or MEK inhibitors currently in clinical use. In summary we have described, for the fi rst time, the essential pathways for Ras transformation in human ovarian surface epithe lial cells. We have reduced these signaling requirements to activation of RalA, RalB and Akt1/2 and have further demonstrated that specific inhibition of these proteins by eith er small interfering RNA or by pharmacologic

PAGE 147

inhibition is potently limiting to Ras transforma tion of an experimentally derived ovarian model. These results validate the Ral-GDS/Ral and the PI3K/Akt pathways as targets for developing novel anti-cancer drugs to combat ovarian cancer. To this end, we have recently identified a selective Akt activation inhibitor, tric iribine monophosphate or API002, which will soon enter phase II clinical tria ls (50). Similarly, our GGTase I inhibitor, GGTI-2418, is undergoing advanced preclinical studies and will soon enter phase I clinical trials. Finally, while we have is olated the Ras signal transduction pathways required for ovarian transformation whether actiavted RalGEFs and/or PI3K isoformes can independently transform ovarian surf ace epithelial cells remains unknown and could further define the central requirements fo r human surface ovarian epithelial cell oncogenesis.

PAGE 148

Acknowledgements The authors would like to acknowledge Norbert Berndt and Channing Der for thoughtful advice in writing this manuscript. Furthermore they would like to acknowledge Gary Reuther for supplying pl asmid used in this study and Gregor Springette for use of his T80 ovarian cell lines.

PAGE 149

References 1. Aguirre-Ghiso, J. A., P. Frankel, E. F. Farias, Z. Lu, H. Jiang, A. Olsen, L. A. Feig, E. B. de Kier Joffe, and D. A. Foster. 1999. RalA requirement for vSrcand v-Ras-induced tumorigenicity and overproduction of urokinase-type plasminogen activator: involvement of metalloproteases. Oncogene 18: 4718-25. 2. Bellacosa, A., D. de Feo, A. K. Godwin, D. W. Bell, J. Q. Cheng, D. A. Altomare, M. Wan, L. Dubeau, G. Scam bia, V. Masciullo, G. Ferrandina, P. Benedetti Panici, S. Mancuso, G. Neri, and J. R. Testa. 1995. Molecular alterations of the AKT2 oncogene in ovarian and breast carcinomas. Int J Cancer 64: 280-5. 3. Borresen, A. L. 1992. Oncogenesis in ovarian ca ncer. Acta Obstet Gynecol Scand Suppl 155: 25-30. 4. Carpten, J. D., A. L. Faber, C. Horn, G. P. Donoho, S. L. Briggs, C. M. Robbins, G. Hostetter, S. Boguslawski, T. Y. Moses, S. Savage, M. Uhlik, A. Lin, J. Du, Y. W. Qian, D. J. Zeckner, G. Tucker-Kellogg, J. Touchman, K. Patel, S. Mousses, M. Bittner, R. Schevi tz, M. H. Lai, K. L. Blanchard, and J. E. Thomas. 2007. A transforming mutation in th e pleckstrin homology domain of AKT1 in cancer. Nature 448: 439-44. 5. Cheng, J. Q., A. K. Godwin, A. Bellacosa, T. Taguchi, T. F. Franke, T. C. Hamilton, P. N. Tsichlis, and J. R. Testa. 1992. AKT2, a putative oncogene encoding a member of a subfamily of pr otein-serine/threoni ne kinases, is amplified in human ovarian carcinom as. Proc Natl Acad Sci U S A 89: 9267-71.

PAGE 150

6. Chien, Y., and M. A. White. 2003. RAL GTPases are linchpin modulators of human tumour-cell proliferation and survival. EMBO Rep 4: 800-6. 7. Cuatrecasas, M., A. Villanueva, X. Matias-Guiu, and J. Prat. 1997. K-ras mutations in mucinous ovarian tumors: a clinicopathologic a nd molecular study of 95 cases. Cancer 79: 1581-6. 8. Elad-Sfadia, G., R. Haklai, E. Ballan, H. J. Gabius, and Y. Kloog. 2002. Galectin-1 augments Ras activation and diverts Ras signals to Raf-1 at the expense of phosphoinositide 3-kinase. J Biol Chem 277: 37169-75. 9. Enomoto, T., C. M. Weghorst, M. Inoue, O. Tanizawa, and J. M. Rice. 1991. K-ras activation occurs frequently in mucinous adenocarcinomas and rarely in other common epithelial tumors of the human ovary. Am J Pathol 139: 777-85. 10. Falsetti, S. C., D. A. Wang, H. Peng, D. Carrico, A. D. Cox, C. J. Der, A. D. Hamilton, and S. M. Sebti. 2007. Geranylgeranyltransferas e I inhibitors target RalB to inhibit anchorage-dependent gr owth and induce apoptosis and RalA to inhibit anchorage-independent growth. Mol Cell Biol 27: 8003-14. 11. Fujita, M., T. Enomoto, M. Inoue, O. Ta nizawa, M. Ozaki, J. M. Rice, and T. Nomura. 1994. Alteration of the p53 tumor suppressor gene occurs independently of K-ras activation and more frequently in serous adenocarcinomas than in other common epithelial tumors of the human ovary. Jpn J Cancer Res 85: 1247-56. 12. Goodman, M. T., and H. L. Howe. 2003. Descriptive epidemiology of ovarian cancer in the United States, 1992-1997. Cancer 97: 2615-30.

PAGE 151

13. Hahn, W. C., C. M. Counter, A. S. Lundberg, R. L. Beijersbergen, M. W. Brooks, and R. A. Weinberg. 1999. Creation of human tumour cells with defined genetic elements. Nature 400: 464-8. 14. Hahn, W. C., and R. A. Weinberg. 2002. Rules for making human tumor cells. N Engl J Med 347: 1593-603. 15. Hamad, N. M., J. H. Elconin, A. E. Karnoub, W. Bai, J. N. Rich, R. T. Abraham, C. J. Der, and C. M. Counter. 2002. Distinct requirements for Ras oncogenesis in human versus mouse cells. Genes Dev 16: 2045-57. 16. Hanahan, D., and R. A. Weinberg. 2000. The hallmarks of cancer. Cell 100: 5770. 17. Hashiguchi, Y., H. Tsuda, K. Yamamoto, T. Inoue, O. Ishiko, and S. Ogita. 2001. Combined analysis of p53 and RB path ways in epithelial ovarian cancer. Hum Pathol 32: 988-96. 18. Hennessy, B. T., D. L. Smith, P. T. Ram, Y. Lu, and G. B. Mills. 2005. Exploiting the PI3K/AKT pathway fo r cancer drug discovery. Nat Rev Drug Discov 4: 988-1004. 19. Hinoi, T., S. Kishida, S. Koyama, M. Ikeda, Y. Matsuura, and A. Kikuchi. 1996. Post-translational modifications of Ra s and Ral are important for the action of Ral GDP dissociation stimulator. J Biol Chem 271: 19710-6. 20. Ichikawa, Y., M. Nishida, H. Suzuki, S. Yoshida, H. Tsunoda, T. Kubo, K. Uchida, and M. Miwa. 1994. Mutation of K-ras prot ooncogene is associated with histological subtypes in human mucinous ovarian tumors. Cancer Res 54: 335.

PAGE 152

21. Kendall, S. D., C. M. Linardic, S. J. Adam, and C. M. Counter. 2005. A network of genetic events sufficient to convert normal human cells to a tumorigenic state. Cancer Res 65: 9824-8. 22. Khosravi-Far, R., M. A. White, J. K. Westwick, P. A. Solski, M. Chrzanowska-Wodnicka, L. Van Aelst, M. H. Wigler, and C. J. Der. 1996. Oncogenic Ras activation of Raf/mitogen-activated protein kinase-independent pathways is sufficient to cause tumo rigenic transformation. Mol Cell Biol 16: 3923-33. 23. Khwaja, A., P. Rodriguez-Viciana, S. Wennstrom, P. H. Warne, and J. Downward. 1997. Matrix adhesion and Ras transformation both activate a phosphoinositide 3-OH kinase and protein kinase B/Akt cellular survival pathway. Embo J 16: 2783-93. 24. Kolasa, I. K., A. Rembiszewska, A. Janiec-Jankowska, A. DansonkaMieszkowska, A. M. Lewandowska, B. Konopka, and J. Kupryjanczyk. 2006. PTEN mutation, expression and LOH at its locus in ovarian carcinomas. Relation to TP53, K-RAS and BRCA1 mutations. Gynecol Oncol 103: 692-7. 25. Kupryjanczyk, J., A. D. Thor, R. Beauch amp, V. Merritt, S. M. Edgerton, D. A. Bell, and D. W. Yandell. 1993. p53 gene mutations and protein accumulation in human ovarian cancer. Proc Natl Acad Sci U S A 90: 4961-5. 26. Li, W., T. Zhu, and K. L. Guan. 2004. Transformation potential of Ras isoforms correlates with activ ation of phosphatidyli nositol 3-kinase but not ERK. J Biol Chem 279: 37398-406.

PAGE 153

27. Lim, K. H., A. T. Baines, J. J. Fiordalisi, M. Shipitsin, L. A. Feig, A. D. Cox, C. J. Der, and C. M. Counter. 2005. Activation of RalA is critical for Rasinduced tumorigenesis of human cells. Cancer Cell 7: 533-45. 28. Lim, K. H., K. O'Hayer, S. J. Adam, S. D. Kendall, P. M. Campbell, C. J. Der, and C. M. Counter. 2006. Divergent roles for RalA and RalB in malignant growth of human pancreatic carcinoma cells. Curr Biol 16: 2385-94. 29. Liu, J., G. Yang, J. A. Thompson-Lan za, A. Glassman, K. Hayes, A. Patterson, R. T. Marquez, N. Auersperg, Y. Yu, W. C. Hahn, G. B. Mills, and R. C. Bast, Jr. 2004. A genetically defined mo del for human ovarian cancer. Cancer Res 64: 1655-63. 30. Marks, J. R., A. M. Davidoff, B. J. Kern s, P. A. Humphrey, J. C. Pence, R. K. Dodge, D. L. Clarke-Pearson, J. D. Iglehart, R. C. Bast, Jr., and A. Berchuck. 1991. Overexpression and mutation of p53 in epithelial ovarian cancer. Cancer Res 51: 2979-84. 31. Mayr, D., A. Hirschmann, U. Lohrs, and J. Diebold. 2006. KRAS and BRAF mutations in ovarian tumors: a compre hensive study of invasive carcinomas, borderline tumors and extraova rian implants. Gynecol Oncol 103: 883-7. 32. McFall, A., A. Ulku, Q. T. Lambert, A. Kusa, K. Rogers-Graham, and C. J. Der. 2001. Oncogenic Ras blocks anoikis by activation of a novel effector pathway independent of phosphatidyli nositol 3-kinase. Mol Cell Biol 21: 5488-99. 33. Mok, S. C., D. A. Bell, R. C. Knapp, P. M. Fishbaugh, W. R. Welch, M. G. Muto, R. S. Berkowitz, and S. W. Tsao. 1993. Mutation of K-ras protooncogene

PAGE 154

in human ovarian epithelial tumors of borderline malignancy. Cancer Res 53: 1489-92. 34. Nakayama, K., N. Nakayama, R. J. Kurman L. Cope, G. Pohl, Y. Samuels, V. E. Velculescu, T. L. Wang, and M. Shih Ie. 2006. Sequence mutations and amplification of PIK3CA and AKT2 genes in purified ovarian serous neoplasms. Cancer Biol Ther 5: 779-85. 35. Oxford, G., C. R. Owens, B. J. Titus, T. L. Foreman, M. C. Herlevsen, S. C. Smith, and D. Theodorescu. 2005. RalA and RalB: antagonistic relatives in cancer cell migration. Cancer Res 65: 7111-20. 36. Peng, H., D. Carrico, V. Thai, M. Blaskovi ch, C. Bucher, E. E. Pusateri, S. M. Sebti, and A. D. Hamilton. 2006. Synthesis and evalua tion of potent, highlyselective, 3-aryl-piperazinone inhibitors of protein geranylgeranyltransferase-I. Org Biomol Chem 4: 1768-84. 37. Rangarajan, A., S. J. Hong, A. Gifford, and R. A. Weinberg. 2004. Speciesand cell type-specific requirements for cellular transformation. Cancer Cell 6: 17183. 38. Rapp, U. R., M. D. Goldsborough, G. E. Mark, T. I. Bonner, J. Groffen, F. H. Reynolds, Jr., and J. R. Stephenson. 1983. Structure and biological activity of v-raf, a unique oncogene transduced by a retrovirus. Proc Natl Acad Sci U S A 80: 4218-22. 39. Rodriguez-Viciana, P., P. H. Warne, A. Khwaja, B. M. Marte, D. Pappin, P. Das, M. D. Waterfield, A. Ridley, and J. Downward. 1997. Role of

PAGE 155

phosphoinositide 3-OH kinase in cell transf ormation and control of the actin cytoskeleton by Ras. Cell 89: 457-67. 40. Shields, J. M., K. Pruitt, A. McFall, A. Shaub, and C. J. Der. 2000. Understanding Ras: 'it ain't over 'til it's over'. Trends Cell Biol 10: 147-54. 41. Singer, G., R. Oldt, 3rd, Y. Cohen, B. G. Wang, D. Sidransky, R. J. Kurman, and M. Shih Ie. 2003. Mutations in BRAF and KRAS characterize the development of low-grade ovarian ser ous carcinoma. J Natl Cancer Inst 95: 484-6. 42. Soung, Y. H., J. W. Lee, S. W. Nam, J. Y. Lee, N. J. Yoo, and S. H. Lee. 2006. Mutational analysis of AKT1, AKT2 and AKT3 genes in common human carcinomas. Oncology 70: 285-9. 43. Ulku, A. S., R. Schafer, and C. J. Der. 2003. Essential role of Raf in Ras transformation and deregulation of matrix metalloproteinase expression in ovarian epithelial cells. Mol Cancer Res 1: 1077-88. 44. Varras, M. N., G. Sourvinos, E. Diakomanolis, E. Koumantakis, G. A. Flouris, J. Lekka-Katsouli, S. Michalas, and D. A. Spandidos. 1999. Detection and clinical correlations of ras gene mutations in human ovarian tumors. Oncology 56: 89-96. 45. Ward, Y., W. Wang, E. Woodhouse, I. Linnoila, L. Liotta, and K. Kelly. 2001. Signal pathways which promote inva sion and metastasis : critical and distinct contributions of extracellular signal-regulated kinase and Ral-specific guanine exchange factor pathways. Mol Cell Biol 21: 5958-69.

PAGE 156

46. Webb, C. P., L. Van Aelst, M. H. Wigler, and G. F. Woude. 1998. Signaling pathways in Ras-mediated tumorigenicity and metastasis. Proc Natl Acad Sci U S A 95: 8773-8. 47. Willner, J., K. Wurz, K. H. Allison, V. Galic, R. L. Garcia, B. A. Goff, and E. M. Swisher. 2007. Alternate molecular genetic pathways in ovarian carcinomas of common histological types. Hum Pathol 38: 607-13. 48. Yan, J., S. Roy, A. Apolloni, A. Lane, and J. F. Hancock. 1998. Ras isoforms vary in their ability to activate Raf-1 and phosphoinositide 3-kinase. J Biol Chem 273: 24052-6. 49. Yang, G., J. A. Thompson, B. Fang, and J. Liu. 2003. Silencing of H-ras gene expression by retrovirus-mediated siRNA decreases transformation efficiency and tumorgrowth in a model of hu man ovarian cancer. Oncogene 22: 5694-701. 50. Yang, L., H. C. Dan, M. Sun, Q. Liu, X. M. Sun, R. I. Feldman, A. D. Hamilton, M. Polokoff, S. V. Nicosia, M. Herlyn, S. M. Sebti, and J. Q. Cheng. 2004. Akt/protein kinase B signaling inhibitor-2, a selective small molecule inhibitor of Ak t signaling with antitumor activity in cancer cells overexpressing Akt. Cancer Res 64: 4394-9. 51. Yin, J., C. Pollock, K. Tracy, M. Chock, P. Martin, M. Oberst, and K. Kelly. 2007. Activation of the RalGEF/Ral pathway promotes prostate cancer metastasis to bone. Mol Cell Biol 27: 7538-50. 52. Yuan, Z. Q., M. Sun, R. I. Feldman, G. Wang, X. Ma, C. Jiang, D. Coppola, S. V. Nicosia, and J. Q. Cheng. 2000. Frequent activation of AKT2 and

PAGE 157

induction of apoptosis by inhibiti on of phosphoinositide-3-OH kinase/Akt pathway in human ovarian cancer. Oncogene 19: 2324-30. 53. Zhou, D. J., N. Gonzalez-Cadavid, H. Ahuja, H. Battifora, G. E. Moore, and M. J. Cline. 1988. A unique pattern of proto-oncogene abnormalities in ovarian adenocarcinomas. Cancer 62: 1573-6.

PAGE 158

Chapter 4 Discovery of a Proposed Database of Ral In teracting Proteins: RACK1 Binds Ral and is Required for Hand K-Ras Mediated Transformation By Samuel C. Falsetti 1,2, Sad M. Sebti 1,2,* 1Drug Discovery Program, The H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL 2Departments of Oncological Sciences and Mo lecular Medicine, The University of South Florida, Tampa, FL *Corresponding Author: 12902 Magnolia Drive, Tampa, FL 33612; Tel (813) 745-6734; Fax (813) 745-6748; email:said. sebti@mofitt.org

PAGE 159

Abstract RalA and B proteins mediate Ras malignant transformation in multiple human epithelial cell types and have been determined to be independently transforming in a tissue specific manner. However, very little is known about the protein-binding partners that govern Ral A and B biological activities. He re, we have used a proteomics approach to describe a proposed database of novel Ral interacting proteins. De spite a greater than 85% sequence similarity RalA and RalB inte ract primarily with different proteins. However, the proposed database also cont ains common interacting partners. For example, we have identified and validate d RACK1/GBLP as a novel RalA and RalB interacting protein. Specifically, we have determined that RACK1 interacts with ectopically expressed RalA and RalB as we ll as with endogenous RalA. Furthermore, depletion of RACK1 by siRNA phenocopies the effects of RalA and RalB depletion in reducing the ability of oncogenic H-Ras12V and KRas12V to transform human ovarian epithelial cells. Thes e results both expand the number of proposed Ral A and B protein binding partners. They also led us to identi fy a specific protein interaction with RACK1 and identify RACK1 expression as a central requirement for both H-Ras12V and KRas12V transformation of human ovarian surface epithelial cells.

PAGE 160

Introduction One of the best studied mechanisms of oncogenesis in human cancer is mutational activation of members of the Ras GTPase ge ne family. Mutations in Ras gene family members (H-, K-, and NRas) occur in approximately 30% of human cancers (6). Mutation of Ras isoforms has been associat ed with progression of numerous types of human epithelial cancer such as pancreatic, lung, ovarian and breast (6, 16, 19). Recently, we (Falsetti and Sebti manuscript in subm ission) and others (18, 37) have used genetically defined models of human epithe lial transformation to describe the tissue specific mechanisms of Ras transformati on (18, 37). In these models, sequential inactivation of the tumor s uppressors p53 and retinoblastoma (Rb) protein by ectopic expression of SV40-T-antigen (SV40-T), c oupled with stable expression of human telomerase catalytic subunit (hTERT) in human epithelial cells recapitulates the critical events in the development of human ne oplastic disease and creates a minimally transformed model for the study of oncogenic signaling pathways (15-17, 19). Addition of constitutively activated Hor KRas12V results in a robust tr ansformation response and mimics many of the pathological events of tumor progression (15-17, 19, 28). Multiple studies have used a combination of pharmacological and genetic approaches to determine the tissue specific requirements for Ras transf ormation. Interestingly, all of these studies indicate that the Ras/ Ral guanine nucleotide dissociation stimulator (RalGDS)/Ral signaling pathway is required for human epithelial cell transformation by oncogenic Ras. Both Counter and Weinbergs groups have separately reported th at in sequentially transformed human embryonic kidney (HEK-HT) cells RalGDS/Ral sign aling is required for Ras transformation (18, 37). Counters gr oup has specifically defined a RalGDS/RalA

PAGE 161

signaling axis which is independently transf orming in this system while Weinbergs group has determined that Ras requires both RalGDS and phosphotidyl-inositol 3,4,5triphosphate (PI3K) signaling for transformation in a similar, independently derived, human embryonic kidney cell line. These propos ed mechanisms of Ras transformation are not common to all human epithelial cell line s; for example, in a genetically defined model of human mammary tumorigenesis, the human mammary epithelial cell line (HMEC), Ras transformation re quires activation of at least three downstream effector pathways: RalGEF-Ral, PI3K-Akt and RafMek (37). We have previously used a genetically defined model of human ovarian cancer, human ovarian surface epithelial cells (T80), to determine that Hand K-Ras12V transformation of human ovarian cells requires RalGEF-Ral and PI3K-Akt, but not Raf-Mek, signaling. In further confirmation of a role for the Ral small GTPases in oncogenesis both RalA and RalB have been identified as critical mediators of multiple tumorigenic processes, including metastasis, invasion, anchorage-independent gr owth, survival and cell motility (8, 11, 34, 42, 43, 46). Specifically, depletion of RalA by siRNA has been shown to inhibit anchorage-independent proliferation of multip le human cancer cell lines, such as the cervical and prostate cancer cell lines HeLa and SW680 (8) as well as multiple human pancreatic cancer cell lines (27). Also, RalA has also been shown to be required for anchorage independent growth of sequentially Ras transformed human ovarian (Falsetti and Sebti, manuscript in submission) and kidne y cells (26). Similarly, stable depletion of RalA, but not RalB, by shRNA has been shown to inhibit tumor formation and metastasis of multiple human pancreatic cancer cell lines in athymic nude mice (27). Further confirming a role for RalA in metastasis, stable overexpression of constitutively activated

PAGE 162

RalA has been shown to promote both st andard and experimental metastasis in vivo (41, 42) and in human prostate cancer cells stable expression of activated RalA promotes bone, but not brain, metastasis (47). Even in non-epithelial cancer s there is emerging evidence that RalA is involved in tumo rigenicity; for example, in human HT1080 fibrosarcoma cells stable overexpression of activated RalA promotes anchorageindependent growth (46). In agreement with these findings two cellu lar processes thought to promote metastasis and anchorage independent growth, cell motility and invasion, are inhibited by siRNA mediated de pletion of RalA in multiple human renal cancer cell lines (34). While RalA has been well validated to play an essential role in anchorage independent processes in multiple tissue types the role of RalB has been determined to be far more tissue specific. For example, while de pletion of either RalA or RalB inhibits anchorage independent growth in human ovarian epithelial cells transformed by either Hor K-Ras12V (Falsetti et al, manuscript in submission) only depletion of RalA but not RalB inhibits H-Ras12V transformation of HEK-HT cells ( 26). We, as well as others have previously determined a role for RalB, but not RalA in the survival of multiple human cancer cell lines (8, 11). Specifically, Whites group has found that depletion of RalB by siRNA induces apoptosis in both HeLa and SW680 human cancer cell lines (8). Similarly, we have previously determined us ing a chemical biology approach that a RalB, but not RalA, survival pathway unde rlies the apoptotic response to geranylgeranyltransferase I inhibitors ( GGTIs) (11). Specifically, by ectopically expressing a GGTI-resistant RalB mutant in human pancreatic carcinoma cells (MiaPaCa2) we have shown that GGTI-inducti on of apoptosis requires inhibition of

PAGE 163

RalB, but not RalA, processing (11). Taken as a whole, these studies strongly suggest a role for RalB tumor survival in a wide arra y of tissue types as well as a tissue specific role in anchorage independent growth. Despite the wealth of information rega rding the role of Ral small GTPases in transformation very little is known about th e pathways through which Ral exerts these effects. For example, while RalA and RalB bear a high similarity to Ras over 20 Ras effectors (6, 39),but only 6 Ral effectors have been described to date. RalA and/or RalB have been previously shown to interact with phospholipase D1 (PLD1), the exocyst components Sec5/ Exo84, Filamin A, ZO-1 N-terminally associated binding protein (ZONAB) and Ral binding protein-1 (RalBP1/RLIP ) (12). Precise roles for these proteins in the various transformation specific processes that are regulated by RalA and RalB remain poorly defined. Further underscoring th e need for an understanding of Ral protein interactions, while RalA and RalB are commonly thought of as Ras effector proteins both RalA and RalB are found in the hyperactivated state inde pendently of Ras in human pancreatic tumors (26, 27). Thus novel mean s of Ral activation and inactivation may constitute an important and undescribed m echanism of transformation. However, no Ral GTPase activating proteins (RalGAPs), which would negatively regu late RalA and RalB, have been described to date. Also, RalA and RalB are known to regulate diverse physiological processes such as signaling via STAT3, NFB, JNK and AFX (12). However, the intermediate proteins in thes e signal transduction pathways through which RalA and RalB exert these functions remain unknown. In an effort to more fully understand the protein interactions that govern the biological activities of RalA and RalB we ha ve used proteomic analysis to describe a

PAGE 164

proposed database of Ral interacting proteins We have uncovered several potential RalA and RalB interacting partners and used a sy stems biology approach to analyze the themes that emerge from this proposed database. We have further characterized one of these proteins, receptor for activa ted C-kinase-1 (RACK1, also known as Guanine nucleotide binding protein-1 [GBLP1]), as a novel RalA and RalB interacting protein that is required for both oncogenic Hand KRas mediated transformation of human ovarian surface epithelial (T80) cells. In addition to uncoveri ng a critical role fo r RACK1 expression in both Hand K-Ras mediated transformation this study also constitutes the first largescale analysis aimed at uncovering the biological differences and similarities between the Ral isoforms and provides an important da tabase of potential pr oteins through which RalA and RalB may mediate their biological functions.

PAGE 165

Materials and Methods Cells and culture Human ovarian surface epithelial ce lls T80 and T80H were cultured as described previously (28). The T80Kstably expressing cell lines were maintained in similar culture conditions as the T80 and T80H cells. Human HEK-293T cells were a kind gift from Dr. Gary Reut her and were maintained in Dulbeccos modified minimum essential media (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin/ streptomycin at 10% CO2 and 37C. Immunoprecipitation of FLAG-tagged protein FLAG-Ral proteins were isolated using FLAG-agarose beads according to the manuf acturers recommendations (Sigma-Aldrich, St. Louis, MO). Briefly, transfected cells were lysed using Cell Lytic-M lysis byffer (Sigma-Aldrich) containing protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO) and lysates were incubated ove rnight with FLAG-agarose beads at 4C while rocking; after incubation the beads were washed four times with an excess of lysis buffer. Subsequently, samples were boiled at 100C for 10 minutes in 2X Laemmli sample buffer to elute bound proteins. Immunoprecipitation of RalA protein Endogenous RalA was isolated using anti-human RalA mouse monoclonal antibody RalA (61022, BD Biosciences Pharmingen, Franklin Lakes, NJ). Briefly, T80, T80H or T80K cells were lysed using Cell Lytic-M lysis byffer (Sigma-Aldrich, St. Louis, MO) containing pr otease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO) and lysates were incubate d overnight with anti -RalA antibody at 4C while rocking; after incuba tion protein A anti-IgG agarose beads were added and incubated for 2 hours at 4C then washed f our times with an excess of lysis buffer.

PAGE 166

Following this, samples were boiled at 100C for 10 minutes in 2X Laemmli sample buffer to elute RalA and RalA binding proteins. Identification of Ralinteracting proteins Following immunoprecipitation and SDSPAGE, gel bands were visualized using GelC ode Blue (Promega, Madison, WI), excised and washed once with water and twice w ith 50 mM ammonium bicarbonate in 50% aqueous methanol. Samples were digested overnight with modifi ed sequencing grade trypsin (Promega, Madison, WI). Peptides were extracted from the gel slices and concentrated under vacuum centrifugation. A nanoflow liquid chromatograph (LC Packings/Dionex, Sunnyvale, CA) coupled to an electrospray ion trap mass spectrometer (LTQ Orbitrap, Thermo, San Jose, CA) was used for tandem mass spectrometry peptide sequencing experiments. Peptides were sepa rated with a C18 reve rse phase column (LC Packings C18Pepmap, 75 um ID x 15 cm) using a 40 minute gradient from 5% B to 50% B (A: 2% acetonitrile/0.1% formic acid; B: 90% acetonitrile/0.1% formic acid). Five tandem mass spectra were acquired for each MS1 scan (spray voltage 2.5 kV, 30% normalized collision energy, scanning m/z 450-1,600). Sequences were assigned using Sequest (Thermo) and Mascot (www.matrixscience.com) databa se searches against NCBI or SwissProt protein entries of the appropriate species. Oxidized methionine, deamidation, and carbamidomethyl cysteine were selected as variab le modifications, and as many as 2 missed cleavages were allowe d. Assignments were manually verified by inspection of the tandem mass spectra a nd coalesced into Scaffold reports ( www.proteomesoftware.com ). RalA and RalB interacting proteins were identified from this initial list by subtractive analysis of the vector transfected cells. Specifically, to qualify as a proposed FLAG-Ral interacting part ner the protein had to fulfill one of two

PAGE 167

criteria: the protein had to be identified by at least 2 pe ptides at greater than 94% probability in at least one of two experiments and could not be present in any of the three vector-transfected controls; or the protein had to be identified by at least 2 peptides at greater than 94% probab ility in both experiments and could not be present in more than one of three vector transfected controls. Small interfering RNA (siRNA) sequences SiRNA sequences targeting RalA RalA(5'GACAGGUUUCUGUAGAAGAdTdT3') and RalB (5'GGUGAUCAUGGUUGGCAGCdTdT-3) were purchased from Dharmacon using previously described sequences(8). Pre-desi gned chemically synthesized siRNA targeting RACK1 (Cell Signaling Technology Inc., Da nvers, MA) were used according to manufacturers recommendation. Small interfering RNA (siRNA) transfection procedure Cells were grown to 50% confluence and transfected with Oligofectamine transf ection reagent (Invitrogen, Carlsbad, CA) according to the manufactur ers instructions. Briefly, 5 l of Oligofectamine reagent was suspended per 1mL of OPT I-MEM medium (Invitrogen, Carlsbad, CA) and allowed to equilibrate at 24-27C. A 100 nM final concentration of siRNA was suspended and complexed with the liposomes. Cells were briefly washed with OPTI-MEM and the transf ection mix was plated on top of the cells and incubated at 37C. Following an 8 h incubation, 2mL of complete growth medium, without penicillin-streptomycin, was added and th e cells were further incubated at 37C overnight. The media was replaced after 24 h with complete growth media for 24 hours. Experiments were performed in triplicate.

PAGE 168

Western blotting Cells were harvested, and lysed in HEPES lysis buffer as described previously (11). Proteins were then resolved by 11.5% SDS-PAGE gel and immunoblotted with antibodies against phosphorylated-Ser473 Akt1/2 (9217, Cell Signaling Technologies Inc., Danvers, MA), phosphorylated-Thr202/Tyr 204 Erk1/2 (9101, Cell Signaling Technologies Inc., Danve rs, MA), Akt1/2 (N-19, Santa Cruz Biotechnology, Santa Cruz, CA), Erk1/2 (p44/p42 MAP Kinase, 9102, Cell Signaling Technologies Inc., Danvers, MA), RalA (61022, BD Biosciences Pharmingen, Franklin Lakes, NJ), H-Ras (C-20, Santa Cruz Biotechnology, Santa Cruz, CA), K-Ras (F234, Santa Cruz Biotechnology, Sant a Cruz, CA), RalB (04037, Mill ipore, Billerika, MA), actin (AC15, Sigma-Aldrich, St. Louis, MO ) and anti-FLAG M2 monoclonal antibody (Sigma-Aldrich, St. Louis, MO). The ECL blotting system (NEN Life Science Products, Boston, MA) was used for detection of pos itive antibody reactions. Experiments were performed in triplicate. Soft agar clonogenicity assay For soft agar growth assays, the cell lines were seeded at a cell density of 5000/well in tr iplicate in 12-well culture dishes in 0.3% agar over a 0.6% bottom agar layer as previously described(11). Cultures were fed once weekly until colonies grew to a suitable size for observ ation (approximately 14 days). Colonies were photographed after overnight in cubation with 1 mg/ml MTT in cell growth media. The colony number was visually determined and quantified. Experiments were performed in triplicate.

PAGE 169

Results Use of an ectopically expressed FLAG-tagge d Ral-Tandem MS system to isolate and identify novel Ral binding partners: Determination of a potential database of Ral interacting proteins Although RalA and RalB are required for mutant Ras malignant transformation in a variety of human cancers, the mechanism by which they mediate Ras transformation is not known. This prompted us to identify a database of potential RalA and RalB interacting proteins with the ultimate goal of discovering effectors of Ral A and Ral B that mediate Ras transformation. To this e nd we ectopically expr essed N-terminal FLAGtagged RalA72L and RalB72L in human embryonic kidney cells (HEK-293T) then isolated potential Ral-interacting proteins th rough subsequent FLAG immunoprecipitation, separation of protein complexes by SDS-PAGE and protein identification in the various gel segments by tandem MS and subsequent SCAFFOLD analysis as described under Material and Methods. Figures 17A and 17B s how that transient tr ansfection resulted in high levels of expression of FLAG-RalA and RalB, respectively, as demonstrated in both whole cell lysate as well as FLAG immunopreci pitates as detected by both Ral antibody western blots and GelCode Blue stain

PAGE 170

Figure 17. Expression and imm unoprecipitation of FLAG-Ral72LFigure 17.Expression and immunoprecipitationof FLAG-Ral72L. Retroviral pBABEplasmid s containgeither RalA72L (A) or RalB72L(B) along with an empty vector sequence were transiently transfected into HEK-293T for 48 hours. Cells were collected, lysed, and immunoprocipitatedfollowing which protein was analyzed for western blot analysis or GelCodeBlue. V RalA V RalA FLAG IP WCL V RalA GelCodeBlue V RalB V RalB FLAG IP WCL V RalB GelCodeBlue A. B. FLAG IP FLAG IP Blot: FLAG-RalA FLAGRalA Blot: RalA Stain: GelCodeBlue RalA GelCode Blue

PAGE 171

Using a proteomic approach we have identified 68 RalA and 28 RalB potential interacting proteins. Of these the known Ral-in teracting protein Filamin A, as well as the closely related Filamin C, were identified as potential RalA and RalB interacting partners in this system (see tables 2 and 3, follo wing pages). Additiona lly, other predicted members of known Ral interacting complexes, such as ZO-1 and SC22B, were also identified as potential Ral-interacting prot eins (see tables 2 and 3, following pages). These results provide evidence in support of the overall feas ibility of our approach in identifying both proposed direct, as well as indirect, Ral-inte racting partners.

PAGE 172

3 GTPB1 GTP-binding protein1 6 7 GBLP guanine-nucleotide binding protein 2 FLNC filaminC 5 FLNA filaminA 2 FEN1 flap endonuclease1 2 FA98B protein FAM98B 2 EF1B elongation factor 1B 2 CT116 uncharacterized protein 3 CSK1 casein kinase II subunit-1 4 CISY citrate sythase, mitochondrial precursor 3 CAPZB F-actincapping protein 2 ARP10 actinrelated protein-10 2 AP3M1 AP-3 complex subunit-Mu-1 2 ACTZ alpha-centractin 3 ACOT9 acylcoenzyme A-thioesterase-9 2 6PGD 6-phosphogluconate dehydrogenase 3 1433G 14-3-3 Gamma Experiment 2 Experiment 1 Peptides Matched (>94%) Protein ID Gene name RalA SCAFFOLD Results Table 2. Proposed RalA interacting proteins (pt1)3 VIME vimentin 2 VAPB vessicle-associated membrane 2 SPIN3 spindlin-3 1 3 SPIN1 spindlin-1 2 SC22B vesicle-trafficking protein 3 RTN4 reticulon-4 2 7 RSSA 40S ribosmoalprotein SA 6 2 RS6 40S ribosomal protein S6 3 RS26 60S ribosmalprotein S26 2 RS20 40S ribosomal S20 2 RS17 40S ribosomal S17 2 1 RS16 40S ribosomal S16 2 RL7A 60S ribosomal L7A

PAGE 173

5 PHB prohibitin 4 PGK1 phosphoglyceratekinase-1 11 PA2G4 proliferation asso ciated protein-2G4 2 NPS3A protein NipSnap3 2 NELFE negative elongation factor 3 NELFE negative elongation factor 2 MPCP phosphate carrier protein 2 LUC7L putative RNA-binding protein-7L 6 LRC47 leucinerich repeat protein-47 4 2 KCRU creatinekinase mitochondrial precursor 6 2 KCIE casein kinase I isoform epsilon 2 K2C8 keratin type 2 cytoskeletal-8 7 K1C18 keratin type 1 cytoskeletal-18 10 ILF2 interleukin enhancer factor 2 3 IF4H eukaryotic translation initiation factor 3 IDHP isocitratedehydrogenase(NADP) 4 HNRPG heterogeneous nuclear protein-G 3 2 HNRPF heterogeneous nuclear protein-F Experiment 2 Experiment 1 Peptides Matched (>94%) Protein ID Gene name RalA SCAFFOLD Results Table 2. Proposed RalA interacting proteins (pt2)2 RL17 60S ribosomal protein-17 5 RL10A 60S ribosomal protein-10A 3 RCN1 reticulocalbin-1 2 RAN ran GTP binding protein 14 10 RALA ralA 5 PUR6 multifunctional protein ADE2 2 PSMD8 26S proteasomeregulatory subunit-8 2 PSD13 26S proteasomenon-ATPase 7 PSD11 26S proteasomeregulatory subunit 2 PSA1 proteasomesubunit alpha 2 PP1B serine-threoninephosphatase-1B

PAGE 174

2 2 PRDX1 peroxiredoxin-1 3 PCBP2 poly(rC) binding protein-2 3 P5CR2 pyrroline-5-caraboxylate 2 ODPA pyruvatedehydrogenaseE1 2 NUPL2 nucleoporin-like 2 4 KCRU creatinekinase 3 K1C16 keratin, Type I cytoskeletal-16 12 IF32 eukaryotic translation initiation 2 2 HNRPK heterogeneous nuclear riboprotein 4 GTPB1 GTP binding protein 5 4 GBLP guanine-nucleotide binding 3 FLNC filaminC 3 FLNA filaminA 2 ENAH protein enabled homologue 6 DDX5 probable ATP-dependent RNA 2 ARP2 actin-like protein 2 4 AHNK neuroblastdifferentiation factor 5 1433E 14-3-3 epsilon Experiment 2 Experiment 1 Peptides Matched (>94%) Protein ID Gene name RalB SCAFFOLD Results Table 3. Proposed RalB interacting proteins6 ZO1 tight junction protein 3 XAB1 XPA-binding protein 2 VIME vimentin 2 TBB2C tubulinbeta-2 chain 3 SFRS7 splicing factor-S7 7 2 RS6 40S ribosomal protein-S6 5 RS26 40S ribosomal protein-S26 1 3 ROA1 heterogeneous nuclear 2 RL18 60S ribosomal protein-18 6 4 RALB RalB

PAGE 175

RACK1/GBLP interacts with ectopically ex pressed RalA and RalB in HEKT cells One interesting protein our analysis rev ealed to be a potential RalA and RalB interacting proteins was guanine nucleot ide-binding protein GBLP, also known as receptor for activated C-kinase1 (RACK1), which is a scaffolding protein able to interact simultaneously with several signaling molecule s (32). In order to confirm and validate that RalA and RalB interact with RA CK1 we ectopically expressed FLAG-tagged RalA72L and RalB72L in HEK-293T cells and used FLAGaffinity immunoprecipitation to isolate Ral protein complexes. Both RalA72L and RalB72L were expressed at high levels in HEK-293T cells (figs. 18A and 18B, followi ng page). In confirmation of our MS/MS results both ectopically expressed RalA72L and RalB72L interacted with endogenous RACK1 (figs. 18A and 18B). These results demonstrate that a RACK1/Ral complex is formed within HEK-293T cells. While these results are strongly suggestive that Ral proteins interact with a RA CK1 containing complex, if not directly with RACK1, we wanted to further validate this interaction using endogenous RalA.

PAGE 176

Figure 18. RACK1/ GBLP interacts with ectopically expressed Ral72L Ectopic RalA V RalA WCL FLAG RACK1 Ectopic RalB V RalB WCL FLAG RACK1 FLAG IP FLAG IP V RalA V RalB FLAG RACK1 FLAG RACK1 A. B. Figure 18. RACK1 interacts with ectopically expressed Ral72L. Retroviral pBABEplasmid s containgeither RalA72L (A) or RalB72L(B) along with an empty vector sequence were transiently transfectedinto HEK-293T for 48 hours. Cells were collected, lysed, and immunoprocipitatedfollowing which protein was analyzed for western blot analysis or GelCodeBlue.

PAGE 177

RACK1/GBLP interacts with endogenously expressed RalA in ovarian cancer cells We next evaluated whether RACK1 binds RalA in human cancer cells where Ral proteins are critical to Ras transformation. We have previously reported (Falsetti and Sebti, manuscript in submission) that RalA and RalB expression is required for both Hand K-Ras12V transformation of human ovarian surface epithelial cells, which are sequentially immortalized by stable expression of human telomerase catalytic subunit (hTERT) and simian virus-40 large T-antigen (T80 cells). Therefore, we used T80 cells transformed with H-Ras12V (T80H) and K-Ra s12V (T80K) to evaluate the binding of endogenous RACK1 to endogenous RalA protei n and to subsequently determine the importance of these interactions to Ras-medi ated transformation in this human ovarian cancer model. Figure 3 shows that in bot h the non-transformed and Ras-transformed T80 isogenic cell lines RalA st rongly interacted with endogenous RACK1 but not with beta-actin (Figure 19, following page). Im portantly, expression levels of RalA and RACK1 did not vary between cell lines. Also, since RalA is GTP-bound in both T80H and T80K (Falsetti and Sebti, manuscript in submission) we expected that if RalA binding to RACK1 was GTP-dependent there wo uld be increased association in the Ras transformed cell lines. However, RalA binding to RACK1 appeared to be GTPindependent and only varied accord ing to the levels of RalA isolated in each experiment.

PAGE 178

Figure 19. RACK1/GBLP interacts with endogenous RalA protein T80 T80H T80K WCL RalA Actin RACK1 Endogenous RalA RalA IP T80 T80H T80K RalA Actin RACK1 Figure 19. RACK1 interacts with endogenously expressed RalA. Protein lysatewas collected from T80, T80H and T80K cel ls and used for immunoprecipitationof endogenous RalA using anti-RalA monoclonal antibody.

PAGE 179

RACK1 depletion phenocopies the effects of Ral depletion in Ras transformed ovarian epithelial cells Since we have previously characterized both RalA and RalB expression as being required for Hand KRas12V-mediated transformation of T80 cells and since RalA and RalB were found to interact with RACK1 we endeavored to further define a role for RACK1 in Ras-mediated transformation of human epithelial cells. One widely used method of determining similar roles for proteins is to determine if the effects of depletion of one gene will phenocopy the effects of the other. To that end, we used siRNA to specifically and potently deplete T80H and T80K cells of RalA, RalB and RACK1 and then examined the effects of single deplet ion of each gene on Ras transformation. As demonstrated in figures 20A and 20B (followi ng page) we were able to specifically and potently deplete the cells of RalA, RalB or RACK1 through transient transfection with siRNA targeting regions unique to each ge ne. Furthermore, depletion of RACK1 phenocopies the effects of Ral depletion on Ras transformation in both T80H and T80K cells. Specifically, RalA inhibited T80H clonogenicity in soft agar by 54.9% (p value <0.001) and T80K clonogenicity by 51.7% (p value <0.001). Depletion of RalB by siRNA inhibited T80H clonogenicity by 53.8% (p value <0.001) and T80K clonogenicity by 65.2% (p value <0.001). In support of a role for RACK1 in the Ral-dependent Rasmediated transformation of T80 cells RA CK1 depletion by siRNA inhibited T80H clonogenicity by 59.9% (p value <0.001) a nd T80K clonogenicity by 34.3% (p value <0.01).

PAGE 180

Figure 20. RACK1/ GBLP depletion phenocopies the effects of Ral depletion in Ras-transformed ovarian epithelial cells 0.0 25.0 50.0 75.0 100.0 125.0 = T80-H = T80-K% Control C. N.C.RalARalBRACK1 *** *** *** ****** ** Figure 20 Ras transformation of T80 cells requires expression ofRalA/B and Akt1/2 but not Raf1. Small interfering RNA (s iRNA) specific to RalA, RalB, and RACK1 or a negative contro l siRNA were transfected for48 hours into T80, T80H and T80K cells. Cells were collected, counted using trypan blue dye exclusion assay and collected for western blot analysis(A) and an equal number of cells were plated for so ft agar assay (B) and scored manually at the end of two weeks; results are re presentative of thr ee experiments each performed in triplicate. Actin RalB RalA RACK1 N.C. RalA RalB RACK1 T80H A. B.N.C. RalA RalB RACK1T80K Actin RalB RalA RACK1 [ siRNA [ siRNA

PAGE 181

Discussion While Ral proteins are well known to be important components of Ras-mediated transformation of multiple hum an epithelial cell types very little is known about the protein-binding partners that govern Ral biologic al activity. Here, we have used a proteomics approach to describe, for the first time, a proposed database of novel Ralinteracting proteins. Despite a greater than 85% sequence identity RalA and RalB interact primarily with different proteins across a wide range of gene ontology sub-groups. Furthermore, we have identified and valid ated RACK1/ GBLP as a novel RalA, and RalB, interacting protein. Specifically, we have determined that RACK1 interacts with ectopically expressed RalA and RalB as we ll as with endogenous RalA. Furthermore, depletion of RACK1 by siRNA phenocopies the effects of RalA and RalB depletion in reducing the ability of multiple Ras isoforms to transform human ovarian epithelial cells. These results both dramatic ally expand the number of proposed Ral protein binding partners and identify a specific protein, RACK 1, which is at least partially required for Ras transformative activity. While use of proteomics technology has allo wed us to rapidly identify a multitude of potential individual protein interactions with both RalA and RalB like all systems biology analyses confirmation and validation of these results will require extensive efforts to individually characte rize each of these potential pa rtners. Nevertheless, we felt a further analysis of these potential interac ting partners by sorting these proteins using SWISS-PROT GO annotation would be helpfu l in revealing potentially emergent biological themes in RalA a nd RalB biological function. In particular, these analyses revealed cytoskeletal regulati on as a potential biological pr ocess that could be governed

PAGE 182

by both RalA and RalB. Importantly, both RalA and RalB potentially interacted with multiple critical regulatory proteins with well -described roles in the regulation of this process (9, 13, 24, 40). Our proteomics an alysis suggests that both RalA and RalB interact with both Filamin A (FLNA) and Filamin C (FLNC), known regulators of the cytoskeleton. Importantly, RalA has been previo usly shown to interact with Filamin A in a GTP-dependent manner (33). The filamin fa mily consists of three genes: Filamin-A, B, and C; of these Filamin-A is the most we ll understood. (13). Filamins play a critical role in cross-linking the cortical actin cy toskeleton into orthogonal networks and link these three dimensional actin structures to the inner leaflet of the cellular membrane (40). Furthermore, Filamins modulate the response of cells to their extra-cellular environment by regulating changes in shape and motility (38). Interestingly, our analyses revealed vimentin (VIME, see tables 2 and 3), an impor tant intermediate filament in mesenchymal cells which is commonly overexpressed in can cer cells as part of the epithelial to mesenchymal transition (22), to be a potential RalA and RalB interacting protein. Another potential RalB associat ed protein. tubulin beta-2 chain (TBB2C, see table 3), is known to form a supramolecular complex with vimentin (9). Certainly, a further analysis of the roles of vimentin and filamin in Raldriven actin cytoskeletal changes would be warranted given these results. We validated RACK1 as interacting with both ectopically expressed RalA72L and RalB72L as well as with endogenous RalA in an activation state independent manner. Interestingly, there exists a paradigm for small GTPase interaction with RACK1. Rac2, a small GTPase highly homologous to RalA a nd RalB, has been previously shown to interact indirectly with RACK 1 and this interaction has been shown to regulate natural

PAGE 183

killer cell adhesion (30). This is particularly interesting given that Ral negatively regulates Rac proteins via the Rac-GAP protein RalBP1 (23, 35). Ral and Rac proteins regulate similar processes such as cell motility, transformation, invasion, cell-cell adhesion and actin reorganizati on (4, 5, 12, 44). Perhaps, the ability of RalA and RalB to subvert Rac processes occurs via both negative regulation of Rac GTP levels and through subsequent association with Rac down stream effectors such as RACK1. We have also demonstrated, using siR NA, that RACK1 expression was at least partially required for both Hand K-Ras12V-mediated transformation of T80 cells and that the effects depletion of RACK1 was similar to the effects of RalA or RalB depletion. This is consistent with previous work show ing that a putative dominant negative isoform of GBLP (or RACK1) has been previously shown to reverse K-Ras12V transformation in NIH-3T3 cells (2) and to restor e contact inhibition and stre ss fiber formation (3). In contrast, paradoxically, overe xpression of wild-type RACK 1 in transformed NIH-3T3 cells reduces both anchorage-independent and dependent proliferation (2). Nevertheless, other studies support the impli cation of RACK1 in oncogenesi s. For example, studies have used antisense siRNA to inhibit RACK1 expression in NIH3T3 cells and have demonstrated a requirement for RACK1 in ce ll spreading and cell proliferation (20). Furthermore, RACK1 expression is drama tically increased in both human carcinomas and during human ovarian morphogenesis (1). In addition to playing apparently contra dictory roles in anchorage independent proliferation, RACK1 is known to play a role in a variety of cell pr ocess such as coupling of signal transduction to contro l of protein transcription and translation (32), recruitment of ribosomes to local sites of translation (s uch as focal adhesions) (32), regulation of cell

PAGE 184

spreading (3), cytoskeletal organization (20), and cell cycle pr ogression (7). RACK1 associates with a variety of signal transduc tion molecules that may account for some of these biological activities, such as Protein Kinase C(PKC), -integrin and c-Src (29). Most interestingly, multiple mechan isms exist through which RACK1 and RACK1interacting proteins, such as PKC isoforms, may regulate Ras. RACK1 has been previously shown to associate with p120RasGAP (24), a we ll-characterized negative regulator of Ras activity (31) though the exact significance of the RACK1 interaction in RasGAP biological function has not yet been established. RACK1 is a well-characterized regulator of multiple PKC isoforms (29) which are known to regulate Ras guanyl nucleotide releasing proteins (RasGRPs), pos itive regulators of Ra s signaling (14). PKC phosphorylation of K-Ras in the hypervariable C-terminal region results in accumulation of K-Ras in the outer mitochondrial me mbrane and promotes apoptosis (36). Additionally, phorbol esters, which activate PK C isoforms, are known to cooperate with Ras in transformation and deregulation of diacylglycerol (DAG) has been observed in Ras-transformed cells, indicating that Ras and PKC might cooperate during transformation (10, 21, 25, 45). Certainly, our results support a role for RACK1 expression in regulating Ras transformation of human ovarian surface epithelial cells and suggest that further study of both the Ral/RACK1 associatio n and the role of RACK1 in human ovarian epithelial cell transformation will prove to be interesting avenues of future research.

PAGE 185

Acknowledgements The authors would like to thank John Kooman for his expert advice and well considered opinions in the analysis of proteomic data for this manuscript. The authors would also like to acknowledge the support from the National Cooperative Drug Discovery Group funding as well as the Moffitt Proteomics F acility, which is supported by the National Functional Genomics Center grant from the U. S. Department of Defense and the Moffitt's Cancer Center Support Grant from the National Cancer Institute.

PAGE 186

References 1. Berns, H., R. Humar, B. Hengerer, F. N. Kiefer, and E. J. Battegay. 2000. RACK1 is up-regulated in angiogenesi s and human carcinomas. Faseb J 14:254958. 2. Bjorndal, B., L. M. Myklebust, K. R. Rosendal, F. D. Myromslien, J. B. Lorens, G. Nolan, O. Bruland, and J. R. Lillehaug. 2007. RACK1 regulates Ki-Ras-mediated signaling and morphologica l transformation of NIH 3T3 cells. Int J Cancer 120:961-9. 3. Buensuceso, C. S., D. Woodside, J. L. Huff, G. E. Plopper, and T. E. O'Toole. 2001. The WD protein Rack1 mediates prot ein kinase C and integrin-dependent cell migration. J Cell Sci 114:1691-8. 4. Burridge, K., and K. Wennerberg. 2004. Rho and Rac take center stage. Cell 116:167-79. 5. Bustelo, X. R., V. Sauzeau, and I. M. Berenjeno. 2007. GTP-binding proteins of the Rho/Rac family: regulation, eff ectors and functions in vivo. Bioessays 29:356-70. 6. Campbell, S. L., R. Khosravi-Far, K. L. Rossman, G. J. Clark, and C. J. Der. 1998. Increasing complexity of Ras signaling. Oncogene 17:1395-413. 7. Chang, B. Y., K. B. Conroy, E. M. Machleder, and C. A. Cartwright. 1 998. RACK1, a receptor for activated C kinase and a homolog of the beta subunit of G proteins, inhibits activity of src tyrosine kinases a nd growth of NIH 3T3 cells. Mol Cell Biol 18:3245-56.

PAGE 187

8. Chien, Y., and M. A. White. 2003. RAL GTPases are linchpin modulators of human tumour-cell proliferation and survival. EMBO Rep 4:800-6. 9. Clarke, E. J., and V. Allan. 2002. Intermediate filaments: vimentin moves in. Curr Biol 12:R596-8. 10. Dotto, G. P., L. F. Parada, and R. A. Weinberg. 1985. Specific growth response of ras-transformed embryo fibrobl asts to tumour promoters. Nature 318:472-5. 11. Falsetti, S. C., D. A. Wang, H. Peng, D. Carrico, A. D. Cox, C. J. Der, A. D. Hamilton, and S. M. Sebti. 2007. Geranylgeranyltransferas e I inhibitors target RalB to inhibit anchorage-dependent gr owth and induce apoptosis and RalA to inhibit anchorage-independent gr owth. Mol Cell Biol 27:8003-14. 12. Feig, L. A. 2003. Ral-GTPases: approaching thei r 15 minutes of fame. Trends Cell Biol 13:419-25. 13. Feng, Y., and C. A. Walsh. 2004. The many faces of filamin: a versatile molecular scaffold for cell motility and signalling. Nat Cell Biol 6:1034-8. 14. Griner, E. M., and M. G. Kazanietz. 2007. Protein kinase C and other diacylglycerol effectors in cancer. Nat Rev Cancer 7:281-94. 15. Hahn, W. C., C. M. Counter, A. S. Lundberg, R. L. Beijersbergen, M. W. Brooks, and R. A. Weinberg. 1999. Creation of human tumour cells with defined genetic elements. Nature 400:464-8. 16. Hahn, W. C., and R. A. Weinberg. 2002. Modelling the mol ecular circuitry of cancer. Nat Rev Cancer 2:331-41.

PAGE 188

17. Hahn, W. C., and R. A. Weinberg. 2002. Rules for making human tumor cells. N Engl J Med 347:1593-603. 18. Hamad, N. M., J. H. Elconin, A. E. Karnoub, W. Bai, J. N. Rich, R. T. Abraham, C. J. Der, and C. M. Counter. 2002. Distinct requirements for Ras oncogenesis in human versus mouse cells. Genes Dev 16:2045-57. 19. Hanahan, D., and R. A. Weinberg. 2000. The hallmarks of cancer. Cell 100:5770. 20. Hermanto, U., C. S. Zong, W. Li, and L. H. Wang. 2002. RACK1, an insulinlike growth factor I (IGF-I) receptor-i nteracting protein, modulates IGF-Idependent integrin signaling and prom otes cell spreading and contact with extracellular matrix. Mol Cell Biol 22:2345-65. 21. Hsiao, W. L., G. M. Housey, M. D. Johnson, and I. B. Weinstein. 1989. Cells that overproduce protein kinase C are mo re susceptible to transformation by an activated H-ras oncogene. Mol Cell Biol 9:2641-7. 22. Ivaska, J., H. M. Pallari, J. Nevo, and J. E. Eriksson. 2007. Novel functions of vimentin in cell adhesion, migratio n, and signaling. Exp Cell Res 313:2050-62. 23. Jullien-Flores, V., O. Dorseuil, F. Romero, F. Letourneur, S. Saragosti, R. Berger, A. Tavitian, G. Gacon, and J. H. Camonis. 1995. Bridging Ral GTPase to Rho pathways. RLIP76, a Ral effector with CDC42/Rac GTPase-activating protein activity. J Biol Chem 270:22473-7. 24. Koehler, J. A., and M. F. Moran. 2001. RACK1, a protein kinase C scaffolding protein, interacts with the PH do main of p120GAP. Biochem Biophys Res Commun 283:888-95.

PAGE 189

25. Lacal, J. C., T. P. Fleming, B. S. Warren, P. M. Blumberg, and S. A. Aaronson. 1987. Involvement of functional protein kinase C in the mitogenic response to the H-ras oncogene product. Mol Cell Biol 7:4146-9. 26. Lim, K. H., A. T. Baines, J. J. Fiordalis i, M. Shipitsin, L. A. Feig, A. D. Cox, C. J. Der, and C. M. Counter. 2005. Activation of RalA is critical for Rasinduced tumorigenesis of human cells. Cancer Cell 7:533-45. 27. Lim, K. H., K. O'Hayer, S. J. Adam, S. D. Kendall, P. M. Campbell, C. J. Der, and C. M. Counter. 2006. Divergent roles for RalA and RalB in malignant growth of human pancreatic car cinoma cells. Curr Biol 16:2385-94. 28. Liu, J., G. Yang, J. A. Thompson-L anza, A. Glassman, K. Hayes, A. Patterson, R. T. Marquez, N. Auersperg, Y. Yu, W. C. Hahn, G. B. Mills, and R. C. Bast, Jr. 2004. A genetically defined mode l for human ovarian cancer. Cancer Res 64:1655-63. 29. McCahill, A., J. Warwicker, G. B. Bolg er, M. D. Houslay, and S. J. Yarwood. 2002. The RACK1 scaffold protein: a dyna mic cog in cell response mechanisms. Mol Pharmacol 62:1261-73. 30. Meng, X., O. Krokhin, K. Cheng, W. Ens, and J. A. Wilkins. 2007. Characterization of IQGAP1-containing complexes in NK-like cells: evidence for Rac 2 and RACK1 association during ho motypic adhesion. J Proteome Res 6:74450. 31. Miao, W., L. Eichelberger, L. Baker, and M. S. Marshall. 1996. p120 Ras GTPase-activating protein interacts w ith Ras-GTP through specific conserved residues. J Biol Chem 271:15322-9.

PAGE 190

32. Nilsson, J., J. Sengupta, J. Frank, and P. Nissen. 2004. Regulation of eukaryotic translation by th e RACK1 protein: a platform for signalling molecules on the ribosome. EMBO Rep 5:1137-41. 33. Ohta, Y., N. Suzuki, S. Nakamura, J. H. Hartwig, and T. P. Stossel. 1999. The small GTPase RalA targets filamin to i nduce filopodia. Proc Natl Acad Sci U S A 96:2122-8. 34. Oxford, G., C. R. Owens, B. J. Titus, T. L. Foreman, M. C. Herlevsen, S. C. Smith, and D. Theodorescu. 2005. RalA and RalB: antagonistic relatives in cancer cell migration. Cancer Res 65:7111-20. 35. Park, S. H., and R. A. Weinberg. 1995. A putative effector of Ral has homology to Rho/Rac GTPase activating proteins. Oncogene 11:2349-55. 36. Philips, M. R. 2005. Compartmentalized signalling of Ras. Biochem Soc Trans 33:657-61. 37. Rangarajan, A., S. J. Hong, A. Gifford, and R. A. Weinberg. 2004. Speciesand cell type-specific requirements for cel lular transformation. Cancer Cell 6:17183. 38. Robertson, S. P. 2005. Filamin A: phenotypic diversity. Curr Opin Genet Dev 15:301-7. 39. Shields, J. M., K. Pruitt, A. McFall, A. Shaub, and C. J. Der. 2000. Understanding Ras: 'it ain't over 'til it's over'. Trends Cell Biol 10:147-54. 40. Stossel, T. P., J. Condeelis, L. C ooley, J. H. Hartwig, A. Noegel, M. Schleicher, and S. S. Shapiro. 2001. Filamins as integrators of cell mechanics and signalling. Nat Rev Mol Cell Biol 2:138-45.

PAGE 191

41. Tchevkina, E., L. Agapova, N. Dyakova, A. Martinjuk, A. Komelkov, and A. Tatosyan. 2005. The small G-protein RalA stimul ates metastasis of transformed cells. Oncogene 24:329-35. 42. Ward, Y., W. Wang, E. Woodhouse, I. Linnoila, L. Liotta, and K. Kelly. 2001. Signal pathways which promote inva sion and metastasis : critical and distinct contributions of extracellular signal-regulated kinase and Ral-specific guanine exchange factor pathways. Mol Cell Biol 21:5958-69. 43. Webb, C. P., L. Van Aelst, M. H. Wigler, and G. F. Woude. 1998. Signaling pathways in Ras-mediated tumorigenicity and metastasis. Proc Natl Acad Sci U S A 95:8773-8. 44. Wennerberg, K., and C. J. Der. 2004. Rho-family GTPases: it's not only Rac and Rho (and I like it). J Cell Sci 117:1301-12. 45. Wolfman, A., and I. G. Macara. 1987. Elevated levels of diacylglycerol and decreased phorbol ester sens itivity in ras-transforme d fibroblasts. Nature 325:35961. 46. Yamazaki, Y., Y. Kaziro, and H. Koide. 2001. Ral promotes anchorageindependent growth of a human fi brosarcoma, HT1080. Biochem Biophys Res Commun 280:868-73. 47. Yin, J., C. Pollock, K. Tracy, M. Chock, P. Martin, M. Oberst, and K. Kelly. 2007. Activation of the RalGEF/Ral pathway promotes prostate cancer metastasis to bone. Mol Cell Biol 27:7538-50.

PAGE 192

Chapter 5 Summary and Implications By Samuel C. Falsetti 1,2, Sad M. Sebti 1,2,* 1Drug Discovery Program, The H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL 2Departments of Oncological Sciences and Mo lecular Medicine, The University of South Florida, Tampa, FL *Corresponding Author: 12902 Magnolia Drive, Tampa, FL 33612; Tel (813) 745-6734; Fax (813) 745-6748; email:said. sebti@mofitt.org

PAGE 193

In this thesis we have determined novel roles for RalA and RalB in human ovarian neoplasia, defined a database of potential Ral interacting proteins, validated RACK1 as a novel Ral interacting protei n which is required for Ras ovarian transformation. Furthermore, in the second chapter of this thesis we have used an elegant experimental chemical biology rescue system to determine the roles of inhibition of RalA and RalB prenylation in the GGTI-anticancer mechanisms of action. Here we determined that both RalA and RalB are validat ed targets of GGTIs and that inhibition of RalA prenylation is required for GGTI-inh ibition of anchorage independent growth whereas inhibition of RalB prenylation is required for GGTI-mediated suppression of proliferation and induction of apoptosis. Re scue of the anti-p roliferative and proapoptotic effects of GGTI-2417 by expression of GGTI-resistant RalB also rescued from induction of p27kip1, a well known CDK inhibitor of the cell cycle and suppression of survivin, a well known inhibitor of apoptosis. Future work will undoubtedly further define a role for RalB in regulation of the survivin gene expression. Data not presente d in this thesis ce rtainly indicate the importance of RalB in regulating survivin expression. Specifically, we have depleted MiaPaCa2 pancreatic cancer cells of survivin by Ral B but not Ral A siRNA. This depletion has dramatically reduced the level of survivin expr ession. Surprisingly, depletion of RalB by siRNA does not result in an induction of p27kip1. However, this illustrates the complex mechanism of action of GGTIs. GGTIs inhibit phosphorylation of Akt in a non-Ral dependent manner. This is relevant in that Ras governs p27kip1 expression in two synergistic pathways: Akt dependent phosphorylation of the AFX transcription factor and Ral depende nt phosphorylation of AFX. These two

PAGE 194

phosphorylation events regulate, respectiv ely, the nuclear import and DNA binding activity of AFX, a well characterized regul ator of p27kip1 transcription. Further work should establish if dual inhibition of Akt and RalB expression is sufficient to induce p27kip1 expression and if GGTIs affect AFX phosphorylation and transc riptional activity. Certainly, these results sugge st new avenues for research in the GGTI anti-tumor mechanism of action. In Chapter 3 of this thesis, we have both di scovered new roles for, and reaffirmed the importance of, both RalA and RalB in human cancer. We have defined Ral, as well as Akt, as central mediators of Ras-driven ovari an oncogenesis. In c ontrast to previous systems, both RalA and RalB were found to be required for ovarian transformation. Surprisingly, Raf1 and Mek1/2 are not requir ed for Ras-transformation in this system. Indeed, K-Ras mutants incapable of activating Raf/Mek/Erk signaling were still able fully recapitulate the same degree of transforma tion as fully active oncogenic K-Ras. These results suggest that therapeutic intervention in ovarian cancer may be greatly assisted by future efforts to target RalGEF /Ral and PI3K/Akt signaling. There remains much work to be done in defining a role for Ral signaling in ovarian cancer development. Of particular interest, the lack of a role for Raf-1 signaling in this system suggests that the T80 cell syst em is a more appropriate model for ovarian cancer sub-types such as mucinous ovarian and high-grade serous ovarian but not lowgrade serous ovarian. Low-grad e serous is characterized by a non-overlapping pattern of K-Ras and B-Raf mutation which can be interpreted as clear evidence that these two genes operate in a redundant fashion. Howeve r, the incidence of K-Ras mutation in highgrade serous and mucinous coupled with a complete lack of B-Raf mutations and

PAGE 195

strikingly different disease et iology, as compared to low-gr ade serous, indicates non-Raf oncogenesis may determine the progression of these neoplasms. In order to more fully expand on this concept further work will need to determine a role for B-Raf in Ras transformation. Furthermore, further work w ill need to determine a role for individual RalGEFs and PI3K/Akt genes in promoting tran sformation of T80 cells. If these cells are capable of forming tumors in intra-ovarian nude mouse xenograft models, it will be of great interest to determine which sub-type s of ovarian cancers these cells will most closely resemble. Also, assessing patterns of Ral activation and pot ential mutation of RalGEFs in human ovarian tumors, divided by subtype, could serve as further evidence of a role for Ral in ovarian cancer. In the fourth chapter of th is thesis we have used a subtractive proteomics method to determine a database of potential Ral in teracting proteins. On e of these proteins, RACK1, was validated as a Ral interacti ng protein. The expression of RACK1 was required for both Hand K-Ras oncogenesis in T80 cells, a system where RalA and RalB are also required. These results indicate a role for RACK1 in Ras-driven ovarian oncogenesis and, potentially, in th e activity of Ral proteins. Interestingly, RACK1 is a plieotropic s caffolding molecule which interacts with other small GTPases, such as Rac1, and serves as a platform for PKC isoforms to interact with target substrates. The provocative questi on then is does RACK1 serve as a platform for Ral phosphorylation by PKC? If so what is the role of phosphorylation of Ral? These questions are made all the more relevant by the observation that Aurora A phosphorylates RalA and that phosphorylati on regulates Ral localiza tion and fuction. Similarly,

PAGE 196

About the Author Samuel C. Falsetti received his bachelo r’s degree from the University of Tampa in 2002 where he majored in Biochemistry, as well as Marine Biology. Prior to working in the Sebti lab he has worked in the labs of Ping Dou, Ph.D and Gary Litman, Ph.D. While at USF he has published three papers and has two more in various stages of submission. He has served as the chief financ ial officer for the Association of Medical Science Graduate Students and is currently the Director of Student Recruitment for the Florida Chapter of the American Medical Writing Association. He has received awards for excellence in research from USF Health and the USF-IGERT program. Currently, he is employed as Associate Scientific Director for a medical communications company. He enjoys traveling, spending time at home with hi s wife and (crazy) pets, as well as reading Dilbert.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200433Ka 4500
controlfield tag 001 002058224
005 20100201134934.0
008 100201s2008 xx a bm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002307
035
(OCoLC)503001781
040
FHJ
c FHJ
049
FHJA
096
QZ 200
b F197r 2008
1 100
Falsetti, Samuel C.
4 245
The role of RalA and RalB in cancer /
by Samuel C. Falsetti.
260
2008.
300
v, 187 leaves :
ill. ;
29 cm.
500
Includes vita.
Also available online.
502
Dissertation (Ph.D.)--University of South Florida, 2008.
504
Includes bibliographical references.
590
Advisor: Sad M. Sebti, Ph.D.
520
ABSTRACT: Ras genes are frequently mutated in human cancers and present compelling targets for therapeutic intervention. While previous attempts to directly inhibit oncogenic Ras function have largely been unsuccessful use of targeted agents to inhibit the three primary oncogenic pathways activated by mutated Ras: RalGEF-Ral, PI3K-Akt and Raf- MEK-Erk, is an area of intense investigation. Here, we describe the ability of a novel pharmacological inhibitor of geranylgeranyltransferase I, GGTI-2417, to inhibit Ral prenylation and localization. We further used a Ral rescue system to selectively preserve RalA and RalB function and localization during GGTI-2417 treatment and determine the precise roles for inhibition of Ral prenylation in the GGTI anti-cancer response. Specifically, we determined inhibition of RalA is required for GGTI-attenuation of anchorage independent growth whereas inhibition of RalB is required for inhibition of proliferation, induction of apoptosis, suppression of survivin and induction of p27Kip1. We next determined the role of RalGEF-Ral signaling as well as PI3K-Akt and Raf-MEKErk signal transduction pathways in an in vitro model of human ovarian surface epithelial (T80 HOSE) cell Ras-dependent transformation. Using both small interfering RNA (siRNA) and pharmacological inhibitors of Ral, PI3K and MEK we determined that Ras signaling via Ral and PI3K but not MEK is required for ovarian oncogenesis. Furthermore, stable expression of Ras mutants unable to activate Raf-MEK-Erk signaling were able to robustly transform T80 cells. Since we had confirmed the importance of Ral proteins to human epithelial malignancies we next sought to explore the molecular interactions governing Ral transformation using a proteomics approach to rapidly identify proposed Ral interacting partners. Using immunoprecipition of transiently overexpressed FLAG-tagged RalA and RalB followed by 1D-gel separation and tandem MS/MS analysis we determined a database of proposed Ral interacting proteins. One of these, RACK1, is a validated RalA and RalB interacting protein which is at least partially required for Ras and Ral transformation. These results provide both a strong impetus and a solid basis for future studies into the mechanisms of RalA- and RalB- dependent transformation.
653
Ras
Geranylgeranyltransferase I inhibitors
RACK1
Ovarian cancer
Proteomics
2 650
Intracellular Signaling Peptides and Proteins
Neoplasm Proteins.
Ovarian Neoplasms.
Receptors, Cell Surface.
Proteomics.
Genes, ras.
Alkyl and Aryl Transferases.
Protein Prenylation.
Leucine
x analogs & derivatives.
ral GTP-Binding Proteins.
0 690
Dissertations, Academic
z USF
Biochemistry and Molecular Biology
Doctoral.
773
t USF Electronic Theses and Dissertations.
856
u http://digital.lib.usf.edu/?e14.2307