Genetic Variation and Complex Disease: the Examination of an X-Linked Disorder and a Multifactorial Disease

GENETIC VARIATION AND COMPLEX DISEASE: THE EXAMINATION OF AN X-LINKED DISORDER AND A MULTIFACTORIAL DISEASE

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Catherine Elise Cottrell

*****

The Ohio State University 2007

Dissertation Committee: Approved by Dr. Julie M. Gastier-Foster, Advisor

Dr. John Bauer ______Dr. Gail Herman Advisor Graduate Program in Dr. Susan R. Mallery Integrated Biomedical Sciences

ABSTRACT

Two genetic studies are presented in this dissertation. The studies share a common theme of the influence of genetic variability in complex disease. The first study is entitled, “The Association of Antioxidant-Related Gene Polymorphisms and Second

Primary Malignant Neoplasms in Pediatric Hodgkin Lymphoma.”

Improved treatment strategies in pediatric Hodgkin Lymphoma (HL) have

resulted in a cure rate approaching 95%, yet the development of a second primary

malignant neoplasm (SMN) is a risk for long-term survivors. In a pediatric HL

population, between 6-26% of patients will develop a SMN within 30 years following

treatment. The etiology of HL is still largely unknown, as is the cause of certain types of

SMNs.

Oxidative stress has been linked to the development of cancer due to the

damaging effects of reactive oxygen species (ROS) on DNA, lipids, and proteins. During

periods of extended oxidative stress, the damaging effects of ROS are likely to increase

which in turn raises the risk of genetic change. In a multi-stage model of tumorigenesis,

genetic change is considered to be the initiating event of cancer development. Under

normal circumstances, antioxidant enzymes within the cell mitigate the effects of ROS by

converting reactive species into non-toxic molecules. Mechanisms within the cell

maintain a steady state balance between ROS and antioxidant enzyme levels. We

ii hypothesize that individuals predisposed to lower levels of antioxidant enzyme activity

due to polymorphic variants within those genes may be at risk for increased damage

caused by ROS. Although decreased amounts of antioxidant enzymes have been found in

a variety of cancers, a correlation between polymorphisms in antioxidant genes and

predisposition to secondary cancer has not been studied.

The purpose this study was to assess the association of antioxidant gene alleles with the risk of developing a SMN in HL survivors. DNA samples were obtained from

768 HL patients enrolled in the Childhood Cancer Survivor Study (CCSS). The samples were genotyped for 90 polymorphisms in antioxidant related genes including SOD, GPX,

NOS, CAT, and CYP2C9. Statistical analysis methods to determine risk of developing a

SMN included association, haplotype, and multiple regression models.

In our HL cohort, 131 patients developed a SMN. An additional 117 patients

developed nonmelanoma skin cancer, and 34 patients developed 2 SMNs. Out of the 36

SNPs that were included in the final analysis, 4 SNPs in the GPX1, GPX3, GPX4, and

SOD2 genes, were potentially suggestive (p<0.05) of an association between genotype

and the development of a SMN. A general trend observed in the genotyping data

suggested that an increased risk of SMN was conferred by the presence of the minor

allele, while a decreased risk of SMN was conferred by the presence of the major allele in

the population. Replication studies are necessary, though it is notable that polymorphisms within the GPX family may be associated with the development SMN in our cohort. Additionally, hazard ratios were only moderately increased in this population suggesting that the development of a SMN in a pediatric HL is likely multi-factorial in nature. The elucidation of critical genes involved in SMN development could influence

iii patient follow-up and intervention strategies, and potentially aid in the prevention of

SMN.

The second study in this thesis is entitled, “Atypical X-Inactivation in an X;1

Translocation Patient Diagnosed with Otopalatodigital Syndrome.” X-chromosome inactivation (XCI) is an epigenetic process used to regulate gene dosage in mammalian

females by silencing one X-chromosome. While the pattern of XCI is typically random

in normal females, abnormalities of the X-chromosome may result in skewing due to

disadvantaged cell growth. We describe a female patient with an X;1 translocation

[46,X, t(X;1)(q28;q21)dn] and unusual pattern of XCI who was clinically diagnosed with

Otopalatodigital syndrome (OPD) type 1. There was complete skewing of XCI in the

patient, along with the atypical findings of an active normal X-chromosome and an

inactive derivative X. An X-linked disorder, OPD1 is characterized by multiple

congenital anomalies including skeletal abnormalities, craniofacial defects, and hearing

loss. Mutations within the FLNA gene (Xq28) are known to cause OPD, though none

were detected in our patient. Additionally, no abnormalities in FLNA mRNA or protein

were detected in our patient. Characterization of the translocation revealed that the

patient’s Xq28 breakpoint interrupts the DKC1 gene, located 400kB distal to FLNA.

Molecular analysis of the breakpoint region revealed functional disomy of Xq28 genes

distal to DKC1. No monosomy of 1q genes was detected. Possible explanations for the

patient’s phenotype include a position effect due to the translocation breakpoint, an

undetected FLNA-related mutation, or altered gene dosage due to consequences of

atypical XCI.

Dedicated to Andy Thank you for your love, constant support, and words of encouragement.

To my parents, Charles and Denise Thank you for inspiring me, fostering a sense of creativity, and for a lifetime of love and support.

ACKNOWLEDGMENTS

It is with deepest gratitude that I would like to thank my advisor, Dr. Julie

Gastier-Foster for her dedication and commitment to me on both a personal and professional level. You are a constant source of inspiration and an exceptional mentor.

Thank you for your patience, guidance, and support over the years. Most of all, I thank you for your time, for the countless hours spent teaching me, and for the direction you provided so that I could mature on both a scientific and personal level.

I would like to thank my committee members, Dr. Gail Herman, Dr. John Bauer, and Dr. Susan Mallery for your commitment to teaching, dedication to students, and exceptional guidance.

I would like to express my sincere appreciation for Dr. Steve Qualman. Thank you for guiding me to choose a wonderful lab in which to complete my dissertation.

Thank you for introducing me to Julie, because without her none of this would be possible. Finally, thank you most of all for your dedication, your commitment to research and teaching, and your passion in supporting Children’s Hospital.

I would like to thank Dr. Allan Yates and Christine Kerr for your support of the

IBGP program, for your devotion to the students, and for your sincere desire to see each of us succeed.

vi I would especially like to thank all of the technologists and staff within the

Molecular and Cytogenetics Laboratories at Children’s. Thank you for accepting me into

the lab, and for sharing your time and resources with me. Thank you especially to Jessica

who made me feel welcome when I first began and for your help along the way. Thanks

to Beth and Morgan for teaching me the basics of FISH, cell harvests and for your

expertise in taking wonderful photos on the scope. A special thank you goes to Eileen for

your unwavering support, your friendship, and your sense of humor. I don’t think I could have made it through this without you.

Finally, I would like to thank my family and friends for your support throughout

this process. Thank you to Jen for understanding what I was going though when no one

else did. Thank you for graduating ahead of me, so then I could ask you all sorts of

questions on finishing up. Thank you for extended chats over coffee and for going to

Buckeye football games with me. Thank you to Colleen for being a great sister. Thank you for your support and kind words, and for feeding the cats when I was working late or away on weekends. Thank you especially to Andy for your love and encouragement.

Thank you for driving long hours to see me, for cheering me up during hard times, and for always making me laugh, especially when this day seemed very far away. Thank you to my parents for sparking my interest in science early on. Thank you for those kitchen science experiments, grade school science fair ideas, and for taking me to your lab when I was very young. Most of all, thank you for your love and guidance over the years.

vii

VITA

August 3, 1979...... Born – Columbus, OH

June 2001...... B.S. Molecular Genetics The Ohio State University

2001 – 2007...... Graduate Fellow The Ohio State University.

AWARDS

2004 ...... Graduate Student Research Award Columbus Children’s Research Institute Annual Research Conference

FELLOWSHIPS

2001 ...... University Fellowship The Ohio State University

2001-2005 ...... Molecular Life Sciences Fellowship The Ohio State University

2002-2007 ...... General Mason Fellowship Columbus Children’s Hospital

FIELD OF STUDY

Major Field: Integrated Biomedical Sciences Specialization: Genetics

viii

TABLE OF CONTENTS

P a g e

Abstract...... ii

Dedication...... v

Acknowledgments ...... vi

Vita ...... viii

List of Tables...... xiv

List of Figures ...... xv

Introduction ...... 1

Section I: The Association of Antioxidant Related Gene Polymorphisms and Second Primary Malignant Neoplasm in a Pediatric Hodgkin’s Lymphoma Population

Chapters:

1. Hodgkin Lymphoma and the Childhood Cancer Survivor Study ...... 5

1.1 Introduction to Hodgkin Lymphoma ...... 5

1.2 Classification of Hodgkin Lymphoma ...... 6

1.3 Risk Factors and Causative Agents in Hodgkin Lymphoma ...... 6

1.4 Treatment of Hodgkin Lymphoma: Maximizing Efficacy while Minimizing Risk ...... 8

ix 1.5 The Childhood Cancer Survivor Study ...... 9

1.5.1 Study Design and Eligibility ...... 9

1.5.2 Study Participation ...... 10

1.5.3 HL Patient Demographics within the CCSS ...... 11

1.6 Genetic Susceptibility to SMN in Cancer Survivors ...... 11

2. Introduction to Reactive Species, Antioxidant Enzymes and the Relationship between Oxidative Stress and Malignancy ...... 20

2.1 Reactive Species and Antioxidants ...... 20

2.1.1 Reactive Oxygen Species ...... 20

2.1.2 Antioxidant Enzymes ...... 21

2.1.3 Reactive Nitrogen Species ...... 22

2.2 Types of Antioxidant Enzymes ...... 23

2.2.1 Superoxide Dismutase ...... 23

2.2.2 Glutathione Peroxidase...... 23

2.2.3 Catalase...... 25

2.3 Oxidative Stress and Cancer ...... 25

2.3.1 A Multi-Stage Model of Carcinogenesis ...... 25

2.3.2 Initiation ...... 26

2.3.3 Promotion ...... 26

2.3.4 Reactive Species: A Balancing Act of Tumor Promotion or Tumor Suppression . . . 26

2.3.5 Progression ...... 27

2.3.6 Antioxidant Enzymes and Malignancy...... 27

x 3. The Association of Antioxidant Related Gene Polymorphisms and Second Primary Malignant Neoplasms in a Pediatric Hodgkin Lymphoma Population ...... 37

3.1 Abstract ...... 37

3.2 Introduction ...... 39

3.3 Materials and Methods ...... 41

3.3.1 The Childhood Cancer Survivor Study ...... 41

3.3.2 Buccal Cell Collection ...... 42

3.3.3 Genotyping Methodology...... 43

3.3.4 Statistical Analysis Methods...... 45

3.4 Results ...... 47

3.5 Discussion ...... 51

Section II: Atypical X Chromosome Inactivation in an X;1 Translocation Patient Diagnosed with Otopalatodigital Syndrome

4. X Chromosome Inactivation and X;Autosome Translocations ...... 88

4.1 X Chromosome Inactivation ...... 88

4.1.1 Introduction ...... 88

4.1.2 Genes which Escape X-Inactivation ...... 89

4.2 X;Autosome Translocations ...... 90

4.2.1 Introduction to Translocations ...... 90

4.2.2 Balanced X;Autosome Translocations ...... 91

4.2.3 Phenotypic Consequences Translocation Breakpoint Location . . .91

4.2.4 XCI in a t(X;A) Carrier ...... 92

4.2.5 The Impact of the Mode of Inheritance in a t(X;A) Carrier...... 93

xi 5. Otopalatodigital Syndrome Spectrum Disorders and Filamin A ...... 94

5.1 Introduction to Otopalatodigital Syndrome ...... 94

5.2 Phenotypic Features of OPD ...... 94

5.3 Molecular Characterization of the OPD-Spectrum Disorders ...... 95

5.4 Filamin A Protein ...... 97

5.5 OPSD Genotype – Phenotype Correlations ...... 99

5.6 X-Chromosome Inactivation and OPSD...... 99

6. Atypical X Chromosome Inactivation in an X;1 Translocation Patient Diagnosed with Otopalatodigital Syndrome ...... 105

6.1 Abstract ...... 105

6.2 Introduction ...... 106

6.3 Materials and Methods ...... 108

6.3.1 Subjects and Consent ...... 108

6.3.2 Tissue Sources, Cell Lines, DNA and RNA Preparations ...... 108

6.3.3 Molecular and Cytogenetic Breakpoint Mapping ...... 109

6.3.4 X-Inactivation Analysis ...... 109

6.3.5 Expression – RT-PCR ...... 110

6.3.6 Sequencing ...... 110

6.3.7 Northern Blot ...... 111

6.3.8 Expression Array Analysis ...... 111

6.3.9 Comparative Genomic Hybridization ...... 112

6.3.10 Western Blot Analysis ...... 112

6.4 Results ...... 113

xii 6.4.1 Clinical Report ...... 113

6.4.2 Breakpoint Characterization ...... 114

6.4.3 X-Chromosome Inactivation Status ...... 115

6.4.4 RT-PCR Analysis ...... 115

6.4.5 Molecular Analysis of FLNA ...... 116

6.4.6 Expression Array Analysis ...... 118

6.4.7 Comparative Genomic Hybridization ...... 118

6.5 Discussion ...... 118

Appendix ...... 154

References ...... 155

xiii

LIST OF TABLES

Table Page

1.1 Composition of Participants within the CCSS Cohort ...... 16

1.2 Type of SMN in HL Survivors within the CCSS Cohort...... 18

3.1 Type of SMN/NMSC within the Subset of Genotyped HL Survivors. . .66

3.2 SNPs Included in the Final Statistical Analysis of the HL Cohort . . . . . 68

3.3 Associations between SMN or NMSC and Genotype by Cox Regression Analysis ...... 71

3.4 A Summarized List of Associations by Cox Regression Analysis . . . . . 76

3.5 Associations between SMN Classification and Genotype by Cox Regression Analysis ...... 78

3.6 A Summary of Association Results by Cox Regression Analysis Classified by Type of SMN or NMSC...... 81

3.7 Associations between SMN or NMSC and Genotype by Poisson Multivariate Regression Analysis...... 83

3.8 Haplotype Analysis using Poisson Regression Modeling ...... 85

5.1 Phenotypic Features of FLNA-Related Disorders ...... 103

6.1 Primer Sequences...... 145

6.2 Refinement of the X;1 Translocation Breakpoint by FISH Mapping. . .148

6.3 RT-PCR Analysis ...... 150

6.4 Combined RT-PCR and Array Results for Xq28 Genes...... 152

xiv

LIST OF FIGURES

Figure Page

1.1 Annual Incidence Rates of Hodgkin Lymphoma ...... 14

2.1 Chemical Reactions Involving ROS and Antioxidant Enzymes...... 29

2.2 Reactive Species Interact to Form an Oxidizing Free Radical ...... 31

2.3 A Multi-Stage Model of Tumorigenesis Relative to Oxidative Stress. . .33

2.4 Cellular Response to ROS During Times of Oxidative Stress...... 35

3.1 SNPlex Genotyping Chemistry ...... 58

3.2 GeneMapper Graphical Plots ...... 60

3.3 GeneMapper Samples Plot ...... 62

3.4 Causes of SNP Attrition in the SNPlex Genotyping Method...... 64

5.1 A Model of the Filamin A Dimer and Location of Pathogenic Mutations ...... 101

6.1 X;1 Translocation Breakpoint...... 125

6.2 Clinical Features of the Patient...... 127

6.3 Fluorescence In-Situ Hybridization Analysis...... 129

6.4 Map of the Xq28 Breakpoint Region...... 131

6.5 X-Chromosome Inactivation Analysis...... 133

6.6 Disomic Expression of the FLNA Gene...... 135

6.7 Sequence Analysis of a 3’UTR FLNA Polymorphism ...... 137

6.8 Mfold Predictions of Optimal Secondary Structure...... 139

6.9 Northern Blot Analysis of FLNA Transcripts...... 141

6.10 Western Blot Analysis of Filamin A ...... 143

xvi

LIST OF ABBREVIATIONS

α Alpha

β Beta

AB Applied Biosystems

°C degrees Celsius

CAT Catalase

CCSS Childhood Cancer Survivor Study

CI Confidence Interval

der(X) Derivative X Chromosome

GEE Generalized Estimating Equation

GPX Glutathione Peroxidase

GST Glutathione S-Transferase

H2O2 Hydrogen Peroxide

HL Hodgkin Lymphoma

HRS Hodgkin Reed-Sternberg Cell

NHL Non-Hodgkin Lymphoma

NMSC Non-Melanoma Skin Cancer

NO• Nitric Oxide

xvii NOS Nitric Oxide Synthase

O2 Oxygen

•– O2 Superoxide Anion

•OH Hydroxy Radical

ONOO– Peroxynitrite Anion

OPD Otopalatodigital Syndrome

OPSD Otopalatodigital Syndrome Spectrum Disorders

RNS Reactive Nitrogen Species

ROS Reactive Oxygen Species

RS Reactive Species

SIR Standardized Incidence Ratio

SMN Second Primary Malignant Neoplasm

SMR Standardized Mortality Ratio

SNP Single Nucleotide Polymorphism

SOD Superoxide Dismutase t(X;A) X;Autosome Translocation

XCI X-Chromosome Inactivation

Xi Inactive X-Chromosome

Xa Active X-Chromosome

xviii

INTRODUCTION Thesis Composition and Objectives

Beginning with the Watson and Crick discovery of the double helix in 1953, major milestones in the field of genetics have slowly begun to reveal the secrets of the genome. More recent advances such as the completion of the Human Genome Project

(2003) resulted in a thorough guide to our genetic code, and subsequent studies using

HapMap data have been used to assess population variance. From data compiled so far, findings have been particularly striking in regard to genetic variation. Results suggest that the genomes between two individuals may differ by at least 1%. This indicates that the human genome contains far more variability than previously thought. Variability at the level of a single nucleotide polymorphism, a short tandem repeat, or even a chromosomal rearrangement adds to the complexity and function of the human genome.

It remains a challenge for investigators to decode the secrets of the genome, and in turn to decipher how variability influences phenotype. Of perhaps the greatest clinical relevance is to determine how variability influences the risk of genetic disease. Progress on this front could improve patient screening, genetic counseling, and comprehensive care.

This thesis is a compilation of two separate research projects that were completed

during the course of my graduate study. Though each project is vastly different, a

common theme of genetic variation in complex disease is present.

1 Section one of this thesis relates to the Hodgkin Lymphoma project. Background

information regarding HL, antioxidants, and SMNs are included in chapters 1 and 2 to

introduce the reader to the field of study prior to presenting the results and conclusions.

The HL project began in 2004 using samples that were collected from pediatric cancer

survivors enrolled in the Childhood Cancer Survivor Study (CCSS). Cancer survivors within the CCSS represent one of the largest and most well characterized cohorts within

North America. HL was studied exclusively, as survivors have the greatest risk of

developing a SMN following cancer treatment, and because we were granted access to

patient buccal cell DNA from the CCSS. While previous genetic studies of HL and SMN

focused on genes involved in DNA repair and detoxification, we chose to focus on a

genetic analysis of polymorphisms within antioxidant related genes. The results of the

study are presented in chapter 3.

Section two of my thesis relates to the OPD project. This research project has

evolved greatly since its inception in 2002. At that time the causative gene for OPD

syndrome had not yet been cloned, but was localized to Xq28. By chance, a patient

diagnosed with OPD1 and having an X;1 translocation was currently being followed by

the genetics clinic and presented the opportunity to attempt to clone the gene. The

project began following parental and IRB approval. Concurrent with our efforts to clone

the gene, another research group in Europe was working toward the same goal. In 2003

Robertson et al. reported mutations within the FLNA gene to be causative for the OPD

spectrum disorders. Based on this finding, the focus of our project shifted to determine

the etiology of OPD1 in our patient. An introduction to X chromosome inactivation and

X;autosome translocations are presented in chapter 4, while background information

2 regarding OPD and Filamin A is presented in chapter 5. The results of this study are presented in Chapter 6.

The objectives of this thesis are to provide a thorough explanation of each independent project, to draw relevant conclusions from the results, and to identify future aims of this research.

3 SECTION I

THE ASSOCIATION OF ANTIOXIDANT RELATED GENE POLYMORPHISMS AND SECOND PRIMARY MALIGNANT NEOPLASM IN A PEDIATRIC HODGKIN LYMPHOMA POPULATION

CHAPTER 1

INTRODUCTION TO HODGKIN LYMPHOMA AND THE CHILDHOOD CANCER SURVIVOR STUDY

1.1: Introduction to Hodgkin Lymphoma

Hodgkin Lymphoma (HL) is a malignancy of the lymphatic system that first

presents as a painless swelling of the lymph nodes. Fever, night sweats, weight loss and

itching are other common symptoms. Disease progression occurs through the lymphatic

vessels as malignant cells spread into other lymph nodes or adjacent organs. The age of

incidence of HL follows a bimodal distribution which peaks first in young adults (15-34

years) and again in older individuals (>60 years) (Figure 1.1).1 Children and adolescents under the age of 20 represent approximately 12% of all HL diagnoses.1 Presently, the

five year event free survival of HL patients is greater than 90%.2, 3

The etiology of HL is unknown but may be caused by interactions between genes, the environment and infectious agents. Risk factors for the disease include age, exposure to Epstein-Barr virus, immune deficiency or a family history of HL. Unlike many other leukemias or lymphomas, an underlying cytogenetic abnormality is not detected in HL patients, though this may be due to ascertainment bias. The neoplastic component of HL is believed to be the Hodgkin and Reed-Sternberg (HRS) cell, which represents only 0.1-

5 1.0% of cells in tumor tissue.4, 5 The cellular origin of the HRS cell was only recently elucidated, as immunological studies revealed a clonal rearrangement of the immunoglobulin genes indicative of B-cell origin.6-8 HRS cells are thought to be highly interactive within the cellular environment. It has been hypothesized that the ability of

HRS cells to mimic antigen-presenting immune cells results in the pathogenesis of HL.9

1.2: Classification of Hodgkin Lymphoma

Two major subtypes of HL exist under the World Health Organization’s

classification scheme; classical HL and nodular lymphocyte predominant HL.10, 11

Classical HL is further divided into four groups including, nodular sclerosing (60% of pediatric cases), mixed cellularity (30% of pediatric cases), lymphocyte rich (10% of pediatric cases), and lymphocyte depleted (extremely rare in children).12 In most cases, all forms of classical HL are treated similarly, usually including multi-drug chemotherapy and radiation.13, 14 Nodular lymphocyte predominant HL represents only 5% of HL cases and typically presents early on with only local involvement. Very few or no HRS cells are seen in this type of HL.

1.3: Risk Factors and Causative Agents in Hodgkin Lymphoma

No definitive causative agents have been identified in HL, though certain risk factors have been linked to the disease. Exposure to the Epstein-Barr virus is associated with about 40% of HL cases, but it is unclear how the virus is involved in the development of disease.15 Serology studies have demonstrated that HL patients have

higher antibody titers to EBV as compared to controls.16 EBV RNA and protein have been found to localize to HRS cells in HL patients and the virus is known to be transcriptionally active.17 Additional studies have demonstrated that EBV encodes an

6 oncogene (LMP-1) capable of transforming cells in-vitro which gives some insight as to

how the virus may function to promote a malignant phenotype in vivo.18

Family history is another potential risk factor in the development of HL. In a study by

Mack et al., concordance for HL was demonstrated in monozygotic twins, but not in dizygotic twins, which suggests a genetic component is involved in the etiology of disease. The study revealed that monozygotic twins had a 99 fold excess risk for disease

(Standardized Incidence Ratio (SIR) 99; 95% Confidence Interval (CI) 48-182).19 In an analysis of the Childhood Cancer Survivor Study cohort, non-twin full siblings of a HL survivor had a 2 fold increased risk for the development of disease (SIR 2.3; 95% CI 1.6-

3.1).20

A dysregulation of immune function is associated with HL. Patients affected with

HIV/AIDS and those taking immunosuppressive drugs because of organ transplant or

auto-immune disease are at greater risk for HL.10 Additionally, HL patients have been shown to demonstrate a lack of immune response known as anergy. This was first observed as HL patients failed to generate a cuteneous reaction to tuberculin despite having an active form of the disease.21 The frequency of anergy has been correlated with

HL disease severity in patients, with a lack of immune response observed more

frequently in patients in the later stages of disease. Upon remission of HL, it has been

documented that patients recover the ability to react to an antigen.22 The delayed immune response in HL patients is of clinical importance as patients are extremely susceptible to the development of infection. Despite extensive study, it is unclear whether impaired cellular immunity serves as a predisposing factor or result of HL.22

7 1.4: Treatment of HL: Maximizing Efficacy While Minimizing Risk

With 5 year event free survival rates of greater than 90%, challenges in the treatment of HL have shifted to minimize secondary effects of therapy without reducing overall efficacy. Late effects of treatment are associated with increased morbidity and mortality of HL survivors, and frequently include infertility, cardiopulmonary disease, and second primary malignant neoplasms (SMN), in addition to other complications.23

The highest mortality risk in HL patients is due to a SMN.24 Reports of cumulative risk

for a SMN vary from 6% to 26%.14, 23, 25-32 The difference in risk is likely due to self-

reporting of SMNs, varying length of follow-up time, inclusion or exclusion of non-

melanoma skin cancer as a SMN, and differences in therapy regimens.

HL survivors face treatment-related sequelae caused by radiation and

chemotherapy agents used to treat their primary disease. With several decades of

treatment and research experience, investigators have become acutely aware of the

complications associated with late effects and have devised strategies to minimize

exposure to toxic agents. It is particularly desirable to limit exposure of the most

damaging agents including mechlorethamine and other alkylating drugs, anthracyclines,

and radiation. Elimination or dosage reduction of certain drugs may improve sequelae;

such as second hematological malignancy (associated with mechlorethamine), sterility or

premature ovarian failure (associated with alkylating agents), and cardiopulmonary

disease (associated with anthracyclines).33-36 Radiotherapy treatment effects such as growth impairment of bone and soft tissue, thyroid dysfunction, and second malignancy

(particularly breast cancer) could be ameliorated by limiting radiation dose and field or by omission of this modality altogether.27, 37-40

8 Current treatment strategies for pediatric HL utilize combination chemotherapy

and low dose radiotherapy to involved lymph nodes.41 Low risk patients may now be placed on abbreviated chemotherapy regimens with radiotherapy omitted if complete response occurs.14 Intermediate and high risk patients are placed into very intensive

chemotherapy regimens with or without radiation treatment.41

1.5: The Childhood Cancer Survivor Study

Funded by the National Cancer Institute, the Childhood Cancer Survivor Study

(CCSS) was initiated in 1993 to examine the long term effects of cancer and its treatment

in a cohort of pediatric survivors. Objectives of the study were twofold: to accrue

knowledge in order to facilitate the design of new treatment protocols and intervention

strategies, and to educate survivors about their diagnosis, treatment, appropriate follow-

up and long term health risks.

1.5.1: Study Design and Eligibility

The CCSS represents the collaborative effort of 27 participating clinical centers

in the US and Canada and was established to follow a cohort of pediatric and adolescent

survivors of cancer. In order to assemble the cohort, participating institutions contacted

eligible patients who met the following criteria: (a) diagnosis of leukemia, central

nervous system malignancy, HL, non-HL, neuroblastoma, soft tissue sarcoma, kidney or

bone cancer; (b) date of diagnosis between January 1, 1970 and December 31, 1986; (c)

age 0-21 at the time of diagnosis; (d) survival 5 years past initial diagnosis; (e) English or

Spanish speaking (for data collection purposes); and (f) a US or Canadian resident.

Patient contact and recruitment began in 1994.42

9 Following recruitment into the study, survivors or family members (if the patient

was deceased, or a minor) were sent a comprehensive baseline questionnaire containing

289 questions. Subject matter was wide-ranging and included questions about patient

demographics, health information, depression, pain, primary cancer recurrence,

development of a SMN, marital status, pregnancy history, education, employment,

insurance, and family history. Siblings were also contacted from a random sampling of

study participants in order to serve as a comparison group.

1.5.2: Study Participation

Based on the initial subject registry completed by the clinical centers, a total of

20,276 survivors met eligibility requirements for the cohort. Of that group 17,280

subjects were successfully contacted, with the remaining survivors lost to follow-up. A

total of 14,054 survivors completed the study questionnaire representing 69.3% of the eligible cohort.42 Participating survivors were asked to sign a consent form to allow for medical record abstraction. Detailed information regarding cancer treatment, including chemotherapy, radiation, and surgical procedures was documented for each member of

the cohort. In the time since the baseline questionnaire was completed, additional follow-

up questionnaires have been sent to study participants to update cohort characteristics.

Demographic characteristics within the cohort indicated that greater than half (54%) of

CCSS participants were male, and the majority (87%) were classified as Caucasian.42

Detailed participant information including cancer diagnosis and treatment as well as cohort demographic data is shown in table 1.1.

10 1.5.3: HL Patient Demographics within the CCSS

HL patients represented 13% of the eligible CCSS cohort with a total of 2718

survivors. Of this group, 1906 HL survivors chose to participate in the study, and

complete medical records plus baseline data was available for 1815 HL survivors.42

Significantly increased morbidity and mortality was observed in the HL cohort.

Within this group, deaths occurred 8 times more often than expected relative to US

mortality rates (Standardized Mortality Ratio [SMR] 8.3; 95% CI 7.4-9.2)24. Aside from primary cancer recurrence, the highest rate of mortality was due to a SMN. Out of all cancer diagnoses within the CCSS, patients with HL were found to be at the greatest risk for death caused by a SMN (SMR 24.0; 95%CI 19.2-29.7).24

Of 1815 HL survivors within the CCSS cohort, 111 (6%) had a confirmed

SMN.23 Observed SMNs included leukemia, non-Hodgkin lymphoma (NHL), CNS tumor, breast cancer, bone cancer, sarcoma, thyroid cancer and melanoma (Table 1.2).

The most common solid tumor in HL survivors was breast cancer, representing 32% of all SMNs. Little is known about the differences between HL patients who develop a

SMN in comparison to those who do not.

1.6: Genetic susceptibility to SMN in Cancer Survivors

Aside from primary cancer recurrence, the greatest threat to pediatric cancer survivors is the development of a SMN. In a review of the CCSS cohort, 91.9% of individuals without a SMN were alive at time of study (2001), in comparison to 59.4% who developed a SMN.23 In order to improve screening practices, follow-up care and to enhance the quality of life for cancer survivors, investigators have attempted to determine the etiology of SMNs.

11 The development of some SMNs is thought to be related to a patient’s response to

therapy. Acute myelogenous leukemia (AML) is observed within 4-7 years of primary

cancer diagnosis and is associated with chemotherapy.23, 36 Solid tumors such as breast and thyroid cancer are associated with radiotherapy and typically occur greater than 5 years past primary cancer diagnosis. Non-Hodgkin Lymphoma does not appear to be associated more strongly with either chemotherapy or radiation, and like many other

SMNs, the etiology is unknown.23 Family history may also impact the risk of SMN in

cancer survivors. In a recent study by Nichols et al., a positive family history of cancer

(3 or more relatives with cancer and spanning 2 generations) was observed in 21% of HL

survivors who developed a SMN in comparison to 4% who did not develop a SMN.43

Findings such as these support the hypothesis that an underlying genetic predisposition to malignancy may exist.

Therapy-related SMNs have been studied to determine if there is an underlying genetic susceptibility in affected patients. Of particular interest are genes involved in

DNA repair and drug metabolism. In a study by Mertens et al, polymorphisms within the

XRCC1 and glutathione S-transferase (GST) genes were studied in relation to radiation therapy related SMNs in HL patients enrolled in the CCSS.44

XRCC1 functions within the DNA base excision repair pathway to correct

damaged bases and to repair single strand breaks.45 Individual bases may be damaged by

deamination, alkyation, or oxidation caused by a number of factors, including radiation

and xenobiotic exposure (particularly alkylating agents).45, 46 Polymorphisms within

XRCC1 resulting in amino acid substitutions, including Arg399Gln, have been previously

12 studied. Results have been contradictory, with the glutamine allele associated with a

decreased DNA repair capacity in some studies and a protective effect in others.47-50

GST genes code for enzymes involved in conjugation reactions to detoxify the highly reactive intermediate species formed during drug metabolism. The GST gene family includes 5 isoenzymes, two of which (GSTM1 and GSTT1) are commonly deleted resulting in no functional protein production. In a Caucasian population, 50% of individuals have a homozygous deletion of GSTM1, and 20% of individuals have a homozygous deletion of GSTT1.51, 52

The genotype of the XRCC1 polymorphism (Arg399Gln) and the presence or absence of GSTM1 and GSTT1 alleles were determined in HL survivors.44 A

nonsignificant increased risk of breast cancer was observed in HL survivors heterozygous

for the glutamine allele (Arg/Gln) at position 399 in XRCC1. Individuals with a homozygous deletion of GSTM1 were at an elevated risk of SMN with an odds ratio of

1.5 (95% CI 1.0-2.3). Increased risk for SMN in relation to XRCC1, GSTM1, and GSTT1 genotype was low, causing the authors to conclude that sensitivity to DNA damage is likely multigenic in this cohort.

The Mertans et al. study mirrors the results of many similar studies with few

significant findings correlating genotype and increased risk of SMN in a variety of

primary cancer backgrounds.47, 53-59 This would lead us to believe that risk of malignancy

is dependent on the additive effects of multiple genes in a variety of pathways including

DNA repair, detoxification, drug metabolism, and potentially antioxidant related genes.

Figure 1.1: Annual Incidence Rates of Hodgkin Lymphoma Age is a risk factor in the development of Hodgkin Lymphoma. Age of incidence occurs in a bimodal distribution with adolescents and young adults (15-34y) and individuals over the age of 60 years at the greatest risk. Incidence rates among men and women are nearly equal until age 35, at which point the disease prevalence decreases in women and remains constant in men. (Created using SEER incidence data1)

14 Figure 1.1

Table 1.1: Composition of Participants within the CCSS Cohort Demographic characteristics and cancer diagnosis and treatment information is listed for participants in the CCSS cohort. Participants included pediatric cancer survivors who met CCSS eligibility criteria and completed a baseline questionnaire for enrollment into the cohort. (Demographic data available from www.stjude.org/ccss)

16 Table 1.1 Composition of Participants within the CCSS cohort

Participant Breakdown Characteristic Number and (Percentage)

Total number of Participants 14054 (100)

Sex Male 7541 (54) Female 6513 (46)

Race Caucasian 11679 (87) Hispanic 635 (5) African American 291 (2) Asian/Pacific Islander 165 (1) Other 615 (5)

Diagnosis ALL 4233 (29) AML 357 (3) Other Leukmia 137 (2) Astrocytomas 1158 (8) Medulloblastoma/PNET 366 (3) Other CNS tumors 305 (2) Hodgkin Disease 1906 (14) Non-Hodgkin Disease 1054 (7) Kidney tumors 1214 (9) Neuroblastoma 928 (7) Soft tissue sarcoma 1222 (9) Ewing's sarcoma 393 (3) Osteosarcoma 727 (5) Other bone tumors 53 (<1)

Age at Diagnosis <1 975 (7) 1-3 3539 (25) 4-7 3157 (22) 8-10 1538 (11) 11-14 2373 (17) 15-20 2472 (18)

Treatment Modality Chemotherapy + Radiation + Surgery 5491 (44) Chemotherapy + Radiation 1443 (12) Chemotherapy + Surgery 2237 (18) Radiation + Surgery 1477 (12) Chemotherapy Only 797 (6) Surgery Only 901 (7) Radiation Only 32 (<1)

Table 1.2: Type of SMN in HL Survivors within the CCSS Cohort The distribution by type of second and subsequent malignant neoplasms in HL survivors within the CCSS cohort is shown in the table. Radiation induced cancers such as breast and thyroid carcinoma are the most common SMNs in survivors of HL. This table represents SMNs within the CCSS HL cohort as of 2001.

18 Table 1.2

Second and Subsequent Primary Malignant Neoplasms in HL Survivors within the CCSS cohort

Second Malignant Neoplasm Number of Cases

Breast Cancer 35 Thyroid Cancer 17 Sarcoma 9 Leukemia 9 Non-Hodgkin Lymphoma 7 Melanoma 6 Bone Cancer 5 CNS Tumor 5 Other Malignancies 18 Total 111

Subsequent Malignant Neoplasm Number of Cases

Breast Cancer 4 Thryroid Cancer 3 Non-Hodgkin Lymphoma 1 Melanoma 1 Other Malignancies 1 Total 10

CHAPTER 2

INTRODUCTION TO REACTIVE SPECIES, ANTIOXIDANT ENZYMES, AND THE RELATIONSHIP BETWEEN OXIDATIVE STRESS AND MALIGNANCY

2.1: Reactive Species and Antioxidants

2.1.1: Reactive Oxygen Species

Reactive oxygen species (ROS) are natural byproducts of cellular metabolism that

have the potential to damage DNA, oxidize proteins, and initiate lipid peroxidation,

resulting in mutagenesis or cell death.60 Common forms of ROS include superoxide

•– • anion (O2 ), hydroxyl radical ( OH), and hydrogen peroxide (H2O2). Both exogenous and endogenous sources contribute to the generation of ROS. Cellular sources of ROS include mitochondria, peroxisomes, cytochrome P450 metabolism, and inflammatory cytokine activation.61 Because mitochondria serve as the site of oxidative phosphorylation, they are thought to generate the greatest amount of ROS within the cell.62 Superoxide radicals are regularly produced in mitochondria as a result of the electron transport chain and may subsequently be converted to H2O2 or other forms of

ROS.63 Exogenous sources, such as carcinogens, xenobiotics, and radiation, are also associated with the generation of ROS.61, 64-67

20 Chemotherapeutic drugs, such as those used in the treatment of HL, are known to

generate ROS during metabolism. The anthracycline class of drugs (which includes

doxorubicin), are particularly associated with ROS generation because they are known

redox-cycling agents.68, 69 Anthracyclines are quinone based drugs, and during

activation, the quinone molecule serves an as electron acceptor and is reduced.

Biological reducing agents such as NADPH, NADH, glutathione, and flavoproteins serve

as electron donors and become oxidized during the reaction.69 The reduced form of anthracycline may become re-oxidized in the presence of molecular oxygen by the donation of an electron to O2, resulting in the generation of superoxide anion. This redox

cycle continues as long as reducing agents and molecular oxygen are available.

Continued redox cycling may be cytotoxic due to the increasing amounts of ROS, and

depletion of reducing agents, such as glutathione, one of the most important molecules

involved in the cellular antioxidant system.68

2.1.2: Antioxidant Enzymes

The levels and actions of ROS are controlled by the presence of antioxidant enzymes within the cell. These enzymes function to mitigate the effects of ROS by detoxifying reactive species so that they can safely be removed from the body.

Antioxidants include the family of superoxide dismutases (SODs), glutathione peroxidases (GPXs), and catalase (CAT) enzymes. For proper biological function to occur, it is critically important to maintain a steady state balance between ROS production and neutralization by antioxidant enzymes. A process known as oxidative stress occurs when ROS overwhelm the capabilities of antioxidant enzymes leading to oxidative damage.

21 Antioxidant enzymes drive chemical reactions within the cellular environment to

convert ROS into non-toxic molecules. Chemical reactions involving ROS and

antioxidant enzymes are shown in figure 2.1. While each enzyme has a specialized

function, the system is complex and highly integrated. SOD enzymes function by

•– catalyzing the dismutation of the superoxide anion (O2 ) to form H2O2 and O2. GPXs convert H2O2 to H2O, by the oxidation of glutathione to glutathione disulfide, and

61, 70, 71 catalase converts H2O2 to H2O and O2. Although substrate affinities, conversion

rates, and subcellular distributions differ, GPX and catalase have somewhat overlapping functions, and the loss of catalase may be partially compensated by GPX activity.72

Additionally, dietary intake of antioxidants, such as vitamins C and E, as well as carotenoids and flavonoids, may help counteract the damaging effects of ROS.73-75

2.1.3: Reactive Nitrogen Species

Reactive nitrogen species (RNS) are another type of reactive molecule produced during normal physiological processes. Examples of RNS include, nitric oxide (NO•), a signaling molecule generated by nitric oxide synthase (NOS), as well as peroxynitrite anion (ONOO-), a powerful oxidizing free radical.61 Though NO• is a short-lived molecule (with a half life of mere seconds), it serves as a critical messenger involved in a

wide variety of biological processes.76 It is an endogenously produced gas involved in regulating neurotransmission, immune function, blood pressure, and smooth muscle relaxation.61, 77, 78 In combination with superoxide anion, NO• has the potential to form

ONOO-(figure 2.2). This reaction typically occurs during inflammation, as immune cells

•– • produce an oxidative burst (releasing O2 and NO ) when they come into contact with an antigen.79 The reactive potential of peroxynitrite anion is much greater than that of NO•.

22 As an oxidizing free radical, peroxynitrite anion is capable of damaging DNA and

inducing lipid peroxidation. A process known a nitrosative stress occurs when the steady

state balance between RNS production and neutralization is unstable leading to an

overload of RNS and the potential for cellular damage.80 As both reactive oxygen

species and reactive nitrogen species have the potential to cause oxidative damage, they

may be referred to collectively as reactive species (RS).

2.2: Types of Antioxidant Enzymes

2.2.1: Superoxide Dismutase

SOD is an essential antioxidant enzyme that defends cells against potentially

damaging superoxide radicals. There are three known human isoforms of SOD, two of

which are copper-zinc containing enzymes (CuZnSOD), and one manganese containing

enzyme (MnSOD). SOD1 (CuZnSOD) is found in the cytoplasm and nucleus in the form

of a dimer.81 SOD2 (MnSOD) is a tetrameric protein that functions in the mitochondria,

and SOD3 (CuZnSOD) is a tetrameric, extracellular form of the enzyme82, 83 While each enzyme performs a critical function, SOD2 is particularly important due to its location within the mitochondria. Over 95% of cellular oxygen is metabolized in the

mitochondria during oxidative phosphorylation, making these organelles particularly

susceptible to oxidative stress.84, 85

2.2.2: Glutathione Peroxidase

Members of the GPX family are selenoproteins, meaning they incorporate selenocysteine into their primary protein structure. Incorporation occurs at the active site of the protein, and selenocysteine is known to act as an efficient biological catalyst.86-88

23 The main function of GPX is to catalyze the reduction of hydroperoxides to water and the

respective alcohols, while oxidizing glutathione (GSH) to glutathione disulfide (GSSG).

GSH is an important antioxidant that functions as a scavenger for ROS, including

• - 89 OH, ONOO , H202 and lipid peroxyl radicals. It is synthesized in the liver from

glutamate, cysteine, and glycine to form a tripeptide. Within the cellular environment,

glutathione cycles between a reduced form (GSH) and an oxidized form (GSSG) with the ratio of GSH to GSSG greater than 500 to1.90 Glutathione sulfide reductase is responsible for converting GSSG to GSH, and uses NADPH as a cofactor in this reaction.91 GSH and GSSG serve as the major redox couple within the cell, and are

important determinants in the total cellular antioxidant capacity.89, 90

In humans, four distinct isoforms of GPX have been identified. Expression of

each isoform is ubiquitous, though levels vary by tissue type.71 GPX1 is found in the cytoplasm and mitochondria of most cells, and is highly expressed in the liver, kidney,

92 lung, and red blood cells. This isoform functions to catalyze the reduction of H2O2 and some organic peroxides and is similar in both function and homology to GPX2 and

GPX3. The cytosolic isoform, GPX2, is found mainly in the liver and gastrointestinal tract, and the unique distribution of this enzyme suggests that it may protect against ingested lipid hydroperoxides.92, 93 Produced mainly by the kidney, GPX3, is secreted

into the extracellular environment and is most highly detected in plasma.94 Finally,

GPX4 is found in the cytoplasm, nucleus, and mitochondria of many cells, and is most highly expressed in the testis.95

Studies involving GPX4 have demonstrated that it differs from the other GPX enzymes in several important ways. Unlike the other isoforms, GPX4 is capable of 24 directly reducing phospholipid hydroperoxide within the cell, including lipid peroxides

derived from cholesterol.96, 97 Additionally, it functions as a monomer, rather than as a tetramer as seen in the other isoforms.98 Finally, GPX4 is translated as either a long or short form of the protein, with the long form containing a mitochondrial targeting sequence at the N terminus of the protein.99, 100 Once imported into the mitochondrion, the long form of the protein is modified to remove the targeting sequence, resulting in a protein of equal length to the short form. Expression of the short form was found to occur in the nucleus, endoplasmic reticulum and cytosol, but was not observed in the

101 mitochondrion. While all GPX isoforms and catalase share a common substrate, H2O2,

GPX is considered to be the major enzyme responsible for the reduction of lipid and organic peroxides.71

2.2.3: Catalase

Catalase is an antioxidant enzyme that is concentrated mainly in peroxisomes (an organelle involved in the breakdown of fatty acids). This enzyme functions as a tetramer and contains a ferriprotoporphyrin group in each of the four subunits71 As the name implies, this enzyme catalyzes the conversion of H2O2 to water and oxygen. While catalase is clearly a critical enzyme, GPX has the ability to compensate partially for low or absent catalase function.72

2.3: Oxidative Stress and Cancer

2.3.1: A Multi-Stage Model of Carcinogenesis

There is mounting evidence that persistent oxidative stress is a major contributor to the development of cancer.61, 65, 70-72 In a multi-stage model of carcinogenesis, the development of cancer is based on the cumulative effects of successive events within a

25 single cell.102, 103 The model may be described by three stages including the initiation,

promotion and progression of tumorigenesis. The effects of oxidative stress within a cell

can impact each stage of the model as shown in figure 2.3.

2.3.2: Initiation

Genetic change is the initiating event in the multi-stage model of carcinogenesis.

During periods of oxidative stress, damage to DNA has the potential to result in genetic

change. Oxidizing free radicals may cause irreparable damage including base

modification, DNA strand breaks, and DNA-protein cross-links.60, 104 The presence of modified base products, such as 7,8-dihydro-8-oxy-2’-deoxyguanosine (8-OH-dG) serve as good indicators of oxidative base damage. Studies involving cancers of the breast, lung, and kidney all report significant increases of 8-OH-dG in tumor vs. normal tissues from the same patient.73, 74, 105

2.3.3: Promotion

Following cellular damage, an initiated cell may gain a pre-neoplastic phenotype due to the dysregulation of cell signaling pathways and tumor suppressors, and an inhibition of apoptosis. The clonal expansion of initiated cells represents the promotion stage of carcinogenesis. Promotion is considered to be a reversible event, as a constant exposure to tumor promoting factors is required to induce a malignant phenotype.61, 106

2.3.4: RS: A Balancing Act of Tumor Promotion or Tumor Suppression

The effects of oxidative damage within a cell represent a paradox between tumor

suppression and tumor promotion due to the nature of RS (Figure 2.4). The collective

cellular environment (including cell type, levels of antioxidants, types of RS, relative

amounts of RS, and total exposure) ultimately sways the balance in one direction.70 On

26 one end of the spectrum, tumor suppression may occur because of the severe oxidative

damage that results from an overload of RS. Suppression of a malignant phenotype

occurs due to senescence, necrosis, or induction of apoptosis in affected cells.

Alternately, a low to moderate increase in RS may have a tumor promoting effect.

Escalating cellular damage is caused by oxidative stress, but not to the point of inducing

apoptosis. In turn, oxidative damage may result in the dysregulation of cellular processes

and contribute to proliferation and inhibition of apoptosis.107

2.3.5: Progression

The final stage in the model of carcinogenesis is progression. Pre-neoplastic cells

from the promotion stage undergo further genetic damage and are no longer able to

respond to growth inhibition or apoptotic signals.61 Evidence suggests that the involvement of RS in tumorigenesis may also include roles in angiogenesis and metastasis. RS may activate matrixmetalloproteinases, which degrade the extracellular matrix and contribute to the invasive potential of the tumor.108, 109 RS are also known to regulate vascular endothelial growth factor signaling in relation to angiogenesis.110, 111

2.3.6: Antioxidant Enzymes and Malignancy

In relation to oxidative stress and cancer, a variety of studies involving antioxidant enzyme levels and malignancy have been performed. In studies of SOD activity, tumor cells nearly always show a decrease in SOD2 expression, and quite often a decrease in SOD1 as well.112-114 In knock-out experiments, mice lacking SOD1

demonstrated an increased rate in the development of liver cancer, while SOD2 null mice exhibit a neonatal lethal phenotype.115, 116 Interestingly, mice with a single copy of SOD2 survive, but are at increased risk for the development of cancer later in life.117 Catalase

27 activity is generally lower in tumor cells than in normal tissue, while glutathione

peroxidase activities have been found to be variable.113, 118 In mice lacking the catalase enzyme, an increased rate of spontaneous mammary tumor formation has been observed.119 Additionally, in a simultaneous knock-out experiment of both GPX1 and

GPX2, mice were found to develop intestinal cancer.120 While a causal relationship

between oxidants and cancer has yet to be proven, it is important to note the many

correlations that have been observed to date. Further research in this field will improve

our understanding of the role of oxidative stress in the development of cancer.

Figure 2.1: Chemical Reactions Involving ROS and Antioxidant Enzymes In reaction [1] superoxide anion may be converted to a secondary ROS after catalysis by SOD during the dismustase reaction. H202 is a substrate for both the GPX and CAT enzymes as shown in reactions [2] and [3]. A redox reaction occurs in [3] as reduced glutathione (GSH) is converted to its oxidized form (GSSG) in turn reducing H202 to H2O.

29 Figure 2.1

Figure 2.2: Reactive Species Interact to Form an Oxidizing Free Radical • •– Nitric oxide (NO ) in combination with superoxide anion (O2 ) forms the oxidatively active peroxynitrite anion (ONOO-). This reaction is known to occur during the oxidative burst triggered by an inflammatory reaction within the immune system.

31 Figure 2.2

· ·- - NO + O2 ONOO

Figure 2.3: A Multi-Stage Model of Tumorigenesis Relative to Oxidative Stress Oxidative damage from RS may serve as the initiating event in carcinogenesis. During the promotion stage, RS may cause either tumor promoting or tumor suppressing effects dependent on the level of oxidative stress. Following progression, RS may contribute to the invasiveness and metastic potential of tumor cells.

33 Figure 2.3

Figure 2.4: Cellular Response to ROS During Times of Oxidative Stress ROS are constantly generated during normal physiological processes in response to endogenous and exogenous stimuli. Inherently, a steady state balance is maintained within a cell as antioxidant enzymes counteract and neutralize ROS. During times of oxidative stress, the balance has shifted due to an increase in ROS production, or a decrease in antioxidant enzyme buffering capacity. As levels of ROS increase, so does the potential for oxidative damage to occur. While small amounts of damage are quickly mitigated by cellular repair processes, increasing damage over an extended time period is likely to be detrimental. Low to moderate levels of ROS may be tumor promoting as typical cellular processes are disrupted due to the cumulative effects of oxidative damage. The inappropriate activation of genes, phosphorylation of proteins, and increase in ROS availability for cell signaling may result in a proliferative phenotype. Alternately, when concentrations of ROS are too high, the results are catastrophic due to overwhelming oxidative damage and ultimately cell death.

35 Figure 2.4

CHAPTER 3

THE ASSOCIATION OF ANTIOXIDANT RELATED GENE POLYMORPHISMS AND SECOND PRIMARY MALIGNANT NEOPLASMS IN A PEDIATRIC HOGDKIN LYMPHOMA POPULATION

This work was performed in collaboration with Dr. Kim McBride (The Ohio State University/Nationwide Children’s Hospital), Dr. Stella Davies (Cincinnati Children’s Hospital Medical Center), Dr. John Whitton and Dr. Wendy Leisenring (Fred Hutchinson Cancer Research Center).

3.1: Abstract

Improved treatment strategies in pediatric Hodgkin Lymphoma (HL) have

resulted in a cure rate approaching 95%, yet the development of a second primary

malignant neoplasm (SMN) is a risk for long-term survivors. In a pediatric HL

population, between 6-26% of patients will develop a SMN within 30 years following

treatment. The etiology of HL is still largely unknown, as is the cause of certain types of

SMNs.

Oxidative stress has been linked to the development of cancer due to the damaging effects of reactive oxygen species (ROS) on DNA, lipids, and proteins. During periods of extended oxidative stress, the damaging effects of ROS are likely to increase which in turn raises the risk of genetic change. In a multi-stage model of tumorigenesis, genetic change is considered to be the initiating event of cancer development. Under

37 normal circumstances, antioxidant enzymes within the cell mitigate the effects of ROS by

converting reactive species into non-toxic molecules. Mechanisms within the cell

maintain a steady state balance between ROS and antioxidant enzyme levels. We

hypothesize that individuals predisposed to lower levels of antioxidant enzyme activity