Src Kinase Inhibitors for the Treatment of Sarcomas: Cellular and Molecular Mechanisms of Action Audrey Cathryn Shor University of South Florida

University of South Florida Scholar Commons

Graduate Theses and Dissertations Graduate School

2007 Src kinase inhibitors for the treatment of sarcomas: Cellular and molecular mechanisms of action Audrey Cathryn Shor University of South Florida

Follow this and additional works at: http://scholarcommons.usf.edu/etd Part of the American Studies Commons

Scholar Commons Citation Shor, Audrey Cathryn, "Src kinase inhibitors for the treatment of sarcomas: Cellular and molecular mechanisms of action" (2007). Graduate Theses and Dissertations. http://scholarcommons.usf.edu/etd/2364

This Dissertation is brought to you for free and open access by the Graduate School at Scholar Commons. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Scholar Commons. For more information, please contact [email protected].

Src Kinase Inhibitors for the Treatment of Sarcomas:

Cellular and Molecular Mechanisms of Action

Audrey Cathryn Shor, M.P.H.

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

Co-Major Professor: Richard Jove, Ph.D. Co-Major Professor: W. Jack Pledger, Ph.D. Denise R. Cooper, Ph.D. Larry P. Solomonson, Ph.D. Daniel M. Sullivan, M.D G. Douglas Letson, M.D. Carlos A. Muro-Cacho, M.D., Ph.D.

Date of Approval: March 8, 2007

Keywords: tyrosine kinase inhibitor, dasatinib, cancer therapy, microarray, gene expression profile

Dedication

I dedicate this to my father, Richard C. Burns for his love and devotion – inspiring

my yearning to seek more.

And to my wonderful, caring husband Adam. For his endless love and support

throughout our journey.

Acknowledgements

I would like to acknowledge Sean Yoder and the rest of the members of the H. Lee

Moffitt Microarray Core for completing the microarrays and aiding in the analysis. I

would also like to graciously thank Samuel Falsetti for the coffee breaks and brainstorming sessions throughout all of our hurdles. Ralf Buettner for helping me get started in the beginning, laying the foundation and molding me into the scientist I am today. Bonnie Goodwin, for being there for me during the most difficult year of my

life. Tania Mesa for keeping me sane with our long runs every weekend during the final stages. Lastly, I would like to thank my wonderful mentors, Drs. Rich Jove and

Jack Pledger for having faith in me, Drs. Doug Letson and Carlos Muro-Cacho for

their constant support.

Table of Contents

List of Tables ...... vi

List of Figures...... viii

List of Abbreviations ...... xi

Abstract...... xiv

Introduction...... 1

Epidemiology of Sarcomas...... 1

Risk Factors Investigated for the Development of Sarcoma ...... 4

Limitations of Current Treatment Options ...... 8

Role of Signal Transduction in Cancer...... 9

Applications of Tyrosine Kinase Inhibitors in Sarcomas ...... 14

c-KIT and PDGFR ...... 14

EGFR/HER-2/neu...... 17

VEGFR ...... 19

Other TKIs ...... 21

History of Src: 1911- Present...... 23

Development of Src Inhibitors for Clinical Trials ...... 31

Objectives ...... 35

Aim 1: Evaluate the Biological Response of Three Src Kinase Inhibitors in

Sarcoma Cell Lines...... 35

i Aim 2: Identify a Molecular Signature that Predicts Response to Dasatinib by

Induction of Apoptosis in Sarcoma Cell Lines...... 36

Aim 3: Verify the Presence of the Molecular Signature in Primary Human

Sarcoma Specimens ...... 37

Materials and Methods...... 38

Cells and Reagents...... 38

Preparation of Cell Extracts and Western Blotting...... 39

Immunohistochemistry ...... 40

Src Family Kinase PCR ...... 41

Wound Healing Assay for Cell Migration...... 41

Cell Invasion Assay ...... 42

TUNEL Assay...... 42

siRNA Transfections...... 43

Statistical Analysis...... 43

Microarray Sample Preparation...... 43

Isolation of RNA...... 44

Preparation of Labeled RNA Targets for Hybridization...... 44

Array Hybridization and Scanning ...... 46

Data Analysis...... 47

Preparation of Samples for Quantitative Real-Time PCR Analysis ...... 48

cDNA Reactions ...... 48

Real-Time PCR Reactions ...... 49

Personal Health Identifier...... 49

ii Involvement of Human Subjects...... 50

Introductory Preclinical Data...... 51

Results...... 53

Expression of SFKs in Sarcoma Cell Lines...... 53

STAT3 is Activated in Human Sarcomas and Sarcoma Cell Lines...... 53

PD180970 Inhibits Src and STAT3 Signaling in Human Sarcoma

Cell Lines...... 58

PD180970 Inhibits Cell Viability and Induces Apoptosis in Sarcoma

Cell Lines...... 60

SKI-606 Does Not Inhibit Src-STAT3 Signaling in Human Sarcoma

Cell Lines...... 63

Dasatinib Does Not Inhibit Src-STAT3 Signaling in Human Sarcoma

Cell Lines...... 64

Discussion...... 66

Dasatinib Inhibits Migration and Invasion in Diverse Human Sarcoma Cell Lines and

Induces Apoptosis in Bone Sarcoma Cells Dependent on Src Kinase

for Survival ...... 69

Results...... 71

Src Kinase is Activated in Human Sarcomas and Sarcoma

Cell Lines...... 71

Dasatinib Inhibits Src Kinase Activity in Human Sarcoma Cell Lines.....73

Dasatinib Selectively Blocks Src Downstream Signaling ...... 77

Dasatinib Blocks Cell Motility and Invasion by Sarcoma Cells...... 78

iii Dasatinib Induces Apoptosis of Bone Sarcoma Cell Lines ...... 81

Src is required for survival of bone sarcoma cell lines...... 84

Discussion...... 88

Gene Expression Profile of Sarcoma Cell Lines Serves as Preliminary Signature

Predictive of Response to Treatment with Dasatinib...... 91

Results...... 93

Unsupervised Clustering Identified Three Main Classes, with Five

Subgroups in Relation to Cell Line Types...... 93

A Molecular Signature Distinguishes Response to Dasatinib as Defined by

Induction of Apoptosis...... 95

Identification of Two Probesets with Greater Fold Changes May Provide

Further Insight into the Prediction of Response ...... 105

Testing the Molecular Signature with Cell Lines of Unknown Response

Reveals that the Molecular Signature Can Accurately Predict Response

to Dasatinib in Cell Lines ...... 112

Discussion...... 126

Validation of Gene Expression Profile in Primary Human Sarcomas...... 129

Results...... 131

Unsupervised Clustering Identified Diversity Among Human Sarcoma

Specimens ...... 131

Identification of the Cell Line Molecular Signature in Human

Sarcomas...... 131

iv Molecular Signature that Predicts Response to Dasatinib in Cell Lines

Can be Used to Group Tumors by Potential Response...... 137

Analysis of Two Probesets with Greater Fold Changes May Provide

Further Insight into the Prediction of Response in Tumors...... 141

Discussion...... 145

Conclusions...... 148

Clinical Significance...... 157

References...... 159

About the Author ...... End Page

List of Tables

Table 1 Relative Frequency of Subtypes of Sarcomas...... 3

Table 2 Incidence of Soft Tissue Sarcomas Diagnosed in the US by Race,

SEER, 1998-2002...... 3

Table 3 Potential Risk Factors Evaluated by Epidemiological Studies...... 6

Table 4 Development Stage of TKI Targeted Agents in Sarcoma...... 13

Table 5 SFK Gene Expression Status in Human Sarcoma Cell Lines...... 55

Table 6 Summary of Cell Line IC50 Values and Responses to Dasatinib...... 87

Table 7 Fold Changes for Key Signature Genes in Sarcoma Cell Lines

Compared to Median Intensity...... 99

Table 8 Quantitative RT-PCR Expression of Key Signature Genes in

Sarcoma Cells ...... 101

Table 9 Fold Changes for Key Signature Genes in Sarcoma Cells Compared

to Median Intensity of Non-Responders...... 104

Table 10 Fold Changes for Ephrin-A1 and Dapper in Sarcoma Cell Lines

Compared to Median Intensity...... 108

Table 11 Microarray Fold Changes for Ephrin-A1 and Dapper in Sarcoma Cell

Lines Compared to Median Intensity of Non-Responders ...... 109

Table 12 Quantitative RT-PCR Values for Ephrin-A1 and Dapper in Sarcoma

Cell Lines...... 111

vi Table 13 Fold Changes for Key Signature Genes in Test Cells Compared to

Median Intensity ...... 117

Table 14 Relative qRT-PCR Fold Changes Key Signature Genes in Test Cells

Compared to Median Intensity...... 117

Table 15 Fold Changes for Ephrin-A1 and Dapper in Test Cells Compared to

Median Intensity ...... 125

Table 16 Microarray Fold Changes for Ephrin-A1 and Dapper in Test Cells

Compared to Median Intensity of Non-Responders ...... 125

Table 17 Classification of Sarcomas using Predictive Signature from Cell

Lines...... 134

Table 18 Fold Changes for Key Signature Genes in Sarcomas ...... 140

Table 19 Fold Changes for Ephrin-A1 and Dapper in Sarcomas ...... 144

vii

List of Figures

Figure 1 Illustration of Associated Tyrosine Kinases and Targeted Therapies ...... 12

Figure 2 Organization of Src Kinase...... 26

Figure 3 Src Activation and Signaling Conformations ...... 27

Figure 4 Structure of Three ATP-Competitive Src Kinase Inhibitors...... 33

Figure 5 Evaluation of SFK Expression in Human Sarcoma Cell Lines ...... 54

Figure 6 STAT3 Activation Status in Human Sarcoma Tissues and

Cell Lines...... 56

Figure 7 PD180970 Inhibits Src and STAT3 Signaling in Sarcoma

Cell Lines...... 59

Figure 8 PD180970 Inhibits Viability and Induces Apoptosis in Sarcoma

Cells Lines ...... 61

Figure 9 PD180970 Induces Apoptosis in Sarcoma Cell Lines as Measured by

TUNEL ...... 62

Figure 10 SKI-606 Does Not Inhibit Src-STAT3 Signaling in Sarcoma

Cell Lines...... 63

Figure 11 STAT3 Signaling is Independent of Src Kinase Activity in Human

Sarcoma Cell Lines...... 65

Figure 12 Src Kinase is Activated in Sarcoma Tissues and Cell Lines...... 72

viii Figure 13 Dasatinib Inhibits Src Activation and Downstream Signaling in Sarcoma

Cell Lines...... 74

Figure 14 Dasatinib Does Not Induce c-Src mRNA Expression...... 76

Figure 15 Dasatinib Inhibits Cell Motility and Invasion...... 79

Figure 16 Dasatinib Induces Apoptosis in Bone Sarcoma Cell Lines...... 82

Figure 17 Src is Required for Bone Sarcoma Cell Line Survival...... 85

Figure 18 Hierarchical Clustering of Sarcoma Cell Lines Based on GEP...... 94

Figure 19 Molecular Signature Generated Using the Median Value for Each Probe

Set as Comparison Group...... 97

Figure 20 Quantitative RT-PCR Validations of Histone H3, FAF1, α-Catenin

and δ-Catenin Genes...... 100

Figure 21 Molecular Signature Generated Using the Median Value for Each Probe

Set in the Non-Responders as Comparison Group...... 102

Figure 22 Expression of Ephrin-A1 and Dapper in Sarcoma Cell Lines Using the

Median of All Samples as Comparison Group...... 106

Figure 23 Expression of Ephrin-A1 and Dapper in Sarcoma Cell Lines Using the

Median of the Non-Responders as Comparison Group...... 107

Figure 24 Quantitative RT-PCR Validations of Ephrin-A1 and Dapper

Genes...... 110

Figure 25 Molecular Signature that Predicts Response to Dasatinib Generated

Using the Median Value for Each Probe Set as Comparison

Group ...... 114

ix Figure 26 Dasatinib Inhibits Src Activation and Signaling in HOS and SW1353

Cells ...... 118

Figure 27 Dasatinib Induces Apoptosis in HOS but Not SW1353 Sarcoma

Cells...... 120

Figure 28 Expression of Ephrin-A1 and Dapper in HOS and SW1353 Cells

Using the Median of All Samples as Comparison Group...... 121

Figure 29 Expression of Ephrin-A1 and Dapper in HOS and SW1353 Cells

Using the Median of the Non-Responders Comparison Group...... 123

Figure 30 Hierarchical Clustering of Primary Sarcoma Specimens...... 132

Figure 31 Heatmap of Molecular Signature Expression in Sarcomas...... 135

Figure 32 Expression of Histone H3, FAF1, α-Catenin and δ-Catenin in

Sarcomas...... 138

Figure 33 Expression of Ephrin-A1 and Dapper in Sarcomas ...... 142

List of Abbreviations

AJ Adherens Junctions

AS Adenosarcoma

ATP Adenosine Triphosphate

CAS Crk and Src Associated Substrate

Chk Csk homologous kinase

CS Chondrosarcoma

DFSP Dermatofibrosarcoma Protuberans

DMSO Dimethyl sulfoxide

EGFR Epidermal Growth Factor Receptor

EMSA Electrophoretic Mobility Shift Assay

EWS Ewing’s sarcoma

FAF1 Fas Associated Factor 1

FAK Focal Adhesion Kinase

FBS Fibrosarcoma

Fig Figure

GEP Gene Expression Profile

GIST Gastroinstestinal Stromal Tumor h Hour

HSP Heat Shock Protein

xi IC50 50% Inhibitory Concentration

ICD International Classification of Diseases

IGFR-1 Insulin-like Growth Factor Receptor-1

IHC Immunohistochemistry

LMS Leiomyosarcoma

LPS Liposarcoma

MAS Affymetrix Microarray Suite

MHF Malignant Fibrous Histiocytoma

MFS Myxoid Fibrosarcoma

NF1 Neurofibromatosis type 1

N/R No Response

NRTK Non-receptor Tyrosine Kinase

OSA Osteosarcoma

PDGFR Platelet Derived Growth Factor Receptor qRT-PCR Quantitative Real-Time Polymerase Chain Reaction

Rb Retinoblastoma

RD Rhabdomyosarcoma

RSV Rous sarcoma virus

RTK Receptor Tyrosine Kinase

RT-PCR Real-Time Polymerase Chain Reaction

SAM Statistical Analysis of Microarrays

Ser Serine

SCS Synovial Cell Sarcoma

xii SFK Src Family Kinase

SH Src homology siRNA Silencing RNA

STAT Signaling and Transducer of Activation Transcription

STS Soft Tissue Sarcoma

TK Tyrosine Kinase

TKI Tyrosine Kinase Inhibitor

Tyr Tyrosine

U Units

VEGFR Vascular Endothelial Growth Factor Receptor

Y Ty

xiii

Src Kinase Inhibitors for the Treatment of Sarcomas:

Cellular and Molecular Mechanisms of Action

Audrey Cathryn Shor

ABSTRACT

Sarcomas are rare mesenchymally-derived tumors with limited treatment options.

Tyrosine kinases may serve as potential targets for sarcoma therapy because many are mutated or overexpressed in sarcomas and cell lines. One potential molecular target for sarcoma treatment is the Src tyrosine kinase. Three independently synthesized Src kinase inhibitors were evaluated in human sarcoma cell lines. Of the three, dasatinib, provided promising results as a potential sarcoma therapy. Until this study, dasatinib activity had not been characterized in sarcoma cells. Based on our previous findings of Src activation in human sarcomas, we evaluated the effects of dasatinib in twelve sarcoma cell lines.

Dasatinib inhibited Src activity and downstream signaling at nanomolar concentrations.

Inhibition of Src signaling was accompanied by blockade of cell migration and invasion.

Moreover, apoptosis was induced in a subset of bone sarcomas at nanomolar concentrations of dasatinib. Inhibition of Src protein expression by siRNA also induced apoptosis, indicating that these bone sarcoma cell lines are dependent on Src activity for survival. These results demonstrate that dasatinib inhibits migration and invasion of diverse sarcoma cell types, and selectively blocks the survival of bone sarcoma cells.

xiv Therefore dasatinib may provide therapeutic benefit by preventing the growth and

metastasis of sarcomas.

Microarray analysis of the sarcoma cell lines lead to the identification of a

molecular signature that successfully predicts response to dasatinib by induction of apoptosis. Components of this molecular signature are expressed in primary human

sarcomas. Furthermore, expression of the molecular signature in sarcomas can be

utilized to cluster tumors based on theoretical response to dasatinib. While the prediction of response in tumors is theoretical, there is encouraging evidence to support further endeavors into validating the potential of this molecular signature to predict response in patients.

Together, these studies reveal that, in cell lines, both constitutive Src activation

and the presence of a molecular signature that predicts response to dasatinib are

important parameters to consider when selecting dasatinib as a treatment for.

Furthermore, novel therapeutic approaches that inhibit Src signaling may selectively

induce apoptosis in tumor cells and sensitize to chemotherapy those tumors that contain

the relevant molecular signature.

Introduction

Epidemiology of sarcomas

Sarcomas comprise a relatively rare and diverse group of malignant tumors that

arise from mesenchymally-derived connective tissues including bone, fat and muscle.

There are more than 50 different subtypes of sarcomas, with approximately 12,000 new cases diagnosed nationwide each year (1, 2). While this represents a fraction of all

cancers diagnosed in the United States (3), sarcomas account for approximately 20% of

newly diagnosed pediatric solid tumor malignancies (4) and are among the cancers that

pose the greatest risks of mortality and morbidity in children and young adults (5-9).

Histologically, sarcomas are divided into two sub-categories, soft-tissue and bone

sarcomas. There are over 9,000 cases of soft tissue sarcomas expected to be diagnosed in

the United States this year and 5,500 Americans are expected to die as a result this

disease (10). Less than 1,000 soft tissue sarcomas are diagnosed in children each year in the United States, with an annual incidence of approximately 11 per million (11). The most commonly diagnosed soft tissue sarcomas (STS) include malignant fibrous histiocytomas (MFH), liposarcomas (LPS) and leiomyosarcomas (LMS) (Table 1) (10-

12). In 2006, over 2,760 new cases of bone sarcoma are expected to be diagnosed in the

United States and 1,260 Americans are expected to die from complications from these tumors (13). Bone sarcomas may occur at any age and account for less than 0.2% of all cancers diagnosed each year, with 30% observed in children and young adults (13).

1 Osteosarcoma (OSA) is the most commonly diagnosed primary bone tumor followed by

Ewing’s sarcoma (EWS), chondrosarcoma (CS) and fibrosarcoma (FBS) (13, 14).

In the United States, sarcomas affect men more often than women, and white men have the highest overall incidence among all citizens (Table 2) (4,17). There is a high frequency of sarcomas reported in Uganda and Zimbabwe which is related to the high prevalence of HIV/AIDS infections in those countries (18). Israel has the highest prevalence of all non-African nations with an elevated age-standardized incidence of diagnoses that is consistent with previous reports of increased incidence of sarcomas in men of Jewish background (19, 20).

The detection of specific molecular abnormalities has contributed significantly to sarcoma classification while eliciting an interest in targeted therapies (1). Several well- described genetic mutations that predispose individuals to specific types of sarcomas have been described. The most common are mutations in the retinoblastoma (Rb) pathway, which increase the risk of developing osteogenic sarcomas (21), Li-Fraumeni syndrome, which is caused by a p53 mutation, increases the risk of a variety of sarcomas

(22-25), deletion of the neurofibromatosis type 1 (NF1) gene increases the risk of malignant nerve sheath tumors (26) and c-kit gene mutations are responsible for the development of gastrointestinal stromal tumors (GIST) (27).

2 Table 1: Relative frequency of subtypes of sarcomas

Histological Subtypes % Malignant fibrous histiocytoma 28.15 Liposarcoma 15.16 Leiomyosarcoma 11.94 Synovial sarcoma 10.08 Malignant peripheral nerve sheath tumor 5.81 Rhabdomyosarcoma 4.84 Unclassified sarcoma 11.29 Fibrosarcoma 3.06 Ewing’s sarcoma 2.02 Angiosarcoma 2.02 Osteosarcoma 1.13 Epthelioid sarcoma 1.13 Chondrosarcoma 1.05 Clear cell sarcoma 0.97 Alveolar soft part sarcoma 0.56 Malignant hemangiopericytoma 0.40 Adapted from (15)

Table 2: Incidence of soft tissue sarcomas diagnosed in the US by race, SEER, 1998-2002

Race/Ethnicity Men Women All Races 3.5 per 100,000 men 2.4 per 100,000 women White 3.6 per 100,000 men 2.4 per 100,000 women Black 3.5 per 100,000 men 2.6 per 100,000 women Asian/Pacific Islander 2.7 per 100,000 men 1.6 per 100,000 women Hispanic 3.0 per 100,000 men 2.3 per 100,000 women Adapted from (16)

3 Risk factors investigated for the development of sarcomas

Few risk factors have been associated in the literature with the development of sarcomas (reviewed in Table 3). Epidemiologic studies have identified radiation, certain chemotherapies and vinyl chloride as significant factors while data is conflicting with regard to pesticides and herbicides. Currently, phenoxyacetic acid and chlorophenol have been accepted as causal agents in the development of sarcomas, although the data generated from epidemiologic studies investigating this association do not support this conclusion, (28-38). Dioxin and Agent Orange, more potent forms of pesticides and herbicides, have not been shown to play a significant role in the induction of sarcomas

(39-49). There have been no published reports suggesting an association between viruses and the development of sarcomas in humans to date although ongoing investigations are trying to better understand the role of viruses in sarcomagenesis (50,

51).

The relationship between traumatisms, burns and scars and sarcomagenesis is weak. Case series and isolated reports are the only studies available to assess the relevance of these risk factors and a measure of the magnitude of an association has not been established so far. Even in case-control studies a temporal relationship is difficult to establish. Nevertheless, there have been numerous reports of sarcoma development in the location of previous injuries. However the lack of objective quantitation and the uncertainty in the temporal sequence of events has prevented the establishment of a definitive exposure-outcome.

Many difficulties plague epidemiological studies of sarcoma. One example is the previous faults inherent of the International Classification of Diseases (ICD) system. The

4 identification of sarcomas was complicated by inefficient coding rules, which until

recently, classified sarcomas according to visceral organ or connective tissue of origin,

rather than placing sarcomas in one specific ICD category. As a result, many sarcomas

were not accurately reported and this has complicated retrospective analyses. It is

expected that recent improvements in the methodology will improve recognition.

Another concern of epidemiology studies has been the poor correlation between the diagnoses recorded on death certificates and hospital records.

The rarity of sarcomas leads to a greater chance of misdiagnosis and

misclassification. This is complicated by discordant diagnostic opinions even among

experts. Furthermore, methodological deficiencies in study design to investigate

potential risk factors have been common. An example is the difficulty in determining a

precise exposure to a risk factor and, often, surrogate measurements are used to estimate

average exposures. Also, in the case of pesticides and herbicides, it has been difficult to

precisely determine the exact chemical composition of the substances of interest in a

given cocktail. Thus, it is speculated that that the observed risk from exposure to phenoxy herbicides may become insignificant when controlling for some of these concomitant exposures. Thus, further studies are needed to establish a relationship between exposure to phenoxyacetic acids and chlorophenols and development of sarcomas. The known epidemiologic risk factors of sarcomagenesis have had major

public health implications but more research is still needed. Sarcoma is an extremely rare

cancer and there are insufficient cases to conduct valid, classic epidemiologic studies.

5 Table 3: Potential risk factors evaluated by epidemiological studies.

Risk Factor Type of Study Conclusion Reference Case-series and (38, 52-76) Radiation Definitive Case-reports (77-80) Thorotrast Case-control Probable (81-84) Breast Conservation Case-series and Probable (85, 86) Therapy Case-reports Chemotherapy Retrospective cohort Definitive (87) Case-series Probable (88) Case-control Probable (73) Tamoxifen Case-series Possible (89-92) Vinyl Chloride Retrospective cohort Definitive (93-97) Retrospective cohort Unclear (98) Case-series Probable (99) Pesticides & Herbicides Retrospective cohort No Association (34, 37, 38) Case-control No Association (32, 36) Retrospective cohort Possible (35) Case-Control Possible (31, 33) Occupational Study Possible (28-30) Dioxin Retrospective cohort Possible (41, 44) Prospective cohort No Association (40) Agent Orange Case-control No Association (39, 42, 43, 45-49) Lymphadema Case-report Possible (100) Case-series Possible (101) Trauma Case-series Possible (102) Case-control Possible (103) Foreign Body Implantation Case-report Probable (104-111)

Burns & Surgical Scars Case-series Possible (112-119)

Viruses Retrospective cohort No Association (50, 51)

6 A molecular epidemiology approach guided by classic principles may provide a

more appropriate venue. The data reported in the literature supports causal roles for

radiation, vinyl chloride, certain chemotherapies and pesticides and herbicides containing phenoxyacetic acid and chlorophenol exposures, as well as previous cases of lymphedema. This evidence has restricted occupational exposures to vinyl chloride, pest- and herbicides. Radiation therapy has been re-evaluated and lesser doses are currently used to try to prevent the occurrence of future sarcomas. Physicians are monitoring patients who are receiving treatments such as tamoxifen in order to detect adverse effects before a sarcoma develops. Furthermore, the problems evident in the ICD coding of sarcomas have prompted a more sophisticated method of classifying rare diseases. This will aid in resolving the difficulties in detecting each case and allow for more thorough and complete studies of rare disease such as sarcomas in the future.

There have been recent advances in the understanding of sarcoma molecular biology. Employing what is currently known of the epidemiologic risk factors and the knowledge gained at the bench will help tailor design future studies. These studies should correlate environmental exposures with the molecular biology of the disease.

Collectively, epidemiologists and basic scientists can achieve a more thorough understanding of sarcomagenesis. Together with a more efficient classification system, this will provide a more precise approach in the design of future clinical trials and

treatments, as well as a greater understanding of potential risk factors of sarcomas yet to

be identified.

7 Limitations of current treatment options

The primary approach to treating sarcomas is surgery. Radiation therapy is also used following surgery for patients with unresectable or residual tumors. Systemic chemotherapy, such as doxorubicin and ifosphamide are used to treat patients with metastatic disease. Pediatric and certain cases of chemosensitive sarcomas undergo chemotherapy treatment following surgery and/or radiation treatment as an effort to prevent recurrence or metastasis. However, because many sarcomas harbor Rb mutations and exposure to radiation increases the risk of developing sarcomas, variations in the method of treatment is required to prevent further complications. While this approach has proven effective in treating patients with localized disease, in many cases only modest degrees of success are achieved. Furthermore, treatment for patients with advanced and/or metastatic disease remains limited and chemotherapy is palliative in nature at best.

Doxorubicin was first employed as a treatment option for metastatic soft tissue sarcomas in 1972 at MD Anderson, Houston, TX (120) and ifosphamide was introduced in the late 1980s (121). The outcome for most adult patients with advanced sarcoma still remained bleak, however. Recently, the power of molecular biology was clearly demonstrated with the introduction of Gleevec as a treatment option for GISTs (122,

123). Revealing a new method of approaching sarcoma treatment. To fully appreciate this, the role of signal transduction in cancer and most specifically in sarcomas must first be reviewed.

8 Role of signal transduction in cancer

Protein phosphorylation plays a key role in nearly every aspect of cellular molecular biology. It is controlled by kinases, which phosphorylate, and phosphates, which de-phosphorylate protein-serine, -threonine and -tyrosine residues. To date, 90 tyrosine kinases - 58 transmembrane receptors and 32 non-receptor proteins (124) and

107 protein-tyrosine phosphatase genes (125) have been identified in the human genome.

Signals are communicated in the cell either through external or internal stimuli. External stimulation is elicited through cell receptors and internal stimulation is processed through non-receptors tyrosine kinases which activate specific proteins and transmit the signal to other proteins within the cell, ultimately transducing the signal to the nucleus, regulating gene transcription. Cancer arises as a result of aberrant activation or stimulation of signal transduction pathways.

Genetically, sarcomas fall into two subgroups, those with complex karyotypes characteristic of severe genomic instability and those characterized by near-diploid karyotypes. Sarcomas with simple, near-diploid karyotypes usually possess specific chromosomal translocations. Sarcomas with complex karyotypes have high frequencies of p53 and Rb mutations as well as impairments in DNA repair and severe chromosomal instability. This group includes some of the more commonly diagnosed sarcomas; EWS,

LMS, rhabdomyosarcoma (RD) and OSA. Many sarcomas that fall into both genetic subgroups also possess abnormalities in growth factor signaling and signal transduction pathways. Tyrosine kinases (TKs) make up the majority of defective signaling pathways in sarcomas, including mutations in the platelet-derived growth factor receptor (PDGFR), c-KIT, vascular endothelial growth factor (VEGF) and insulin-like growth factor-1

9 receptor (IGF1-R) signaling pathways. For example, GISTs, EWS, dermatofibrosarcoma

protuberans (DFSP), synovial sarcoma and Kaposi’s sarcoma all have been shown to

have mutations which elicit c-kit overexpression and/or PDGFR over-stimulation (1).

Tyrosine kinase inhibitors (TKIs) are a class of novel therapeutics that are

effective alone and in combination with conventional chemotherapeutics in treating a

variety of cancer subtypes. There are two basic classes of TKs, receptor tyrosine kinases

(RTKs) and non-receptor tyrosine kinases (NRTKs). Upon TK activation, a cascade of protein interactions occur releasing signals of positive and negative regulators of a variety of cellular processes including cell cycle regulation, proliferation, adhesion, migration,

invasion, transcription and survival. Under normal conditions, cellular signaling tightly

regulates activated TKs. The induction of TK signaling in the oncogenic state overcomes

controlled regulation and signaling becomes activated by a myriad of cellular

mechanisms including mutation and over-expression of the TK receptors or receptor

ligands. All TKs rely on adenosine triphosphate (ATP) to mediate the transfer of energy

in the kinase domain and to elicit downstream signaling cascades via pathway

intermediates. TKIs employ two strategies: (1) antibodies which act as receptor

antagonists or to sequester the TK ligand, preventing the ligand from binding to the RTK,

and (2) small molecule inhibitors which act by competing for the ATP-binding domain in

the catalytic site of the enzyme (Figure 1).

The use of TKIs for the treatment of sarcomas is predicated on the hypothesis that

malignant cells rely more heavily on TK signaling than neighboring normal cells. GIST

tumors were previously untreatable with our current chemotherapeutic options. Advances

in understanding the genetic nature of cancer have lead to the development of new

10 treatment for sarcoma. GISTs that harbor activating mutations in the c-kit tyrosine kinase

are sensitive to treatment with imatinib mesylate (Gleevec) as well as sunitinib malate,

whereas those without c-kit mutations are insensitive. Proof-of-principle of this concept has already been obtained in the sarcomas in the case of GISTs. TKIs currently under

investigation as potential therapies for sarcomas are reviewed in Table 4.

Induction Ligands Inhibition Her-2/neu Trastuzumab VEGF Bevacizumab

Imatinib RTKs NRTKs Imatinib Sunitinib PDGFR c-KIT Sunitinib Semaxanib EGFR Src Gefinitib IGF-1R Downstream Signaling VEGFR Dasatinib NVP-AEW541

Ras Sorafenib Semaxanib SU6668 CEP-5214 CEP-7055 Sunitinib

Figure 1. Illustration of associated tyrosine kinases and targeted therapies. Several TKIs have been evaluated as potential therapeutics in sarcomas. TKI can inhibit TK signaling and activation by preventing ligand binding to receptors or by directly inhibiting the TK by binding in the catalytic domain of the kinase. Here, the targets of several TKIs are depicted.

12 Table 4. Development stage of TKI targeted agents in sarcoma

Agent Company Targets Stage of Development References

Small Molecule Inhibitors

Phase III GIST, (126, 127) Imatnib Phase II DFSP, Kaposi’s sarcoma (128) Novartis c-KIT, PDGFR (Gleevec) Preclinical osteosarcoma, MFH, (129) Ewing’s sarcoma

Phase I Ewing’s sarcoma Gefitinib (130) AstraZeneca EGFR family Preclinical osteosarcoma, (Iressa) rhabdomyosarcoma (131, 132)

Phase II STS VEGFR, (132) Semaxanib Pfizer Preclinical neurogenic sarcomas, PDGFR Ewing’s sarcoma (132, 133)

SU6668 Sugen VEGFR Preclinical Ewing’s sarcoma (133)

CEP-5213 Cephalon VEGFR Preclinical angiosarcoma (134)

CEP-7055 Cephalon VEGFR Preclinical angiosarcoma (135) Preclinical osteosarcoma, NVP-AEW541 Novartis IGF1-R Ewing’s sarcoma, (136) rhabdomyosarcoma Brisol-Myers Dastinib Src Phase II solid tumors (137, 138) Squibb

Sorafenib Bayer Ras Phase II Kaposi’s sarcoma (127, 136) Cell membrane Perifosine Keryx Phase II STS (139, 140) signaling Pan TKI, Sunitinib Pfizer PDGFR, Phase III refractory GIST (136, 141, 142) (Sutent) VEGFR Antibody Targeted Herceptin Genentech HER-2/neu Preclinical Ewing’s sarcoma (132) (Tratusumab)

Phase II osteosarcoma, STS, (136, 141, 142) Ewing’s sarcoma, Bevacizumab Genentech VEGF Kaposi’s sarcoma, (Avastin) alveolar soft part sarcoma Preclinical Ewing’s sarcoma (132)

13 Applications of tyrosine kinase inhibitors in sarcomas

c-KIT and PDGFR

GISTs are one of the most common mesenchymal malignancies of the

gastrointestinal system. Furthermore, advanced, unresectable or malignant GISTs are fatal and highly resistant to conventional chemotherapies (143, 144). Mutation in c-kit is

a critical event in the development of this malignancy. It is estimated to occur in 75% to

92% of GISTs diagnosed (145, 146). Imatinib mesylate (Gleevec), a c-KIT, Bcr-ABL and

PDGFR TKI, was originally utilized as a therapy for chronic myelogenous leukemia and

is now used to treat GISTs. The first GIST patient received imatinib treatment in 2000.

Within weeks this patient demonstrated a clinical response and remained stable for 18

months (122). Clinical trials were initiated in the United States and Europe following the

dramatic results of this case-report. The first trials showed responses that ranged from 59

to 69%. However, approximately 12% of the patients on the initial trials experienced

resistance to imatinib (147-150). Duensing et al. and Heinrich et. al. demonstrated that

GISTs lacking c-kit mutations have a higher likelihood of demonstrating imatinib

resistance, while tumors with c-kit mutations experienced at least a 50% reduction of

tumor volume (149, 150).

A clinical trial consisting of 147 GIST patients randomly assigned to receive

either 400 or 600 mg of imatinib daily, demonstrated that 53.7% of patients had a partial

response, 27.9% had stable disease, and 0% achieved a complete response (151). A

follow-up randomized clinical trial conducted by Verweij et al., consisted of 946 patients

randomly allocated 400 mg of imatinib either once or twice a day. Patients remained

enrolled for an average of 760 days, and 56% of the patients that received imatinib once a

14 day progressed compared to the 50% of the patients who received the regiment twice- daily. The estimated hazard ratio for this study was 0.82 [95% CI 0.69-0.98]; p-0.026.

There was a similar response with no significant difference between treatment arms: 5% of patients achieved a complete response, 47% achieved a partial response, and 32% achieved stable disease. This study concluded that the once daily 400 mg dose of imatinib is a sufficient treatment option for treating advanced or metastatic GISTs (152).

Together, these studies provide the necessary evidence to recommend treating patients with advanced or metastatic GISTs containing c-kit and PDGFR mutations with imatinib for extended periods. There are numerous ongoing studies evaluating the role TKI; imatinib and sunitinib can be used in neoadjuvant, adjuvant, and metastatic settings alone or in combination with other agents.

DFSP is a slow growing sarcoma that is locally aggressive and has a high likelihood of local recurrence. Most DFSPs have a characteristic translocation, t(17;22).

This places the regulation of PDGFβ, a ligand for PDGFR, under the control of the collagen 1A1 promoter, inducing overexpression of PDGFβ. In vivo experiments conducted in nude mice carrying tumors induced by DFSP-transformed cells treated with imatinib demonstrate significant inhibition of tumor growth (153).

A case-series of ten patients with advanced DFSP with the characteristic t(17;22) translocation were evaluated for clinical response to imatinib. Patients were treated with

400 mg of imatinib two times daily. Eight patients with local disease experienced a clinical response, four of which had a complete clinical response of an average duration of 220 days. The patients with metastatic disease had more complex karyotypes than those of the localized DFSPs. One patient with metastatic disease and the t(17;22)

15 translocation achieved a partial response (198 of 383 days of follow-up) but experienced disease progression seven months following treatment. The other patient with metastatic disease did not have the t(17;22) translocation and did not experience a clinical response.

Based on this, imatinib is a viable option for the treatment of DFSP.

AIDS-related Kaposi’s sarcoma is associated with HIV and herpes virus/human

herpes virus-8 co-infections usually found on the skin (126). Kaposi’s sarcoma has been

shown to express high levels of both c-KIT and PDGFR, and imatinib has been evaluated

as a potential therapeutic option for this sarcoma. In a study ten male patients with

AIDS-related Kaposi’s sarcoma received 300 mg of imatinib twice daily for four weeks.

Of the ten patients, five had partial clinical responses. The remaining five participants

had stable disease at the end of the study, two of them with histological disease

regression. This study demonstrates that imatinib has potential effectiveness as a

treatment option for patients with AIDS-related Kaposi’s sarcoma (126).

TKIs inhibit proliferation over a range of sarcoma-derived cell lines in vitro and

in vivo as measured by xenograft models in nude mice. Studies using imatinib have been

completed in rat OSA and MFH cell lines expressing high levels of PDGFRα. Imatinib

inhibited 20% and 40% of cellular proliferation, respectively when OSA and MFH cell

lines were treated with 10 μM imatinib (154). OSAs have been shown to express high levels of PDGFR. However, preclinical studies have not shown imatinib to achieve anti-

tumor activity within clinically relevant or achievable doses (155, 156).

The effects of imatinib were evaluated on a panel of eight different EWS cell lines

and found that imatinib inhibited proliferation by 50% and induced apoptosis at IC50s ranging from 10-12 μM. Furthermore, in xenograft models, imatinib treatment resulted

16 in regression and/or stabilization of primary Ewing’s tumors (157). Druker et al., also

determined a synergistic effect on growth inhibition when 10 μM imatinib was combined

with increasing doses of doxorubicin (134). Phase I clinical trials have demonstrated that the maximally tolerated dose of imatinib is 1000 mM/day (148, 158, 159), which corresponds to a concentration between 6-10 μM in the blood (160). This concentration is below the IC50 required to inhibit the proliferation and induce apoptosis of EWS cell

lines (161). However, a study conducted on a panel of eight EWS cell lines with high

levels of c-KIT expression, found the cell lines to be resistant to imatinib at

concentrations ranging from 0.1 to 10 μM as determined by proliferation assays (162).

The preclinical data is conflicting in EWS and may be caused by inconsistencies among

the basic science experiments performed or may be inherent in the nature of the cell lines

utilized in these studies. Therefore, the preclinical data suggests that c-KIT is not a

critical target for EWS survival and thus may not serve a single agent treatment option.

EGFR/HER-2/neu

Gefitinib (Iressa), an epidermal growth factor receptor (EGFR) inhibitor, has

shown potential anti-tumor effects in several sarcoma cell lines when used in

combination with irinotencan, a topoisomerase I inhibitor (129). Murine xenograft

models have been used to assess the effects of gefitinib on implanted tumor activity.

Gefitinib was orally administered at 100 mg/kg either once or twice daily for 5

days/week. The anti-tumor activity of the combination was greater than additive in one

of the seven cell line-derived xenografts and enhanced the activity of irinotecan in three

of the seven xenografts. These data also reported an increased bioavailability of

irinotecan when used in combination with gefitinib (129).

17 A phase I trial of gefitinib was completed on 25 children with refractory solid

tumors. Twelve of the 25 patients enrolled were sarcoma patients. Gefitinib was

administered once daily for a course of 28 consecutive days starting at 150 mg/m2 and escalated to 500 mg/m2 a median of 54 courses were delivered. A recurrent EWS patient

experienced a partial response after one course of treatment which lasted for ten weeks.

The maximum tolerated dose in this study was 400 mg/m2/day and was tolerated well

(128).

There is conflicting evidence in the literature evaluating HER-2/neu as an

effective target in treating sarcoma, thus the potential role of HER-2/neu as a therapeutic

target for sarcoma is controversial. One group evaluated five EWS cell lines and 13

archival primary EWS samples for HER-2/neu gene amplification and protein expression.

While several of the EWS cell lines and tumor samples had high HER-2/neu protein expression, none had HER-2/neu gene amplification. In addition, upon assessment of

trastuzumab, (Herceptin) an anti-HER-2/neu monoclonal antibody that inhibits HER-

2/neu expression and blocks tumorigenesis induced by HER-2/neu, the group determined that trastuzumab had minimal inhibitory effects on cell growth, survival or colony formation of the EWS cell lines evaluated (140). Furthermore, others have evaluated trastuzumab alone and in combination with conventional chemotherapeutics used to treat sarcomas in two EWS cell lines. These studies have not been presented anti-tumor activity at clinically achievable doses (139) and it was concluded that HER-2/neu was not a major therapeutic target for the treatment of EWS. These results argue that HER-2/neu is not a critical pathway for the pathogenesis of EWS and targeting this pathway may have little therapeutic benefit in EWS patients.

18 VEGFR

Semaxanib is a dual VEGFR and PDFGR inhibitor that has been evaluated in neurogenic sarcoma cell lines and human tumor explants. Semaxanib has no effect on cell lines in vitro at doses up to 200 μM; however, it reduces proliferation in tumor explants by 54.8% in mice treated with 25 mg/kg/day for eight days. This growth reduction was due to decreased tumor angiogenesis, which led to decreased proliferation and increased apoptosis (131). These data suggest that VEGF may serve as a viable target in preventing angiogenesis in neurogenic sarcomas.

VEGFR inhibitors have been shown to be potent inhibitors of EWS growth in mouse models. Several inhibitors of VEGF signaling have been evaluated in EWS and have shown significant reduction of tumor growth in mouse models. The VEFGR inhibitors, SU6668 at 25 mg/kg/day and semaxanib at 100 mg/kg/day and anti-VEGF agents bevacizumab (avastin) at 10 mg/kg twice weekly and VEGF Trap at 2.5 or 25 mg/kg twice weekly significantly reduced tumor growth at clinically achievable doses

(132). Another VEGF signaling inhibitor, CEP-5214, and its pro-drug CEP-7055 inhibit tumor growth in a mouse angiosarcoma model. CEP-5214 administration at 1-3 mg/kg/day for 10 days achieved the minimum effective dose and the maximum efficacy was observed by treating with 23.8 mg/kg CEP-7055 for 10 days (133). These preliminary data demonstrate the anti-tumor activity of these drugs and offer evidence to further investigate the possible therapeutic potential of these inhibitors for the treatment of angiosarcoma.

Therapeutics that inhibit TK signaling through the use of antibodies are currently undergoing clinical trials in sarcomas. Bevacizumab, a recombinant VEGF monoclonal

19 antibody, has shown clinical promise by significantly increasing disease-free progression and overall survival in several malignancies including breast, renal cell and colorectal cancers (136). A phase II trial combining doxorubicin at 75 mg/m2 with 15mg/kg bevacizumab intravenously every three weeks was completed in patients with metastatic soft tissue sarcomas. A 12% response rate was observed in this study, and 65% of the patients experienced stable disease for four or more cycles (142).

A case-report of a nine-year-old boy demonstrated that bevacizumab has potential activity for treating alveolar soft part sarcoma. The patient was treated with bevacizumab intravenously for 1 hour at 5 mg/kg biweekly for four cycles, then increased to 10 mg/kg with no substantial side effects. He experienced a reduction of tumor load at all sites, including lung metastases and the patient was in good condition after 26 cycles of treatment (141). Additionally, there are several ongoing clinical trials investigating the use of bevacizumab as a possible therapy for sarcoma. The National Cancer Institute

(NCI) is sponsoring clinical trials evaluating bevacizumab as a single agent and in combination with sorafenib (another TKI) in Kaposi’s sarcoma. In addition, there is an open-label phase II trial evaluating bevacizumab as a single agent in the treatment of angiosarcoma (136). Further investigation of bevacizumab in combination with conventional chemotherapies for the treatment of metastatic sarcoma is therefore warranted.

20 Other TKIs

Several additional TKIs are currently under investigation. IGF-1R is a target of great interest for the treatment of sarcomas. The therapeutic potential of NVP-AEW541, an IGF-1R inhibitor, was evaluated in a panel of 8 OSA, 10 EWS and 5 RD cell lines.

NVP-AEW541 induced cell cycle arrest in a dose-dependent manner in the cell lines tested when treated with 300 nM, 1 μM and 3 μM for 24 hours. Apoptosis was induced in all of the EWS cell lines and in many of the OSA and RD cell lines, which correlated with IGF-1R inhibition. (134).

Perifosine targets cell membrane signaling pathways and has undergone phase I and phase II clinical trials in patients with advanced, metastatic solid tumors, including sarcomas. In the phase I study, 5 of 42 patients enrolled were sarcoma patients. One

LMS patient was the only patient of the 42 total to experience a partial response. Six months following therapy, the patient reported elevated energy levels and symptom relief, and she experienced no significant disease progression once off all anticancer therapies (138). A phase II clinical trial consisting of 16 adult patients with locally advanced or metastatic STS were treated with perifosine. There were no clinical responses observed in this study. Four patients had stable disease which lasted 1.3 to 8.2 months (137). The preliminary clinical data may have suggested potential therapeutic benefit from treatment with perifosine; however, the phase II trial does not provide compelling evidence to suggest a high degree of anti-tumor activity.

The Raf kinase inhibitor sorafenib is currently under investigation in National

Cancer Institutes sponsored studies for the treatment of metastatic and locally advanced sarcoma (136). Sunitinib (Sutent), a pan-TKI that elicits inhibitory effects on a variety

21 of TKs including PDGFR α and β and VEFGR 1-3 has recently been approved for

refractory GIST patients (136). A double-blind, phase III trial conducted to determine

the efficacy of sunitinib in patients with GISTs that failed imatinib treatment demonstrated that patients had a significant increased progression free duration of 6.3 months compared to 1.5 months in the placebo controls (127). Semaxanib, a reversible

VEGFR2 and PDGFR inhibitor, underwent phase II clinical trials investigating its use in patients with relapsed soft tissue sarcoma. Anti-tumor activity was encouraging with

23% observed responses when treated with 145 mg/m2 twice weekly (130).

There is considerable evidence to suggest that TKIs may serve as effective

treatments options for sarcoma either alone or in combination with other chemotherapies.

A protein kinase that serves as a point of convergence for many of these targeted TK

signaling pathways is Src kinase. Our preliminary data has shown that Src is

constitutively activated in sarcoma tissues. Therefore, Src may be a worthwhile target for

treating sarcomas.

22 History of Src: 1911- present

Nearly a century Peyton Rous discovered a “filterable agent”, or virus, that caused sarcoma in chickens (163, 164). This virus, later called Rous sarcoma virus (RSV), has remained in mainstream cancer research ever since its identification. However, it was some 30 to 40 years after its initial discovery that the scientific community would come to accept that idea that a virus could contain genetic information which induced tumors.

The road to establishing the foundation of identifying v-Src, the oncogene that causes the

RSV has been long. In 1941 it was shown that chicken embryo fibroblasts growing in culture infected with RSV adopted round and spindle-shaped morphology, similar in characteristic to Rous sarcoma cells (165). The focus assay was developed to quantify the number of cells infected with RSV in the early 1950s. By the late 1950s, RSV was shown to induce tumors in mammals by many groups (166-170). These findings later, proved significant in identifying the transforming ability of Src protein. The transformation abilities of RSV mutants were identified in the 1960s, in which mutations induced spindle-shaped cells, rather than rounded cells (171). This discovery, among others, lead to the 1970 identification of the oncogene responsible for cellular transformation of RSV, v-src. Then in 1977, a 60 kDa phosphoprotein immunoprecipitated from RSV tumor-bearing rabbits was identified as the src gene product (172). A year later two independent groups working on Src demonstrated that

Src was a protein kinase (173, 174) and more specifically, a tyrosine kinase in 1979 by

Tony Hunter (175).

The discovery that RSV required a specific gene for its transforming capability shed light on a new way of studying the role genes play in cancer. The idea of the

23 existence of “oncogenes”, transforming genes found in RNA tumor viruses, lead to the

idea that proto-oncogenes may exist as normal cellular counterparts. Proto-oncogenes

are precursors of viral transforming genes, which they themselves lack the ability to

transformation cells unless mutated or overexpressed. The first of these to be identified

was c-src (176, 177), the discovery of which initiated a fury of research into the role of

proto-oncogenes in cancer.

Upon sequencing c-src, it was revealed that the cellular and viral versions differ

in that v-src has carboxy-terminal substitutions and deletions, which were later linked to

the aberrant activity of v-Src protein (176). Upon cloning of the c-src gene lead to

comparing the biological activity of both forms of the gene. Except when overexpressed,

c-Src has lower transforming activity in mammalian and avian fibroblast cell lines (178-

180). Determining how the transforming ability of c-Src was activated lead to characterization of the c-src protein product. The human Src gene encodes a 536 amino acid protein, three more amino acids than the chicken ortholog.

It is now known that Src is one of 11 members of the human Src-family Kinases

(SFK), which also includes Blk, Brk, Frk, Fyn, Hck, Lck, Lyn, Srm and Yes (124). Fyn,

Src and Yes are expressed in all cell types, Srm is found in keratinocytes and Blk, Fgr,

Hck, Lck and Lyn are primarily found in hematopoietic cells. While Frk is expressed in bladder, brain, breast colon and lymphoid cells, and Brk is expressed in colon prostate and the small intestine (181).

SFKs share similar homology and protein organization. The organization of Src protein from N- to C- termini contains a 14-carbon myristoyl group, which facilitates the attachment of Src to membranes and is required for normal function, is attached to an

24 SH4 (Src Homology) domain. This unique domain is followed by an SH3 and SH2 domain, an SH2 kinase linker, a protein kinase domain containing the active site, and the

C-terminal regulatory domain (Fig. 2). The SH2 and SH3 domains have several important functions, including constraint of enzyme activity via intramolecular contacts during periods of inactivity and binding to ligands with SH2 and SH3 domains to attract them to specific locations within the cell. There are two vital tyrosine phosphorylation sites on Src, Tyr530 and Tyr419. When phosphorylated, Tyr530 binds intramolecularly to the SH2 domain, stabilizing the restrained enzyme and preventing interaction of SH2 and SH3 with other ligands (Fig. 3). In normal cells Src activity is tightly regulated, 90-

95% of Src is phosphorylated on Try530 and in an inactive state and becomes transiently activated during specific cellular event such as mitosis (182). Csk and Csk homologous kinase (Chk) are two cytosolic tyrosine kinases responsible for phosphorylation of

Tyr530. Tyr419 is an autophosphorylation site that is located in the active site of the enzyme and promotes Src kinase activity. Phosphorylation of Try419 stabilizes the active confirmation. Thus Tyr419 is the activation phosphorylation site and Tyr530 is the inhibitory phosphorylation site of Src which is missing from the v-Src protein (183).

When phosphorylated on Tyr530, Src cannot undergo autophosphorylation, Tyr530 must first be dephosphorylated. However, when phosphorylated on Tyr419 and Tyr530 simultaneously, the Tyr419 phosphorylation overrides the inhibitory Tyr530 and remains active (184).

Src activation is initiated by several cellular mechanisms including receptor tyrosine kinases, integrin receptors, cytokine receptors, G-protein receptors and steroid hormone receptors (185). Once activated, Src signals to the nucleus through a variety of

25 downstream effectors including STAT3 (186, 187), PI3K (188, 189) and Ras (190). Src

plays a role in a myriad of cellular functions including proliferation, survival, cell

adhesion, migration, cellular morphology and bone resorbtion. For example, under normal conditions, when epidermal growth factor receptor (EGFR) is activated, Src becomes activated during G2/M transition through the cell cycle. Src activity has also

been shown to be an important characteristic of osteoporosis. Mice with Src -/- null

mutations die within weeks of birth and express an osteoporosis phenotype, resulting

from increased bone resorbtion caused by a defect in osteoclasts (191). Thus, Src kinase

inhibitors currently serve as targeted therapy option for treating osteoporosis.

MB U SH3 SH2 Linker Catalytic Domain Reg c-Src Gly P Myr Tyr

H2N COOH 2 530 536 v-Src

Myr Gly P Tyr

H2N COOH 2 419 529

Figure 2. Organization of Src kinase. Structural features of c-Src compared to v-Src protein. Both proteins are myristoylated at Gly2 contain membrane binding (MB), unique (U), SH3, SH2, linker and catalytic domains. Notice only c-Src possesses the inhibitory Tyr530, or regulatory (Reg) domain at the C-termini. c-Src is usually phosphorylated on Tyr530 and in the inhibitory conformation, while v-Src is phosphorylated at Tyr419 and constitutively active because the regulatory domain is missing. Other differences between the two proteins include amino acid substitutions in v-Src. 26

Inactive Conformation Src Stimulation Active Conformation

SH3 BindingSignaling Protein Protein

SH3 SH2 Dasatinib Tyr 530 P Tyr 419 K P in i SH2 a n Tyr 530 om a D s Tyr 419 se e (BMS‐354825) ina D K o m a in Downstream Signaling Cascade

Figure 3. Src activation and signaling conformations. The left panel represents the inactive conformation of the c-Src protein. When phosphorylated, Tyr530 binds to the SH2 domain and prevents phosphorylation of the active site, Tyr419, while the SH3 domain interacts with the linker domain, further stabilizing the conformation. Upon activation of the active site the SH2 domain release Tyr530, SH3 releases the linker domain and they are now available to interact with other proteins containing SH domains. Furthermore, the active site is open or active, available to elicit signals (right side of panel).

Considerable evidence suggests that Src may serve as an option for targeted

therapeutics in cancer as well. Src kinase activity has been shown to be constitutively

activated in several types of human cancers including breast, prostate, colon, pancreatic

and just recently sarcoma as compared to normal tissues. For example, Src is overexpressed in ~70% of breast cancers (192). Furthermore, the increased activity of

Src is strongly implicated in the development, growth, progression and metastasis of cancer. Mutations in the c-Src gene are very rare, therefore overexpression may be induced by increased activation and association with RTKs. Reduced levels of Csk or

Chk activity and decreased phosphatase activity within the cell may also be potential mechanisms for increased Src activity in cancer (193).

The increased activity of Src in cancer induces dramatic alterations in signal transduction of downstream transcriptional events. Many of these downstream targets have been linked to tumorigenicity and metastasis (194). One widely studied pathway involved in tumorigenesis and downstream of Src activation is STAT3, a member of the

signal transducers and activators of transcription (STAT) family. STATs are mediators of Src induced cellular transformation. This model of evidence introduced in 1995, showed that the DNA-binding activity of STATs was enhanced in cells transformed with

v-Src (187). Upon activation, Src phosphorylates STAT3, which homodimerizes and

translocates to the nucleus where it regulates the transcription of genes involved in

cellular proliferation, survival and migration. It is not surprising considering that STAT3 is activated by Src to learn that STAT3, along with other STAT proteins, are constitutively activated in diverse human cancer cell lines and tissues. In cancer, STAT3

28 regulates the transcription of genes involved in uncontrolled tumor cell proliferation, resistance of apoptosis, evasion of the immune system and angiogenesis, among other malignant phenotypes (195). Consequently, STAT proteins and their activators have emerged as potential targets for cancer therapy (196, 197).

In addition to STATs, increased Src activity has been is associated with adhesion changes characteristic of mesenchymal transition and is suggested to promote cancer cell migration and metastasis. Focal adhesion kinase (FAK), along with Crk-associated substrate (CAS) are substrates of Src and serve as focal adhesion proteins vital for integrin signaling. FAK, also a tyrosine kinase, modulates the formation and turnover of focal adhesions, the intracellular structures that link the extracellular matrix to the actin cytoskeleton (198). Src and FAK associate upon autophosphorylation of the Tyr397 residue on FAK, after which Src becomes autophosphorylated and can then phosphorylate several residues on FAK, including Tyr576/577 and Tyr925.

Phosphorylation of FAK by Src enhances FAK activity.

Mutational studies have shown that the inactivation of FAK is embryonic lethal

(199) and selective inhibition of fak expression in skin keratinocytes suppressed chemically induced tumor formation in vivo (200). Others have shown that inhibition of fak expression induced apoptosis in vivo and in vitro in keratinocytes (200) and endothelial cells in vitro (201). Furthermore, the interplay between Src and FAK is important for modulating the gene expression of transcription factors and cell motility proteins, as elucidated in deletion and rescue experiments (202). Overexpression of FAK suppresses p53-mediated apoptosis by binding to the transactivation domain of p53 and limiting transcription activity (203). While, FAK itself does not function as an oncogene,

29 elevated expression in cancer correlates with increased cell motility, invasiveness and

proliferation (202, 204). RNAi-mediated knockdown of FAK in an aggressive breast

carcinoma cell line did not affect cellular proliferation, but cells lacking FAK lacked

invasive characteristics in vitro and lacked spontaneous lung metastasis in vivo (205). In

addition, FAK expression, along with Src, increases with colon cancer invasiveness

(206).

The FAK/Src complex also phosphorylates CAS, a cytoskeletal adapter protein.

Upon Src-mediated phosphorylation, CAS acts as docking protein for multiple protein –

protein interaction domains and is important for the recruitment of adapter proteins,

including Crk, to the substrate domain of CAS (207). CAS activity has been shown to

promote the invasiveness of Src-transformed cells (208) and play a vital role in cell

motility and survival. While the precise mechanism of CAS mediated apoptosis is

unknown, phosphorylation of CAS and its association with FAK have been shown to be essential in the survival pathway. Furthermore, expression of antisense CAS partially reversed the transformation activity of v-Src, suggesting that the inhibition of CAS plays a direct role in cellular transformation (209).

30 Development of Src inhibitors for clinical trials

Models have demonstrated that overexpression and activation of Src is associated with many oncogenic characteristics such as disruption of the cell cycle, increased migration and invasion and protection from apoptotic stimuli (193). Given the history of

Src and because it plays such a pivotal role in many oncogenic processes, it is surprising that Src has only recently been widely considered as a potential therapeutic target for cancer therapy. There are three variety of Src inhibitors; SH2/SH3 blocking inhibitors,

Src destabilizing agents and ATP-competitive inhibitions. SH2/SH3 blocking inhibitors prevent the SH domain mediated interactions and prevents Src protein-protein interactions. These inhibitors have demonstrated poor transport and uptake properties and only inhibit a subset of Src protein interactions, limiting the clinically efficacy of this class of Src inhibitors (210-212). Src destabilizing agents interfere with the association of

Src and its molecular chaperone Heat Shock Protein (HSP) 90. These agents have proven to be non-specific and results in the disruption of many HSP interactions, which could result in multiple adverse side-effects in patients (213, 214). ATP competitive Src kinase inhibitors target the active site and prevent ATP from binding and initiating the phosphotransferase activity of the enzyme. There is significant homology in the structure and sequence of ATP-binding domains of kinases, which has made the narrowing the specificity of ATP competitors difficult.

However, several ATP competitive inhibitors are currently under various stages of investigation as potential therapies for cancer. PP1 and PP2 were among the first Src kinase inhibitors utilized to study the role of Src activity in cellular events. The use of

PP2 has shown that inhibition of Src results in a loss of downstream Src signaling

31 through pathways including FAK, Akt and STAT3 and elicits anti-tumorigenic

phenotypes (193). These findings were later proven to be through Src inhibition using

siRNA to knockdown Src expression (215). Since the discovery that inhibition of Src can lead to anti-tumorigenic effects in cancer models several ATP-competitive Src kinase

inhibitors have been synthesized and studied. Three such inhibitors include PD180970

(Pyridol[2,3-d] pyrimidine), SKI-606 (4-anilino-3-quinolinecarbonitrile) and dasatinib

(BMS-354825, [N-(2-chloro-6-methylphenyl)-2-(6-(4-(2-hydroxyethyl) piperazin-1-yl)-

2-methylpyrimidin-4-ylamino)thiazole-5-carboxamide) (Fig. 4), all of which have been thoroughly studied in epithelial and hematopoietic cancers.

PD180970 was the first of the three to undergo investigation as a Src kinase inhibitor. Several groups have shown that PD180970 can inhibit Src at nanomolar doses

(216-219). Most significantly, others have shown that the inhibition of Src-STAT3 signaling in cancer cells can lead to the induction of apoptosis (216, 219). However it is a non-specific kinase inhibitor, only soluble in DMSO and is highly unstable, which leaves this compound with little to no clinical future. SKI-606 is an orally available compound with dual specificity for SFKs and Abl tyrosine kinase (220). Studies have shown that SKI-606 decreases growth and motility of colorectal cancer cells by preventing activation of Src and downstream signaling (221). Furthermore, SKI-606 has been shown to have anti-proliferative effects in colon cancer and myelogenous leukemia cancer models (220, 222). These data have lead to the initiation of clinical trials in patients with hematopoietic and epithelial cancers. Dasatinib, is also an orally available

Src inhibitor that is currently undergoing clinical trial in epithelial cancers. Preclinical data have shown dasatinib to be a potent inhibitor of Src activity in prostate, head and

Figure 4. Structure of three ATP-competitive Src kinase inhibitors. A, PD180970, B, SKI-606 and C, dasatinib were independently developed and provided by three different pharmaceutical companies. All three inhibitors act as ATP-competitors for the catalytic domain of Src kinase and prevent the activation of Src by phosphorylation of Tyr419.

33 neck and lung cancers (223-225). Nam et al., showed dasatinib inhibition of Src at

nanomolar concentrations inhibited downstream signaling of STAT3 and FAK which

subsequently inhibited the migration and invasive properties of prostate cells (223).

While others have shown that in addition to inhibiting migration and invasion, dasatinib also induces cell cycle arrest and apoptosis in head and neck (224) and lung cancer cell lines (225).

There is remarkable evidence to suggest that Src kinase activity may serve as an important target for the treatment of cancer. Until the culmination of this dissertation research, the role of Src activation in human sarcoma had not yet been elucidated. In addition, the responses and mechanism of action of ATP-competitive Src kinase inhibitors in mesenchymally-derived tumor cell lines have not been described previously.

There is considerable evidence to suggest a role for Src in sarcomagenesis. Vigneron et

al., have shown that increased Src expression in sarcoma cell lines reduced sensitivity to

conventional chemotherapies used to treat sarcomas. This chemo-resistance was

associated with a failure of cells to up-regulate p21 in response to adriamycin (226).

Furthermore, increased Src activity has also been shown to contribute to anoikis

resistance in human OSA cell lines. Inhibition of Src activity restored anoikis sensitivity

and cells underwent apoptosis upon detachment (227). Thus, these three independently

synthesized compounds were evaluated in sarcoma cell lines to identify the role of Src

kinase activation in the sarcoma cell lines. The objective of this dissertation was to

determine the role of Src activation in sarcomas and determine whether Src may serve as a viable option for the treatment of sarcoma patients.

Objectives

The overall objective of the studies conducted for this dissertation was to

investigate the role of Src kinase activation in the survival and malignant phenotype of sarcomas. Src kinase was the original oncogene identified in Rous Sarcoma Virus (172).

Previous research has suggested a possible role for Src activation in tumor survival, in that Src is involved in a myriad of cellular mechanisms such as survival, migration and invasion. However, the role of Src in human sarcomas had not been determined.

Results from preliminary experiments suggested that Src kinase is activated in a variety of human sarcomas. The recent development and interest of Src kinase inhibitors aided in evaluating Src as a possible target of sarcoma treatment. These findings in addition to the overwhelming evidence suggesting that Src may serve a possible target essential for the survival of sarcomas lead to the development of the hypothesis that Src is required for the migration, invasion and survival of sarcomas. To verify this hypothesis, the following aims were pursued.

Aim 1: Evaluate the biological response of three Src kinase inhibitors in sarcoma cell lines.

A. To determine the role of Src kinase activation in sarcomas cells lines. Src

kinase has been shown to be activated to varying degrees in primary sarcomas.

However, the degree of activation and significance of Src activation in sarcoma

35 cell lines has not been determined. Because Src was originally identified as an

oncogene significant in the development of sarcoma in chickens, we sought to

determine to role of Src kinase activity in sarcoma cell lines using three different

Src kinase inhibitors.

B. To determine if STAT3 activation is required Src activated migration,

invasion and survival of sarcomas. STAT3 is among multiple downstream

signals initiated by Src activation. STAT3 was evaluated as a possible signaling

protein involved in the cellular events elicited by Src activation in sarcoma cell

lines.

Aim 2: Identify a molecular signature that predicts response to dasatinib by induction of apoptosis in sarcoma cell lines.

A. Identify a gene expression profile unique to sarcoma cell lines that undergo

apoptosis when treated with dasatinib. Src kinase has previously been shown

to be essential for tumor survival in epithelial derived tumors (185, 207, 228-230).

The data from aim 1 has shown that Src kinase activity is required for survival in

a subset of sarcoma cell lines. Therefore, microarray analysis was performed on

sarcoma cell lines to identify a molecular signature that predicts response to

dasatinib by induction of apoptosis in sarcoma cells.

B. Confirm the predictability of the molecular signature to predict response to

dasatinib in an independent dataset. It is hypothesized that a molecular

signature may be identified in tumors and tumor cell lines which may predict the

response to specific therapies. To test the effectiveness of the molecular

36 signature that predicts response to dasatinib two new cell lines of unknown

response status were utilized. The cell lines were used to determine whether the

molecular signature extracted from the responsive cell lines was successful in

identifying new cell lines that would or would not respond to dasatinib by

induction of apoptosis.

Aim 3: Verify the presence of the molecular signature in primary human sarcoma specimens.

A. Determine if the molecular signature identified in the sarcoma cell lines is

present in untreated, primary human sarcoma specimens. Cell lines lack the

microenvironment interactions of human tumors and one could not be certain that

the molecular signature identified in the cell lines is expressed in human tumors.

Untreated sarcoma specimens were utilized to establish whether the molecular

signature unique to the cell lines that respond to dasatinib was present in human

tumors. Microarray analysis was performed on a diverse set of human sarcomas

to validate the expression of the cell line signature in tumors.

B. Establish whether the molecular signature can cluster sarcomas based on

expression and theoretically predict response to dasatinib. Once the

molecular signature was identified in primary sarcomas, the signature was used to

predict theoretical response to dasatinib. The molecular signature was further

tested to determine if tumors clustered into the respective theoretical categories

based on the expression of the molecular signature.

Materials and Methods

Cells and reagents

SaOS-2, U 2 OS, MG-63, SK-ES-1, A673, RD, SK-LMS-1, HT-1080, SW-872,

HOS and SW1353 sarcoma cell lines were obtained from the American Type Culture

Collection. The LM2 and LM7 OSA cell lines were provided by Dr. Eugenie S.

Kleinerman of MD Anderson Cancer Center (Houston, TX). The TC-71 EWS cell line was provided by Dr. Timothy Triche of the University of Southern California (Los

Angeles, CA). The SaOS-2, MG-63, LM2, LM7, TC-71, SK-LMS-1, HT-1080 and SW-

872 cell lines were maintained in MEM supplemented with Eagle’s salts, 10% fetal bovine serum, 2-fold MEM vitamins, 1 mM sodium pyruvate, 1 mM non-essential amino acids and 2 mM L-glutamine. The SK-ES-1 and U 2 OS cell lines were maintained in

McCoy’s 5A medium supplemented with 10% fetal bovine serum. The A673 cell line was maintained in DMEM supplemented with 10% fetal bovine serum. The RD and

RD18 cell lines were maintained in DMEM/F12 (1:1) supplemented with 10% fetal bovine serum. All cells were maintained at 37°C in 5% CO2. All experiments were

performed using exponentially proliferating cells unless otherwise noted.

Polyclonal antibodies to phosphorylated p-Src (Y419), p-FAK (Y576/577), p-

p130CAS (Y410), pSTAT3 (Y705) proteins, and to total FAK and PARP proteins were

obtained from Cell Signaling Technologies (Cambridge, MA). Polyclonal antibodies to

p130CAS and β-actin were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

Monoclonal antibodies to total Src (clone GD11) and hILP/XIAP were obtained from

38 Upstate Biotechnology (Lake Placid, NY) and BD Biosciences (San Diego, CA),

respectively. Dasatinib was provided by Bristol-Myers Squibb Pharmaceutical Research

Institute (Princeton, NJ). Dasatinib was synthesized by the addition of methylprimidine to

the 2-amino group of thiazole, followed by a reaction with hydroxyethyl piperazine

(231).

Preparation of cell extracts and Western blotting

Adherent sarcoma cells were washed with ice cold 1x PBS, followed by washing

and scraping from plates in ice cold 1x PBS containing 5 mM sodium fluoride and 1 mM

sodium orthovanadate (1x PBS + inhibitors). Cells were harvested by centrifugation,

washed with 1x PBS + inhibitors and harvested again by centrifugation. Pellets were

resuspended in RIPA buffer containing 0.1 mM Na3VO4, NaF, DTT and 1:100 fold

dilution of protease inhibitor cocktail obtained from Sigma (St. Louis, MO). Samples

were vortexed for 30 min at 4°C and insoluble material was removed by centrifugation.

For Western blotting, 50 μg of protein was resolved by SDS-PAGE and transferred to nitrocellulose membranes. Blots were blocked in 1x TBS containing 0.01% Tween-20 and 5% non-fat dry milk for 30 min, and then incubated with primary antibody in 1x TBS containing 0.01% Tween-20 and 5% BSA obtained from Sigma (St. Louis, MO) overnight at 4°C with rocking. Protein-bound primary antibodies were detected using respective horseradish peroxidase-coupled secondary antibodies (Amersham anti-rabbit

for polyclonal and anti-mouse for monoclonal, obtained from GE Healthcare Unlimited,

Buckinghamshire, UK) diluted 1:10,000 in 1x TBS containing 0.01% Tween-20 and 5%

non-fat dry milk and incubated for 2 h at room temperature. Bound secondary antibodies

39 were detected using Amersham ECL PLUS Western blotting detection reagents obtained

from GE Healthcare Unlimited. Densitometry was performed on p-Src (Y419) Western

blots performed for dose responses in each of the cell lines and then analyzed using

ImageQuant 5.2 (Molecular Dynamics) software. Percent inhibition of Src kinase

activity as measured by Src (Y419) phosphorylation was determined by nonlinear

regression analyses and data were reported as the inhibitory concentration required to

achieve 50% inhibition relative to control reactions (IC50) in Table 1. Data are the

averages of triplicate determinations.

Immunohistochemistry

Human tissues were obtained through the Moffitt Cancer Center Tumor Bank

using IRB-approved protocols. Tissues were fixed in formalin within 15-20 min from the

moment of surgical excision to preserve the phosphorylation status of proteins such as

Src and STAT3. Formalin-fixed, paraffin-embedded tissue sections of 3 μm thickness

were deparaffinized by an initial warming to 60°C, followed by two xylene changes of 10

min each, two series of 30 dips in absolute alcohol, 30 dips in 95% alcohol, and 20 dips

in deionized water. Slides were placed for 5 min in TBS/Tween and processed on a

DAKO Autostainer using the Dako LSAB+ peroxidase detection kit (Carpinteria, CA).

Endogenous peroxidase was blocked with 3% aqueous hydrogen peroxide followed by 20 dips in deionized water. The anti-p-Src (Y419) or pSTAT3 (Y705) was applied at 1:100 dilution for 30 min after microwave antigen retrieval with 0.1 mol/L citrate buffer (pH

6.0; Emerson 1,100 W microwave, high to boiling, then 20 min on power level 5). The

chromogen 3,3'-diaminobenzidine was used for detection. Counterstain was done with

40 modified Mayer's hematoxylin. Slides were dehydrated through graded alcohol, cleared with xylene, and mounted with resinous mounting medium.

Src Family Kinase PCR

A multiGene-12 RT-PCR Profiling Kit was obtained from SuperArray Bioscience

Corporation to determine the status of gene expression for 11 SFK members in eight

sarcoma cell lines; SaOS-2, LM-2, LM-7, U2 OS, MG-63, SK-LMS-1, HT1080 and

SW872 cell lines. RNA was isolated from exponentially growing cells using TRIzol

Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. First

strand cDNA synthesis was completed using the SuperScript First Strand Synthesis

System for RT-PCR (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol.

The remainder of the RT-PCR protocol was completed according to the SuperArray

protocol. Once resolved in a 1% agarose gel, photographs were taken for each of the cell

lines to compare gene expression.

Wound healing assay for cell migration

Cells were plated in 24 well tissue culture plates, grown to confluency and serum

starved in 0.1% FBS overnight. Monolayer wounds were produced using a pipette tip

scratched through the center of the well. Photomicrographs were taken of the initial

wound for comparison. Cells were then treated with either DMSO alone, as vehicle

control or escalating doses of dasatinib and allowed to migrate into the denuded areas for

24 h. Following incubation, cells were briefly stained with Coomassie blue. Cell

migration was visualized at 10x magnification and digitally photographed. The distance

41 of migration was measured as pixel units and compared to time zero. The average

number of pixel units measured in the denuded area were determined from wound healing assays for dose responses in each of the cell lines analyzed. Percent inhibition of migration was determined by nonlinear regression analyses and data were reported as the inhibitory concentration required to achieve 50% inhibition relative to control reactions

(IC50). Data are the averages of triplicate determinations.

Cell invasion assay

Cell invasion assays were performed following the BioCoat Matrigel Invasion

Chambers protocol, obtained from BD Biosciences (Bedford, MA). Briefly, cells were

trypsinized and washed once with 1x PBS and twice using serum-free medium. Cell suspensions were prepared at 5 x 104 cells/mL in 0.5 mL containing DMSO alone as

vehicle control or escalating doses of dasatinib in serum-free medium and added to the

chamber insert. Chambers were incubated for 22 h at 37 °C and 5% CO2. Following

incubation, non-invading cells were removed using a cotton-tipped swab. The cells that

invaded to the lower surface of the membrane were stained using a Diff Quick Staining

Kit from Fisher Scientific (Pittsburgh, PA), digitally photographed and counted. Each

experiment was completed in triplicate.

TUNEL assay

Apoptosis was detected by the TUNEL assay using the In Situ Cell Death

Detection Kit obtained from Roche Molecular Biochemicals (Indianapolis, IN) according

42 to the manufacture’s protocol. Cells were treated with either DMSO alone as vehicle

control or escalating doses of dasatinib for 48 h.

siRNA transfections

siRNA directed specifically against c-Src and a non-targeting siRNA control were obtained from Dharmacon RNA Technologies (Chicago, IL). Cells were plated on 6 cm tissue culture plates in complete media (5 x 105 cells per plate) and allowed to attach

overnight. siRNA was transfected in escalating doses (50 nM and 100 nM) using

Oligofectamine obtained from Sigma-Aldrich (St. Louis, MO). The transfection

incubation time for the siRNA/Oligofectamine complexes was 24 h, and total incubation

time before harvesting cell lysates was 72 h.

Statistical analysis

Descriptive statistics, such as mean values and standard deviation, were calculated

for the biological effects of dasatinib on invasion by dose levels (nM). To determine

statistical significance between pair-wise dose levels, the exact Wilcoxon two-sample test

or T-test were used, considering the small sample sizes. One-sided tests at a significance

level of 0.05 were examined. All data were analyzed using the SAS software (version 9.1,

SAS Institute, Cary, NC).

Microarray sample preparation

Three consecutive passages of untreated, exponentially growing SaOS-2, LM2,

LM7, U 2 OS, MG-63, SK-ES-1, SK-LMS-1, HT1080, A673, RD, RD|18, HOS and

43 SW1353 were utilized for microarray analysis. Sarcoma tissue was collected from the

H. Lee Moffitt Tumor Bank. Tissues were untreated primary malignancies that were

snap-frozen in liquid nitrogen within 15 minutes following removal from the patient to

minimize cellular degradation.

Isolation of RNA

Total RNA was excised from sarcoma tissue specimens and cell lines using the

TRIzol Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol.

The aqueous phase containing the RNA separated from the TRIzol reagent was further

purified using the RNeasy clean-up procedure (Qiagen Inc., Valencia, CA). The quality

of total RNA was assessed by agarose gel electrophoresis and the total RNA from the

tissues was further analyzed on a Agilent 2100 Bioanalyzer.

Preparation of labeled RNA targets for hybridization

Within the total RNA pool the poly(A) RNA was specifically converted to cDNA

and then amplified and labeled with biotin following the procedure initially described by

Van Gelder et al. (232). First-strand cDNA synthesis was carried out using the

Superscript Choice System (Invitrogen, Carlsbad, CA) and the T7 promoter/oligo (dT)

primer (5’-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGG-(dT)24-3’),

(Genset Corp., La Jolla, CA). Following annealing the rest of the cDNA synthesis reaction was prepared such that the final reaction contains 5 μg RNA, 100 pmol T7-(T)24 primer, 500 μM each dNTP, 10 mM DTT, 50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM

MgCl2, and 200 U of Superscript II reverse transcriptase (Invitrogen Corporation, 44 Carlsbad, CA). The reaction was incubated for 1 hr at 42°C. A second-strand cDNA

synthesis was performed at 16°C for 2 hr in a total volume of 150 μL, using 10U of

E.coli DNA ligase, 40 U of E. coli DNA polymerase I, and 2 U of E. coli RNase H in the

presence of 200 μM of each dNTP, 10 mM (NH4)SO4, 1.3 mM DTT, 26.7 mM Tris-HCl,

+ pH 7.0, 100 mM KCl, 5 mM MgCl2, and 150 μM β-NAD (Invitrogen). Following the

second-strand DNA synthesis, 10 U of T4 DNA Polymerase (Invitrogen) was added and the samples incubated an additional 5 min at 16°C. The reaction was stopped by the addition of 0.5 M EDTA and subsequently extracted with an equal volume of phenol/chloroform/isoamyl alcohol. The double-stranded DNA (dsDNA) will then be precipitated with the addition of 0.5 volumes of 7.5 M NH4 Acetate and 2.5 volumes of

ice-cold 100% ethanol. The dsDNA then serves as a template for a transcription reaction performed with the GeneChip IVT Labeling kit according to manufacturer's instructions

(Affymetrix Corp., Santa Clara, CA) which incorporates biotinylated UTP into the transcripts. The Biotin-labeled RNA was purified using RNeasy columns (Qiagen) and fragmented to a size of 35 to 200 bases by incubating at 940 C for 35 minutes in

fragmentation buffer (40 mM Tris-acetate, pH 8.1/100 mM potassium acetate/30 mM magnesium acetate). The integrity of the starting material and the products of each

reaction were monitored on agarose gels to assess the size distribution of the products and

compare them to the starting material.

45 Array hybridization and scanning

The hybridization solution consisted of 20 µg of fragmented RNA and 0.1 mg/ml sonicated herring sperm DNA, in 1x MES buffer (containing 100 mM MES, 1 M Na+, 20 mM EDTA, and 0.01% Tween 20). In addition the hybridization solutions were spiked with known concentrations of RNA from the bacterial genes, BioB, BioC, and BioD, and one phage gene, Cre, as hybridization standards. The hybridization mixtures were heated to 99°C for 5 min followed by incubation at 45°C for 5 min before injection of the sample into a probe array cartridge. All hybridizations were carried out at 45°C for 16–17 h with mixing on a rotisserie at 60 rpm. Following hybridization, the solutions were removed and the arrays were rinsed with 1x MES. Subsequent washing and staining of the arrays was carried out using the GeneChip Fluidics station protocol EukGE_WS2, which consists of 10 cycles of 2 mixes per cycle with non-stringent wash buffer (6x SSPE,

0.01% Tween 20) at 25°C followed by 4 cycles of 15 mixes per cycle with stringent wash buffer (100 mM MES, 0.1 M Na+, and 0.01% Tween 20) at 50°C. The probe arrays were then stained for 10 min in streptavidin-phycoerythrin solution (SAPE) [1x MES solution,

10 µg/ml SAPE (Molecular Probes, Eugene, OR), and 2 µg/µl acetylated BSA

(Invitrogen)] at 25°C. The post-stain wash was 10 cycles of 4 mixes per cycle at 25°C.

The probe arrays were treated for 10 min with an antibody solution [1x MES solution, 2

µg/µl acetylated BSA, 0.1 µg/µl normal goat IgG (Sigma Chemical, St. Louis, MO), 3

µg/µl biotinylated goat-anti-streptavidin antibody, (Vector Laboratories, Burlingame,

CA)] at 25°C. The final wash consisted of 15 cycles of 4 mixes per cycle at 30°C.

Following washing and staining, probe arrays were scanned once at 1.5-µm resolution using the Affymetrix GeneChip Scanner 3000. 46

Data analysis

Scanned output files were visually inspected for hybridization artifacts and then analyzed by using Affymetrix Micro Array Suite (MAS) 5.0 software. Arrays were scaled to an average intensity of 500 and analyzed independently. The MAS 5.0 software uses a statistical algorithm to determine the signal intensity of a transcript from the behavior of

11 different oligonucleotide probes designed to detect the same gene (233). Probe sets that yield a change p-value less than 0.05 were identified as changed. Gene changes were selected using one of two methods the Significance Analysis of Microarrays (SAM) technique of Tusher et al. (234) or by filtering the probe sets to contain only probe sets with a maximum/minimum intensity value ratio of less then 2.0 for responders and non- responders. Probe sets that fit both criteria were further examined for use in the molecular signature.

To test the reliability of the signature and determine the response status of two cell lines with unknown response to dasatinib the fold change of the probe sets that comprise the molecular signature was determined as compared to either the median intensity for each probe set for all samples as the comparison group. A decision on response was determined based on how the fold change compared to the fold changes for the molecular signature in the original dataset.

Hierarchical clustering analysis and data transformation for microarray analysis were performed using Cluster version 2.11 (235). Microarray data visualization using

47 heatmaps was prepared using Java Treeview (236), where red depicts an increased expression and green depicts a decreased expression in a given gene.

Preparation of samples for Quantitative Real-Time PCR analysis

For src analysis SaOS-2, U 2 OS and SK-LMS-1 cells were treated with 30, 100,

300 and 1000 nM of dasatinib or DMSO for six hours. Cells were washed with ice cold

1x PBS, RNA was isolated using the TRIzol Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. The aqueous phase containing the RNA separated from the TRIzol reagent was further purified using the RNeasy cleanup procedure (Qiagen Inc., Valencia, CA). The quantitative Real-Time PCR analysis for histone H3, FAF1,

α-catenin, δ-catenin, ephrin- A1 and dapper was completed using the same RNA samples used for the microarray analysis.

cDNA reactions

Reverse Transcriptase reactions were random hexamer-primed using Applied

Biosystems’ (Foster City, CA) High Capacity cDNA Archive Kit. (All Reverse

Transcriptase reactions were done at the same time so that the same reactions could be used for all gene studies.) For the construction of standard curves, serial dilutions of pooled sample RNA were used (50, 10, 2, 0.4, 0.08, and 0.016 ng) per reverse transcriptase reaction. One “no RNA” control was included. Additionally, one ‘no reverse transcriptase’ control was included for the standard curve and for each sample.

48 Real-Time PCR reactions

TaqMan® Gene Expression Assays (Applied Biosystems) were used. The assay primer & probe sequences are proprietary. The probe is labeled with 6-carboxy- fluorescein as the reporter on the 5’ end, and a non-fluorescent quencher plus a minor- groove binder on the 3’-end. Each assay is supplied as a 20X mix of primers and probe.

Real-time quantitative PCR analyses were performed using the ABI PRISM 7900HT

Sequence Detection System (Applied Biosystems). All standards, the no template control

(H2O), and the no amplification control (Bluescript plasmid) were tested in quadruplicate wells (2 wells/plate x 2 plates). All samples were tested in triplicate wells. The no RT controls were tested in duplicate wells. PCR was carried out with the TaqMan Universal

PCR Master Mix (Applied Biosystems) using 2 µl of cDNA and 1X primers & probe in a 20-µl final reaction mixture. After a 2-min incubation at 50°C, AmpliTaq Gold was activated by a 10-min incubation at 95°C, followed by 40 PCR cycles consisting of 15 s of denaturation at 95°C and hybridization of probe and primers for 1 min at 60°C. Data were analyzed using SDS software version 2.2.2 and exported into an Excel spreadsheet.

The 18s data were used for normalizing the gene values - ng gene/ng 18s per well.

Personal health identifier

Tissues from patients treated at the Sarcoma clinic and collected by the tissue procurement lab at the H. Lee Moffitt Cancer Center were used for this study. The specimens were handled and stored by the tissue procurement lab under the supervision of Dr. T. Hoover. Specimens were released for RNA extraction and microarray analysis

49 after the removal of all PHIs. The specimens were coded and could not be linked to the

donor any individual involved in the analyses of the said samples.

Involvement of human subjects

No procedures were performed on patients. Only pathologic material that was previously collected for research purposes and that was not necessary for a complete pathologic diagnosis was utilized in this study. A separate informed consent beyond what is routine for surgery at H. Lee Moffitt Cancer Center was not necessary.

Introductory Preclinical Data

Studies have demonstrated that STAT proteins participate in essential cellular

functions including cellular immune function, development, differentiation and proliferation (237-242). STATs are a family of latent cytoplasmic transcription factors that associated with RTKs and NRTKs. Upon association with activated tyrosine kinases and activation by tyrosine phosphorylation they can either homdimerize or heterodimerize with other activated STAT proteins via SH2 domain-phosphotyrosine interacting with other STAT family members. Activated STAT dimers translocate to the nucleus and regulate gene transcription by binding to elements within gene promoters

(243, 244).

An accumulation of evidence has demonstrated a critical role of STAT3 in the molecular pathology of cancer, including tumor formation (245, 246). Constitutive activation of STATs has been shown to play a significant role in the malignant transformation of cells (245, 247, 248). Specifically, constitutive activation of STAT3 is associated with transformation by v-Src and other viral oncoproteins that activate

tyrosine kinase signaling pathways (187, 217, 249, 250) and is required for the

transcription of genes involved in v-Src-induced cellular transformation (186, 251).

STATs have been demonstrated to be constitutively activated in epithelial and

hematopoietic human cancers. However, despite overwhelming evidence to suggest a

potential role of STATs in sarcomagenesis, few direct studies have been completed to 51 evaluate the role of STAT3 activation in human sarcomas. The most significant evidence to imply a role for STATs in sarcomagenesis is the overexpression and/or mutation of several TKs in sarcomas. Recall that PDGFR, c-KIT, IGFR-1 and VEGFR have been shown to be overexpressed in many sarcomas of various histologies. STATs are involved in transducing TK signaling for all of these pathways under normal and malignant conditions. Furthermore, Src, another essential protein involved in activating STATs and eliciting the activation of the aberrantly activated TKs in sarcomas also activates STAT3.

Src was originally identified as the viral oncoprotein responsible for RSV, however the role of Src has yet to be established in human sarcomas. Therefore, the evaluation of the

Src-STAT3 signaling pathway in human sarcoma cell lines was initiated.

Signaling by STAT proteins is associated with activation of Src tyrosine kinase

(187, 249), thus Immunohistochemistry (IHC) was utilized to determine activation status of STAT3 in human sarcoma specimens and STAT3 DNA-binding activity in nuclear extracts from human sarcoma cell lines. Pharmacological inhibitors of Src were used to evaluate its role in STAT3 activation in human sarcoma cell lines.

52 Results

Expression of SFKs in sarcoma cell lines

RT- PCR was performed to determine which SFKs were expressed in eight sarcoma cell lines. Of the 11 recognized SFK members, three including Src, Fyn and

PTK2 were expressed in all eight cell lines examined (Fig. 5A-H; Table 5). Yes1 was expressed in all but the HT-1080 cell line (Fig. 5G). Hck was not expressed in any of the cells lines, while only barely detectable levels of Lck were expressed in U 2 OS cells

(Fig. 5D). The remaining SFK members were expressed in some cell lines, but not others without an apparent preference for tumor type, i.e. STS or OSA.

STAT3 is activated in human sarcomas and sarcoma cell lines

Immunohistochemistry for the activated form of STAT3 protein was performed on human sarcoma specimens using antibodies specific for phosphorylated STAT3

Tyr705 (pSTAT3). STAT3 is activated upon phosphorylation of Tyr705, which induces dimerization, translocation to the nucleus and DNA binding (248). Results show activated STAT3 in sarcomas of diverse subtypes (Fig. 6), including high grade OSA (A), pleomorphic LPS (B) and pleomorphic, undifferentiated high grade sarcoma (C). Since activation of STAT3 was found in the majority of sarcoma tissues examined (Fig. 6D), we determined the status of STAT3 activation in seven sarcoma cell lines using EMSA.

Results show STAT3 was activated in all seven cell lines to varying degrees (Fig. 6E);

MG-63, an OSA cell line, had the greatest and HT1080, a FBS cell line, had the least amount of STAT3 activation.

1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 A E

B F

C G

D H

Figure 5. Evaluation of SFK expression in human sarcoma cell lines. The expression of SFKs were determined in eight sarcoma cell lines using the SFK Superarray multiGene- 12 RT-PCR Profiling Kit. The kit was performed on eight cell lines and the PCR products were resolved on 1% agarose gels for SaOS-2 (A), LM-2 (B), LM-7 (C), U 2 OS (D), MG-63 (E), SK-LMS-1 (F), HT1080 (G) and SW-872 (H). The lane numbers correspond to the specific SFK for that well and the numbering system for can be found on Table 5.

Table 5. SFK gene expression status in human sarcoma cell lines

SFK Gene Expression Status GAPD (12) PTK2B (8) YES1 (11) PTK2 (7) SYK (10) HCK (4) HCK LCK (5) LYN (6) BLK (1) FGR (2) FYN (3) Cell Line SRC (9)

SaOS-2 x x x x x x x LM-2 x x x x x x x LM-7 x x x x x x x x U 2 OS x x x x x x x x MG-63 x x x x x x x x x SK-LMS x x x x x x x x HT1080 x x x x x x SW872 x x x x x x x x x

Table 5. PCR was performed for 11 SFK family members in eight sarcoma cell lines. The presence of an ‘x’ on the table denotes expression of the corresponding SFK. The SFK genes are listed at the top of the table, the number in parentheses corresponds to the lane number in Figure 6; panels A-H.

A B C

140 160 Osteosarcoma 200 240 270 120 Chondrosarcoma 270 5 EWS 20 120 5 MPNST 140 40 Angiosarcoma 80 30 40 60 Undifferentiated 140 High Grade 160 Sarcoma 160 180 240 270 5 40 60 60 80 Liposarcoma 100 100 150 180 240 10 Leiomyosarcoma 20 80

0 50 100 150 200 250 300

pStat3 Expression Index

E SW872 HT1080 M-G63 SaOS-2 U 2 OS LM-2 SaOS-2 Supershift SK-LMS-1

STAT3

Figure 6. STAT3 activation status in human sarcoma tissues and cell lines. A-C, immunohistochemistry for p-STAT3 (Y705) reveals that Src is activated in human sarcoma tissues; high grade OSA (A), pleomorphic LPS (B) and pleomorphic, undifferentiated high grade sarcoma (C). D, STAT3 is activated in several diverse sarcomas, this graph depicts the average intensity of STAT3 phosphorylation as measured by IHC. E, Cell-free extracts were prepared from untreated cells grown in 10% FBS and EMSA was performed to determine the activation status of STAT3 in several sarcoma cell lines. Supershift was completed on all cell lines in previous experiments to prove that STAT3, specifically, is activated in the cell lines. Supershift was only completed on SaOS-2 in lane one of this experiment and is denoted by the *.

57 PD180970 inhibits Src and STAT3 signaling in human sarcoma cell lines

PD180970 has previously been shown to directly inhibit the kinase activity of purified Src protein in vitro with an IC50 of 16.8 nM (252). To evaluate the effect of

PD180970 on Src kinase activity in intact sarcoma cells, we treated the SK-LMS-1,

SaOS-2, LM2 and U2 OS cell lines with escalating doses of PD180970 (125, 250 and

500 nM) for 24 h. Western blot analysis was performed to evaluate p-Src (Tyr419) levels and EMSA was performed to determine the effect on STAT3 activation. Dose-response results for p-Src are shown in Figure 7A for a representative cell line (SK-LMS-1) and

STAT3 activation for SK-LMS-1, SaOS-2 and LM2 are shown in Figure 7B. Src phosphorylation is diminished with 125 nM and completely inhibited with 250 nM

PD180970. The kinetics of STAT3 inhibition is not as straightforward as the Src response to PD180970. EMSA shows that STAT3 activation is partially inhibited with all doses of PD180970 in three of the four cell lines displayed (Fig. 7B). SaOS-2 is the most sensitive with STAT3 inhibition markedly decreased with 125 nM PD180970. SK-

LMS-1 and LM2 demonstrate maximum inhibition of STAT3 activation with 500 nM

PD180970. U 2 OS cells had no response to PD180970 as measured by Src phosphorylation and STAT3 activation. The activity of Src (data not shown) and STAT3 remained constitutively activated at all doses of PD180970 (Fig. 7B). Time course experiments were performed to determine the inhibitory kinetics of PD180970 on STAT3 activation. As shown in representative results with SK-LMS-1 cells, inhibition of STAT3 activation is reduced at 3 h and is completely inhibited at 24 h with 500 nM treatment of

PD180970 (Fig. 8A). However, STAT3 activation increases at 36 h and later time points, although activation does not return to untreated levels.

A PD180970 125 nM 250 nM 500 nM DMSO

pSrc Tyr419

ß‐Actin

B PD180970 DMSO 125 nM 250 nM 500 nM

SK-LMS-1

SaOS-2

LM-2 U2 OS

Figure 7. PD180970 inhibits Src and STAT3 signaling in sarcoma cell lines. A, SK- LMS-1 cells were treated with PD180970 in a dose-dependent manner for 6 h. Cell-free extracts were immunoblotted with antibodies specific to p-Src (Y419) and β-actin. B, SK- LMS-1, SaOS-2, LM-2 and U 2 OS cells were treated with PD180970 in a dose response for 24 h. EMSA was performed to evaluate STAT3 activation. Western blot analysis and EMSA were performed as described. DMSO was used as a vehicle control in all experiments.

59 PD180970 inhibits cell viability and induces apoptosis in sarcoma cell lines

To determine the effect of PD180970 on sarcoma cell survival over time, we performed growth curve analyses in cell lines treated with 500 nM PD180970 in a time course. These assays suggested that the sarcoma cell lines responded to PD180970 by induction of apoptosis in a time-dependent manner (Fig. 8B). SaOS-2 cells experienced reduced viability with increased exposure to 500 nM PD180970. By 36 h, less than 55% of the cells were viable (p< 0.05) and by 72 h less than 10% of the cells were viable

(p<0.001). To further validate the induction of apoptosis, Western blot analysis was performed for PARP cleavage. SK-LMS-1 and SaOS-2 cells were treated with 500 nM of PD180970 increasing periods of time and PARP cleavage was evaluated in both cell lines (Fig. 8C). PARP cleavage, an indicator of apoptosis, was evident in both cell lines after 8 h of treatment and increased with time. Moreover, TUNEL assays performed on

SK-LMS-1, SaOS-2 and LM2 cells confirmed that increasing numbers of the cells underwent apoptosis 48 h after treatment with 500 nM of PD180970 (Fig. 9A-C respectively). Therefore, PD180970 induces apoptosis in sarcoma cell lines with doses that correspond to the inhibition of Src and STAT3 activation by PD180970.

DMSO 3H8 12 24 36 48 72 Stat3 EMSA

120

100

80 y

60 * * % Viabilit

20 **

0 DMSO81224364872 Hours

C DMSO 3H8 12 24 36 48 72 PARP SK-LMS-1 PARP Cleavage PARP SaOS-2 PARP Cleavage

Figure 8. PD180970 inhibits viability and induces apoptosis in sarcoma cell lines. A, PD180970 inhibits STAT3 activation in a time-dependent manner. SK-LMS-1 cells were treated with 500 nM PD18097 for increasing periods of time. Cell-free extracts were utilized for EMSA to determine STAT3 activation status. B, SaOS-2 were treated with 500 nM PD18097 in a time course and cell viability was determined by trypan blue exclusion assays in triplicate. C, SK-LMS-1 and SaOS-2 cells were treated with 500 nM PD18097 in a time course. Cell-free extracts were immunoblotted with antibodies specific to PARP to measure induction of apoptosis. DMSO was used as vehicle control for all experiments.

DMSO 48 Hr

Figure 9. PD180970 induces apoptosis in sarcoma cell lines as measured by TUNEL. A- C, apoptosis was further verified by TUNEL assay as described. SK-LMS-1, SaOS-2 and LM-2 cells were plated in 12-well tissue culture plates and treated with DMSO or 500 nM PD180970 for 48 h. After completing the TUNEL assay, cells were visualized using light microscopy and photographed at 20x magnification.

62 SKI-606 does not inhibit Src-STAT3 signaling in human sarcoma cell lines

SKI-606 has previously been shown to directly inhibit the kinase activity of purified Src protein in vitro with an IC50 of 3.8 nM (253) and 100 nM for intact cells

(254). To evaluate the effect of SKI-606 on Src kinase activity and STAT3 activation in

intact sarcoma cells, we treated several sarcoma cell lines with escalating doses of SKI-

606 (300, 1000, 3000 and 10,000 nM) for 24 h. Western blot analysis was performed to

evaluate p-Src (Tyr419) levels and EMSA was performed to determine the effects on

STAT3 activation. There were no apparent effects on pSrc (Y419) in any of the cell lines

evaluated (data not shown). Furthermore, dose-response results for STAT3 activation for

SK-LMS-1 and U2 OS are shown in Figure 10. There was no inhibition of STAT3 activation by SKI-606 in any of the cell lines examined. Furthermore, there were no observed anti-proliferative or pro-apoptotic effects from SKI-606 at any of the doses tested.

vSRCNIN3T3

SKLMS-1 U2 OS DMSO 0.3 1310DMSO0.313 µM

Stat3

Figure 10. SKI-606 does not inhibit Src-STAT3 signaling in sarcoma cell lines. SK- LMS-1 and U 2 OS cell were treated with SKI-606 in a dose-response for 24 h. EMSA was performed to measure STAT3 activation in response to SKI-606 treatment. DMSO was used as vehicle control.

63 Dasatinib does not inhibit Src-STAT3 signaling in human sarcoma cell lines

Dasatinib has previously been shown to directly inhibit the enzymatic activity of

Src with an IC50 of 0.5 nM (231). The effects of Src activity are described in the next chapter of this dissertation. However, to evaluate the effect of dasatinib on STAT3 activation in sarcoma cells, several sarcoma cell lines were treated with escalating doses of dasatinib (30, 100, 300 and 1000 nM) for 6 h. EMSA was performed to determine the effects on STAT3 activation six cell lines. There was no inhibition of STAT3 activation by dasatinib in any of the cell lines examined Fig. 11A. Western blot analysis was performed to confirm these observations, Figure 11B and C show that pSTAT3 expression does not decrease with increased dasatinib concentration in SaOS-2 and U 2

OS cell lines, respectively. Furthermore, there were no observed effects of STAT3 activation by dasatinib in SK-LMS-1 cell lines when the doses escalated to 3 and 10 µM for 24 h (Figure 11D).

A DMSO 30 100 300 1000 nM SaOS-2

LM-2

U 2 OS

MG-63

SK-LMS-1

HT1080

B DMSO 30 100 300 1000 nM pStat3 (Y705)

ß-actin

C DMSO 30 100 300 1000 nM pStat3 (Y705)

ß-actin

D DMSO1 10 25 50 100 125 250 500 1000 3000 10,000 nM

Stat3

Figure 11. STAT3 signaling is independent of Src kinase activity in human sarcoma cell lines. A, SaOS-2, LM-2, U 2 OS, MG-63, SK-LMS-1 and HT1080 were treated with dasatinib in a dose response for 6 h. EMSA was performed to evaluate STAT3 activation status in response to treatment with dasatinib. B and C, SaOS-2 and U 2 OS cells were treated with escalating doses of dasatinib for 6 h and cell free extracts were immunoblotted for pSTAT3 (Y705) and β-actin as loading control. D, SK-LMS-1 cells were treated with dasatinib in a dose response that achieved higher doses of dasatinib for 24 h. EMSA analysis was performed to evaluate STAT3 activation status in response to treatment with dasatinib.

65 Discussion

Constitutive activation of STAT3 has been observed and demonstrated to play an

essential role for the tumorigenesis in many solid tumor and hematopoietic malignancies

(195, 196, 245, 246). The status of STAT3 activation in sarcomas has not been explored

previously. Our results show that STAT3 is constitutively activated in many of the

human sarcoma specimens analyzed and most of the sarcoma cell lines examined. To

elucidate the role of STAT3 activation in human sarcoma cell lines, we investigated the activity of three different Src kinase inhibitors. We have shown the PD180970 inhibits

Src and STAT3 signaling in several of the cell lines utilized. Although the IC50 for

STAT3 inhibition is markedly higher than the IC50 required to inhibit Src activity for

most of the responsive cell lines (Fig. 7). However, cell viability was decreased and

apoptosis was induced by 500 nM PD108970 treatment in a time-dependent fashion.

Furthermore, the dose of PD180970 corresponded to complete inactivation of Src and

STAT3 kinases is all responsive cell lines.

These data highlight the potential promise of employing a Src or STAT3

inhibitors for the treatment of sarcomas. There is little incentive to perform additional

experiments with PD180970 because the stability of this compound is limited and

obtaining reproducible data is difficult. This compound also has little to no future in the

clinic as it is only soluble in DMSO and is highly nonspecific. In addition, several more

stable, more specific, bioavailable Src kinase inhibitors with more promising clinical

futures have been made available. Two such Src kinase inhibitors are SKI-606 and

66 dasatinib were further investigated to determine a more precise role for Src and STAT3

activity in sarcomas.

The Src kinase inhibitor, SKI-606, is currently undergoing several phase 2 clinical

trials for malignancies including breast, pancreatic, colon and non-small cell lung cancers

and chronic myelogenous leukemia (255). We evaluated the activity of this compound

in sarcoma cell lines to investigate the effects of Src and STAT3 signaling. To our disappointment, SKI-606 had no inhibitory activity on either of these proteins in sarcomas. There were no apparent proliferation or survival effects generated by exposure to this compound either. One possible explanation for these results may be that the compound is not crossing the cell membrane and gaining access to the cell. Src kinase may also have a greater affinity for PD180970 or dasatinib, which may further explain the lack of response to SKI-606.

Dasatinib was the last Src inhibitor to be obtained and evaluated, interestingly enough it was also the compound that exhibited the most potential as a possible therapeutic option for sarcomas. However, the promising anti-tumor effects of dasatinib do not appear to be mediated via a Src-STAT3 pathway. Dasatinib was shown to potentially inhibit Src activity in sarcoma cell lines, as will be more thoroughly demonstrated in the following chapter. Surprisingly, inhibition of Src activity by dasatinib had no effect on STAT3 activation in sarcoma cell lines (Fig. 11). These results do not rule out a possible role for constitutive STAT3 activation in sarcomas, but suggests that the Src and STAT3 signaling pathways may be uncoupled in sarcomas. The data generated using PD180970 suggested a higher IC50 was required to inhibit STAT3 as

67 compared to Src activity, suggesting that Src and STAT3 may be cross-talking with other pathways in sarcoma cell lines. In spite of this, these results support the possibility of a dependence of Src activity for survival in sarcomas. Therefore, we next set out to determine the role of Src activity in sarcomas by investigating other Src dependent pathways.

Dasatinib Inhibits Migration and Invasion in Diverse Human Sarcoma Cell Lines

and Induces Apoptosis in Bone Sarcoma Cells Dependent on

Src Kinase for Survival

One potential molecular target for sarcoma treatment is the Src tyrosine kinase.

Dasatinib, a small-molecule inhibitor of Src kinase activity, is a promising cancer

therapeutic agent with oral bioavailability. Dasatinib exhibits anti-tumor effects in

cultured human cell lines derived from epithelial tumors including prostate and lung

carcinomas. However, the action of dasatinib in mesenchymally-derived tumors has yet

to be demonstrated. Dasatinib was originally selected as a Src kinase inhibitor and then

shown to inhibit Bcr-Abl as well as other tyrosine kinases. There have been several

studies demonstrating the activity of dasatinib against Bcr-Abl-positive leukemic cell

lines as well as epithelial tumor cell lines (43-48). In addition, early phase clinical trials

have established the safety and efficacy of dasatinib for treatment of imatinib-resistant

chronic myelogenous leukemia patients. However, the responses and mechanisms of

action of dasatinib in mesenchymally-derived tumor cell lines have not been described

previously. We report that dasatinib inhibits Src and downstream FAK signaling at

nanomolar concentrations, blocks cell migration and invasion in many diverse human

sarcoma cell lines and induces apoptosis in bone sarcomas. Furthermore, knockdown of

Src expression by siRNA in bone sarcoma cells also induces apoptosis, suggesting that

the observed response to dasatinib in these cells is conveyed through inhibition of Src- mediated signaling. Together, these findings indicate that dasatinib is a promising

69 therapeutic agent for preventing growth and metastasis of a wide diversity of soft-tissue and bone sarcomas. Based on our previous findings of Src activation in human sarcomas, we evaluated the effects of dasatinib in twelve cultured human sarcoma cell lines derived from bone and soft-tissue sarcomas.

70 Results

Src kinase is activated in human sarcomas and sarcoma cell lines

Immunohistochemistry for activated Src protein was performed on human sarcoma specimens for activated Src protein using antibodies to phospho-Src (p-Src).

Levels of p-Src on tyrosine residue 419 (Y419) due to autophosphorylation reflect Src kinase activities in intact cells and tissues. Results show activated Src in sarcomas of diverse subtypes (Fig. 12), including leiomyosarcoma (A), high-grade OSA (B), and LPS

(C). Since autophosphorylated Src was found in a majority of the human sarcomas examined, including diverse soft-tissue and bone sarcomas (data not shown), we determined the level of Src activation in a panel of human sarcoma cell lines by Western blot analysis for p-Src (Y419) and total Src protein levels. Src was detectably activated in all but one (HT-1080) of the cell lines examined, albeit to different extents (Fig. 12D).

Total Src protein expression does not correlate with levels of phosphorylated Src in every case, indicative of different levels of Src kinase activation among the sarcoma cell lines.

In addition, the level to which Src kinase is activated (p-Src levels) does not correlate with specific sarcoma histological sub-types (Fig. 12 and Table 6).

AB C

D RD SK-ES-1 A673 LM7 RD18 TC-71 MG-63 HT-1080 U-2 OS U-2 SK-LMS-1 LM2 SaOS-2

p-Src (Y419) Total Src ß-Actin

Figure 12. Src kinase is activated in human sarcoma tissues and cell lines. A-C, immunohistochemistry for p-Src (Y419) reveals that Src is activated in human sarcoma tissues (leiomyosarcoma A, high grade osteosarcoma B and liposarcoma C). D, Src is activated in all but one (HT-1080) of the cell lines utilized for these experiments. Cell- free extracts were prepared from untreated cells grown in 10% FBS and immunoblotted with antibodies specific for p-Src (Y419), total Src or β-actin.

72 Dasatinib inhibits Src kinase activity in human sarcoma cell lines

Dasatinib has previously been shown to directly inhibit the kinase activity of

purified Src protein in vitro with an IC50 of 3 nM (223). To evaluate the effect of

dasatinib on Src kinase activity in intact sarcoma cells, we treated the above cell lines

with escalating doses of dasatinib (30, 100, 300 and 1000 nM) for 6 h and Western blot

analysis was performed to evaluate p-Src levels. Dose-response results for the two

representative cell lines (SaOS-2 and U 2 OS) are shown in Figure 13 (A and B), the IC50

values for inhibition of p-Src by dasatinib range from 3 to 68 nM for all cell lines

analyzed (summarized in Table 6). Time course experiments were performed to

determine the kinetics at which Src phosphorylation is inhibited by dasatinib. As shown

in representative results with U 2 OS cells, inhibition of Src phosphorylation is complete

by 15 min following treatment with 100 nM of dasatinib and persists for at least 24 h

(Fig. 13C). Interestingly, total Src protein expression was increased in a dose- and time-

dependent manner in a subset of the cell lines treated with dasatinib. In particular, all but

one bone sarcoma cell line (MG-63) produced increases in total Src protein expression,

yet this effect was not observed in the soft-tissue sarcoma cell lines (Fig. 13A-C, and data

not shown). However, this increased protein expression of Src was not accompanied by

an increase of c-Src mRNA, as shown in both cell lines which an increased expression of total Src was and was not observed (Fig. 14). These data suggest a positive feedback mechanism for compensation of Src kinase inhibition with increased levels of Src protein

expression in the bone sarcoma cells.

A DMSO 30 100 300 1000 nM p-Src (Y419) Src ß-Actin

B DMSO30 100 300 1000 nM p-Src (Y419) Src ß-Actin

C DMSO0.25.501 2 4 6 12 24 H p-Src (Y419)

Src ß-Actin

DMSO30 100 300 1000 nM p-FAK (Y397)

p-FAK (Y576/577)

p-FAK (Y925)

FAK

p-p130CAS

p130CAS

ß-Actin

Figure 13. Dasatinib inhibits Src activation and downstream signaling in sarcoma cell lines. A and B, SaOS-2 and U 2 OS cells were treated with dasatinib in a dose-dependent manner for 6 h. Cell-free extracts were immunoblotted with antibodies specific to p-Src (Y419) and total Src. C, U 2 OS cells were treated with 100 nM of dasatinib in a time- dependent manner. Western blot analysis was performed as described. DMSO was used as a vehicle control and β-actin was immunoblotted for as a loading control in all experiments. D, dasatinib specifically blocks tyrosyl phosphorylation of FAK (Y576/577, Y925) and 130CAS (Y410), but not FAK Y397. SaOS-2 cells were treated with dasatinib for 6 h in a dose-dependent manner. Cell-free extracts were immunoblotted with antibodies specific to p-FAK (Y397, Y576/577, Y925), total FAK, p130CAS and p- p130CAS (Y410).

160 SaOS-2 U 2 OS 140 SK-LMS-1 120

l 100

80 % Contro 60

0 0 30 100 300 1000 dasatinib, nM

Figure 14. Dasatinib does not induce c-Src mRNA expression. Quantitative RT-PCR was performed on SaOS-2, U 2 OS and SK-LMS-1 cells to determine the effect of dasatinib on c-Src mRNA expression. Cells were treated with dasatinib in a dose-response and RNA was isolated, purified and qRT-PCR for c-Src was performed as described.

76 Dasatinib selectively blocks Src downstream signaling

Src kinase has been shown to regulate cellular activities through a number of downstream signaling pathways. One such pathway is FAK, a non-receptor tyrosine kinase found to be increased in a variety of epithelial cancers including those arising from

prostate, cervical and colon (256-259). Furthermore, increased FAK expression is

associated with tumor progression in a mouse model of skin carcinogenesis (200). FAK,

in turn, has been implicated in the activation of CRK-associated substrate, p130CAS,

which together with Src and FAK plays a vital role in cell adhesion, migration,

proliferation and survival (209). To investigate the effect of dasatinib on these Src downstream signaling pathways, sarcoma cell lines were treated in culture with escalating doses of dasatinib for 6 h. As representative results using SaOS-2 cells, Figure 13D shows Western blot analysis performed using antibodies to total FAK protein, phosphorylated FAK (Y397, Y576/Y577, and Y925), p130CAS and phosphorylated

CAS p130 (Y410). The IC50 values for inhibition of phosphorylated FAK (Y576/Y577 and

CAS Y925) and p130 were between 30 to 100 nM, consistent with the IC50 values for

inhibition of Src kinase activity in these cells (Fig. 13 and Table 6). FAK

autophosphorylation (Y397) was not substantially inhibited until higher doses of

dasatinib (1000 nM), indicating that dasatinib does not directly inhibit FAK kinase

activity. Surprisingly, total p130CAS protein was diminished with dasatinib treatment

while total FAK protein was not affected (Fig. 13D), suggesting that p130CAS protein is

subject to negative feedback regulation in these cells. By contrast, dasatinib did not

inhibit STAT3 signaling in sarcoma cell lines (data not shown), another signaling

pathway that has been shown to act downstream of Src in cells of other tumor types

77 (260). Thus, dasatinib selectively blocks FAK and p130CAS signaling downstream of Src

in sarcoma cell lines.

Dasatinib blocks cell motility and invasion by sarcoma cells

Both FAK and p130CAS activity are involved in regulating cell migration and

invasion downstream of Src kinase. The effect of dasatinib on cell migration was

evaluated using “wound healing” assays (by scratching cell monolayers with a pipette tip)

and treating with drug. Cells were plated in 0.1% serum medium prior to inducing the

wound to ensure that migration rather than cell growth was measured. The width of the

wound was determined at T0 and then cells were treated with escalating doses of

dasatinib in 0.1% serum medium for 24 h. Representative results are shown with the

SaOS-2 cells, which were digitally photographed and the width of denuded area in the wound was measured in pixels (Fig. 15A and B). Wound healing was dramatically inhibited by dasatinib in a dose-dependent manner, with detectable inhibition at 30 nM and substantial inhibition at 100 nM dasatinib. To evaluate the effect of dasatinib on cell invasion, SaOS-2 and U 2 OS cells were treated with dasatinib in a dose-response manner for 22 h in Matrigel Invasion Chambers. Dasatinib significantly inhibited cellular invasion in both cell lines (Fig. 15C). The SaOS-2 cell line was more sensitive to inhibition of invasion by dasatinib compared to the U 2 OS cell line. The IC50 values for

inhibition of tumor cell invasion in this assay ranged from 30 to 100 nM. These IC50

values for blockade of cell migration and invasion are consistent with the IC50 values for

inhibition of Src kinase as well as downstream FAK and p130CAS signaling (compare

with Table 6 and Fig. 13D).

T0 DMSO 30 nM

100 nM 300 nM 1 µM

800.0 * 700.0

600.0 * * 500.0

400.0 Pixe

300.0

200.0

100.0

0.0 T0 0 30 100 300 1000 dasatinib, nM

C 120

100

40 ** Invasion (% control) (% Invasion 20 * **

0 0 30 100 300 1000 dasatinib, nM

Figure 15. Dasatinib inhibits cell motility and invasion. A, wound healing assays were performed to determine the effects of dasatinib on inhibiting cell migration. SaOS-2 cells were plated in 12-well tissue culture plates, grown to confluency and serum-starved overnight in medium containing 0.1% FBS. Wounds were introduced on cell monolayers using a pipette tip. Cells were washed with 1x PBS to remove non-adherent cells and treated with DMSO or dasatinib in a dose-dependent manner for 24 h. Cell migration was visualized at 10x magnification by light microscopy and photographed with a digital camera. B, width of voided area versus dasatinib dose concentration was graphed to express the degree of inhibition of cell migration. The number of pixels within the denuded area were the units used to demonstrate inhibition of cell migration induced by dasatinib (*p<0.001, n=3). C, matrigel Invasion Chambers were used to measure dasatinib inhibition of cellular invasion. SaOS-2 (solid bars) and U 2 OS (shaded bars) were treated with dasatinib in a dose-dependent manner. Cells were diluted in 500 μL of serum-free medium and placed over the inner chamber of the insert in a 24-well tissue culture plate and 500 μL of complete medium was placed in the lower chamber of the insert. After incubating for 22 h, cells that invaded through the Matrigel were stained, visualized using light microscopy, photographed and counted. Each experiment was performed in triplicate. The mean values of invasive cells were graphed versus dasatinib concentration (*p<0.001, **p<0.01, n=3). To confirm that invasion was measured rather than induction of apoptosis, trypan blue exclusion assays were performed with SaOS-2 and U 2 OS cells treated in a dose-dependent manner with dasatinib for 24 h. Greater than 90% of the cells were viable after dasatinib treatment (data not shown).

80 Dasatinib induces apoptosis of bone sarcoma cell lines

To determine the effect of dasatinib on sarcoma cell survival, we performed

growth curve analyses in cell lines treated with increasing concentrations of dasatinib.

These assays suggested that the subset of cell lines derived from bone sarcomas responded to dasatinib by induction of apoptosis in a dose-dependent manner (data not shown). To further validate the induction of apoptosis in this subset of cell lines, Western blot analysis was performed for apoptotic markers. SaOS-2 and U 2 OS cells were treated with escalating doses of dasatinib for 72 h, and PARP cleavage and XIAP expression were evaluated in both cell lines (Fig. 16A and C). PARP cleavage, an indicator of apoptosis, is evident at 30 nM dasatinib and increased with escalating doses of dasatinib. Furthermore, expression of XIAP, an inhibitor of apoptosis, was diminished by dasatinib treatment with IC50 values ranging from 30 nM to 100 nM (Fig. 16A and C).

A time-course analysis with 100 nM dasatinib was performed in the SaOS-2 cell line to determine when apoptosis was induced as measured by PARP cleavage. PARP cleavage is evident as early as 6 h following treatment with dasatinib and increased with time (Fig.

16B). Moreover, TUNEL assays performed on SaOS-2 cells confirmed that increasing numbers of the cells were undergoing apoptosis by 48 h after treatment with escalating doses of dasatinib (Fig. 16D). Therefore, dasatinib induces apoptosis in the bone sarcoma subset with IC50 values corresponding to those of inhibition of Src kinase and

downstream signaling by dasatinib (Table 6).

A DMSO30 100 300 1000 nM XIAP PARP PARP cleavage ß-Actin

B DMSO0.25.501 2 4 6 12 24 36 48 72 H

PARP PARP cleavage

ß-Actin

C DMSO 30 100 300 1000 nM XIAP PARP PARP Cleavage ß-Actin

D DMSO 30 nM

100 nM 300 nM

Figure 16. Dasatinib induces apoptosis in bone sarcoma cell lines. A and C, dasatinib induces apoptosis in a dose-dependent manner. SaOS-2 and U-2 OS cells were treated with dasatinib for 72 h with escalating doses. Cell-free extracts were immunoblotted with antibodies specific to XIAP and PARP. B, SaOS-2 cells were treated with 100 nM of dasatinib in a time-dependent manner and immunoblotted with an antibody specific for PARP. D, apoptosis was further verified by TUNEL analysis as described. SaOS-2 cells were plated in 12-well tissue culture plates and treated with dasatinib for 48 h in a dose- dependent manner. After completing the TUNEL assay, cells were visualized using light microscopy and photographed at 20x magnification.

83 Src is required for survival of bone sarcoma cell lines

To determine if inhibition of Src kinase by dasatinib is sufficient to induce

apoptosis in the bone sarcoma cell lines, we transfected these cell lines with siRNA to c-

Src. Two representative bone sarcoma cell lines, SaOS-2 and U 2 OS, underwent

induction of apoptosis in a dose-dependent manner as measured by PARP cleavage in

response to siRNA against c-Src but not to control siRNA. Src protein expression was

inhibited by transfection with 50 nM and 100 nM siRNA against c-Src, corresponding to

induction of PARP cleavage in both cell lines (Fig. 17A and B). MG-63, an OSA cell line

that does not undergo apoptosis when treated with dasatinib, also does not undergo

apoptosis when transfected with siRNA to Src (data not shown). These data demonstrate that a subset of the bone sarcoma cell lines rely on Src kinase for survival, indicating that inhibition of Src kinase activity by dasatinib is sufficient to induce apoptosis in these cells.

si Src Buffer Control si Control 50 100 nM Src PARP PARP Cleavage ß-actin

Growth Factors

Growth Factor Receptor Extracellular Matrix Cellular Membrane

FAK dasatinib Y419 P Src p130CAS

Survival Migration Invasion Induction of Apoptosis

Figure 17. Src is required for bone sarcoma cell line survival. A and B, siRNA to c-Src induces apoptosis in bone sarcoma cell lines. SaOS-2, A, and U 2 OS, B, cells were plated in 6 cm tissue culture plates, transfected with 50 and 100 nM of siRNA to c-Src (si Src) and harvested after 72 h. Cell-free extracts were immunoblotted with antibodies specific for Src and PARP to measure the efficiency of knockdown by siRNA and to determine if apoptosis was induced upon depletion of Src protein expression. C, Src activation and signaling are inhibited by dasatinib. Inhibition of Src signaling by dasatinib prevents cellular migration and invasion in sarcoma cell lines. Upon inhibition of Src phosphorylation or c-Src expression, a sub-set of bone sarcoma cell lines undergo an induction of apoptosis.

Table 6. Summary of cell line IC50 values and responses to dasatinib

Induction of pSrc (Y419) IC50, nM Cell Line Tumor Type pSrc Parp Cleavage, Expression Migration (Y419) nM SaOS-2 Osteosarcoma ++ 46 65 30 LM2 Osteosarcoma ++ 26 35 30 LM7 Osteosarcoma ++ 68 24 100 U 2 OS Osteosarcoma ++++ 57 4 30 MG-63 Osteosarcoma +++ 28 58 N/R SK-ES-1 Ewing’s Sarcoma +++ 11 4 30 TC-71 Ewing’s Sarcoma + 3 4 30 SK-LMS-1 Leiomyosarcoma +++ 46 44 N/R HT-1080 Fibrosarcoma - N/R N/R N/R A673 Rhabdomyosarcoma ++++ 26 23 N/R RD Rhabdomyosarcoma ++ 45 29 N/R RD18 Rhabdomyosarcoma + 50 277 N/R +/- depicts relative p-Src (Y419) expression N/R No Response

Table 6. Summary of responses to dasatinib in human sarcoma cell lines. With the exception of one soft-tissue sarcoma cell line (HT1080) all of the cell lines examined respond to dasatinib by inhibition of Src phosphorylation on Y419 and migration at IC50 values consistent with Src kinase inhibition. A subset of bone sarcoma cell lines respond to dasatinib treatment by induction of apoptosis. Induction of apoptosis was not observed in the soft-tissue sarcomas cell lines and one osteosarcoma (MG-63) cell line.

87 Discussion

After nearly a century since the discovery of the Rous sarcoma virus, which subsequently was shown to induce sarcomas by capturing and mutationally activating the cellular gene encoding the Src tyrosine kinase, targeted Src kinase inhibitors are now entering clinical trials for solid tumors. Sarcomas comprise a highly diverse set of human tumors that frequently occur among pediatric cancer patients and for which there are limited treatment options. Based on our observation of Src kinase activation in sarcoma clinical specimens, we sought to determine the action of dasatinib, a potent and orally bioavailable inhibitor of Src kinase, on human sarcoma cell lines.

Our findings demonstrate that dasatinib inhibits Src kinase activity, as measured by autophosphorylation at Tyr419, in a dose-dependent manner in sarcoma cells.

Furthermore, in 11 out of 12 sarcoma cell lines examined dasatinib inhibits cell migration and invasion. The single cell line that did not respond to dasatinib (HT-1080) was also

the only one that lacked detectable Src kinase activity in this panel (Table 1). Moreover,

suppression of cell migration and invasion was associated with inhibition of downstream

Src signaling through FAK and p130CAS, proteins known to be involved in mediating

CAS these cellular processes (261-264). The IC50 values for inhibition of Src/FAK/p130 signaling as well as migration and invasion are all in the range of approximately 30 nM to

100 nM regardless of histological type (Table 1). Taken together, our findings suggest a model for the mechanism of dasatinib action in which blockade of Src and downstream signaling suppresses migration and invasion of sarcoma cells (Fig. 5C).

Significantly, dasatinib induces apoptosis in the majority of bone sarcoma cells lines, including OSA and EWS, but not in any of the soft tissue sarcoma cell lines in our

88 panel. Genetic inhibition of Src using siRNA also induced apoptosis in bone sarcoma cell lines that respond to dasatinib with apoptosis, but not in the only OSA cell line (MG-

63) in which dasatinib did not induce apoptosis. Thus, dasatinib induces apoptosis in bone sarcoma cell lines dependent on Src kinase for survival. A major Src signaling pathway involved in preventing apoptosis in other cellular contexts is STAT3 (245, 265).

While most of the sarcoma cell lines in this study harbor activated STAT3, with the sole exception of SK-ES-1, dasatinib did not inhibit STAT3 activation, indicating that this pathway is not involved in the dasatinib-mediated apoptosis response in sarcomas (our unpublished results). On the other hand, the IC50 values for induction of apoptosis by

dasatinib are in the same range required for blockade of FAK and p130CAS signaling in

these cell lines. Because FAK and p130CAS have been implicated in tumor cell survival, in addition to cell migration and invasion, it is possible that these pathways are involved

in dasatinib-mediated apoptosis in sarcoma cells.

It is notable that levels of Src activation do not correlate with IC50 values of

dasatinib responses in terms of cell migration, invasion or apoptosis (Table 1). This

finding may be explained by the possibility that low levels of Src kinase activation are

sufficient to induce these biological properties. Alternative explanations are the

possibilities that other SFK members or unidentified targets are involved in the responses

to dasatinib. Similar results have been observed for other molecular targeted-therapeutic

agents, such as Iressa, where clinical response to this EGFR inhibitor is not correlated

with levels of EGFR expression or activation (225). In the specific case of Iressa, EGFR

mutations have been shown to influence response to Iressa; however, mutations in the c-

Src gene are extremely rare in human cancers (230). Thus, selection of patients for

89 dasatinib treatment on the basis of Src expression or activation levels may not predict the optimal clinical responses. It remains to be determined whether any of the known genetic sub-types of sarcomas are more sensitive to dasatinib than others.

Earlier preclinical laboratory studies pointed to the promise of dasatinib in the

treatment of Gleevec-resistant chronic myeloid leukemia, a prediction that has been borne

out in clinical trials (266-272). On the basis of more recent preclinical laboratory studies, several human solid tumor sites have shown promise for clinical trials, including prostate, lung, pancreatic, and head and neck cancers (223-225). We have established that Src is activated in a wide variety of human sarcoma clinical specimens, including STS and bone sarcomas. Furthermore, our data demonstrate that dasatinib inhibits Src kinase and downstream signaling, leading to blockade of cell migration and invasion of sarcoma cell lines of diverse origins. In the subset of bone sarcomas, dasatinib also induces apoptosis.

Taken together, our results suggest that dasatinib will provide clinical benefit to soft tissue and other sarcomas by preventing metastasis, which may be further augmented in bone sarcomas by induction of apoptosis.

Gene Expression Profile of Sarcoma Cell Lines Serves as Preliminary Signature

Predictive of Response to Treatment with Dasatinib

It has recently been shown that unlike epithelial cancers, sarcomas are better

defined by their molecular pathology rather than the organ of origin (1). Identified

molecular alterations and cytogenetic analysis of specific subtypes of sarcomas have

proven that previous classifications based on the site of the tumor are less important than

the molecular phenotype of the tumor. Targeted therapies have aided in reaching this

conclusion. Patients with specific molecular phenotypes have been shown to respond

better to specific targeted therapies, as was the case with Gleevec and c-kit mutations in

GISTs. These findings have had an important impact on the approaches in treating sarcomas. However, there are still many sarcomas subtypes with little to no improved treatment options.

Microarray analysis can help identify significant genes involved in sarcomagenesis and better classify tumor subtypes by molecular phenotype. Gene expression profiles (GEP) generated from microarray analysis have proven to successfully identify signatures that can predict prognosis and response to chemotherapies for breast cancer (273-276). Microarray analysis has recently provided insight into receptor tyrosine kinase expression patterns in sarcomas that may serve as potential targets for novel therapeutics . The examination of GISTs by microarray analysis lead to the identification of c-KIT and PDGFR as potential targets of therapy and

91 provided prognostic signatures that presented a biological basis for the differential

responses exhibited to Gleevec. Furthermore, molecular signatures predicting response to therapy have been published for EWS and OSA (277). Consequently, the use of gene

expression profiling by microarray analysis cannot only rapidly provide potential

therapeutic targets, but may also serve as a screening mechanism used to predict response

to a specific therapy, thereby preventing unnecessary patient exposure to various

chemotherapeutics.

The second aim of this dissertation research was to identify a candidate molecular

signature that predicts response to dasatinib by induction of apoptosis in human sarcoma

cell lines. The response status of 12 cell lines was determined in aim one and was used to classify the cells into responsive and non-responsive categories. RNA was extracted

from three consecutive passages of each cell line and purified for microarray analysis. A

total of 36 Human Genome U133 Plus 2.0 Arrays were utilized for the initial analysis and

six more for the second phase of this aim. Once a molecular signature was identified and

validated, the accuracy of the signature to predict response in cell lines was tested using two new cell lines of unknown response status, to prevent bias. The same approach for

RNA collection and purification was used for the test cell lines. Microarray analysis was performed independently of the molecular analysis and verification of response to dasatinib for each cell line. Once both analyses were completed, the results were compared. Amazingly, the results from both analyses were the identical; one cell line responded by induction of apoptosis to dasatinib, while the other did not. Therefore, a molecular signature that successfully predicts response to dasatinib by induction of apoptosis was identified in sarcoma cell lines.

92 Results

Unsupervised clustering identified three main classes, with five subgroups in relation to cell line types

Using an unsupervised hierarchical clustering approach, we tried to identify

natural subclasses of cell lines as determined by gene expression profiles. We performed

unsupervised clustering on a low-level filtered probe list using genes with median

intensities of greater than or equal to 2000 to avoid using background noise in this

analysis. The clustering results are shown in Figure 18A. Interestingly, three main

classes were defined; the first included the EWS and RD cell lines, the second consisted

of the STS and two OSA cell lines, while the third included three OSA cell lines. The

first class is further sub-grouped by EWS and RD. Intriguingly, the A673 cell line, which

was originally characterized as a RD, but later identified to possess an EWS

translocation, was sub-grouped with the EWS arm. The second class consists of two sub-

groups, one arm with two STS cell lines, the other arm with two OSA cell lines. Not surprising, the third class includes three cell lines which were derived from one of the cell lines in the sub-group.

A second hierarchical clustering approach was completed on the data set to determine if the cell lines could group according to response. This cluster was performed using a filtered probe of genes with median intensities greater than or equal to 2000 and had a significant (p-value ≤ 0.05) difference fold change between responders and non- responders. To calculate the fold change the median signal intensity of the non- responders was compared to the average signal intensity of each probe set and the cluster is shown in Figure 18B. This analysis provided two classes, one with all but one of the

93 non-responsive cell lines (A673) and the other with all but one responsive cell line (U 2

OS), suggesting that the cell lines can be separated into two distinct groups based on

GEP.

A B

Figure 18. Hierarchical clustering of sarcoma cell lines based on GEP. Microarray analysis was performed on 12 sarcoma cell lines in triplicate. The GEP of the cell lines was determined by calculating the average intensity per triplicate sample set and the fold change for each probe set using the median intensity for each probe set as the comparison group. The data were filtered for probe sets with ≥ 2000 intensity values, the remaining gene fold changes were transformed using Log2 and an unsupervised cluster was completed. A, shows the dendogram associated with this cluster. B, another hierarchical clustering analysis was performed, but the median intensity of the non-responders was used as the comparison group rather than the median of all samples. Data was further filtered for genes with fold changes that were significantly (p-value ≤ 0.05) different between responders and non-responders and had median signal intensities ≥ 2000.

94 A molecular signature distinguishes response to dasatinib as defined by induction of

apoptosis

In order to stringently identify genes that might represent the molecular signature

that predicts response to dasatinib, namely to select genes which have higher discrimination between the lists previously identified by the unsupervised analysis, we employed two different approaches of analysis. Our first approach was a SAM

(statistical analysis of microarrays) analysis of the responsive versus the non-responsive

cell lines as two groups. To prevent extracting a bone sarcoma specific signature, because those are the only cell lines that responded to dasatinib, we filtered out the bone

sarcoma specific overlapping genes from the list. This provided a list of more than 1000

genes. Our second approach was to complete the standard MAS5 analysis. Once the

signal intensities were determined for each probe the ratio for the maximum to minimum

value for each probe set and each group was determined. Ratios less than 2.0 were

included for additional analysis. This was completed to utilize the most consistent data

for further analysis. Next, the fold change of responders was calculated. To complete

this, the ratio of the average of each sample set for the responders to the median for all

samples was calculated for each probe set. Probe sets with a significant fold change (p-

value ≤0.05) were identified and selected to be compared to the probe sets which appeared on list generated using SAM. There were 26 probe sets that appeared on both

lists, representing 22 different genes. Theses probe sets were selected as the molecular

signature that predicts response to dasatinib (Figure 19A).

Probe sets for four genes: Histone H3, Fas associated factor-1 (FAF1), α-catenin

and δ-catenin were further validated using qRT-PCR because more than one probe set

95 appeared on the molecular signature for these genes. The fold change values for these four genes are shown in Table 7 and Log2 transformed fold changes are graphically depicted in Figure 19B and validated using qRT-PCR, the results of which are displayed in Table 8 and Log2 fold changes are graphed in Figure 20. Histone H3 and FAF1 had higher fold expression, while α-catenin and δ-catenin had lower fold expression in the responsive cell lines.

The initial analysis of the 22 gene list provides the fold change of the responders compared to the median of all the cell lines analyzed and may dilute the precise fold change between responders and non-responders for this molecular signature. To determine the exact fold change between the two groups we also calculated the fold change using the ratio of the average responder signal for each cell line to the median non-responder signal intensity using the same probe set list. The heatmap generated for this analysis is shown in Figure 21A. This analysis provides a more pronounced fold change between the two groups for the molecular signature. Furthermore, the fold change for Histone H3, FAF1, α-catenin and δ-catenin are also shown in Table 9 and the

Log2 transformed fold changes are graphically depicted in Figure 21B and the trends of expression are upheld.

2 Histone H3 Fas Associated Factor α-Catenin 1.5 δ-Catenin

0.5

Relative Fold Change Fold Relative -0.5

-1

-1.5

SaOS-2 LM-2 LM-7 U 2 OS MG-63 SK-ES-1 TC-71 SK-LMS HT1080 A673 RD RD|18 -2

Figure 19. Molecular signature was generated using the median value for each probe set as comparison group. A, heatmap of 26 different probe sets representing 22 genes, the average was determined for each sample set and fold change was calculated using median of all samples for the comparison group. The probe sets on this list were significantly different among responders and non-responders (p-value ≤ 0.05) and appeared on the list for both analyses. B, graph expressing fold changes of four genes, Histone H3, FAF1, α- catenin and δ-catenin that appeared several times on the molecular signature. The fold changes calculated for A were transformed by Log2 and the values graphed to depict accurate fold changes between samples.

Table 7. Fold changes for key signature genes in sarcoma cell lines compared to median intensity

Fas Associated Cell Line Histone H3 α-Catenin δ-Catenin Factor SaOS-2 1.14 ± 0.21 1.62 ± 0.15 0.982 ± 0.13 0.640 ± 0.21 LM-2 1.30 ± 0.088 1.62 ± 0.089 0.925 ± 0.14 0.653 ± 0.024 LM-7 1.46 ± 0.11 1.49 ± 0.35 0.865 ± 0.10 0.478 ± 0.072 U 2 OS 1.09 ± 0.17 0.905 ± 0.14 0.909 ± 0.069 0.569 ± 0.089 MG-63 0.919 ± 0.11 0.947 ± 0.12 1.71 ± 0.19 1.27 ± 0.051 SK-ES-1 1.51 ± 0.12 0.965 ± 0.17 0.908 ± 0.080 0.732 ± 0.088 TC-71 1.33 ± 0.39 1.00 ± 0.12 0.684 ± 0.30 1.00 ± 0.0053 SK-LMS 0.530 ± 0.03 0.823 ± 0.057 1.25 ± 0.20 1.77 ± 0.25 HT1080 0.641 ± 0.10 0.771 ± 0.057 1.26 ± 0.40 1.35 ± 0.45 A673 1.01 ± 0.13 0.678 ± 0.037 1.52 ± 0.13 1.48 ± 0.27 RD 0.680 ± 0.12 1.03 ± 0.078 1.01 ± 0.15 1.29 ± 0.17 RD|18 0.690 ± 0.23 1.03 ± 0.25 0.913 ± 0.12 1.32 ± 0.44

4 Histone H3 Fas Associated Factor α-Catenin 3 δ-Catenin

0 Relative Fold Change(Log2) -1

-2

-3 U2OS HT1080 SaOS-2 LM-2 LM-7 MG-63 SK-ES-1 TC-71 SK-LMS A673 RD RD-18

Figure 20. Quantitative RT-PCR validation of Histone H3, FAF1, α-catenin and δ- catenin gene expression. Bars represent Log2 fold changes for the selected genes. Positive fold change represents up-regulated, and negative fold change represents down-regulated in sarcoma cells.

100

Table 8. Quantitative RT-PCR expression of key signature genes in sarcoma cells

Fas Associated Cell Line Histone H3 α-Catenin δ-Catenin Factor SaOS-2 0.865 ± 0.016 1.04 ± 0.14 1.10 ± 0.13 1.10 ± 0.0058 LM-2 1.45 ± 0.010 2.13 ± 0.12 1.08 ± 0.19 1.14 ± 0.0070 LM-7 3.43 ± 0.025 4.17 ± 0.18 1.13 ± 0.062 1.04 ± 0.025 U2OS 0.227 ± 0.018 0.187 ± 0.18 1.13 ± 0.037 1.03 ± 0.036 MG-63 0.500 ± 0.0089 0.483 ± 0.18 1.13 ± 0.10 1.07 ± 0.015 SK-ES-1 1.33 ± 0.0049 0.637 ± 0.20 1.15 ± 0.015 1.01 ± 0.016 TC-71 2.62 ± 0.010 1.21 ± 0.082 1.06 ± 0.17 1.13 ± 0.020 SK-LMS 2.95 ± 0.0077 4.29 ± 0.15 1.11 ± 0.15 1.11 ± 0.014 HT1080 0.160 ± 0.0043 0.334 ± 0.19 1.14 ± 0.12 1.08 ± 0.035 A673 0.812 ± 0.0073 0.551 ± 0.19 1.14 ± 0.10 1.07 ± 0.016 RD 0.337 ± 0.020 0.559 ± 0.23 1.17 ± 0.013 1.01 ± 0.054 RD|18 0.818 ± 0.087 1.23 ± 0.011 1.01 ± 0.28 1.22 ± 0.053

101

102

1.5 Histone H3 Fas associated factor α-Catenin δ-Catenin 1

0.5

-0.5 Relative Fold Change (Log2) Change Fold Relative

-1

-1.5 SaOS-2 LM-2 LM-7 U 2 OS MG-63 SK-ES-1 TC-71 SK-LMS-1 HT1080 A673 RD RD|18

Figure 21. Molecular signature generated using the median value for each probe set in the non-responders as comparison group. A, heatmap of 26 different probe sets representing 22 genes, the average was determined for each sample set and fold change was calculated using median of the non-responders as comparison group for each probe set. The probe sets on this list were significantly different among responders and non- responders (p-value ≤ 0.05) and appeared on the list for both analyses. B, graph expressing fold changes of f Histone H3, FAF1, α-catenin and δ-catenin. The fold changes calculated for A were transformed by Log2 and the values graphed to depict accurate fold changes between samples.

103

Table 9. Fold Changes for key signature genes in sarcoma cells compared to median intensity of non-responders

Fas Associated Cell Line Histone H3 α-Catenin δ-Catenin Factor SaOS-2 1.63 ± 0.31 1.95 ± 0.18 0.854 ± 0.11 0.482 ± 0.16 LM-2 1.86 ± 0.12 1.96 ± 0.10 0.804 ± 0.13 0.493 ± 0.018 LM-7 2.09 ± 0.15 1.81 ± 0.42 0.752 ± 0.090 0.360 ± 0.054 U 2 OS 1.56 ± 0.24 1.10 ± 0.16 0.792 ± 0.061 0.429 ± 0.067 MG-63 1.311 ± 0.15 1.15 ± 0.15 1.49 ± 0.17 0.955 ± 0.038 SK-ES-1 2.16 ± 0.18 1.18 ± 0.21 0.789 ± 0.069 0.552 ± 0.067 TC-71 1.91 ± 0.56 1.20 ± 0.14 0.594 ± 0.26 0.756 ± 0.0040 SK-LMS 0.756 ± 0.037 0.993 ± 0.069 1.08 ± 0.18 1.33 ± 0.19 HT1080 0.915 ± 0.15 0.931 ± 0.067 1.10 ± 0.35 1.02 ± 0.34 A673 1.44 ± 0.18 0.820 ± 0.046 1.32 ± 0.12 1.12 ± 0.21 RD 0.970 ± 0.17 1.26 ± 0.090 0.877 ± 0.13 0.971 ± 0.13 RD|18 0.984 ± 0.32 1.25 ± 0.30 0.793 ± 0.10 0.997 ± 0.33

104 Identification of two probe sets with greater fold changes may provide further insight into

the prediction of response

To identify possible probe sets that may convey a greater fold change between

response groups, we altered our search criteria by increasing the ratio of the maximum to

minimum ratio value for each probe set to 10 and searched for probe sets with ≥ 5.0 or ≤ -

5.0 fold change between groups. Of the genes provided from this search two probe sets

with interesting potential were followed-up; Ephrin-A1, the ligand of a known target of

dasatinib, EphA2 receptor, and Dapper, an antagonist of catenin, both of which have

recently been shown to play unique roles in the malignant phenotype. When clustered

based on the expression of these two genes, the heatmap generated from this search

shows that the responders and non-responders completely cluster apart on separate arms

when the fold change is calculated using the median probe set intensity as comparison or

the median non-responder probe set intensity (Figures 22A and 23A). In addition, a

graphic depiction of the Log2 transformed fold changes for these two genes shows greater expression of both Ephrin-A1 and Dapper in the responsive cell lines as compared to the

non-responsive cell lines using either median comparison group (Figures 22B and 23B).

The values of relative fold change are also shown on Tables 10 and 11 for both analyses.

Quantitative RT-PCR was completed on the cell lines for ephrin-A1 and dapper to validate the microarray results. Figure 24 shows the graphic expression of these two genes compared to the median expression of all samples, and Table 11 shows the absolute values of the average expression for ephrin-A1 and dapper in the cell lines generated from the analysis. The validation concludes that the microarrays accurately revealed an increased expression of Ephrin-A1 and Dapper in the responsive cell lines.

105

3 Ephrin-A1 Dapper 2

-1

-2

(Log2) Change Fold Relative -3

-4

-5 SaOS-2 LM-2 LM-7 U 2 OS MG-63 SK-ES-1 TC-71 SK-LMS HT1080 A673 RD RD|18

Figure 22. Expression of Ephrin-A1 and Dapper in sarcoma cell lines using the median of all samples as comparison group. A, heatmap of Ephrin-A1 and Dapper fold changes in 12 sarcoma cell lines. The average was determined for each sample set and the fold change was calculated by using the median of all samples as the comparison group for each probe set. B, graph of Ephrin-A1 and Dapper fold change between sample sets. The fold changes were calculated by transforming the fold changes calculated in A by Log2.

106

5 Ephrin-A1 Dapper 4

-1

(Log2) Change Fold Relative -2

-3

-4 U 2 OS SK-ES-1 TC-71 SK-LMS A673 SaOS-2 LM-2 LM-7 MG-63 HT1080 RD RD|18

Figure 23. Expression of Ephrin-A1 and Dapper in sarcoma cell lines using the median of the non-responders as comparison group. A, heatmap of Ephrin-A1 and Dapper fold changes in 12 sarcoma cell lines. The average was determined for each sample set and the fold change was calculated by using the median of the non-responders as the comparison group for each probe set. B, graph of Ephrin-A1 and Dapper fold change between sample sets. The fold changes were calculated by transforming the fold changes calculated in A by Log2.

107

Table 10. Fold changes for ephrin-a1 and dapper in sarcoma cell lines compared to median intensity

Cell Line Ephrin-A1 Dapper SaOS-2 3.71 ± 0.66 5.39 ± 1.5 LM-2 3.60 ± 0.93 3.54 ± 0.44 LM-7 2.94 ± 0.87 5.63 ± 1.1 U 2 OS 1.13 ± 0.82 3.02 ± 0.86 MG-63 0.354 ± 0.13 0.252 ± 0.25 SK-ES-1 3.26 ± 0.78 1.06 ± 0.21 TC-71 1.27 ± 0.78 0.914 ± 0.23 SK-LMS 0.136 ± 0.029 0.374 ± 0.047 HT1080 0.159 ± 0.17 0.328 ± 0.092 A673 0.055 ± 0.17 0.954 ± 0.18 RD 0.237 ± 0.31 0.440 ± 0.30 RD|18 0.388 ± 0.078 0.423 ± 0.16

108

Table 11. Microarray fold changes for ephrin-a1 and dapper in sarcoma cell lines compared to median intensity of non-responders

Cell Line Ephrin-A1 Dapper SaOS-2 10.3 ± 1.8 14.9 ± 4.3 LM-2 10.0 ± 2.6 9.77 ± 1.2 LM-7 8.19 ± 2.4 15.5 ± 3.0 U 2 OS 3.16 ± 2.3 8.33 ± 2.4 MG-63 1.38 ± 0.35 0.921 ± 0.69 SK-ES-1 9.10 ± 2.1 2.92 ± 0.13 TC-71 3.55 ± 2.2 2.52 ± 0.62 SK-LMS 4.67 ± 0.081 2.12 ± 0.13 HT0180 0.983 ± 0.47 0.790 ± 0.25 A673 0.705 ± 0.48 3.05 ± 0.49 RD 1.52 ± 0.88 1.02 ± 0.83 RD|18 1.10 ± 0.22 0.772 ± 0.45

109

5.000 Ephrin A1 4.000 Dapper

3.000

2.000

1.000

0.000

-1.000

-2.000

Relative Fold Change (Log2) Change Fold Relative -3.000

-4.000

-5.000

-6.000 SaOS-2 LM-2 LM-7 U 2 OS MG-63 SK-ES-1 TC-71 SK-LMS HT1080 A673 RD RD|18

Figure 24. Quantitative RT-PCR validations of Ephrin-A1 and Dapper genes. Bars represent fold changes for the selected genes. Positive fold change represents upregulated, and negative fold change represents down-regulated in sarcoma cells.

110

Table 12. Quantitative RT-PCR values for ephrin-a1 and dapper in sarcoma cell lines

Cell Line Ephrin-A1 Dapper SaOS-2 1.74 ± 0.11 1.64 ± 0.16 LM-2 4.44 ± 0.28 2.90 ± 0.19 LM-7 5.58 ± 0.40 11.7 ± 0.31 U 2 OS 1.30 ± 0.16 1.76 ± 0.11 MG-63 0.197 ± 0.028 0.228 ± 0.017 SK-ES-1 3.21 ± 0.12 0.603 ± 0.057 TC-71 2.25 ± 0.13 1.68 ± 0.14 SK-LMS 0.474 ± 0.018 1.05 ± 0.022 HT1080 0.212 ± 0.028 0.156 ± 0.011 A673 0.603 ± 0.075 0.954 ± 0.11 RD 0.406 ± 0.077 0.061 ± 0.019 RD|18 0.700 ± 0.12 0.189 ± 0.017

111 Testing the molecular signature with cell lines of unknown response reveals that the

molecular signature can accurately predict response to dasatinib in cell lines

To test the molecular signature of response, two new cell lines of unknown

response status were acquired. Microarray and molecular analysis were independently

performed and the response status generated from the two analyses were compared. To

determine response status of the new cell lines the fold change of each probe set was

calculated by taking the ratio the average intensity for each probe set in the new cell lines

to the median intensity of the original dataset and compared the fold change to the

molecular signature extracted from the original group of cell lines. Upon further

examination of the gene expression profile of the two cell lines and comparison to the

molecular signature, HOS was characterized as a responsive cell line and SW1353 was

characterized as a non-responsive cell line. Each cell line also clustered with its

respective group as is observed in Figure 25A.

In addition, the expression of Histone H3, FAF1 and δ-catenin were significantly

different in each of the cell lines. Histone H3 and FAF1 had greater expression, while δ-

catenin had a decreased expression in the HOS cell line (Figure 25B, Table 13). The

expression of these three genes correlates with the expected expression patterns

according to the predictive molecular signature. However, when q-RT-PCR was

performed to validate the expression of these genes in the HOS and SW1353 cell lines,

Histone H3 and FAF1 were significantly upregulated in the HOS cell line compared to

the SW1353 cell line. The expression of δ-catenin was not significantly different, although it was upregulated in SW1353 as compared to HOS. While, the expression α-

112 catenin was also significantly different in the two cell lines, the opposite expression pattern was observed than expected.

To validate the characterization of the cell lines dose response to dasatinib was completed in both cell lines at 6, 24 and 72 h. Dasatinib inhibited Src activation, signaling and cellular migration in both cell lines (Figure 26A-E). However, HOS was the more sensitive of the two cell lines, in that a lower dose of dasatinib was required to inhibit Src activation. Furthermore, dasatinib also induced apoptosis in the HOS cell line and not the SW1353 cell line (Figure 27A and B). These results confirm the characterization of the cell lines by microarray analysis. Thus, these data successfully validated the use of our signature in the ability to predict the response of human sarcoma cell lines to dasatinib.

On the other hand, Ephrin-A1 and Dapper expression did not correlate with the expected expression pattern in accordance with response. HOS was observed as having significantly lower expression of both genes as compared to SW1353 (Figure 28A, C and

29A, C). Furthermore, when clustered based on the expression of these two genes, HOS and SW1353 clustered with the opposite response groups (Figures 28B and 28B). The values of Ephrin-A1 and Dapper can be further assessed in Tables 15 and 16. Several possibilities could explain the lack of validation of Ephrin-A1 and Dapper as predictive genes and will be further articulated in the discussion.

113

114

1 * Histone H3 Fas Associated Factor 0.8 α -Catenin δ-Catenin

0.6 **

0.4

0.2 * 0 Relative Fold Change (Log2) Change Fold Relative

-0.2

-0.4

HOS SW1353 -0.6

115

C 1.8 Histone H3 ** Fas Associated Factor α-Catenin 1.6 δ-Catenin ** 1.4 **

1.2

0.8

RelativeFold Change (Log 2) 0.6

0.4

0.2

0 HOS SW1353

Figure 25. Molecular signature that predicts response to dasatinib generated using the median value for each probe set as comparison group. A, the same molecular signature was used to create this heatmap as for figure 20A. The only difference in this heatmap is the presence of two new cell lines, HOS and SW1353. The average was determined for each sample set and fold change was calculated using median of all samples for the comparison group to calculate the fold changes utilized for this heatmap. Samples were clustered based on the GEP of the 26 probe set list. B, graph expressing fold changes of Histone H3, FAF1, α-catenin and δ-catenin in HOS and SW1353 cells. The fold changes calculated for A were transformed by Log2 and the values graphed to depict accurate fold changes between samples. C, graph expressing the relative qRT-PCR fold changes of Histone H3, FAF1, α-catenin and δ-catenin in HOS and SW1353 cells using the median expression for comparison. (* p-value ≤ 0.01, ** p-value ≤ 0.0001)

116

Table 13. Fold changes key signature genes in test cells compared to median intensity

Fas Associated Cell Line Histone H3 α-Catenin δ-Catenin Factor HOS 1.16 ± 0.14 0.907 ± 0.072 0.93 ± 0.12 1.04 ± 0.12 SW1353 1.05 ± 0.17 0.508 ± 0.040 0.88 ± 0.086 1.20 ± 0.063

Table 14. Relative qRT-PCR fold changes key signature genes in test cells compared to median intensity

Fas Associated Cell Line Histone H3 α-Catenin δ-Catenin Factor HOS 3.05 ± 0.035 2.46 ± 0.11 2.51 ± 0.12 1.21 ± 0.045 SW1353 1.91 ± 0.017 1.21 ± 0.15 1.77 ± 0.085 1.62 ± 0.0023

117

A DMSO30 100 300 1000 nM p-Src (Y419) Src ß-Actin

B DMSO30 100 300 1000 nM p-FAK (Y397)

p-FAK (Y576/577)

p-FAK (Y925)

FAK

p-p130CAS

p130CAS

ß-Actin

C DMSO30 100 300 1000 nM p-Src (Y419) Src ß-Actin

D DMSO30 100 300 1000 nM p-FAK (Y397) p-FAK (Y576/577) p-FAK (Y925)

FAK p-p130CAS

p130CAS

ß-Actin

118

120.0 HOS SW1353 100.0

80.0 )

60.0

40.0 *

Migration (% Control * 20.0

0.0 T0 0 30 100 300 1000

-20.0 dasatinib, nM

Figure 26. Dasatinib inhibits Src activation and signaling in HOS and SW1353 cells. A and C, HOS and SW1353 cells were treated with dasatinib in a dose-dependent manner for 6 h. Cell-free extracts were immunoblotted with antibodies specific to p-Src (Y419) and total Src. B and D, dasatinib specifically blocks tyrosyl phosphorylation of FAK (Y576/577, Y925) and 130CAS (Y410), but not FAK Y397. HOS and SW1353 cells were treated with dasatinib for 6 h in a dose-dependent manner. Cell-free extracts were immunoblotted with antibodies specific to p-FAK (Y397, Y576/577, Y925), total FAK, p130CAS and p-p130CAS (Y410). β-actin was blotted for as a loading control in all experiments. Dasatinib inhibits cell motility and invasion. E, wound healing assays were performed on HOS and SW1353 to determine the effects of dasatinib on inhibiting cell migration as completed previously. Cell migration was visualized at 10x magnification by light microscopy and photographed with a digital camera. The width of voided area versus dasatinib dose concentration was graphed to express the degree of inhibition of cell migration. The number of pixels within the denuded area were the units used to demonstrate inhibition of cell migration induced by dasatinib (*p<0.001, n=3).

119

A DMSO30 100 300 1000 nM

PARP PARP Cleavage XIAP

ß-Actin

B DMSO30 100 300 1000 nM

PARP PARP Cleavage XIAP

ß-Actin

Figure 27. Dasatinib induces apoptosis in HOS but not SW1353 sarcoma cells. A, dasatinib induces apoptosis in a dose-dependent manner in HOS cell only. Cells were treated with dasatinib for 72 h with escalating doses. Cell-free extracts were immunoblotted with antibodies specific to XIAP and PARP. β-actin was blotted for as a loading control in all experiments.

120

121

4 * Ephrin-A1 Dapper

0 Relative Fold Change (Log2) Change Fold Relative

-1

HOS SW1353 -2

Figure 28. Expression of Ephrin-A1 and Dapper in HOS and SW1353 cells using the median of all samples as comparison group. A, heatmap of Ephrin-A1 and Dapper fold changes in HOS and SW1353 cells. The average was determined for each sample set and the fold change was calculated by using the median of all samples as the comparison group for each probe set. B, integration of HOS and SW1353 into the original heatmap of Ephrin-A1 and Dapper expression. C, graph of Ephrin-A1 and Dapper fold changes in HOS and SW1353. The fold changes were calculated by transforming the fold changes calculated in A by Log2. (* p-value ≤ 0.01)

122

123

4 Ephrin-A1 * 3.5 Dapper

2.5

1.5

Relative Fold Change (Log2) Change Fold Relative 0.5

-0.5 HOS SW1353 -1

Figure 29. Expression of Ephrin-A1 and Dapper in HOS and SW1353 cells using the median of the non-responders comparison group. A, heatmap of Ephrin-A1 and Dapper fold changes in HOS and SW1353 cells. The average was determined for each sample set and the fold change was calculated by using the median of the non-responders as the comparison group for each probe set. B, integration of HOS and SW1353 into the original heatmap of Ephrin-A1 and Dapper expression. C, graph of Ephrin-A1 and Dapper fold changes in HOS and SW1353. The fold changes were calculated by transforming the fold changes calculated in A by Log2. (* p-value ≤ 0.001)

124

Table 15. Fold changes for ephrin-a1 and dapper in test cells compared to median intensity

Cell Line Ephrin A1 Dapper HOS 0.526 ± 0.10 1.03 ± 0.23 SW1353 0.926 ± 0.25 10.1 ± 1.2

Table 16. Microarray fold changes for ephrin-a1 and dapper in test cells compared to median intensity of non-responders

Cell Line Ephrin A1 Dapper HOS 0.534 ± 0.30 1.48 ± 0.31 SW1353 1.33 ± 0.38 4.80 ± 0.17

125 Discussion

Microarray analysis of gene expression profiles in sarcomas is a valuable

technique that provides a comprehensive survey of activated molecular pathways and has

been shown to successfully identify potential therapeutic targets. In this aim, we attempted to identify a molecular signature that will predict response to dasatinib by induction of apoptosis in sarcoma cell lines. As a model, we initially compared the GEP of 12 sarcoma cell lines characterized by response to dasatinib. The unsupervised analysis of GEP data clearly separated the cell lines by tumor type (Figure 19A). Using statistical filtering to identify genes significantly different between the responders and non-responders, the GEP data analyzed created two distinct classes, with all but one cell line from each group clustering within their corresponding classes (Figure 19B). Then, using two robust statistical filtering procedures, a 22 gene signature with a high discrimination power among the resistant and non-resistant groups was identified (Figure

20A and 21A). Interestingly, many of the genes present in the signature are found within

close proximity on chromosome one, particularly Histone H3 and FAF1 are located on

chromosome 1q21, which is an area that has been shown to be amplified in OSA. Of the

22 genes in this signature four genes were selected for further analysis because multiple

probe sets for these genes appeared on the signature with consistent fold changes between

classes; Histone H3, FAF1, α-catenin and δ-catenin. Histone H3 and FAF1 were

upregulated, while α-catenin and δ-catenin were downregulated in the responsive group

as a whole (Figure 19B and 21B; Tables 7 and 9) and these findings were further

validated by qRT-PCR (Table 8, Figure 20).

126 Two other probe sets were identified as potential predictors of response using a variation of our initial analysis; Ephrin-A1 and Dapper. Ephrin-A1, a known target of dasatinib, and Dapper, an antagonist of catenins, were both upregulated in the responsive cell lines (Figures 22A and 23A) and this upregulation was validated by qRT-PCR

(Figure 24, Table 12). Upon identification of an RNA signature, our next goal was to test whether this signature can successfully characterize cell lines into their respective groups based on response to dasatinib. To test the signature we acquired two new cell lines;

HOS, an OSA, and SW1353, a CS cell line of unknown response to dasatinib.

Microarray analysis was performed independently as the molecular biology assays were performed to determine the response status of the two cell lines. The results of both analyses were compared once the results for each were determined. The results of both analyses agreed; the microarray analysis grouped the HOS with the responders and SW1353 with the non-responders based on the expression of the molecular signature

(Figure 25A). Furthermore, Histone H3, FAF1 and δ-catenin expression were significantly different in the two cell lines and corresponded with the expression patterns of their respective classes. The molecular biology analysis determined that HOS responded to dasatinib by induction of apoptosis, while SW1353 merely responded by inhibition of Src signaling and migration (Figure 26A-E and Figure 27A-B).

Alternatively, prediction of response based on the expression of Ephrin-A1 and

Dapper did not successfully classify the cell lines into the correct classes (Figures 28 and

29; Table 15 and 16). In fact, the cell lines were characterized into the complete opposite groups based on the expression of these two genes. One possible explanation for this finding could be that only two cell lines were used to test the predictive characterization

127 of these genes, perhaps if more cell lines were utilized for the validation of Ephrin-A1

and Dapper their ability to be used in predicting response would be more promising.

Another explanation may be that these two genes alone, are not sufficient enough to predict response, but when applied together within the context of the signature may possess more significance.

Nevertheless, the class prediction analysis performed using these 26 probe sets as

a predictor gene list for two cell lines of unknown response successfully classified the

cell lines with their respective response groups. These findings lend evidence to suggest

that this molecular signature can to predict response to dasatinib in cell lines and aid in

identifying cell lines that required Src activity for survival. To further validate this

signature, our next aim was to establish whether the signature was present in primary

human sarcomas. Should the signature be expressed in sarcomas, future clinical trials

with dasatinib could include a component to test the validity of the molecular signature in

predicting response in sarcoma patients.

128

Validation of Gene Expression Profile in Primary Human Sarcomas

The primary objective of this aim is to determine whether the molecular signature that predicts response to dasatinib by induction of apoptosis in cell lines is expressed in human sarcomas. Validating the expression of the signature predicative of response to

Src inhibition may lay the foundation for designing future clinical trials that will help tailor design treatment options for sarcoma patients. In addition to designing trials, the results gained from this research will be useful for the impending clinical trial to evaluate dasatinib as a potential treatment option for sarcomas. If present, the molecular signature could be tested using samples gathered in the future trial and validated as a potential predicative signature. In addition, the results of this study will contribute to our understanding of the response to treatment of tumors where histological evaluation and ancillary studies have proven to be insufficient predictors.

In the present study, we applied microarray gene expression profiling to establish the presence of a candidate molecular signature, from the cell lines that predicts response to dasatinib, in human sarcomas. Our goal was to obtain untreated, primary human sarcoma specimens collected under a general tissue banking consent protocol. The tumor histologies utilized for this aim corresponded to the tumor types represented in the cell line component of the study, including OSA, LMS, FBS and RD. Unfortunately, there was no RNA available for the EWS that were collected under this protocol. To compensate for the lack of samples, sarcomas representing other histologies were also

129 utilized for these investigations. GEPs were established for each specimen and

compared to the molecular signature extracted from the cell lines. The comparison was completed to determine whether components of the cell line derived molecular signature were present in human sarcomas.

To our astonishment, 10 or the 22 sarcomas used for this study expressed significant components of the molecular signature from the cell lines. The expression of the signature also provided a theoretical assignment of potential response to dasatinib.

Furthermore, the expression of the molecular signature was significant enough to cluster

many of the tumors based on their GEPs into their theoretical response groups. These

data lend evidence to suggest that at least components of the cell line signature are

expressed in sarcomas. These results are encouraging and support the implementation of

further investigation into the expression and predictability of this signature to predict

response in primary sarcomas.

130 Results

Unsupervised clustering identified diversity among human sarcoma specimens

Using an unsupervised hierarchical clustering approach, we tried to identify

natural subclasses of tumor specimens as determined by gene expression profiles. We performed the unsupervised clustering on an unfiltered probe list and have shown that

sarcomas are not easily grouped into tumor subclasses based on gene expression alone

(Figure 30A). In fact, the only subgroup that clustered together was the two OSA

specimens. The remaining sarcomas clustered with apparently no preference for tumor

subgroup.

Identification of the cell line molecular signature in human sarcomas

Our next goal was to determine whether components of the 26 gene molecular

signature are present in human sarcoma specimens. To complete this we compared the

intensity of each probe set for the tumors, calculated fold change by comparing the

intensity for each probe set to the median intensity of the dataset and compared the fold change to the molecular signature extracted from the cell lines. Of the 22 tumors analyzed, 10 were characterized as potential responders and 12 were characterized as potential non-responders based on the fold changes of the genes that comprise the molecular signature that predicts response in cell lines (Table 16). A breakdown of the percentage responders and non-responders by tumor subgroups is illustrated in Figure

30B. Figure 30A identifies the potential responsive tumors with red stars above the sample label. Clearly the tumors do not cluster by potential response to dasatinib when

unsupervised and unfiltered.

131

A * *** * * * * * *

132

Predicted Tumor Response

Responders Non- 45% Responders 55%

Responders RD SCS AS CS FBS 0% 10% 0% 10% OSA 10% 20% Non-Responders SCS AS OSA RD LMS 8% 8% 20% 0% 8% CS MFH LPS MFH 17% 20% 10% 8% LPS 0% FBS LMS 17% 34%

Figure 30. Hierarchical clustering of primary sarcoma specimens. The GEP of the tumors was determined by calculating fold change for each probe set, using the average intensity of each probe set across samples as the comparison group. An unsupervised cluster was performed to determine inherent similarities between tumors, A. Sarcomas categorized as potential responders are indicated with *. B, breakdown of tumors categorized as potential responders and non-responders by molecular signature according to sarcoma type.

133 Table 17. Classification of sarcomas using predictive signature from cell lines

Tumor Responder Non-Responder AS T98 R CS T2355 NR CS T2744 NR CS T3109 R FBS T3946 NR FBS T4436 NR FBS T6390 NR LMS T174 R LMS T191 R LMS T294 R LMS T374 NR LMS T466 R LMS T3863 NR LPS T2500 R MFH T1913 NR MFH T5087 NR MFS T6251 NR OSA T1898 R OSA T6357 R RD T395 NR SCS T548 NR SCS T7431 R

Total 10 12

134

A Responders Non-Responders

135

B * * *** ** * * *

Figure 31. Heatmap of molecular signature expression in sarcomas. Fold changes for the molecular signature were calculated by using the average intensity for each probe set as the comparison group. A, heatmap of unclustered tumors arranged by response category. B, heatmap of tumors clustered based on the expression of the molecular signature used to define theoretical response status. Dendogram at top of heatmap shows the relationship of tumors based of the GEP of the molecular signature. Sarcomas categorized as potential responders are indicated with a *.

136 Molecular signature that predicts response to dasatinib in cell lines can be used to group tumors by potential response

An unclustered heatmap of the molecular signature expression in the tumors arranged by potential response does not present a clear separation by expression between the two groups (Figure 31A). Furthermore, when clustered based on the GEP of the molecular signature, the tumors do not completely cluster into their respective groups

(Figure 31B). However, there is some promise to a trend of consistency in the clustering, in that some of the tumors did cluster into their respective groups; 6 out of 10 for the responders clustered together and 8 out of 12 of the non-responders. There are several possibilities to explain this inconsistency and will be addressed in the discussion.

A heatmap was also prepared for the tumor expression of the four genes validated from the predictive molecular signature, Histone H3, FAF1 α-catenin and δ-catenin,

When clustered, again the tumors classified as potential responders and non-responders did not cluster in two separate groups. Although, 70% of responsive tumors clustered together and 50% of the non-responsive tumors clustered together (Figure 31A). The expression patterns of Histone H3, FAF1 α-catenin and δ-catenin are also inconsistent when are pulled from the GEP and compared to the cell line signature (Figure 32A-B,

Table 17). Again, there is a trend of recurrent expression patterns between responsive verses the non-responsive groups, although it is not uniform across all samples.

137

A * ** * * *** * *

138

1.5 Histone H3 Fas Associated Factor α-Catenin δ-Catenin 1

0.5

0 Relative Fold Change (Log2) Fold Relative -0.5

-1

T1913 T3109 T191 T466 T1898 T7431 T2744 T4436 T374 T6251 T548 T98 T174 T294 T2500 T6357 T2355 T3946 T6390 T3863 T5087 T395 -1.5

Figure 32. Expression of Histone H3, FAF1, α-Catenin and δ-Catenin in sarcomas. A, heatmap of Histone H3, FAF1, α-catenin and δ-catenin extracted from molecular signature and clustered based on GEP of these four genes. B, graphic depiction of fold changes associated with Histone H3, FAF1, α-catenin and δ-catenin in tumors. Sarcomas categorized as potential responders are indicated with a *.

139 Table 18. Fold changes for key signature genes in sarcomas

Tumor Histone H3 Fas Associated Factor α Catenin δ Catenin

AS T98 1.22 1.17 0.969 1.17 CS T3109 0.919 0.753 0.875 0.834 LMS T174 1.405 1.15 0.861 0.718 LMS T191 1.24 0.721 1.49 1.59 LMS T294 1.18 0.804 1.09 1.56 LMS T466 1.57 1.31 0.996 1.43 LPS T2500 0.901 0.818 1.06 1.25 OSA T1898 1.23 0.712 1.03 0.965 OSA T6357 1.42 1.24 1.13 0.781 SCS T7431 0.769 0.798 1.84 1.43 CS T2355 1.12 0.661 1.18 0.793 CS T2744 0.611 1.105 0.973 0.902 FBS T3946 0.881 0.940 0.964 1.04 FBS T4436 0.947 1.07 1.40 1.20 FBS T6390 0.859 0.955 0.856 0.693 LMS T374 1.23 1.73 1.83 1.66 LMS T3863 1.04 1.90 0.664 0.898 MFH T1913 0.732 1.87 0.983 0.894 MFH T5087 0.832 0.716 0.963 1.06 MFS T6251 0.611 1.13 0.674 0.686 RD T395 0.789 1.25 1.11 0.901 SCS T548 1.14 1.24 0.740 1.01

140 Analysis of two probe sets with greater fold changes may provide further insight into the prediction of response in tumors

The expression of Ephrin-A1 and Dapper were also analyzed in the tumors specimens. A heatmap with hierarchical clustering produced three classes of tumors based on Ephrin-A1 and Dapper gene expression. While none of the three groups contained all of one response group, the clustering is promising. One group contained one responder and three non-responders, the second group had seven responders and two non-responders, while the third group included two responders and seven non-responders

(Figure 33A). The fold change and expression patterns of Eprhin-A1 and Dapper correlate more with the cell lines, in that more of the tumors that express the potentially responsive signature had greater fold changes for Ephrin-A1 and Dapper as compared to the potentially non-responsive tumors as a whole (Figure 33B, Table 18).

141

A * *** * ** ** *

142

5 Ephrin-A1 Dapper

3 2

1 Relative Fold Change (Log 0

-1

T3109* T191* T466* T1898* T7431* T2744 T4436 T374 T1913 T6251 T548 T98* T174* T294* T2500* T6357* T2355 T3946 T6390 T3863 T5087 T395 -2

Figure 33. Expression of Ephrin-A1 and Dapper in sarcomas. The fold changes of Ephrin-A1 and Dapper were calculated using the median intensity of all samples as the comparison group. A, Hierarchical clustering was performed based on the GEP of Ephrin-A1 and Dapper in sarcomas. B, graphic depiction of Ephrin-A1 and Dapper fold changes between primary sarcoma specimens as compared to the median intensity for each probe set. Sarcomas categorized as potential responders are indicated with a *.

143

Table 19. Fold changes for ephrin-a1 and dapper in sarcomas

Tumor Ephrin-A1 Dapper AS T98 1.48 2.68 CS T3109 2.75 0.953 LMS T174 1.25 1.05 LMS T191 2.67 1.07 LMS T294 0.848 0.469 LMS T466 0.807 3.53 LPS T2500 2.05 3.14 OSA T1898 2.31 3.02 OSA T6357 2.47 0.943 SCS T7431 0.973 0.178 CS T2355 6.46 6.56 CS T2744 0.316 0.666 FBS T3946 3.76 0.707 FBS T4436 0.849 1.05 FBS T6390 1.81 0.447 LMS T374 0.580 1.76 LMS T3863 0.275 0.391 MFH T1913 0.952 0.341 MFH T5087 0.389 1.49 MFS T6251 0.966 1.09 RD T395 1.03 0.627 SCS T548 0.804 0.208

144 Discussion

Traditionally, sarcoma therapy has been selected on the basis of non-molecular considerations such as tumor type, grade and stage. As a result, treatments for sarcomas

have not developed at the same pace as those for other more common tumors, and today

we still accept with fatalism that a fraction of sarcoma patients will respond to a

particular therapy, whereas others will not. We also accept that we have little control over

which outcome will prevail in a given case. New approaches are desperately required to tailor treatments to the individual patient, select from current therapeutic possibilities,

predict therapeutic response and develop novel targeted therapeutic modalities. Previous

studies conducted in our laboratory have demonstrated a GEP in sarcoma cell lines that is

unique to cell lines, which undergo apoptosis when treated with dasatinib at low nanomolar doses. In addition, we have also identified a molecular signature that

successfully predicts response to dasatinib in sarcoma cell lines. In this third aim, we

further investigated the clinical significance of these findings by demonstrating the

expression of the molecular signature in patient tissue samples from 22 diverse sarcoma

specimens.

Intriguingly, this signature theoretically predicted 10 out of 22 tumors to be

responders (Table 16). The 22 gene molecular signature was not able to successfully

cluster all of the tumors according to potential response, however there is promise

because many of the tumors classified with their respective response groups. The fact

that a signature, which was derived from cell lines, can be found in the tumors lends

significant evidence for further research into the possibility of this signature to serve as a predictor of response in sarcomas.

145 While the focus of this dissertation has been placed on the expression of a molecular signature that predicts response to dasatinib, the reverse could also be tested.

A signature that predicts non-response has equally been generated and proven in cell lines. Furthermore, it is expressed in primary sarcomas as well. The flipside to these studies could provide just as much help in designing future studies and identifying other potential therapeutic targets. Either way, the molecular signature of the responders or non-responders can be used to theoretically predict tumors that require Src activity for survival and/or metastasis.

There are several theories that could explain the discrepancies between the GEP of the cell lines and tumors and why there is not a complete correlation with potential response and GEP. One possibility to explain this discrepancy includes the immense differences between gene and protein expression that are inherent when comparing cell lines and tumors. Cell lines are grown independent of a microenvironment and lack interaction with other biological components. Tumors interact with a complex microenvironment and are exposed to a considerable amount of stimuli on an infinite level. These interactions alone (or the lack of) may account for the discrepancies between the GEP. In addition, cell lines may lack the expression of metabolic enzymes present in tumors. Tumors have been shown to become resistant to therapy with time, as in the case with gleevec and GISTs. Cell lines may lack the ability to increase the metabolism of a drug of choice because of the inability of the cells to adapt to their environment as is the case with some tumors. Moreover, it is important to account for the ability of a tumor to become resistant to a therapy as a consequence of altering the expression of genes encoding metabolizing proteins even when a signature is expressed.

146 The purity of the tumors specimens utilized for the microarrays are also of some

concern. While, we tried to utilize samples that are as pure as possible, sarcomas by

nature are heterogeneous both within the tumor sample and among patients. Sarcomas

can be comprised of a variety of cellular types including stromal cells, blood vessels and other undifferentiated cells. The lack of homogeneity of cell type can dilute the signature

in the tumor specimens. There have been advancements in technology which aim at

improving the purity of patient specimens, such as laser capture microdissection. This

would be a wonderful technique to utilize, however it just was not feasible for this study.

The heterogeneity within tumors between sarcoma patients is also a possible explanation

for the molecular signature discrepancies. While, there have been many consistent

mutations and translocations identified in sarcoma specimens, there are still a

considerable amount of diversity among sarcomas, even within the same tumor type.

Plus, it is very difficult to compare tumors among different patients. The diversity among

tumor GEP is as great as the diversity among people in general. However, when a

molecular signature of commonality identified in cell lines can be extracted from a tumor

set as diverse as ours, there is great promise and encouraging evidence to suggest that this

signature may be upheld in sarcomas.

147

Conclusions

Three independently synthesized Src kinase inhibitors were evaluated in these studies. While the three compounds were each designed to target the catalytic domain of

Src, the data generated using dasatinib have presented the most convincing preclinical data to complete further studies with this compound. Dasatinib inhibits Src mediated cellular migration and invasion in all of the Src activated cell lines examined in these studies, however only a subset of bone derived sarcoma cell lines underwent apoptosis in response to dasatinib. These observations suggest that Src associates and interacts with different molecular pathways in sarcoma cell lines. In the bone sarcoma cell lines, Src mediates an induction of a FAK-Cas-migration/invasion/survival pathways, while in the

STS cell lines Src initiates FAK-Cas-migration/invasion only pathways. These findings lend evidence to suggest that Src interacts with diverse signaling molecules which varies with sarcoma type.

The dasatinib data have shown cell lines that respond to dasatinib by induction of apoptosis can be identified using a 22 gene molecular signature. Eleven of 12 human sarcoma cell lines express constitutively activated Src kinase and respond to dasatinib by

inhibition of Src activation and signaling as measured by Western blot analysis and

migration and invasion studies. Of these cell lines, six, which were all derived from bone

sarcomas, were identified as responders to dasatinib as defined by induction of apoptosis.

148 Moreover, blockade of c-Src expression in this subset of sarcoma cells by silencing RNA induces apoptosis, consistent with the critical role of Src-mediated sarcoma cell survival.

Microarray analysis and hierarchical clustering allowed us to identify a 26 probe set list that successfully predicts response to dasatinib. To test the reliability of the molecular signature two cells lines with unknown response to dasatinib were used.

Independent analysis of both the GEP and the molecular response to dasatinib of the cells were compared. The molecular signature that predicts response successfully characterized the cell lines into their respective groups, one cell line, HOS, was a responder and the other, SW1353 was a non-responder. Here, we show a molecular signature predicting response to dasatinib by induction of apoptosis in cell lines.

Furthermore, we performed microarray analysis on 22 human sarcomas of varying sarcoma histology to determine whether components of the molecular signature,

Ephrin-A1 or Dapper expression could be used to characterize the tumors based on potential response to dasatinib. A list of possible responsive and non-responsive tumors was generated based on their GEP. When clustered based on the expression of the 26 probe set molecular signature or Ephrin-A1 and Dapper, the clustered provided encouraging results to suggest that the signature has promise to predict response in tumors. While not all of the sarcomas clustered with their respective groups (potentially, responsive or non-responsive), many of the tumors did cluster with their groups and the predictive trends of Ephrin-A1, Dapper and the molecular signature are up-held. These data suggest that our signature could become useful for the identification of patients eligible for new therapeutic options.

149 Three of the four genes from the molecular signature further validated by qRT-

PCR analysis were upheld as reliable predictors of response; Histone H3, FAF1 and δ- catenin. However, Ephrin-A1 and Dapper were less reliable as predictive markers as reversed expression pattern was evident in these cell lines. The six genes that were validated by qRT-PCR; Histone H3, FAF1, α-catenin, δ-catenin, Ephrin-A1 and Dapper, are all thought to play unique roles in cancer. The functions these genes play in the malignant phenotype lend evidence to further investigate the role of these genes in sarcomas.

Histone H3 and FAF1 were upregulated in the cell lines that respond to dasatinib.

Many types of cancer are associated with translocations or mutations in chromatin- modifying enzymes and regulatory proteins. Histones are integral components of the machinery responsible for regulating gene transcription. Phosphorylation of Histone H3 plays important regulatory signaling roles in the processes of chromosome condensation during mitosis and transcription following external stimulation of gene expression growth factors or stress. Histone H3 is phosphorylated in serine 10 and 28. Fyn, one of many kinases that phosphorylate Ser10. The phosphorylation of Ser10 on Histone H3 is critical for EGF-induced neoplastic cell transformation. This evidence suggests that Histone H3 may serve as a critical target for cancer therapy (278). Interestingly, the overexpression of Histone H3 was upheld in many of the tumors categorized as potential responders.

Histone H3 overexpression in the responsive group may not be easily explained in the context of these studies. However these studies have identified Histone H3 as an overexpressed gene which provides evidence to further investigate its activity in sarcomas and possibly evaluated it as a therapeutic target.

150 While FAF1 overexpression was only upheld in the cell lines, this may be caused

by the inherent difference in phenotype between cell lines and tumors. FAF1 is a pro-

apoptotic factor that is involved with Fas-mediated caspase 8 cleavage and induction of apoptosis (279). The FAF1 protein expression has been demonstrated to be reduced in gastric carcinomas (280). One possible explanation for the overexpression of FAF1 in responsive cell lines may cause the cells to be more sensitive to induction of apoptosis as compared to the non-responsive cell lines. Furthermore, the function of FAF1 has been evaluated only in the context of cell lines, the role of FAF1 may not be consistent in tumors.

Two catenins were underexpressed in the responsive cell lines; α- and δ-catenin.

The expression of these two catenins were less clear cut in the tumors, however there is

evidence to suggest that they may play important function in tumorigenesis. The

formation of tissues and organs largely depends on interactions between neighboring cells. These associations orchestrate the assembly of diverse cell types into organized patterns to establish a complex organism. These interactions also permit adult tissues to perform unique functions, preserve architectural integrity and precisely coordinate the events that enable cells to remodel tissues during normal homeostasis and to synchronize in response to injury to repair tissues. By communicating signals through adhesion receptors, cells can respond and elicit the spatially coordinated events needed to maintain

tissue homeostasis. The imbalance of which is frequently observed in the malignant

phenotype. Catenins are an intracellular anchor proteins that attach cadherins to actin

fibers within the cell and aid in the formation of adherens junctions (AJ), the building

blocks of cellular architecture. α- and δ-catenin are two important AJ catenins that

151 integrate cell to cell junctions and regulate cytoskeletal dynamics by governing signaling pathways involved in morphogenesis, homeostasis and intercellular communication between different cell types within a given tissue (281).

α-catenin binds indirectly to cadherins via interactions with β-catenin, a protein commonly mutated in cancers. Recently, α-catenin has also been found to be an essential in coordinating actin dynamics and inversely correlating cell adhesion with proliferation

(282). It is hypothesized that loss of α-catenin may account for the absence of calcium- dependent cell-cell adhesion in cancers with intact E-cadherin expression. In addition,

CTNNA1, the gene that encodes α-catenin, has been shown to be a tumor suppressing gene (283-285). Furthermore, inactivating mutations of CTNNA1 have been demonstrated in lung, prostate, ovarian, and colon cancer cell lines (286). These cell lines lack normal cell-cell adhesion because the cadherin-catenin interactions are disrupted. IHC has also demonstrated a loss of α-catenin expression in primary tumors

(286). Mutational analysis has also shown that α-catenin expression is decreased in synovial sarcomas (287). Therefore, α-catenin is considered as an invasion suppressor molecule in cancers. The expression of α-catenin was relatively low in all of the sarcoma specimens. This finding may provide evidence to further explore the role of α-catenin in sarcomas. Thus far, α-catenin mutations have been identified in synovial sarcomas, perhaps the expression of α-catenin is so low in the microarray analysis because more sarcoma possess mutations in this gene.

While the responsive cell lines had low levels of δ-catenin expression, the responsive tumors had higher levels as compared to the non-responsive tumors. These findings could lend evidence to suggest differential roles for δ-catenin in cell lines and

152 tumors and could be do to the lack of microenvironment in the cell lines. A role for δ- catenin in cancer has been demonstrated in the literature. Unlike α-catenin whose sequence is considerably different from other catenins, δ-catenin is closely resembles β- catenin. Originally identified as a Src substrate, δ-catenin associates with cadherin- mediated AJs, thus stabilizing them at the membrane. IHC studies have reported loss of delta-catenin in some primary tumors, however mutation of the CTNND1 gene is extremely rare (286).

Ephrin-A1 and Dapper were two other genes identified as potential biomarkers that predict response. Both genes were upregulated in the responsive cell lines.

However, the predictability of response governed by Ephrin-A1 and Dapper expression was not successful when they were tested using the HOS and SW1353 cell lines. On the other hand, Ephrin-A1 and Dapper were overexpressed in the responsive group of tumors compared to the non-responsive group. Furthermore, a role of Ephrin-A and Dapper in cancer has been described.

The Eph family of receptor tyrosine kinases and their cell-presented ligands, the ephrins, are overexpressed in a variety of cancers, including breast and gastrointestinal cancers, melanomas and neuroblastomas. Upon ephrin binding, the Eph receptors are phosphorylated at specific tyrosine residues in the cytoplasmic region, which then serve as docking sites for various signaling molecules. Many SH2 domain-containing proteins

have been found to interact with the phosphotyrosines of activated Eph receptors, including Src and Fyn. These interactions have been implicated in cell adhesion, cell motility, and cytoskeleton rearrangement. However, the mechanisms by which

153 individual molecules and signaling pathways exert specific functions has yet to be

described (288).

Ephrin-A1 is a ligand of EphA2 and is involved in the initiation of cellular

migration, functions as an angiogenic factor in vitro and is essential for TNF-α-induced

angiogenesis. Its expression in tumors correlates with initiating invasion by attracting

endothelial cells. In mice models using Karposi-sarcoma-derived tumors as xenografts, the vasculature invading the tumors expressed increased ephrin-A1 and EphA2 (289).

Further, ephrin-A1 and EphA2 can be detected on the surface of endothelial cells in a

variety of human tumors.

Dapper1, gene mapped to human chromosome 14q22.3 and is deleted in

astrocytoma, was identified as an interacting protein for Disheveled, a Wnt signaling

mediator, and modulates Wnt signaling. WNT signals play key roles in carcinogenesis

and embryogenesis through the specification of cell fate and polarity. Dapper has been

described as an antagonizer Wnt signaling by inducing Disheveled degradation (290).

Consequently, based on the importance of the WNT signaling pathway in cancer, Dapper

is predicted to be potent cancer-associated gene (291).

While Ephrin-A1 and Dapper were not successful at predicting the response of

HOS and SW1353 to dasatinib, the expression of these genes in sarcomas may be of

interest for further studies. Both have significant roles in cancer, and may have a vital

function in sarcomas. Furthermore, EphA1, the receptor for Ephrin-A1, is a known target

of dasatinib. Perhaps the cell lines have provided a vital clue into a significant biomarker

of predicting dasatinib prediction not only in cell lines, but in tumors as well. Only

future studies can tell.

154 One must keep in mind that this signature that predicts response to dasatinib by

induction of apoptosis in sarcoma cell lines is completely based on a genomic expression

analysis. The information gained from a genomic analysis using microarrays is highly

informative, however does not portray the functional events occurring within the system

analyzed. A GEP provides a snapshot of the genes which are expressed at a given point

of time. In comparison, a proteomic analysis would provide a more informative

understanding of the actual events occurring within a system as dictated by the gene

products. Identifying a protein expression profile would present a functional analysis of

the genes expressed and more directly translate the relevant overall effects of the biology

at play within a system and more precisely predict cellular behavior. The information

lost in the translation between gene expression and protein expression may account for

many of the inconsistencies observed throughout the signature analysis within the cell lines and tumors analyzed in these studies. Nevertheless, the identification of a molecular signature that has been proven to predict response to dasatinib in cell lines is an excellent starting point for generating more sophisticated analyses aimed at further identifying a more precise and accurate signature that effectively predicts response to dasatinib in human sarcoma patients.

In conclusion, Src activity is essential for the survival of a subset of sarcomas.

We have identified a molecular signature that can predict the response of cell lines to dasatinib. The molecular signature can successfully identify cell lines that will undergo

apoptosis upon inhibition of Src activity by dasatinib. This molecular signature is also

present to some degree in primary human sarcomas and should be further validated to

determine the efficiency of this signature to predict patient response to dasatinib.

155 Together, these studies reveal that both constitutive Src activation and the expression of a molecular signature that predicts response to dasatinib in cell lines may potentially play a critical role in the response of sarcomas to dasatinib. Thus, novel therapeutic approaches that inhibit Src signaling in sarcoma that express the molecular signature may have the potential to induce apoptosis and sensitize tumors to chemotherapy.

156

Clinical Significance

TKIs have become a viable therapeutic option for the treatment of sarcomas eliciting aberrantly activated tyrosine kinase signaling cascades. Cell line based in vitro and in vivo models have demonstrated mixed, yet promising potential uses of TKIs as single agent or in combination with current treatment options. Proof-of-principle of this concept has been obtained for sarcomas in the case of Gleevec treatment of GISTs and

DFSPs.

Our data offers preclinical evidence to further investigate the response of sarcomas to dasatinib in clinical trials. In that, dasatinib is a promising therapeutic option for treating sarcomas with activated Src kinase because it may prevent metastasis by inhibiting tumor cell migration and invasion, as well as induce apoptosis in a subset of bone derived sarcomas. A molecular signature in the bone subset of sarcomas that predicts apoptosis response to dasatinib has been identified in these studies. The predictive molecular signature of apoptosis response to dasatinib represents a testable hypothesis in clinical trials

TKIs currently in early phase trials have provided promising therapeutic options for treating sarcomas. Further work is required to delineate the role TKs play as effectors of tumorigenesis in sarcoma and differentiate between relevant and irrelevant mutations with regards to upstream and downstream signaling. To be effective, a TKI must target a

157 pathway which is essential and vital for the survival, as in the case of c-KIT in GIST and

PDGFR in DFSP and Src in OSA cell lines.

158 References

1. Helman LJ, Meltzer P. Mechanisms of sarcoma development. Nat Rev Cancer

2003;3:685-694.

2. Jemal A, Tiwari RC, Murray T, et al. Cancer statistics, 2004. CA Cancer J Clin

2004;54:8-29.

3. U.S. Cancer Statistics Working Group [database on the Internet]. United States

Cancer Statistics: 1999-2002 Incidence and Mortality Web–based Report Version.

Atlanta (GA): Department of Health and Human Services, Centers for Disease

Control and Prevention, and National Cancer Institute (US). c2005 [cited 2006

Oct 01]. Available from: www.cdc.gov/cancer/npcr/uscs/

4. Ries L, Harkins D, Krapcho M, et al SEER Cancer Statistics Review, 1975-2003

[database on the Internet]. Bethesda (MD): National Cancer Institute (US). c2005

[cited 2006 Oct 01]. Available from: http://seer.cancer.gov/csr/1975_2003/

5. Albritton KH. Sarcomas in adolescents and young adults. Hematol Oncol Clin

North Am 2005;19:527-546, vii.

6. Bleyer A, Montello M, Budd T, Saxman S. National survival trends of young

adults with sarcoma: lack of progress is associated with lack of clinical trial

participation. Cancer 2005;103:1891-1897.

7. Bleyer A MM, Budd T., et al. Young adults with sarcoma: lack of clinical trials

participation and lack of survival prolongation [abstract 3280]. Proc Am Soc Clin

Oncol 2003;22:816.

159 8. Bleyer W. The adolescent gap in cancer treatment. J Registry Management

1996;23:114-115.

9. Bleyer WA, Tejeda H, Murphy SB, et al. National cancer clinical trials: children

have equal access; adolescents do not. J Adolesc Health 1997;21:366-373.

10. Society AC What are the key statistics about soft tissue sarcomas? MedlinePlus,

2006.

http://www.cancer.org/docroot/cri/content/cri_2_4_1x_what_are_the_key_statisti

cs_for_sarcoma_38.asp?sitearea=cri

11. Gurney JG YJ, Roffers SD, Smith MA, Bunin GR Soft Tissue Sarcomas. SEER

Pediatric Monograph, 1999.

12. Coindre JM, Terrier P, Guillou L, et al. Predictive value of grade for metastasis

development in the main histologic types of adult soft tissue sarcomas: a study of

1240 patients from the French Federation of Cancer Centers Sarcoma Group.

Cancer 2001;91:1914-1926.

13. Society AC. What are the key statistics about bone cancer? 2006.

14. Gurney JG SA, Bulterys M Malignant Bone Tumors. SEER Pediatric Monograph,

1998.

15. American Cancer Society. What are the key statistics about bone cancer?

American Cancer Society, Inc. c2006 [cited 2006 May 16]. Available from:

http://www.cancer.org/docroot/CRI/content/CRI_2_4_1X_What_are_the_key_sta

tistics_for_bone_cancer_2.asp?sitearea=

16. Cancer of soft tissue including heart. SEER [cited 2007 Feb 6], 2002. Available

from: https://seer.cancer.gov/statfacts/html/soft.html

160 17. Pollock R. Soft Tissue Sarcomas. London: Hamilton; 2002.

18. Thomas JO. Acquired immunodeficiency syndrome-associated cancers in Sub-

Saharan Africa. Semin Oncol 2001;28:198-206.

19. Stiller CA, Parkin DM. Geographic and ethnic variations in the incidence of

childhood cancer. Br Med Bull 1996;52:682-703.

20. Zahm SH, Fraumeni JF, Jr. The epidemiology of soft tissue sarcoma. Semin

Oncol 1997;24:504-514.

21. Abramson DH, Ellsworth RM, Kitchin FD, Tung G. Second nonocular tumors in

retinoblastoma survivors. Are they radiation-induced? Ophthalmology

1984;91:1351-1355.

22. Li FP, Fraumeni JF, Jr. Soft-tissue sarcomas, breast cancer, and other neoplasms.

A familial syndrome? Ann Intern Med 1969;71:747-752.

23. Li FP, Fraumeni JF, Jr. Prospective study of a family cancer syndrome. Jama

1982;247:2692-2694.

24. Malkin D The Li-Fraumeni syndrome. In: PPO Updates. Philadelphia: JB

Lipincott, 1995.

25. Malkin D, Li FP, Strong LC, et al. Germ line p53 mutations in a familial

syndrome of breast cancer, sarcomas, and other neoplasms. Science

1990;250:1233-1238.

26. King AA, Debaun MR, Riccardi VM, Gutmann DH. Malignant peripheral nerve

sheath tumors in neurofibromatosis 1. Am J Med Genet 2000;93:388-392.

27. Nishida T, Hirota S, Taniguchi M, et al. Familial gastrointestinal stromal tumours

with germline mutation of the KIT gene. Nat Genet 1998;19:323-324.

161 28. Blair A, Malker H, Cantor KP, Burmeister L, Wiklund K. Cancer among farmers.

A review. Scand J Work Environ Health 1985;11:397-407.

29. Blair A, Zahm SH. Agricultural exposures and cancer. Environ Health Perspect

1995;103 Suppl 8:205-208.

30. Dirix LY, Van Oosterom AT. Diagnosis and treatment of soft tissue sarcomas in

adults. Curr Opin Oncol 1994;6:372-383.

31. Hardell L, Sandstrom A. Case-control study: soft-tissue sarcomas and exposure to

phenoxyacetic acids or chlorophenols. Br J Cancer 1979;39:711-717.

32. Kelly S, Guidotti, TL. Phenoxyacetic acid herbicides and chlorophenols and the

etioloty of lymphoma and soft tissue neoplasms. JAMA 1990;256:1141-1147.

33. Merlo F, Puntoni R. Soft-tissue sarcomas, malignant lymphomas, and 2,3,7,8-

TCDD exposure in Seveso. Lancet 1986;2:1455.

34. Riihimaki V, Asp S, Hernberg S. Mortality of 2,4-dichlorophenoxyacetic acid and

2,4,5-trichlorophenoxyacetic acid herbicide applicators in Finland: first report of

an ongoing prospective cohort study. Scand J Work Environ Health 1982;8:37-42.

35. Saracci R, Kogevinas M, Bertazzi PA, et al. Cancer mortality in workers exposed

to chlorophenoxy herbicides and chlorophenols. Lancet 1991;338:1027-1032.

36. Smith JG, Christophers AJ. Phenoxy herbicides and chlorophenols: a case control

study on soft tissue sarcoma and malignant lymphoma. Br J Cancer 1992;65:442-

448.

37. Wiklund K, Holm LE. Soft tissue sarcoma risk in Swedish agricultural and

forestry workers. J Natl Cancer Inst 1986;76:229-234.

162 38. Wiklund TA, Blomqvist CP, Raty J, et al. Postirradiation sarcoma. Analysis of a

nationwide cancer registry material. Cancer 1991;68:524-531.

39. The association of selected cancers with service in the US military in Vietnam. II.

Soft-tissue and other sarcomas. The Selected Cancers Cooperative Study Group.

Arch Intern Med 1990;150:2485-2492.

40. Bertazzi PA, Consonni D, Bachetti S, et al. Health effects of dioxin exposure: a

20-year mortality study. Am J Epidemiol 2001;153:1031-1044.

41. Bertazzi PA, Zocchetti C, Guercilena S, et al. Dioxin exposure and cancer risk: a

15-year mortality study after the "Seveso accident". Epidemiology 1997;8:646-

652.

42. Brann E. The association of selected cancers with service in the US military in

Vietnam. II. Soft tissue and other sarcomas. Arch Intern Med 1990;150:2485-

2492.

43. Constable JD, Timperi R, Clapp R, Antman K, Boynton B. Vietnam veterans and

soft tissue sarcoma. J Occup Med 1987;29:726.

44. Fingerhut MA, Halperin WE, Marlow DA, et al. Cancer mortality in workers

exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. N Engl J Med 1991;324:212-218.

45. Greenwald P, Kovasznay B, Collins DN, Therriault G. Sarcomas of soft tissues

after Vietnam service. J Natl Cancer Inst 1984;73:1107-1109.

46. Kang H, Enziger F, Breslin P, al e. Soft tissue sarcoma and military service in

Vietnam: a case control study. J Nat Cancer Inst 1987;79:693-699.

163 47. Kang HK, Weatherbee L, Breslin PP, Lee Y, Shepard BM. Soft tissue sarcomas

and military service in Vietnam: a case comparison group analysis of hospital

patients. J Occup Med 1986;28:1215-1218.

48. Pierce S. The purported link between agent orange and cancer. Hum Pathol

1995;26:693-695.

49. Sarma PR, Jacobs J. Thoracic soft-tissue sarcoma in Vietnam veterans exposed to

Agent Orange. N Engl J Med 1982;306:1109.

50. Engels EA, Katki HA, Nielsen NM, et al. Cancer incidence in Denmark following

exposure to poliovirus vaccine contaminated with simian virus 40. J Natl Cancer

Inst 2003;95:532-539.

51. McClain KL, Leach CT, Jenson HB, et al. Association of Epstein-Barr virus with

leiomyosarcomas in children with AIDS. N Engl J Med 1995;332:12-18.

52. Amendola BE, Amendola MA, McClatchey KD, Miller CH, Jr. Radiation-

associated sarcoma: a review of 23 patients with postradiation sarcoma over a 50-

year period. Am J Clin Oncol 1989;12:411-415.

53. Aoki T, Ozeki Y, Watanabe M, et al. Development of primary leiomyosarcoma of

the sternum postirradiation: report of a case. Surg Today 1998;28:1326-1328.

54. Brady MS, Gaynor JJ, Brennan MF. Radiation-associated sarcoma of bone and

soft tissue. Arch Surg 1992;127:1379-1385.

55. Cahan W, Woodard, HQ, Higinbotham, NL, et al. Sarcoma arising in irradiated

bone- report of 11 cases. Cancer 1948;1:3-29.

56. Chang SM, Barker FG, 2nd, Larson DA, Bollen AW, Prados MD. Sarcomas

subsequent to cranial irradiation. Neurosurgery 1995;36:685-690.

164 57. Hatfield P, Schultz, MD. Postirradiation sarcoma, including five cases after x-ray

therapy of breast carcinoma. Radiology 1970;96:593-602.

58. Hawkins MM. Second primary tumors following radiotherapy for childhood

cancer. Int J Radiat Oncol Biol Phys 1990;19:1297-1301.

59. Huvos AG, Woodard HQ, Cahan WG, et al. Postradiation osteogenic sarcoma of

bone and soft tissues. A clinicopathologic study of 66 patients. Cancer

1985;55:1244-1255.

60. Kim JH, Chu FC, Woodard HQ, et al. Radiation-induced soft-tissue and bone

sarcoma. Radiology 1978;129:501-508.

61. Lagrange JL, Ramaioli A, Chateau MC, et al. Sarcoma after radiation therapy:

retrospective multiinstitutional study of 80 histologically confirmed cases.

Radiation Therapist and Pathologist Groups of the Federation Nationale des

Centres de Lutte Contre le Cancer. Radiology 2000;216:197-205.

62. Laskin WB, Silverman TA, Enzinger FM. Postradiation soft tissue sarcomas. An

analysis of 53 cases. Cancer 1988;62:2330-2340.

63. Mark RJ, Bailet JW, Poen J, et al. Postirradiation sarcoma of the head and neck.

Cancer 1993;72:887-893.

64. Mark RJ, Poen J, Tran LM, et al. Postirradiation sarcomas. A single-institution

study and review of the literature. Cancer 1994;73:2653-2662.

65. Martland H. Occupational Poisoning in Manufacture of Luminous Watch Dials.

JAMA 1929;92:466-473.

165 66. Menu-Branthomme A, Rubino C, Shamsaldin A, et al. Radiation dose,

chemotherapy and risk of soft tissue sarcoma after solid tumours during

childhood. Int J Cancer 2004;110:87-93.

67. Mermershtain W, Cohen AD, Koretz M, Cohen Y. Cutaneous angiosarcoma of

breast after lumpectomy, axillary lymph node dissection, and radiotherapy for

primary breast carcinoma: case report and review of the literature. Am J Clin

Oncol 2002;25:597-598.

68. Miracco C, Materno M, De Santi MM, Pirtoli L, Ninfo V. Unusual second

malignancies following radiation therapy: subcutaneous pleomorphic

rhabdomyosarcoma and cutaneous melanoma. Two case reports. J Cutan Pathol

2000;27:419-422.

69. Phillips TL, Sheline GE. Bone Sarcomas Following Radiation Therapy.

Radiology 1963;81:992-996.

70. Scanlon E, Berk, RS, Khandekar, JD. Postirradiation Neoplasia: A Symposium.

Curr Prob Cancer 1978;111:4-35.

71. Taghian A, de Vathaire F, Terrier P, et al. Long-term risk of sarcoma following

radiation treatment for breast cancer. Int J Radiat Oncol Biol Phys 1991;21:361-

367.

72. Tountas AA, Fornasier VL, Harwood AR, Leung PM. Postirradiation sarcoma of

bone: a perspective. Cancer 1979;43:182-187.

73. Tucker MA, D'Angio GJ, Boice JD, Jr., et al. Bone sarcomas linked to

radiotherapy and chemotherapy in children. N Engl J Med 1987;317:588-593.

166 74. Weatherby RP, Dahlin DC, Ivins JC. Postradiation sarcoma of bone: review of 78

Mayo Clinic cases. Mayo Clin Proc 1981;56:294-306.

75. Wei LH, Huang CY, Cheng SP, Chen CA, Hsieh CY. Carcinosarcoma of ovary

associated with previous radiotherapy. Int J Gynecol Cancer 2001;11:81-84.

76. Zejda JE, McDuffie HH, Dosman JA. Epidemiology of health and safety risks in

agriculture and related industries. Practical applications for rural physicians. West

J Med 1993;158:56-63.

77. Huang J, Mackillop WJ. Increased risk of soft tissue sarcoma after radiotherapy in

women with breast carcinoma. Cancer 2001;92:172-180.

78. Karlsson P, Holmberg E, Samuelsson A, Johansson KA, Wallgren A. Soft tissue

sarcoma after treatment for breast cancer--a Swedish population-based study. Eur

J Cancer 1998;34:2068-2075.

79. Rubino C, Shamsaldin A, Le MG, et al. Radiation dose and risk of soft tissue and

bone sarcoma after breast cancer treatment. Breast Cancer Res Treat

2005;89:277-288.

80. Le Vu B, de Vathaire F, Shamsaldin A, et al. Radiation dose, chemotherapy and

risk of osteosarcoma after solid tumours during childhood. Int J Cancer

1998;77:370-377.

81. dos Santos Silva I, Jones M, Malveiro F, Swerdlow A. Mortality in the

Portuguese thorotrast study. Radiat Res 1999;152:S88-92.

82. Mori T, Kido C, Fukutomi K, et al. Summary of entire Japanese thorotrast follow-

up study: updated 1998. Radiat Res 1999;152:S84-87.

167 83. Platz EA, Wiencke JK, Kelsey KT, et al. Chromosomal aberrations and hprt

mutant frequencies in long-term American thorotrast survivors. Int J Radiat Biol

2000;76:955-961.

84. Ron E. Cancer risks from medical radiation. Health Phys 2003;85:47-59.

85. Fodor J, Orosz Z, Szabo E, et al. Angiosarcoma after conservation treatment for

breast carcinoma: our experience and a review of the literature. J Am Acad

Dermatol 2006;54:499-504.

86. Tomasini C, Grassi M, Pippione M. Cutaneous angiosarcoma arising in an

irradiated breast. Case report and review of the literature. Dermatology

2004;209:208-214.

87. Neglia JP, Friedman DL, Yasui Y, et al. Second malignant neoplasms in five-year

survivors of childhood cancer: childhood cancer survivor study. J Natl Cancer Inst

2001;93:618-629.

88. Saggia C, Forti G, Biaggi G, et al. Two cases of secondary soft tissue sarcomas

after radiotherapy and radiochemotherapy. Tumori 2004;90:622-624.

89. Fotiou S, Hatjieleftheriou G, Kyrousis G, Kokka F, Apostolikas N. Long-term

tamoxifen treatment: a possible aetiological factor in the development of uterine

carcinosarcoma: two case-reports and review of the literature. Anticancer Res

2000;20:2015-2020.

90. Kloos I, Delaloge S, Pautier P, et al. Tamoxifen-related uterine carcinosarcomas

occur under/after prolonged treatment: report of five cases and review of the

literature. Int J Gynecol Cancer 2002;12:496-500.

168 91. Martin-Loeches M, Rius J, Orti RM. Uterine sarcoma associated with tamoxifen

use: case report. Eur J Gynaecol Oncol 2003;24:202-203.

92. Yildirim Y, Inal MM, Sanci M, et al. Development of uterine sarcoma after

tamoxifen treatment for breast cancer: report of four cases. Int J Gynecol Cancer

2005;15:1239-1242.

93. Belli S, Bertazzi PA, Comba P, et al. A cohort study on vinyl chloride

manufacturers in Italy: study design and preliminary results. Cancer Lett

1987;35:253-261.

94. Buffler PA, Wood S, Eifler C, Suarez L, Kilian DJ. Mortality experience of

workers in a vinyl chloride monomer production plant. J Occup Med

1979;21:195-203.

95. Byren D, Engholm G, Englund A, Westerholm P. Mortality and cancer morbidity

in a group of Swedish VCM and PCV production workers. Environ Health

Perspect 1976;17:167-170.

96. Elliott P, Kleinschmidt I. Angiosarcoma of the liver in Great Britain in proximity

to vinyl chloride sites. Occup Environ Med 1997;54:14-18.

97. Evans DG, Baser ME, McGaughran J, et al. Malignant peripheral nerve sheath

tumours in neurofibromatosis 1. J Med Genet 2002;39:311-314.

98. Cooper WC. Epidemiologic study of vinyl chloride workers: mortality through

December 31, 1972. Environ Health Perspect 1981;41:101-106.

99. Makk L, Delorme F, Creech JL. [Angiosarcomas of the liver in workers having

had prolonged contact with vinyl chloride: epidemiology and program of research

in the workers]. Union Med Can 1975;104:1833-1835.

169 100. Muller R, Hajdu SI, Brennan MF. Lymphangiosarcoma associated with chronic

filarial lymphedema. Cancer 1987;59:179-183.

101. Stewart F, Treves N. Lymphoangiosarcoma in postmastectomy lymphedema: a

report of six cases of elephantiasis chirurgica. Cancer 1948;1:64.

102. Brennan M, Lewis J. Diagnosis and Management of Soft Tissue Sarcoma.

London: Martin Dunitz; 2002.

103. Operskalski EA, Preston-Martin S, Henderson BE, Visscher BR. A case-control

study of osteosarcoma in young persons. Am J Epidemiol 1987;126:118-126.

104. Ben-Izhak O, Kerner H, Brenner B, al e. Angiosarcoma of the colon developing

in a capsule of a foreign body. Am J of Clin Path 1991;97:416-420.

105. Ben-Izhak O, Vlodavsky E, Ofer A, et al. Epithelioid angiosarcoma associated

with a Dacron vascular graft. Am J Surg Pathol 1999;23:1418-1422.

106. Brand K Foreign body induced sarcomas. In: B. FF (ed.), Cancer. New York:

Plenum, 1975. pp. 485-489.

107. Burns W, Kanhouwa S, Tillman L, Saini N, Herrmann J. Fibrosarcoma occurring

at the site of a plastic vascular graft. Cancer 1971;29:66-72.

108. Dube VE, Fisher DE. Hemangioendothelioma of the leg following metallic

fixation of the tibia. Cancer 1972;30:1260-1266.

109. Fyfe BS, Quintana CS, Kaneko M, Griepp RB. Aortic sarcoma four years after

Dacron graft insertion. Ann Thorac Surg 1994;58:1752-1754.

110. Hayman J, Huygens H. Angiosarcoma developing around a foreign body. J Clin

Pathol 1983;36:515-518.

170 111. Lindeman G, McKay MJ, Taubman KL, Bilous AM. Malignant fibrous

histiocytoma developing in bone 44 years after shrapnel trauma. Cancer

1990;66:2229-2232.

112. Aboulafia AJ, Brooks F, Piratzky J, Weiss S. Osteosarcoma arising from

heterotopic ossification after an electrical burn. A case report. J Bone Joint Surg

Am 1999;81:564-570.

113. Can Z, Yilmaz S, Riza A, Apaydin EI, Kuzu I. Sarcoma developing in a burn

scar: case report and review of the literature. Burns 1998;24:68-71.

114. Davis GB, Blanchard DK, Hatch GF, 3rd, et al. Tumors of the stomach. World J

Surg 2000;24:412-420.

115. Ju DM. Fibrosarcoma arising in surgical scars. Plast Reconstr Surg 1966;38:429-

437.

116. Micali G, Gasparri O, Nasca MR, Sapuppo A. Kaposi's sarcoma occurring de

novo in the surgical scar in a heart transplant recipient. J Am Acad Dermatol

1992;27:273-274.

117. Nakanishi H, Tomita Y, Yoshikawa H, et al. Frequent p53 gene mutations in soft

tissue sarcomas arising in burn scar. Jpn J Cancer Res 1999;90:276-279.

118. Ozyazgan I, Kontas O. Burn scar sarcoma. Burns 1999;25:455-458.

119. Ugurlu K, Turgut G, Kabukcuoglu F, et al. Malignant fibrous histiocytoma

developing in a burn scar. Burns 1999;25:764-767.

120. Gottlieb JA, Baker LH, Quagliana JM, et al. Chemotherapy of sarcomas with a

combination of adriamycin and dimethyl triazeno imidazole carboxamide. Cancer

1972;30:1632-1638.

171 121. Bramwell VH, Mouridsen HT, Santoro A, et al. Cyclophosphamide versus

ifosfamide: final report of a randomized phase II trial in adult soft tissue

sarcomas. Eur J Cancer Clin Oncol 1987;23:311-321.

122. Joensuu H, Roberts PJ, Sarlomo-Rikala M, et al. Effect of the tyrosine kinase

inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor. N

Engl J Med 2001;344:1052-1056.

123. Blanke C, con Mehren H, Joensuu P, al. e. Evaluation of the safety and efficacy of

an oral molecularly-targeted therapy STI571, in patients with unresectable or

metastatic gastrointestinal stromal tumors (GISTS) expressing c-kit (CD117).

ASCO Proc. 2001;20.

124. Manning G, Whyte DB, Martinez R, Hunter T, Sudarsanam S. The protein kinase

complement of the human genome. Science 2002;298:1912-1934.

125. Alonso A, Sasin J, Bottini N, et al. Protein tyrosine phosphatases in the human

genome. Cell 2004;117:699-711.

126. Koon HB, Bubley GJ, Pantanowitz L, et al. Imatinib-induced regression of AIDS-

related Kaposi's sarcoma. J Clin Oncol 2005;23:982-989.

127. McArthur GA, Demetri GD, van Oosterom A, et al. Molecular and clinical

analysis of locally advanced dermatofibrosarcoma protuberans treated with

imatinib: Imatinib Target Exploration Consortium Study B2225. J Clin Oncol

2005;23:866-873.

128. Daw NC, Furman WL, Stewart CF, et al. Phase I and pharmacokinetic study of

gefitinib in children with refractory solid tumors: a Children's Oncology Group

Study. J Clin Oncol 2005;23:6172-6180.

172 129. Stewart CF, Leggas M, Schuetz JD, et al. Gefitinib enhances the antitumor

activity and oral bioavailability of irinotecan in mice. Cancer Res 2004;64:7491-

7499.

130. Kuenen BC, Tabernero J, Baselga J, et al. Efficacy and toxicity of the

angiogenesis inhibitor SU5416 as a single agent in patients with advanced renal

cell carcinoma, melanoma, and soft tissue sarcoma. Clin Cancer Res

2003;9:1648-1655.

131. Angelov L, Salhia B, Roncari L, McMahon G, Guha A. Inhibition of angiogenesis

by blocking activation of the vascular endothelial growth factor receptor 2 leads

to decreased growth of neurogenic sarcomas. Cancer Res 1999;59:5536-5541.

132. Dalal S, Berry AM, Cullinane CJ, et al. Vascular endothelial growth factor: a

therapeutic target for tumors of the Ewing's sarcoma family. Clin Cancer Res

2005;11:2364-2378.

133. Gingrich DE, Reddy DR, Iqbal MA, et al. A new class of potent vascular

endothelial growth factor receptor tyrosine kinase inhibitors: structure-activity

relationships for a series of 9-alkoxymethyl-12-(3-hydroxypropyl)indeno[2,1-

a]pyrrolo[3,4-c]carbazole-5- ones and the identification of CEP-5214 and its

dimethylglycine ester prodrug clinical candidate CEP-7055. J Med Chem

2003;46:5375-5388.

134. Scotlandi K, Manara MC, Strammiello R, et al. C-kit receptor expression in

Ewing's sarcoma: lack of prognostic value but therapeutic targeting opportunities

in appropriate conditions. J Clin Oncol 2003;21:1952-1960.

173 135. Shor AC, Keschman EA, Lee FY, et al. Dasatinib inhibits migration and invasion

in diverse human sarcoma cell lines and induces apoptosis in bone sarcoma cells

dependent on Src kinase for survival. Cancer Res 2007;In Press.

136. Balasubramanian L, Evens AM. Targeting angiogenesis for the treatment of

sarcoma. Curr Opin Oncol 2006;18:354-359.

137. Knowling M, Blackstein M, Tozer R, et al. A phase II study of perifosine (D-

21226) in patients with previously untreated metastatic or locally advanced soft

tissue sarcoma: A National Cancer Institute of Canada Clinical Trials Group trial.

Invest New Drugs 2006;24:435-439.

138. Van Ummersen L, Binger K, Volkman J, et al. A phase I trial of perifosine (NSC

639966) on a loading dose/maintenance dose schedule in patients with advanced

cancer. Clin Cancer Res 2004;10:7450-7456.

139. Guan H, Jia SF, Zhou Z, Stewart J, Kleinerman ES. Herceptin down-regulates

HER-2/neu and vascular endothelial growth factor expression and enhances taxol-

induced cytotoxicity of human Ewing's sarcoma cells in vitro and in vivo. Clin

Cancer Res 2005;11:2008-2017.

140. Ye D, Maitra A, Timmons CF, et al. The epidermal growth factor receptor HER2

is not a major therapeutic target in Ewing sarcoma. J Pediatr Hematol Oncol

2003;25:459-466.

141. Azizi AA, Haberler C, Czech T, et al. Vascular-endothelial-growth-factor

(VEGF) expression and possible response to angiogenesis inhibitor bevacizumab

in metastatic alveolar soft part sarcoma. Lancet Oncol 2006;7:521-523.

174 142. D'Adamo DR, Anderson SE, Albritton K, et al. Phase II study of doxorubicin and

bevacizumab for patients with metastatic soft-tissue sarcomas. J Clin Oncol

2005;23:7135-7142.

143. DeMatteo RP, Lewis JJ, Leung D, et al. Two hundred gastrointestinal stromal

tumors: recurrence patterns and prognostic factors for survival. Ann Surg

2000;231:51-58.

144. Goss GA MP, Manola K, SInger S, Fletcher CD, Demetri GD. Clinical and

pathological characteristics of gastrointestinal stromal tumors (GIST). Prog Proc

Am Soc Clin Oncol 2000;19:abstract.

145. Joensuu H. Gastointestinal stromal tumor (GIST). Annals of Oncology

2006;17:x280-x286.

146. Rubin BP, Singer S, Tsao C, et al. KIT activation is a ubiquitous feature of

gastrointestinal stromal tumors. Cancer Res 2001;61:8118-8121.

147. Blanke CD, Eisenberg BL, Heinrich MC. Gastrointestinal stromal tumors. Curr

Treat Options Oncol 2001;2:485-491.

148. van Oosterom AT, Judson I, Verweij J, et al. Safety and efficacy of imatinib

(STI571) in metastatic gastrointestinal stromal tumours: a phase I study. Lancet

2001;358:1421-1423.

149. Duensing A, Heinrich MC, Fletcher CD, Fletcher JA. Biology of gastrointestinal

stromal tumors: KIT mutations and beyond. Cancer Invest 2004;22:106-116.

150. Heinrich M, Corless C, Blanke C, at. e. Muational status of the KIT oncoprotein

correlates with the clincinal responses to Imatinib mesylate (Gleevec) in patients

with metastatic gastrointestinal stromal tumors (GISTs). ASCO Proc. 2002;21.

175 151. Demetri GD, von Mehren M, Blanke CD, et al. Efficacy and safety of imatinib

mesylate in advanced gastrointestinal stromal tumors. N Engl J Med

2002;347:472-480.

152. Verweij J, Casali PG, Zalcberg J, et al. Progression-free survival in

gastrointestinal stromal tumours with high-dose imatinib: randomised trial. Lancet

2004;364:1127-1134.

153. Greco A, Roccato E, Miranda C, et al. Growth-inhibitory effect of STI571 on

cells transformed by the COL1A1/PDGFB rearrangement. Int J Cancer

2001;92:354-360.

154. Yoshitani K, Honoki K, Morishita T, et al. Growth inhibition of rat osteosarcoma

and malignant fibrous histiocytoma cells by tyrosine kinase inhibitor STI571. In

Vivo 2003;17:255-258.

155. McGary EC, Weber K, Mills L, et al. Inhibition of platelet-derived growth factor-

mediated proliferation of osteosarcoma cells by the novel tyrosine kinase inhibitor

STI571. Clin Cancer Res 2002;8:3584-3591.

156. Matsuyama S, Iwadate M, Kondo M, et al. SB-431542 and Gleevec inhibit

transforming growth factor-beta-induced proliferation of human osteosarcoma

cells. Cancer Res 2003;63:7791-7798.

157. Merchant MS, Woo CW, Mackall CL, Thiele CJ. Potential use of imatinib in

Ewing's Sarcoma: evidence for in vitro and in vivo activity. J Natl Cancer Inst

2002;94:1673-1679.

176 158. Druker BJ, Talpaz M, Resta DJ, et al. Efficacy and safety of a specific inhibitor of

the BCR-ABL tyrosine kinase in chronic myeloid leukemia. N Engl J Med

2001;344:1031-1037.

159. Druker BJ, Sawyers CL, Kantarjian H, et al. Activity of a specific inhibitor of the

BCR-ABL tyrosine kinase in the blast crisis of chronic myeloid leukemia and

acute lymphoblastic leukemia with the Philadelphia chromosome. N Engl J Med

2001;344:1038-1042.

160. Peng B HM, Racine-Poon A, Druker BJ, Talpaz M, Sawyers CL, et al. CLinical

investigtion of the pharmacokinetic/pharmacodynamic relationship for Glivec

(STI571): a novel inhibiotr of signal transduction. Proc ASCO 2001;20:280a.

161. Druker BJ. Taking aim at Ewing's sarcoma: is KIT a target and will imatinib

work? J Natl Cancer Inst 2002;94:1660-1661.

162. Hotfilder M, Lanvers C, Jurgens H, Boos J, Vormoor J. c-KIT-expressing Ewing

tumour cells are insensitive to imatinib mesylate (STI571). Cancer Chemother

Pharmacol 2002;50:167-169.

163. Rous P. A transmissible avian neoplasm. (sarcoma of the common fowl). J. Exp.

Med. 1910;12:696-705.

164. Rous P. A sarcoma of the fowl transmissible by an agent separable from the

tumor cells. J Exp Med 1911;13:397-411.

165. Halberstaedter L, Doljanski L, Tenenbaum E. Br. J. Exp. Pathol. 22 1941.

166. Schmidt CG, Vermathen H, Ritter S. [Electrophoretic and total-protein studies in

Yoshida sarcoma.]. Z Gesamte Exp Med 1959;131:396-406.

177 167. Svet-Moldavsky GJ. Development of multiple cysts and of haemorrhagic

affections of internal organs in albino rats treated during the embryonic or new-

born period with Rous sarcoma virus. Nature 1957;180:1299-1300.

168. Svet-Moldavsky GJ. Sarcoma in albino rats treated during the embryonic stage

with Rous virus. Nature 1958;182:1452-1453.

169. Zilber L, Kryukova IN. Vopr. Virusol. 1958;3:166-170.

170. Zilber LA, Kryukova IN. Haemorrhagic disease of rats due to the virus of chick

sarcoma. Acta Virol 1957;1:156-160.

171. Temin HM. The control of cellular morphology in embryonic cells infected with

rous sarcoma virus in vitro. Virology 1960;10:182-197.

172. Brugge JS, Erikson RL. Identification of a transformation-specific antigen

induced by an avian sarcoma virus. Nature 1977;269:346-348.

173. Collett MS, Erikson RL. Protein kinase activity associated with the avian sarcoma

virus src gene product. Proc Natl Acad Sci U S A 1978;75:2021-2024.

174. Levinson AD, Oppermann H, Levintow L, Varmus HE, Bishop JM. Evidence that

the transforming gene of avian sarcoma virus encodes a protein kinase associated

with a phosphoprotein. Cell 1978;15:561-572.

175. Hunter T, Sefton BM. Transforming gene product of Rous sarcoma virus

phosphorylates tyrosine. Proc Natl Acad Sci U S A 1980;77:1311-1315.

176. Shalloway D, Zelenetz AD, Cooper GM. Molecular cloning and characterization

of the chicken gene homologous to the transforming gene of Rous sarcoma virus.

Cell 1981;24:531-541.

178 177. Takeya T, Hanafusa H. Structure and sequence of the cellular gene homologous to

the RSV src gene and the mechanism for generating the transforming virus. Cell

1983;32:881-890.

178. Iba H, Takeya T, Cross FR, Hanafusa T, Hanafusa H. Rous sarcoma virus variants

that carry the cellular src gene instead of the viral src gene cannot transform

chicken embryo fibroblasts. Proc Natl Acad Sci U S A 1984;81:4424-4428.

179. Parker RC, Varmus HE, Bishop JM. Expression of v-src and chicken c-src in rat

cells demonstrates qualitative differences between pp60v-src and pp60c-src. Cell

1984;37:131-139.

180. Shalloway D, Coussens PM, Yaciuk P. Overexpression of the c-src protein does

not induce transformation of NIH 3T3 cells. Proc Natl Acad Sci U S A

1984;81:7071-7075.

181. Roskoski R, Jr. Src kinase regulation by phosphorylation and dephosphorylation.

Biochem Biophys Res Commun 2005;331:1-14.

182. Bjorge JD, Jakymiw A, Fujita DJ. Selected glimpses into the activation and

function of Src kinase. Oncogene 2000;19:5620-5635.

183. Roskoski R, Jr. Src protein-tyrosine kinase structure and regulation. Biochem

Biophys Res Commun 2004;324:1155-1164.

184. Sun G, Sharma AK, Budde RJ. Autophosphorylation of Src and Yes blocks their

inactivation by Csk phosphorylation. Oncogene 1998;17:1587-1595.

185. Thomas SM, Brugge JS. Cellular functions regulated by Src family kinases. Annu

Rev Cell Dev Biol 1997;13:513-609.

179 186. Turkson J, Bowman T, Garcia R, et al. Stat3 activation by Src induces specific

gene regulation and is required for cell transformation. Mol Cell Biol

1998;18:2545-2552.

187. Yu CL, Meyer DJ, Campbell GS, et al. Enhanced DNA-binding activity of a

Stat3-related protein in cells transformed by the Src oncoprotein. Science

1995;269:81-83.

188. Macara IG, Marinetti GV, Balduzzi PC. Transforming protein of avian sarcoma

virus UR2 is associated with phosphatidylinositol kinase activity: possible role in

tumorigenesis. Proc Natl Acad Sci U S A 1984;81:2728-2732.

189. Sugimoto Y, Whitman M, Cantley LC, Erikson RL. Evidence that the Rous

sarcoma virus transforming gene product phosphorylates phosphatidylinositol and

diacylglycerol. Proc Natl Acad Sci U S A 1984;81:2117-2121.

190. Smith MR, DeGudicibus SJ, Stacey DW. Requirement for c-ras proteins during

viral oncogene transformation. Nature 1986;320:540-543.

191. Martin GS. The road to Src. Oncogene 2004;23:7910-7917.

192. Biscardi JS, Ishizawar RC, Silva CM, Parsons SJ. Tyrosine kinase signalling in

breast cancer: epidermal growth factor receptor and c-Src interactions in breast

cancer. Breast Cancer Res 2000;2:203-210.

193. Chen T, George JA, Taylor CC. Src tyrosine kinase as a chemotherapeutic target:

is there a clinical case? Anticancer Drugs 2006;17:123-131.

194. Irby RB, Yeatman TJ. Role of Src expression and activation in human cancer.

Oncogene 2000;19:5636-5642.

180 195. Yu H, Jove R. The STATs of cancer -- new molecular targets come of age. Nat

Rev Cancer 2004;4:97-105.

196. Buettner R, Mora LB, Jove R. Activated STAT signaling in human tumors

provides novel molecular targets for therapeutic intervention. Clin Cancer Res

2002;8:945-954.

197. Turkson J, Jove R. STAT proteins: novel molecular targets for cancer drug

discovery. Oncogene 2000;19:6613-6626.

198. Burridge K, Chrzanowska-Wodnicka M. Focal adhesions, contractility, and

signaling. Annu Rev Cell Dev Biol 1996;12:463-518.

199. Mitra SK, Schlaepfer DD. Integrin-regulated FAK-Src signaling in normal and

cancer cells. Curr Opin Cell Biol 2006;18:516-523.

200. McLean GW, Komiyama NH, Serrels B, et al. Specific deletion of focal adhesion

kinase suppresses tumor formation and blocks malignant progression. Genes Dev

2004;18:2998-3003.

201. Shen TL, Park AY, Alcaraz A, et al. Conditional knockout of focal adhesion

kinase in endothelial cells reveals its role in angiogenesis and vascular

development in late embryogenesis. J Cell Biol 2005;169:941-952.

202. Parsons JT. Focal adhesion kinase: the first ten years. J Cell Sci 2003;116:1409-

1416.

203. Golubovskaya VM, Finch R, Cance WG. Direct interaction of the N-terminal

domain of focal adhesion kinase with the N-terminal transactivation domain of

p53. J Biol Chem 2005;280:25008-25021.

181 204. Owens LV, Xu L, Craven RJ, et al. Overexpression of the focal adhesion kinase

(p125FAK) in invasive human tumors. Cancer Res 1995;55:2752-2755.

205. Mitra SK, Lim ST, Chi A, Schlaepfer DD. Intrinsic focal adhesion kinase activity

controls orthotopic breast carcinoma metastasis via the regulation of urokinase

plasminogen activator expression in a syngeneic tumor model. Oncogene

2006;25:4429-4440.

206. Frame MC. Src in cancer: deregulation and consequences for cell behaviour.

Biochim Biophys Acta 2002;1602:114-130.

207. Playford MP, Schaller MD. The interplay between Src and integrins in normal

and tumor biology. Oncogene 2004;23:7928-7946.

208. Brabek J, Constancio SS, Shin NY, et al. CAS promotes invasiveness of Src-

transformed cells. Oncogene 2004;23:7406-7415.

209. Bouton AH, Riggins RB, Bruce-Staskal PJ. Functions of the adapter protein Cas:

signal convergence and the determination of cellular responses. Oncogene

2001;20:6448-6458.

210. Bradshaw JM, Waksman G. Calorimetric investigation of proton linkage by

monitoring both the enthalpy and association constant of binding: application to

the interaction of the Src SH2 domain with a high-affinity tyrosyl

phosphopeptide. Biochemistry 1998;37:15400-15407.

211. Gilmer T, Rodriguez M, Jordan S, et al. Peptide inhibitors of src SH3-SH2-

phosphoprotein interactions. J Biol Chem 1994;269:31711-31719.

182 212. Oneyama C, Nakano H, Sharma SV. UCS15A, a novel small molecule, SH3

domain-mediated protein-protein interaction blocking drug. Oncogene

2002;21:2037-2050.

213. Banerji U, Judson I, Workman P. The clinical applications of heat shock protein

inhibitors in cancer - present and future. Curr Cancer Drug Targets 2003;3:385-

390.

214. Uehara Y. Natural product origins of Hsp90 inhibitors. Curr Cancer Drug Targets

2003;3:325-330.

215. Pengetnze Y, Steed M, Roby KF, Terranova PF, Taylor CC. Src tyrosine kinase

promotes survival and resistance to chemotherapeutics in a mouse ovarian cancer

cell line. Biochem Biophys Res Commun 2003;309:377-383.

216. Song L, Turkson J, Karras JG, Jove R, Haura EB. Activation of Stat3 by receptor

tyrosine kinases and cytokines regulates survival in human non-small cell

carcinoma cells. Oncogene 2003;22:4150-4165.

217. Garcia R, Yu CL, Hudnall A, et al. Constitutive activation of Stat3 in fibroblasts

transformed by diverse oncoproteins and in breast carcinoma cells. Cell Growth

Differ 1997;8:1267-1276.

218. Wang YZ, Wharton W, Garcia R, et al. Activation of Stat3 preassembled with

platelet-derived growth factor beta receptors requires Src kinase activity.

Oncogene 2000;19:2075-2085.

219. Niu G, Bowman T, Huang M, et al. Roles of activated Src and Stat3 signaling in

melanoma tumor cell growth. Oncogene 2002;21:7001-7010.

183 220. Golas JM, Arndt K, Etienne C, et al. SKI-606, a 4-anilino-3-quinolinecarbonitrile

dual inhibitor of Src and Abl kinases, is a potent antiproliferative agent against

chronic myelogenous leukemia cells in culture and causes regression of K562

xenografts in nude mice. Cancer Res 2003;63:375-381.

221. Coluccia AM, Benati D, Dekhil H, et al. SKI-606 decreases growth and motility

of colorectal cancer cells by preventing pp60(c-Src)-dependent tyrosine

phosphorylation of beta-catenin and its nuclear signaling. Cancer Res

2006;66:2279-2286.

222. Golas JM, Lucas J, Etienne C, et al. SKI-606, a Src/Abl inhibitor with in vivo

activity in colon tumor xenograft models. Cancer Res 2005;65:5358-5364.

223. Nam S, Kim D, Cheng JQ, et al. Action of the Src family kinase inhibitor,

dasatinib (BMS-354825), on human prostate cancer cells. Cancer Res

2005;65:9185-9189.

224. Johnson FM, Saigal B, Talpaz M, Donato NJ. Dasatinib (BMS-354825) tyrosine

kinase inhibitor suppresses invasion and induces cell cycle arrest and apoptosis of

head and neck squamous cell carcinoma and non-small cell lung cancer cells. Clin

Cancer Res 2005;11:6924-6932.

225. Song L, Morris M, Bagui T, et al. Dasatinib (BMS-354825) selectively induces

apoptosis in lung cancer cells dependent on epidermal growth factor receptor

signaling for survival. Cancer Res 2006;66:5542-5548.

226. Vigneron A, Roninson IB, Gamelin E, Coqueret O. Src inhibits adriamycin-

induced senescence and G2 checkpoint arrest by blocking the induction of

p21waf1. Cancer Res 2005;65:8927-8935.

184 227. Diaz-Montero CM, Wygant JN, McIntyre BW. PI3-K/Akt-mediated anoikis

resistance of human osteosarcoma cells requires Src activation. Eur J Cancer

2006;42:1491-1500.

228. Alvarez RH, Kantarjian HM, Cortes JE. The role of Src in solid and hematologic

malignancies: development of new-generation Src inhibitors. Cancer

2006;107:1918-1929.

229. Trevino JG, Summy JM, Lesslie DP, et al. Inhibition of SRC expression and

activity inhibits tumor progression and metastasis of human pancreatic

adenocarcinoma cells in an orthotopic nude mouse model. Am J Pathol

2006;168:962-972.

230. Yeatman TJ. A renaissance for SRC. Nat Rev Cancer 2004;4:470-480.

231. Lombardo LJ, Lee FY, Chen P, et al. Discovery of N-(2-chloro-6-methyl-

phenyl)-2-(6-(4-(2-hydroxyethyl)- piperazin-1-yl)-2-methylpyrimidin-4-

ylamino)thiazole-5-carboxamide (BMS-354825), a dual Src/Abl kinase inhibitor

with potent antitumor activity in preclinical assays. J Med Chem 2004;47:6658-

6661.

232. Van Gelder R, von Zastrow M, Yool A, et al. Amplified RNA synthesized from

limited quantities of heterogeneous cDNA. . Proc Natl Acad Sci U S A

1990;Mar;87:1663-1667.

233. Liu WM, Mei R, Di X, et al. Analysis of high density expression microarrays with

signed-rank call algorithms. Bioinformatics 2002;18:1593-1599.

234. Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to

the ionizing radiation response. Proc Natl Acad Sci U S A 2001;98:5116-5121.

185 235. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and display of

genome-wide expression patterns. Proc Natl Acad Sci U S A 1998;95:14863-

14868.

236. Saldanha AJ. Java Treeview--extensible visualization of microarray data.

Bioinformatics 2004;20:3246-3248.

237. Akira S. Functional roles of STAT family proteins: lessons from knockout mice.

Stem Cells 1999;17:138-146.

238. Dearolf CR. JAKs and STATs in invertebrate model organisms. Cell Mol Life Sci

1999;55:1578-1584.

239. Hirano T, Ishihara K, Hibi M. Roles of STAT3 in mediating the cell growth,

differentiation and survival signals relayed through the IL-6 family of cytokine

receptors. Oncogene 2000;19:2548-2556.

240. Levy DE. Physiological significance of STAT proteins: investigations through

gene disruption in vivo. Cell Mol Life Sci 1999;55:1559-1567.

241. Mui AL. The role of STATs in proliferation, differentiation, and apoptosis. Cell

Mol Life Sci 1999;55:1547-1558.

242. Smithgall TE, Briggs SD, Schreiner S, et al. Control of myeloid differentiation

and survival by Stats. Oncogene 2000;19:2612-2618.

243. Decker T, Kovarik P. Transcription factor activity of STAT proteins: structural

requirements and regulation by phosphorylation and interacting proteins. Cell Mol

Life Sci 1999;55:1535-1546.

244. Darnell JE, Jr. STATs and gene regulation. Science 1997;277:1630-1635.

186 245. Bowman T, Garcia R, Turkson J, Jove R. STATs in oncogenesis. Oncogene

2000;19:2474-2488.

246. Bowman T, Jove R. STAT Proteins and Cancer. Cancer Control 1999;6:615-619.

247. Catlett-Falcone R, Landowski TH, Oshiro MM, et al. Constitutive activation of

Stat3 signaling confers resistance to apoptosis in human U266 myeloma cells.

Immunity 1999;10:105-115.

248. Garcia R, Jove R. Activation of STAT transcription factors in oncogenic tyrosine

kinase signaling. J Biomed Sci 1998;5:79-85.

249. Cao X, Tay A, Guy GR, Tan YH. Activation and association of Stat3 with Src in

v-Src-transformed cell lines. Mol Cell Biol 1996;16:1595-1603.

250. Chaturvedi P, Sharma S, Reddy EP. Abrogation of interleukin-3 dependence of

myeloid cells by the v-src oncogene requires SH2 and SH3 domains which

specify activation of STATs. Mol Cell Biol 1997;17:3295-3304.

251. Bromberg JF, Fan Z, Brown C, Mendelsohn J, Darnell JE, Jr. Epidermal growth

factor-induced growth inhibition requires Stat1 activation. Cell Growth Differ

1998;9:505-512.

252. Kraker AJ, Hartl BG, Amar AM, et al. Biochemical and cellular effects of c-Src

kinase-selective pyrido[2, 3-d]pyrimidine tyrosine kinase inhibitors. Biochem

Pharmacol 2000;60:885-898.

253. Boschelli DH, Wang YD, Johnson S, et al. 7-Alkoxy-4-phenylamino-3-

quinolinecar-bonitriles as dual inhibitors of Src and Abl kinases. J Med Chem

2004;47:1599-1601.

187 254. Boschelli DH, Ye F, Wang YD, et al. Optimization of 4-phenylamino-3-

quinolinecarbonitriles as potent inhibitors of Src kinase activity. J Med Chem

2001;44:3965-3977.

255. Wyeth Research and Clinical Trials: Project Listing. 2007.

http://www.wyeth.com/research/projects

256. Tremblay L, Hauck W, Aprikian AG, et al. Focal adhesion kinase (pp125FAK)

expression, activation and association with paxillin and p50CSK in human

metastatic prostate carcinoma. Int J Cancer 1996;68:164-171.

257. McCormack SJ, Brazinski SE, Moore JL, Jr., Werness BA, Goldstein DJ.

Activation of the focal adhesion kinase signal transduction pathway in cervical

carcinoma cell lines and human genital epithelial cells immortalized with human

papillomavirus type 18. Oncogene 1997;15:265-274.

258. Cance WG, Harris JE, Iacocca MV, et al. Immunohistochemical analyses of focal

adhesion kinase expression in benign and malignant human breast and colon

tissues: correlation with preinvasive and invasive phenotypes. Clin Cancer Res

2000;6:2417-2423.

259. Lark AL, Livasy CA, Calvo B, et al. Overexpression of focal adhesion kinase in

primary colorectal carcinomas and colorectal liver metastases:

immunohistochemistry and real-time PCR analyses. Clin Cancer Res 2003;9:215-

222.

260. Yu H JR. The STATs of cancer -- new molecular targets come of age. Nat Rev

Cancer 2004;4:97-105.

188 261. McLean GW, Carragher NO, Avizienyte E, et al. The role of focal-adhesion

kinase in cancer - a new therapeutic opportunity. Nat Rev Cancer 2005;5:505-

515.

262. Defilippi P, Di Stefano P, Cabodi S. p130Cas: a versatile scaffold in signaling

networks. Trends Cell Biol 2006;16:257-263.

263. Hanks SK, Ryzhova L, Shin NY, Brabek J. Focal adhesion kinase signaling

activities and their implications in the control of cell survival and motility. Front

Biosci 2003;8:d982-996.

264. Schaller MD. Biochemical signals and biological responses elicited by the focal

adhesion kinase. Biochim Biophys Acta 2001;1540:1-21.

265. Silva CM. Role of STATs as downstream signal transducers in Src family kinase-

mediated tumorigenesis. . Oncogene 2004;23:8017-8023.

266. Oral BMS-354825 rescues Gleevec-resistant CML. Cancer Biol Ther 2004;3:700.

267. Copland M, Hamilton A, Elrick LJ, et al. Dasatinib (BMS-354825) targets an

earlier progenitor population than imatinib in primary CML but does not

eliminate the quiescent fraction. Blood 2006;107:4532-4539.

268. Doggrell SA. BMS-354825: a novel drug with potential for the treatment of

imatinib-resistant chronic myeloid leukaemia. Expert Opin Investig Drugs

2005;14:89-91.

269. O'Hare T, Corbin AS, Druker BJ. Targeted CML therapy: controlling drug

resistance, seeking cure. Curr Opin Genet Dev 2006;16:92-99.

270. Shah N. Loss of response to imatinib: mechanisms and management. Hematology

(Am Soc Hematol Educ Program) 2005:183-187.

189 271. Shah NP, Lee FY, Luo R, et al. Dasatinib (BMS-354825) inhibits KITD816V, an

imatinib-resistant activating mutation that triggers neoplastic growth in most

patients with systemic mastocytosis. Blood 2006;108:286-291.

272. Tokarski JS, Newitt JA, Chang CY, et al. The structure of Dasatinib (BMS-

354825) bound to activated ABL kinase domain elucidates its inhibitory activity

against imatinib-resistant ABL mutants. Cancer Res 2006;66:5790-5797.

273. Chang JC, Wooten EC, Tsimelzon A, et al. Gene expression profiling for the

prediction of therapeutic response to docetaxel in patients with breast cancer.

Lancet 2003;362:362-369.

274. Sorlie T, Tibshirani R, Parker J, et al. Repeated observation of breast tumor

subtypes in independent gene expression data sets. Proc Natl Acad Sci U S A

2003;100:8418-8423.

275. Sotiriou C, Neo SY, McShane LM, et al. Breast cancer classification and

prognosis based on gene expression profiles from a population-based study. Proc

Natl Acad Sci U S A 2003;100:10393-10398.

276. van 't Veer LJ, Dai H, van de Vijver MJ, et al. Expression profiling predicts

outcome in breast cancer. Breast Cancer Res 2003;5:57-58.

277. Nielsen TO. Microarray analysis of sarcomas. Adv Anat Pathol 2006;13:166-173.

278. Dong Z, Bode AM. The role of histone H3 phosphorylation (Ser10 and Ser28) in

cell growth and cell transformation. Mol Carcinog 2006;45:416-421.

279. Ryu SW, Lee SJ, Park MY, et al. Fas-associated factor 1, FAF1, is a member of

Fas death-inducing signaling complex. J Biol Chem 2003;278:24003-24010.

190 280. Bjorling-Poulsen M, Seitz G, Guerra B, Issinger OG. The pro-apoptotic FAS-

associated factor 1 is specifically reduced in human gastric carcinomas. Int J

Oncol 2003;23:1015-1023.

281. Perez-Moreno M, Fuchs E. Catenins: keeping cells from getting their signals

crossed. Dev Cell 2006;11:601-612.

282. Kobielak A, Fuchs E. Alpha-catenin: at the junction of intercellular adhesion and

actin dynamics. Nat Rev Mol Cell Biol 2004;5:614-625.

283. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell

1990;61:759-767.

284. Ewing CM, Ru N, Morton RA, et al. Chromosome 5 suppresses tumorigenicity of

PC3 prostate cancer cells: correlation with re-expression of alpha-catenin and

restoration of E-cadherin function. Cancer Res 1995;55:4813-4817.

285. Van Aken E, De Wever O, Correia da Rocha AS, Mareel M. Defective E-

cadherin/catenin complexes in human cancer. Virchows Arch 2001;439:725-751.

286. Hajra KM, Fearon ER. Cadherin and catenin alterations in human cancer. Genes

Chromosomes Cancer 2002;34:255-268.

287. Saito T, Oda Y, Sakamoto A, et al. Prognostic value of the preserved expression

of the E-cadherin and catenin families of adhesion molecules and of beta-catenin

mutations in synovial sarcoma. J Pathol 2000;192:342-350.

288. Holder N, Klein R. Eph receptors and ephrins: effectors of morphogenesis.

Development 1999;126:2033-2044.

191 289. Ogawa K, Pasqualini R, Lindberg RA, et al. The ephrin-A1 ligand and its

receptor, EphA2, are expressed during tumor neovascularization. Oncogene

2000;19:6043-6052.

290. Zhang L, Gao X, Wen J, Ning Y, Chen YG. Dapper 1 antagonizes Wnt signaling

by promoting dishevelled degradation. J Biol Chem 2006;281:8607-8612.

291. Katoh M, Katoh M. Identification and characterization of human DAPPER1 and

DAPPER2 genes in silico. Int J Oncol 2003;22:907-913.

192

About the Author

Audrey Shor was born in Pennsylvania and lived most of her life in the Northeast. She received an academic scholarship to attend West Chester University, Pennsylvania, where she obtained a Bachelor’s of Science in Cellular and Molecular Biology as well as minors in Chemistry and Studio Arts in 2002. Audrey was among the first students to pursue the

PhD PLUS program at the University of South Florida. This program offered her the opportunity to obtain a Master of Public Health in Epidemiology from the College of

Public Health in 2006 and a Doctoral Degree in Molecular Medicine from the College of

Medicine in 2007. Audrey has completed her doctoral studies under the guidance of

Doctors Richard Jove and W. Jack Pledger. She has been fortunate enough to have the opportunity to present her research findings both nationally and internationally. This research was published in Cancer Research in March 2007.