University of Southern California Dissertations and Theses

Genetic risk factors in multiple myeloma

(USC Thesis Other)

Genetic risk factors in multiple myeloma

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

GENETIC RISK FACTORS IN MULTIPLE MYELOMA

by

Kristin Alyse Rand

_________________________________

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(EPIDEMIOLOGY)

December 2012

Copyright 2012 Kristin Alyse Rand

ii

Dedication

I dedicate this dissertation to my parents, whose unconditional love and constant
support has carried me through this journey. Without their unwavering belief in my
abilities, I never could have accomplished all that I have. I would also like to dedicate
this work to my husband, my rock throughout this experience. He has seen me through
the good times and bad, and has always helped me push forward. His enduring patience
and generosity is inspiring. Thank you so much; I could not have done this without you
guys.

iii

Acknowledgements

First, I want to thank my advisor Dr. Wendy Cozen for her mentoring and
guidance. I am so grateful for her patience during my learning process and am motivated
by her passion for our work and her daily quest for the truth. I would also like to thank
Dr. David Conti for the constant inspiration, support, and wisdom he has provided
throughout my PhD experience. I would like to acknowledge my committee members Dr.
Chris Haiman, Dr. Roberta McKean-Cowdin, and Dr. Clive Taylor for their insightful
suggestions and support. I am extremely fortunate and grateful for the open-door
mentoring and knowledge that Dr. Thomas Mack, Dr. David Van Den Berg, and Dr.
Victoria Cortessis have shared with me over the years.
I would also like to acknowledge my friends who have been there to support me
everyday. Amie Hwang is an amazing woman and I am sure I couldn’t have done this
without her. I want to thank Pedram Razavi, Mohamed Hassanein, Tanmai Saxena, Chris
Edlund, Melissa Frasco, and Ali Ozhand for being the wonderful people they are.
During various stages of my PhD, I have had the most fortuitous experience to
work with my brother, Adam Muench. His eternal optimism and steadfast support has
been invaluable, and I am eternally grateful for all that he has done.

iv

Table of Contents

Dedication ii

Acknowledgements iii

List of Tables vii

List of Figures viii

Abstract ix

Chapter 1: Overview of Multiple Myeloma
1.1 B-Cell Development and Differentiation: The Formation
of Plasma Cells 1
1.1.1 Antigen-Independent B-Cell Development 1
1.1.2 Antigen-Dependent B-Cell Activation and Maturation 2
1.2 Malignancy of the Post Germinal-Center
Plasma Cell: Multiple Myeloma 3
1.2.1 Subtypes of Multiple Myeloma 4
1.2.2 Diagnostic Criteria and Evaluation 5
1.2.3 Treatment 5
1.2.4 Monoclonal Gammopathy of Undetermined Significance
and Smoldering Myeloma 6
1.2.5 Multiple Myeloma and MYC relationship 7
1.3 DNA Repair Pathways and B-Cell Specific Mechanisms 8
1.3.1 B-Cell Specific Mechanisms and Involvement of DNA Repair 9
1.4 Epidemiology of Multiple Myeloma 10
1.4.1 Descriptive Epidemiology 10
1.4.2 Environmental Risk Factors 13
1.5 Review of Genetic Risk Factors and Multiple Myeloma 14
1.5.1 A Summary of Familial Studies in the Literature 14
1.5.2 Candidate Gene Studies and Multiple Myeloma
Risk – Review of the Literature 17
1.5.3 Genome-wide Association Studies (GWAS) and
MM – Review of the Literature 24
1.6 Overall Summary 24
Chapter 1 References 26

Chapter 2: Genotypic Variation in the 8q24 Region and the Risk
of Multiple Myeloma
2.1 Introduction and Rationale 32
2.2 Abstract 33
2.3 Introduction 34
v

2.4 Study Design and Methods 35
2.4.1 SNP Selection and Genotyping 37
2.4.2 Statistical Analysis 38
2.5 Results 39
2.6 Discussion 41
2.7 Tables 46
Chapter 2 References 56

Chapter 3: Functional Polymorphisms in DNA Repair Pathways
and Multiple Myeloma Risk
3.1 Introduction and Rationale 60
3.2 Abstract 60
3.3 Introduction 61
3.4 Methods 63
3.4.1 Participants 63
3.4.2 SNP Selection and Genotyping 66
3.4.3 Statistical Analysis 68
3.5 Results 70
3.6 Discussion 71
3.7 Tables 73
Chapter 3 References 78

Chapter 4: Meta-Analysis of Genome-wide Association Studies of
Multiple Myeloma Risk
4.1 Introduction and Rationale 80
4.2 Abstract 81
4.3 Introduction 81
4.4 Methods 82
4.4.1 Source of Subjects 83
4.4.2 Genotyping and Quality Control 85
4.4.3 Statistical Analysis 86
4.5 Results 88
4.6 Discussion 89
4.7 Tables and Figures 92
Chapter 4 References 101

Chapter 5: Summary and Future Directions 103
Comprehensive References 107
Appendices
Appendix A: Summary of Tables 1a-2b 117
Appendix B: Tables 1a, 1b, 2a, and 2b 117
vi

Appendix C: Linkage Disequilibrium patterns in Whites
and Africans in the 8q24 Region 119
Appendix D: Graphical Depiction of Projects in this Dissertation 120

vii

List of Tables
Table 2.1 Demographic Characteristics of White and African American
Participants in the Studies Contributing to the Pooled Analysis 46

Table 2.2. Published Associations: 8q24 SNPs and Increased Cancer Risk 47

Table 2.3. Allele Frequencies of Selected 8q24 SNPs among Controls 48

Table 2.4. The Association between 8q24 Genotypes and Risk of MM
from a Pooled Analysis in Whites and African Americans 49

Table2.5a. Correlation between SNPs in Individuals of European
Ancestry (1000 Genomes June 2011 data release) 50

Table 2.5b. Correlation between SNPs in Individuals of African
Ancestry (1000 Genomes June 2011 data release) 52

Table 2.6a. 8q24 Genotypes and Risk of Multiple Myeloma among
Whites by Study Site 54

Table 2.6b. 8q24 Genotypes and Risk of Multiple Myeloma among
African Americans by Study Site 55

Table 3.1. Forty-two Functional SNPs 73

Table 3.2. Demographic Characteristics of the Initial Los Angeles Study
and the Multi-Center Replication Set 74

Table 3.3. Top DNA Repair SNP Estimates in the Discovery and
Replication Sets, and the Meta-Analysis Results 75

Table 3.4. Demographic Characteristics for the UCSF and Mayo
Replication Sets 77

Table 4.1. Demographic Characteristics of the USC Multi-center Study 92

Table 4.2. Demographic Characteristics of the UCSF Study 93

Table 4.3. Top 25 SNP Associations, by Study and Meta-analyzed 94

Table 4.4. Replication of Published SNPs Associated with MM Risk
from a GWAS 95

viii

List of Figures

Figure 1.1 Example of an Antibody 1

Figure 1.2 Antigen Dependent B-cell Activation and
Maturation in the Germinal Center 3

Figure 1.3 Current Suggested Approach for Treatment of MM 6

Figure 1.4 Gender-specific Annual Age-specific Incidence Rates
(per 100,000 by Race/Ethnicity in Los Angeles County) 11

Figure 1.5 MM Average Annual Incidence Rate Stratified by
Race and SES 13

Figure 4.1 Eigenvector Plots from Principal Components Analysis 96

Figure 4.2 Q-Q Plot Displaying the Observed versus the Distribution
of the –Log(P-Values) in the Meta-analysis (MAF >0.05) 97

Figure 4.3 LocusZoom Plot for Chromosome 5 - rs7443528 98
Figure 4.4 LocusZoom Plot for Chromosome 6 – rs76580792 99
Figure 4.5a Q-Q plot for SPORE Study 100

Figure 4.5b Q-Q plot for UCSF Study 100

ix

Abstract
Multiple myeloma (MM) is a malignancy of the plasma cells, accounting for 1%
of all cancer deaths and 20% of all hematological cancer deaths. In 2011 alone, 20,500
people will be diagnosed with myeloma and 10,600 people are expected to die from this
disease (www.cancer.org). To date, multiple myeloma remains incurable with a 5-year
survival rate of ~40%. There are few known genetic or environmental epidemiologic risk
factors for this disease. To assess genetic risk factors associated with multiple myeloma,
my dissertation focuses on both a candidate pathway approach, targeting areas that have
previously been implicated in the disease etiology and pathogenesis (such as DNA repair
pathways) as well as an agnostic approach, scanning the whole genome by employing a
genome-wide association study (GWAS) to identify novel pathways in disease etiology.
Although there are known environmental risk factors involved in the etiology of
MM, there is mounting evidence that heritable traits may also play a substantial role in
disease risk. This dissertation examines genetic risk factors to provide insight into the
etiology of multiple myeloma.
Chapter 1: Overview of Multiple Myeloma

1.1 B-Cell Development and Differentiation: The Formation of Plasma Cells
A B-cell is an essential part of the adaptive immune-system response, with a
primary function to secrete antibodies (also known as immunoglobulins) in response to
foreign antigens [1]. Once activated, B-cells are able to respond rapidly in an antigen-
specific manner. B-cell development results in the
production of plasma cells, which produce antibodies
consisting of two heavy chain and light chains
connected in the center at the hinge region (Figure 1.1).
The light chains are defined by the Greek letters kappa
or lambda. There are areas known as variable regions
(at the antigen binding site) involved in antibody diversity,
and constant regions consisting of heavy chains. Variable regions are involved in a
complex genetic process creating antibody specificity, and constant regions define the
isotype: IgM, IgD, IgE, IgA, or IgG.

1.1.1 Antigen-Independent B-Cell Development
B-cells develop from hematopoietic stem cells in the bone marrow. They begin as
precursor B-lymphocytes, and undergo V(D)J rearrangement where the immunoglobulin
heavy chain gene segments of pro-B-cells are rearranged to create pre-B-cells within the
bone marrow. These naive B-cells now express IgM antibodies and will exit the bone
Figure 1.1. Example of an
Antibody
1
marrow, traveling to the spleen to continue the maturation process [1]. This cell has not
been in contact with an antigen, but once exposed, an antigen-dependent maturation
process is activated.

1.1.2 Antigen-Dependent B-Cell Activation and Maturation
Once exposed to an antigen, naive B-cells begin proliferating in lymphoid tissues
including the spleen, lymph nodes, and tonsils. Germinal centers (GC), or specialized
areas where proliferation and differentiation occur, are formed and B-cells begin multiple
rounds of somatic hypermutation, affinity maturation, and class switch recombination [2,
3]. Somatic hypermutation modifies the immunoglobulin variable region through DNA
strand breaks and nucleotide exchanges, deletions, and duplications [3], while class-
switch recombination is a DNA-level mechanism responsible for isotype or “class”
switching. In class-switch recombination (CSR), the heavy chain of the antibody, that up
until this point has been IgM, is changed to IgG, IgA, IgD, or IgE. Both somatic
hypermutation and CSR will be explained in more detail in a future section of this
dissertation. During GC events, a selection process is occurring where newly generated
cells producing an antibody with a low affinity for the antigen undergo apoptosis, while
those with an increased affinity are selected. These cells then exit the GC as memory
cells, a B-cell with a prolonged lifespan and a “memory” for specific antigens, or long-
lived plasma cells that home to the bone marrow (Figure 1.2).
2

Figure 1.2. Antigen Dependent B-cell Activation and Maturation in the Germinal Center
(Adopted from Klein et al., Nature Review, 2008)

1.2 Malignancy of the Post Germinal-Center Plasma Cell: Multiple Myeloma
Multiple Myeloma (MM) is the clonal proliferation of malignant plasma cells in
the bone marrow, which produce monoclonal immunoglobulins [4]. Evidence of somatic
mutation and immunoglobulin class-switching are confirmation that MM exclusively
arises from post- germinal center plasma cells [2]. Normally, plasma cells make up less
than 5% of cells in the bone marrow [4]. MM is characterized by an excessive amount of
plasma cells in the bone marrow (> 10%) and is associated with high levels of
monoclonal (M) protein. Clinically referred to as an M-spike, a monoclonal (M) protein
3
is defined as a single type of immunoglobulin (or part of an immunoglobulin) found in
excess in the blood and/or serum [5]. While plasma cells are quiescent, MM cells have
been shown to have a low proliferative rate, potentially due to their relationship with the
bone marrow microenvironment [5]. Symptoms of MM, particularly in advanced
disease, include anemia, lytic bone lesions, immunodeficiency, and renal impairment [6].

1.2.1 Subtypes of Multiple Myeloma
Overall, MM can broadly be divided into hyperdiploid and non-hyperdiploid
subtypes. Hyperdiploid MM consists of numerous chromosome trisomies and few IgH
(heavy chain) translocations, while non-hyperdiploid MM is characterized by IgH
translocations and is a more aggressive form of the disease [7, 8]. About 50-60% of
patients have hyperdiploid MM with the other ~40% of patients classified in non-
hyperdiploid subcategories by translocation type [6].
Five primary IgH translocations involving 14q32 have been identified in MM, and
have been associated with gene dysregulation by a strong enhancer in the IgH locus [9].
The five recurrent translocations include 11q13 (cyclin D1), 6p21 (cyclin D3), 4p16
(FGFR3 and MMSET), 16q23 (c-MAF), and 20q12 (MAFB)[9]. Additionally, 16q23 is
associated with a more aggressive disease outcome [7, 10]. MM also exhibits secondary
translocations, which are hypothesized to be associated with disease progression [11].
MYC translocations are considered secondary events, and will be described in more detail
in this dissertation.

4
1.2.2 Diagnostic Criteria and Evaluation
The diagnosis of MM is based on the presence of at least 10% of clonal bone
marrow plasma cells, monoclonal protein in the serum or urine (M-spike), and myeloma-
related organ damage assessed by the CRAB criteria: the presence of hypercalcemia,
renal insufficiency, anemia, or bone disease [4]. Diagnostic evaluation includes a
medical history and physical examination, routine testing including a complete blood
count, chemical analysis, serum and urine protein electrophoresis, and quantification of
the monoclonal protein, bone marrow testing, and imaging [4]. The international staging
system for MM includes three risk stages, while there is also high risk and standard risk
MM based on chromosomal abnormalities. Common criteria of high risk MM include
hypodiploidy, t(4;14), or deletion 17p13, while standard risk MM exhibits hyperdiploidy
or t(11;14)

[4].

1.2.3 Treatment
Historically, the first documented case of MM was diagnosed in 1844 and was
treated with rhubarb pill and infusion of orange peel [12]. With the evolution of drug
therapy in the early 1950’s, MM was treated with melphalan (a chemotherapy drug) and
corticosteroids until the mid-1980’s when autologous transplantation was introduced as a
treatment option for patients under the age of 65. In 1999-2002, novel therapeutic agents
such as thalidomide, bortezomib, and lenalidomide were introduced [12]. The current
treatment for MM includes a combination chemotherapy regimen consisting of these
novel therapeutic agents and bone marrow transplantation for eligible patients, with
5
treatment choice mainly related to age [4, 12]. In active MM, the disease should be
treated immediately, but patients with asymptomatic or smoldering myeloma should be
carefully monitored [4]. The following flow chart is the current suggested approach for a
newly diagnosed, symptomatic MM patient (Figure 1.3).

Figure 1.3. Current Suggested Approach for Treatment of MM (Adopted from Palumbo et al. 2011)

1.2.4 Monoclonal Gammopathy of Undetermined Significance and
Smoldering Myeloma
MM arises from an asymptomatic premalignant proliferation of monoclonal
plasma cells, termed monoclonal gammopathy of undetermined significance (MGUS).
These plasma cell clones produce a single antibody without any other symptoms of
multiple myeloma [13]. MGUS is prevalent in approximately 3% of Caucasians over the
6
age of 50 in the general population [12]. Genetic changes, the bone marrow
microenvironment, and infectious agents all may play a role in the progression of MGUS
to MM, but unfortunately, the role of these factors are not yet known [14]. However,
MGUS consistently precedes multiple myeloma [15], which strengthens the argument
that MGUS is a precursor condition in the cascade of events necessary to develop
multiple myeloma. Additional information regarding MGUS will be addressed in the
descriptive epidemiology section.

1.2.5 Multiple Myeloma and MYC relationship
The 8q24 region has been established as a risk region for many different cancers
[16]. This 1.2 Mb intergenic region is defined by five independent linkage
disequilibrium (LD) blocks, where the closest known gene is the oncogene MYC [17, 18].
Residing ~200-kb telomeric to the region, MYC is involved in cell cycle progression and
transformation, and mutations, overexpression, rearrangements, and translocations of this
gene have been associated with hematopoietic cancers [11].
MYC rearrangements are often observed in MM, but are rare in both MGUS and
smoldering MM, which suggests the rearrangements are a late stage event in MM tumor
progression [11]. These translocations are usually complex, involving three
chromosomes and can be non-reciprocal, involving inversions, deletions, duplications, or
amplifications [11]. Because 8q24/IgH translocations rarely occur within or near regions
specific to B-cell specific DNA modification mechanisms, these translocations are
presumed to be independent events which are mediated by mechanisms that have been
7
poorly characterized thus far [11]. MYC rearrangements occur in approximately 15% of
MM tumors, 45% of advanced MM tumors, and 90% of human multiple myeloma cell
lines (HMCL) [11, 19]. HMCL can be viewed as the ultimate stage of tumor progression,
displaying genetic changes accumulated both in vivo and in vitro. HMCL are usually
generated from extramedullary MM, which is a more aggressive form of the disease [11].

1.3 DNA Repair Pathways and B-Cell Specific Mechanisms
DNA is regularly damaged by both endogenous and exogenous mutagens, and
repair is an integral part of cell survival as unrepaired damage can lead to apoptosis or
unregulated cell growth [20]. There are five major DNA repair pathways: 1) direct
damage reversal, 2) base excision repair (BER), 3) nucleotide excision repair (NER), 4)
mismatch repair (MMR), and 5) double-stranded repair through homologous
recombination (HR) or non-homologous end joining (NHEJ) [21].
Direct damage reversal involves O
6
-methylguanine DNA methyltransferase
(MGMT), which removes alkyl or methyl adducts from the O
6
position of guanine [22].
In single-strand DNA damage, the error is excised and the undamaged strand is used as a
template for repair. BER acts on small lesions by removal of the damaged base and a few
neighboring base-pairs, while NER repairs bulky lesions or large adducts, and includes
unwinding, incision, and removal of up to 30 base-pairs [20]. The MMR pathway is
responsible for correcting base substitution mismatches and insertion-deletion
mismatches generated during DNA replication [23].
8
Double-strand breaks are more complicated to repair, as there is no undamaged
template available. With the help of certain proteins, NHEJ pieces the two ends of
broken DNA back together, which results in quick, but error-prone repair [24]. HR is a
more accurate repair method, but needs a sister chromatid so is only able to operate in
S/G2 phases of cell cycle [24]. NHEJ is hypothesized to be the main pathway of repair
for double-strand breaks [24].

1.3.1 B-Cell Specific DNA mechanisms and Involvement of DNA Repair
V(D)J recombination, somatic hypermutation (SHM), and class switch
recombination (CSR) are B-cell specific DNA modifications that are integral to plasma
cell differentiation, regulation, and diversity [11]. These mechanisms contribute to the
various stages of B-cell development, each with specific DNA-repair pathway
involvement, and errors in these pathways are hypothesized to be involved in the
development of MM.
V(D)J recombination - the recombination of variable (V), diversity (D), and
joining (J) gene segments - is a specialized DNA rearrangement to assemble new
immunoglobulin genes from the pre-existing segments [25]. NHEJ is involved in V(D)J
recombination, occurring in early stages where the immunoglobulin heavy chain gene
segments of pro-B-cells are rearranged to create pre-B-cells within the bone marrow [26].
Both SHM and CSR occur in the germinal center and depend upon the activity of
activation-induced cytidine deaminase (AID), which creates a U:G mismatch and
subsequently recruits BER and MMR, thereby activating both events [27]. SHM
9
introduces base-pair changes in the rearranged variable region of the antibody genes,
causing nucleotide changes, deletions, and sometimes a change in the amino-acid
sequence [3]. CSR occurs later in the germinal cell center after the antibody has been
exposed to an antigen [26]. Also referred to as ‘class-switching’, CSR is a mechanism
that changes the heavy chain class of the antibody from IgM to another possible isotype
(IgG, IgA, IgE, IgD) by DNA recombination and the NHEJ process [3].
It is hypothesized that errors in SHM or CSR are involved in the development of MM, as
recurrent translocations associated with the disease usually occur within or near switch
regions or VDJ sequences, and represent primary – or perhaps initiating – oncogenic
events [28].

1.4 Epidemiology of MM

1.4.1 Descriptive Epidemiology
The overall annual age-adjusted incidence rate for multiple myeloma is 5.6 cases
per 100,000 person-years (www.cancer.org), although incidence rates vary greatly across
race/ethnicity and gender. There are also distinct differences in incidence rates by
geographic location, with the lowest incidence rates of MM observed in Asian countries
[29].
Established risk factors include African-American race, male sex, aging, and a
positive family history of multiple myeloma or the precursor condition MGUS [30, 31].
African-Americans are at an approximate ~2- to 3-fold risk of developing both MGUS
10
and MM compared to whites [30]. Although there is a 2- to 3-fold increased prevalence
of MGUS in African-Americans residing in the US and in Africa, progression to MM is
similar across ethnic groups [31]. A population-
based study conducted in Los Angeles, with MM
case ascertainment from the
University of Southern
California Cancer
Surveillance program (USC-
SCP), reported the highest
age-adjusted incidence rates in
African-Americans, followed
by Spanish-surnamed whites
(SSW), non-spanish surnamed
whites (NSSW), and Filipinos,
with similar patterns in males and females [32]. This study also reported 50-60% higher
overall MM incidence rates in males. Figure 1.4 above presents the sex-specific annual
age-specific incidence rates (per 100,000) of MM by race/ ethnicity in Los Angeles
County [32].
Socio-economic status (SES) has also been implicated in myeloma risk, although
results remain inconclusive. Some studies have reported an inverse risk between SES and
both multiple myeloma incidence and survival [32-34], while others have not [35-38].
Figure 1.4. Gender-specific Annual Age-
specific Incidence Rates (per 100,000 by
Race/Ethnicity in Los Angeles County)
11
In a population-based case-control study utilizing cancer registries in three
geographic locations in the United States, Baris et al. found an increased risk in multiple
myeloma associated with lower SES, income, and education among both US African
Americans and Caucasians [38]. After computing population attributable risks, the
authors found occupation-based low SES to account for 37% of multiple myeloma in
African Americans and 17% in Caucasians, suggesting that SES-related factors explain a
proportion of the differential risk seen between these race/ethnicities [38].
Alternatively, utilizing the population-based cancer registry in Los Angeles,
Gebregziabher et al. reported increasing incidence rates of MM with increasing income in
both African American men and women (Figure 1.5). However, only modest trends were
seen in other ethnic/racial groups for men, and none for women [32]. To determine if this
result could be explained by better access to medical care, the incidence of all other
cancers was examined but there were no positive SES gradient trends observed in African
Americans in the USC-CSP registry in other cancers. This led the authors to conclude
that these results were in fact due to a true possible increased risk, not access to better
medical care [32].

12
Figure 1.5. MM Average Annual Incidence Rate Stratified by Race and SES

SES is considered a likely a proxy for occupational exposure, and further studies are
needed to elucidate the independent relationship of these risk factors with MM risk. Case-
control studies have assessed SES in many different ways, including census-tract data
combined with education, longest held occupation, or a combination of income,
education, and measured occupation [32]. There is the potential for misclassification with
these variables in studies relying on recruitment, as persons at both socio-economic
extremes participate in epidemiologic studies at a lower rate [32].

1.4.2 Environmental Risk Factors
Obesity has been repeatedly linked to an increased risk of developing multiple
myeloma and MGUS [39-41]. In 2007, a meta-analysis including 11 cohort studies and
four case-control studies found an increased risk associated with obesity relative to
normal weight (cohort studies: RR=1.27 (1.15-1.41), case-control studies: RR=1.82
(1.47-2.26)) [39]

. A more recent meta-analysis published in 2011 summarized risk across
13
20 prospective studies, 15 of which examined myeloma incidence and five of which
examined mortality [40]. In this study, Wallin et al. reported that individuals classified as
overweight and obese have a higher risk of developing and dying from multiple myeloma
(cohort studies of myeloma incidence: RR=1.21 (1.08-1.35), mortality studies: RR=1.54
(1.35-1.76)).
IL-6 plays an integral role in myeloma cell growth and survival [2]. Because fat
cells produce IL-6, a link between obesity and increased MM risk is biologically
plausible [42, 43]. Alternative mechanisms linking obesity to cancer include increased
inflammation, interference with cellular mechanisms of DNA repair, gene dysregulation,
or a common genetic predisposition to obesity and cancer [43].
Other potential risk factors with conflicting evidence of association with MM risk
include black hair dye, occupational exposures to chemicals such as benzene, ionizing
radiation, HIV and hepatitis C [30].

1.5 Review of Genetic Risk Factors and Multiple Myeloma
Evaluating the familial aggregation of MGUS and multiple myeloma is a way to
assess the evidence of a heritable etiologic component.

1.5.1 A Summary of Familial Studies in the Literature
Landgren et al. conducted a case-control study among 4458 MGUS patients and
14,621 first-degree relative controls compared with 17,505 population-based matched
controls and 58,387 first-degree relative controls. They found that relatives of cases were
14
almost three times as likely to develop MGUS or multiple myeloma (RR=2.8,
95%CI=1.4-5.6 and RR= 2.9, 95%CI=1.9-4.3 respectively) when compared to relatives
of a control. Additionally, when MGUS cases were stratified by Ig type, relatives of
those with the IgG/IgA subtype were at a higher risk of developing both MGUS and
multiple myeloma than those with IgM subtype, when compared to relatives of the
controls. When the risk of MGUS in relatives was assessed in men and women
separately, women relatives were at a higher risk of developing MGUS than men (RR for
women=3.5, 95%CI=1.5-8.5 RR for men=2.2, 95%CI=0.8-5.9) [13].
Another study by Landgren et al. found a significantly increased risk in multiple
myeloma in relatives of multiple myeloma cases; this risk was even higher when the
cases had been age 65 or older when diagnosed. The risk of developing multiple
myeloma among relatives was three times higher when the family member with multiple
myeloma was a female, and was also higher among female relatives of cases [44].
In a matched case-control study conducted in six provinces in Canada using male
incident cases and their matched controls (age and province), McDuffie et al. found an
increased risk of approximately 40% in developing multiple myeloma when any first
degree relative had the cancer (OR=1.38, 95%CI=1.07-1.78) [45].
In a nested case-control analysis, Kerber et al. found an approximate 40%
increased risk in family members of multiple myeloma cases (SIR=1.36 95% CI=1.17-
1.58). They also found that this risk is only present in first-degree relatives, and the
increased risk is not found in second-degree through fifth-degree relatives [46].
15
To examine whether MGUS risk was increased in first-degree relatives of MGUS
or multiple myeloma cases, Vachon et al. compared MGUS prevalence in relatives to
population-based rates, which were obtained by direct standardization to the 2000 United
States census population. They reported an increased risk in first-degree relatives of
multiple myeloma (RR=2.0, 95%CI=1.4-2.8) and MGUS probands (RR=3.3,
95%CI=2.1-4.8). Vachon et al. also found a higher risk in women relatives of MGUS
patients, although this difference was not significant [47].
In a population based study with 37,838 first-degree relatives of 13,896 patients
with multiple myeloma, Kristinsson et al. reported that first-degree relatives were two
times as likely to develop both MGUS and multiple myeloma (OR=2.1, 95%CI=1.5-3.1
and OR=2.1, 95%CI=1.6-2.9 respectively). Kristinsson et al. also found that first-degree
relatives of multiple myeloma patients were at a slightly higher risk for any solid tumor
(RR=1.1, 95%CI 1.0-1.1) and a 30% increased risk of developing bladder cancer
(RR=1.3, 95%CI 1.0-1.5) [48].
Contradicting the above studies, Ogmundsdottir et al. did not find an increased
risk in developing MGUS among first-degree relatives of multiple myeloma cases,
although they did find an increased risk in developing multiple myeloma (RR=2.33
95%CI=1.12-4.26), and a significantly higher risk for females separately (RR for
females=3.23 95%CI=1.17-7.01, and RR for males=1.64 95%CI=0.44-4.17) [49].
After summarizing the published literature, there appears to be a 2-3 fold
increased risk in developing both MGUS and multiple myeloma in first-degree relatives
16
when a family member has been diagnosed with either MGUS or multiple myeloma, and
female relatives have a higher risk than males.

1.5.2 Candidate Gene Studies and Multiple Myeloma Risk – Summary of Previous
Literature
To our knowledge, there are no epidemiologic cohort studies examining genetic
risk factors, although there is a nested case-control study within a cohort [50], therefore
all data is summarized using the odds ratio. In the following summary of the published
literature, all studies had allelic variants with a minor allele frequency greater than 5%,
with the exception of one study which used a minor allele frequency cutpoint of greater
than 10% [51].
Although there is always a potential for bias when hospital-based controls are
used, we have not excluded hospital-based controls in our literature review, but have
summarized the findings with cautious interpretation. Besides case-control studies with
only population-based controls, we have included case-control studies with both
population and hospital-based controls and case-control studies with hospital-based
controls [52, 53]. We excluded case-control studies with less than 100 cases, as we were
concerned these studies would not have adequate power to detect potential associations.
It should be noted that multiple studies included in this review use the same group
of cases and controls, such as the incident cases of Connecticut women from the
Connecticut tumor registry and random digit dialed Connecticut female controls [54-56],
17
and patient samples from the Medical Research Council VII trial [57-59] to explore
various genetic risk factors.
We have summarized the results of the different case-control studies organized by
relevant etiologic biological pathways.
Xenobiotic Metabolism Pathway:
Xenobiotic metabolism includes phase I (activation) enzymes such as CYP2E1,
mEH, MPO, and phase II (detoxification) enzymes such as GSTT1 and NQO1. Genetic
variations in these enzymes have been associated with benzene poisoning, and there has
been some evidence that benzene exposure is associated with an increased risk of
multiple myeloma [53]. Toxicological effects of exogenous chemicals are dependent on
certain genes. For example, AHR affects transcriptional regulation of CYP1A1, CYP1A2,
CYP1B1, and several other genes, which are involved in the metabolism of chemicals
such as pesticides and solvents. Polymorphisms in these genes that affect the metabolism
and biologic response to exogenous chemicals could be associated with the risk of
multiple myeloma [60].
Gold et al. studied common variants of genes involved in the metabolism of
exogenous chemicals and they found an association with CYP1B1 variation and an
increased risk in multiple myeloma in both Caucasians and African Americans. It is
hypothesized that increased activity of CYP1B1 can lead to an increase of toxic
intermediates in the metabolism of exogenous chemicals [60].
In a case-control study utilizing cases from two hospitals and healthy bone
marrow donors as controls, Lincz et al. showed a statistically significant increased risk in
18
developing myeloma in the high activity mEH genotype, and a suggestive risk in the
NQO1 low activity genotype. They also examined the association between increasing
number of high risk genotypes and the risk of multiple myeloma, and found that a
combined inheritance of high risk polymorphisms in genes coding for GSTT1, mEH, and
NQO1 may increase the risk of developing multiple myeloma (P for trend =0.001). There
was no association with variation in the CYP1A1 gene and multiple myeloma risk [53].
Contradicting the above study which was conducted in a Caucasian population,
Kang et al. found a decreased risk in multiple myeloma in both CYP1A1*1/*2A and
CYP1A1*1/*2B, but no association with GSTM1 or GSTT1 in a Korean population. These
findings led Kang and colleagues to believe there could be ethnic differences in multiple
myeloma susceptibility involving these genes [61].
In a separate study conducted in an Italian population, Maggini et al. found no
association between NQO1 or GSTP1 polymorphisms and multiple myeloma risk [62].
In a hospital based case-control study with healthy blood donors as controls, Ortega et al.
found similar frequencies in GSTM1, GSTT1 and P53 genotypes in cases and controls,
with no association with the risk of multiple myeloma in this population. However, they
did find that an inherited presence of the variant codon 72 P53 allele and the absence of
the GSTM1 detoxification pathway seem to act in disease progression [63].
Folate Pathway:
Although the genes involved in the folate pathway have been hypothesized to play
a role in the etiology of multiple myeloma, there have been inconsistent results in the
literature.
19
In a hospital based case-control study, Lima et al. found an increased risk in
multiple myeloma in the variant allele of the MTR gene, which affects enzyme activity
and is involved in homocysteine reduction and DNA hypomethylation. They
hypothesized that this variant allele could be involved in the activation of a proto-
oncogene by DNA hypomethylation [52]. Contradicting the above study, Kim et al.
reported that a polymorphism in the MS gene might play a protective role in multiple
myeloma. This specific MS polymorphism is hypothesized to reduce enzyme activity,
thereby increasing homocycteine levels and DNA hypomethylation [64]. None of these
studies have found any significant association between genotypic variation in the
MTHFR gene and the risk of multiple myeloma [52, 64, 65].
Immune response pathway:
Multiple myeloma cells and bone marrow stromal cells interact in the bone
marrow micro-environment, which causes a paracrine cascade of cellular events,
including the upregulation of IL-6, and expression of IGF and stromal-cell-derived-
factor-1α. These genes are involved in sustained tumor growth and tumor cell migration
to the bone marrow microenvironment [56].
Birmann et al. found suggestive associations between variants of the IGF-1,
IGFBP3, IRS1 and IL6R genes in the IGF-1 and IL-6 signaling pathways and an
increased risk of multiple myeloma [50]. Cozen et al. found an approximate 2-fold risk
in the variant allele of an IL-6 promoter polymorphism when compared to both familial
and population-based controls. There was also evidence of decreased risk of multiple
myeloma in one of the VNTR genotypes [42]. Brown et al. found that when compared
20
with controls, multiple myeloma risk was positively associated with two gene variants
that mediate immunity, IL4R (an anti-inflammatory cytokine gene) and FCGR2A (an
immune related gene) in Connecticut women [56]. In a matched case-control study of
patients with both multiple myeloma and MGUS, Zheng et al. found a significant genetic
variation in the IL-10 promoter region when compared to controls, but no difference
between the MGUS and multiple myeloma patients, suggesting IL-10 could play a role in
the development of multiple myeloma [66].
In functional studies, TNF2 has been shown to have higher constitutional and
inducible expression, which has been previously hypothesized to be a risk factor in
multiple myeloma [59]. Morgan et al. found a statistically significant protective effect
with a rare variant in TNF2 [59]. Davies et al. found a significant increased risk of
MGUS and multiple myeloma in those with a high production of TNFalpha/LTA,
suggesting that these genes play a role in the initiation of the disease, not the progression
[67]. Supporting the above results, Brown et al. also found a significant increase in risk
in the haplotype block covering LTA*TNF when compared to the most common
haplotype frequency in the control group [56].
In a matched case-control study involving both MGUS and multiple myeloma
patients, Zheng et al. found that the frequency of the CTLA-4 microsatellite
polymorphism was significantly different in MGUS and multiple myeloma patients when
compared to healthy controls, suggesting that this polymorphism could be a susceptibility
locus for multiple myeloma and MGUS [68].

21
DNA Repair Pathway:
Class switch recombination (CSR) is a DNA deletion-recombination event where
DNA forms double strand breaks and consequently repairs them. In the repair process,
the DNA repair gene XRCC3 is involved in homologous recombination, and DNA repair
genes XRCC4 and XRCC5 are used in non-homologous end joining. Impaired CSR is
believed to be involved in the pathogenesis of multiple myeloma [51]. DNA repair
pathways and mechanisms were explained in detail earlier in this dissertation.
Hayden et al. did not find an association between XRCC3 and multiple myeloma,
although there was a suggestive relationship between a variant in XRCC4 and increased
multiple myeloma risk. When a particular regulatory single nucleotide polymorphism
(SNP) in XRCC5 was analyzed using a recessive model, there was also a statistically
significant association with multiple myeloma [51].
Non-homologous end joining is a pathway utilized to repair double strand breaks
in DNA. Roddam et al. suggest that germline variation within the NHEJ pathway could
be an important factor in the etiology of multiple myeloma [58]. In a European case-
control study, they examined two polymorphisms in LIG4, as they suggest that as a
consequence of LIG4 polymorphisms, functional variation in the NHEJ pathway can
modulate the risk of developing disease. There was a gene-dose response effect in LIG4,
showing increased protection with each variant allele [58].
Cell proliferation, differentiation, and apoptosis pathways:
There are two main apoptotic pathways in humans, the extrinsic (receptor-
mediated) and intrinsic (mitochondrial) pathways. Both use the caspase enzyme cascade,
22
but caspase-8 and 10 are mostly involved with the extrinsic pathway, caspase-9 with the
intrinsic pathway, and both pathways utilize caspase-3, 6, and 7, which lead to cell death
[55].
In a population-based case control study of African American and Caucasian
women in Connecticut, Hosgood found a statistically significant protective effect of
mutations in both the CASP3 and CASP9 genes. Those with the homozygote variant in
CASP9 were ~50% less likely to develop multiple myeloma, and the literature also
suggested a protective effect in those with the homozygote variant in CASP3 [55]. In a
separate case-control study involving only non-Hispanic Caucasian women, Hosgood et
al. found variant alleles in BAX and RIPK1 to be associated with a decreased risk of
multiple myeloma, while a variant allele in CASP9 was associated with an increased risk
of the disease [54].
The NF-kB proteins are structurally related proteins that are involved in normal
cellular processes such as immune inflammatory response, cellular growth, and
apoptosis. NF-kB plays a role in multiple myeloma pathogenesis, as it protects against
apoptosis. Studies have shown enhanced NF-kB activity in multiple myeloma cells [57].
Spink et al. found differences in the NFKBIA gene frequencies between cases and
controls. They suggest that variation in the NF-kB complex not only influences risk of
developing multiple myeloma, but also plays a role in treatment and progression of the
disease. The NF-kb pathway is important in multiple myeloma, as two therapeutic
treatments (bortezomib and thalidomide) target this pathway interfering with activation
and enhancing cellular apoptosis [57].
23

1.5.3 Genome-wide Association Studies (GWAS) and MM – Review of the Literature
Broderick et al. conducted a genome-wide association study of 1,675 cases and
5,903 controls to identify risk variants for multiple myeloma [8]. The study incorporated
two GWAS: a UK-GWAS (1,321 cases, 5,199 controls) and a German-GWAS (354
cases, 704 controls). After quality control measures were employed, a meta-analysis of
422,839 SNPs was performed. Two SNPs reached genome-wide significance (p-
value<5x10
-8
), and a promising association for a third SNP was detected. These SNPs
were genotyped in a UK replication set consisting of 169 cases and 927 controls. The
risk loci identified were as follows: 3p22.1 (rs1052501, OR=1.32; p=7.47x10
-9
), 7p15.3
(rs4487645, OR=1.38; p=3.33x10
-15
), and a marginally significant finding at 2p23.3
(rs6746082, OR=1.29; p=1.22x10
-7
). rs1052501 is a missense SNP (results in an amino
acid change) located in the ULK4 autophagy gene, while the other SNPs are in intronic
regions. This study is the first published GWAS in the literature.

1.6 Overall Summary
This dissertation evaluates genetic factors associated with MM risk by employing
both candidate-pathway and agnostic approaches. It targets specific pathways that have
previously been implicated in disease etiology and pathogenesis, while also utilizing an
agnostic approach to scan the entire genome to identify novel genetic factors associated
with the disease.
24
Chapter one summarizes the biology and epidemiology of multiple myeloma.
This chapter includes an explanation of B-cell development and differentiation into
plasma cells, as MM is a neoplasm of plasma cells. A detailed description of the disease,
DNA repair pathways and their role in B-cell specific mechanisms are described. The
literature review covers MM familial studies, candidate-gene and SNP studies, and the
sole MM published genome wide association study (GWAS).
Chapter two examines genetic variation in 16 SNPs located within the 8q24
region and MM risk in a pooled analysis that includes five studies: The Los Angeles
SEER case-control study, the Seattle/Detroit SEER case-control study, and three nested
case-control studies from the Multiethnic Cohort (MEC), the Nurses’ Health Study
cohort, and the Health Professionals Follow-Up Study cohort.
Chapter three examines functional polymorphisms in DNA repair pathways and
MM risk in a multiethnic Los Angeles case-control study with both relative and
population-based controls as a discovery set, with three additional replication sets from
the USC/Dana Farber Cancer Institute SPORE study, the University of California, San
Francisco and the Mayo clinic.
Chapter four evaluates the contribution of common genetic variation to MM risk
by examining the association of SNPs across the genome and MM risk in whites. Study
data will then be combined in a meta-analysis with a GWAS from the University of
California, San Francisco. This project differs from the previous two, as it employs an
“agnostic” approach in identifying associations.
25
Chapter 1 References
1. Shapiro-Shelef, M. and K. Calame, Regulation of plasma-cell development. Nat
Rev Immunol, 2005. 5(3): p. 230-42.
2. Anderson, K.C. and R.D. Carrasco, Pathogenesis of myeloma. Annu Rev Pathol,
2011. 6: p. 249-74.
3. Klein, U. and R. Dalla-Favera, Germinal centres: role in B-cell physiology and
malignancy. Nat Rev Immunol, 2008. 8(1): p. 22-33.
4. Palumbo, A. and K. Anderson, Multiple myeloma. N Engl J Med, 2011. 364(11):
p. 1046-60.
5. Hideshima, T., et al., Understanding multiple myeloma pathogenesis in the bone
marrow to identify new therapeutic targets. Nat Rev Cancer, 2007. 7(8): p. 585-
98.
6. Tonon, G., Molecular pathogenesis of multiple myeloma. Hematol Oncol Clin
North Am, 2007. 21(6): p. 985-1006, vii.
7. Fonseca, R., et al., International Myeloma Working Group molecular
classification of multiple myeloma: spotlight review. Leukemia, 2009. 23(12): p.
2210-21.
8. Broderick, P., et al., Common variation at 3p22.1 and 7p15.3 influences multiple
myeloma risk. Nat Genet, 2011. 44(1): p. 58-61.
9. Fendly, B.M., et al., Characterization of murine monoclonal antibodies reactive
to either the human epidermal growth factor receptor or HER2/neu gene product.
Cancer Res, 1990. 50(5): p. 1550-8.
10. Bommert, K., R.C. Bargou, and T. Stuhmer, Signalling and survival pathways in
multiple myeloma. Eur J Cancer, 2006. 42(11): p. 1574-80.
11. Gabrea, A., P. Leif Bergsagel, and W. Michael Kuehl, Distinguishing primary
and secondary translocations in multiple myeloma. DNA Repair (Amst), 2006.
5(9-10): p. 1225-33.
12. Kyle, R.A. and S.V. Rajkumar, Multiple myeloma. Blood, 2008. 111(6): p. 2962-
72.
13. Landgren, O., et al., Risk of plasma cell and lymphoproliferative disorders among
14621 first-degree relatives of 4458 patients with monoclonal gammopathy of
undetermined significance in Sweden. Blood, 2009. 114(4): p. 791-5.
26
14. Kyle, R.A. and S.V. Rajkumar, Monoclonal gammopathy of undetermined
significance and smouldering multiple myeloma: emphasis on risk factors for
progression. Br J Haematol, 2007. 139(5): p. 730-43.
15. Landgren, O., et al., Monoclonal gammopathy of undetermined significance
(MGUS) consistently precedes multiple myeloma: a prospective study. Blood,
2009. 113(22): p. 5412-7.
16. Al Olama, A.A., et al., Multiple loci on 8q24 associated with prostate cancer
susceptibility. Nat Genet, 2009. 41(10): p. 1058-60.
17. Ghoussaini, M., et al., Multiple loci with different cancer specificities within the
8q24 gene desert. J Natl Cancer Inst, 2008. 100(13): p. 962-6.
18. Jia, L., et al., Functional enhancers at the gene-poor 8q24 cancer-linked locus.
PLoS Genet, 2009. 5(8): p. e1000597.
19. Dib, A., et al., Characterization of MYC translocations in multiple myeloma cell
lines. J Natl Cancer Inst Monogr, 2008(39): p. 25-31.
20. Goode, E.L., C.M. Ulrich, and J.D. Potter, Polymorphisms in DNA repair genes
and associations with cancer risk. Cancer Epidemiol Biomarkers Prev, 2002.
11(12): p. 1513-30.
21. Haiman, C.A., et al., Comprehensive association testing of common genetic
variation in DNA repair pathway genes in relationship with breast cancer risk in
multiple populations. Hum Mol Genet, 2008. 17(6): p. 825-34.
22. Han, J., et al., Genetic variation in DNA repair pathway genes and
premenopausal breast cancer risk. Breast Cancer Res Treat, 2009. 115(3): p. 613-
22.
23. Kunkel, T.A. and D.A. Erie, DNA mismatch repair. Annu Rev Biochem, 2005.
74: p. 681-710.
24. Burma, S., B.P. Chen, and D.J. Chen, Role of non-homologous end joining
(NHEJ) in maintaining genomic integrity. DNA Repair (Amst), 2006. 5(9-10): p.
1042-8.
25. Gellert, M., V(D)J recombination: RAG proteins, repair factors, and regulation.
Annu Rev Biochem, 2002. 71: p. 101-32.
26. Tibaldi, J.M., et al., Postprandial hypoglycemia in islet beta cell hyperplasia with
adenomatosis of the pancreas. J Surg Oncol, 1992. 50(1): p. 53-7.
27
27. Chahwan, R., et al., Mismatch-mediated error prone repair at the
immunoglobulin genes. Biomed Pharmacother, 2011. 65(8): p. 529-36.
28. Chng, W.J., et al., Genetic events in the pathogenesis of multiple myeloma. Best
Pract Res Clin Haematol, 2007. 20(4): p. 571-96.
29. Ferlay, J., International Agency for Research on Cancer., and International
Association of Cancer Registries., CI5VII electronic database of Cancer
incidence in five continents, vol. VII, in IARC cancerBase no 21997, International
Agency for Research on Cancer,: Lyon, France. p. 2 computer disks.
30. Alexander, D.D., et al., Multiple myeloma: a review of the epidemiologic
literature. Int J Cancer, 2007. 120 Suppl 12: p. 40-61.
31. Landgren, O. and B.M. Weiss, Patterns of monoclonal gammopathy of
undetermined significance and multiple myeloma in various ethnic/racial groups:
support for genetic factors in pathogenesis. Leukemia, 2009. 23(10): p. 1691-7.
32. Gebregziabher, M., et al., Risk patterns of multiple myeloma in Los Angeles
County, 1972-1999 (United States). Cancer Causes Control, 2006. 17(7): p. 931-8.
33. Blattner, W.A., A. Blair, and T.J. Mason, Multiple myeloma in the United States,
1950--1975. Cancer, 1981. 48(11): p. 2547-54.
34. Velez, R., V. Beral, and J. Cuzick, Increasing trends of multiple myeloma
mortality in England and Wales; 1950-79: are the changes real? J Natl Cancer
Inst, 1982. 69(2): p. 387-92.
35. Koessel, S.L., et al., Socioeconomic status and the incidence of multiple myeloma.
Epidemiology, 1996. 7(1): p. 4-8.
36. Renshaw, C., et al., Trends in the incidence and survival of multiple myeloma in
South East England 1985-2004. BMC Cancer. 10: p. 74.
37. Kristinsson, S.Y., et al., Socioeconomic differences in patient survival are
increasing for acute myeloid leukemia and multiple myeloma in sweden. J Clin
Oncol, 2009. 27(12): p. 2073-80.
38. Baris, D., et al., Socioeconomic status and multiple myeloma among US blacks
and whites. Am J Public Health, 2000. 90(8): p. 1277-81.
39. Larsson, S.C. and A. Wolk, Body mass index and risk of multiple myeloma: a
meta-analysis. Int J Cancer, 2007. 121(11): p. 2512-6.
40. Wallin, A. and S.C. Larsson, Body mass index and risk of multiple myeloma: a
meta-analysis of prospective studies. Eur J Cancer, 2011. 47(11): p. 1606-15.
28
41. Landgren, O., et al., Obesity is associated with an increased risk of monoclonal
gammopathy of undetermined significance among black and white women. Blood,
2010. 116(7): p. 1056-9.
42. Cozen, W., et al., Interleukin-6-related genotypes, body mass index, and risk of
multiple myeloma and plasmacytoma. Cancer Epidemiol Biomarkers Prev, 2006.
15(11): p. 2285-91.
43. Lichtman, M.A., Obesity and the risk for a hematological malignancy: leukemia,
lymphoma, or myeloma. Oncologist, 2010. 15(10): p. 1083-101.
44. Landgren, O., et al., Familial characteristics of autoimmune and hematologic
disorders in 8,406 multiple myeloma patients: a population-based case-control
study. Int J Cancer, 2006. 118(12): p. 3095-8.
45. McDuffie, H.H., et al., Clustering of cancer among families of cases with
Hodgkin Lymphoma (HL), Multiple Myeloma (MM), Non-Hodgkin's Lymphoma
(NHL), Soft Tissue Sarcoma (STS) and control subjects. BMC Cancer, 2009. 9: p.
70.
46. Kerber, R.A. and E. O'Brien, A cohort study of cancer risk in relation to family
histories of cancer in the Utah population database. Cancer, 2005. 103(9): p.
1906-15.
47. Vachon, C.M., et al., Increased risk of monoclonal gammopathy in first-degree
relatives of patients with multiple myeloma or monoclonal gammopathy of
undetermined significance. Blood, 2009. 114(4): p. 785-90.
48. Kristinsson, S.Y., et al., Patterns of hematologic malignancies and solid tumors
among 37,838 first-degree relatives of 13,896 patients with multiple myeloma in
Sweden. Int J Cancer, 2009. 125(9): p. 2147-50.
49. Ogmundsdottir, H.M., et al., Familiality of benign and malignant
paraproteinemias. A population-based cancer-registry study of multiple myeloma
families. Haematologica, 2005. 90(1): p. 66-71.
50. Birmann, B.M., et al., Insulin-like growth factor-1- and interleukin-6-related gene
variation and risk of multiple myeloma. Cancer Epidemiol Biomarkers Prev,
2009. 18(1): p. 282-8.
51. Hayden, P.J., et al., Variation in DNA repair genes XRCC3, XRCC4, XRCC5 and
susceptibility to myeloma. Hum Mol Genet, 2007. 16(24): p. 3117-27.
52. Lima, C.S., et al., Polymorphisms of methylenetetrahydrofolate reductase
(MTHFR), methionine synthase (MTR), methionine synthase reductase (MTRR),
29
and thymidylate synthase (TYMS) in multiple myeloma risk. Leuk Res, 2008.
32(3): p. 401-5.
53. Lincz, L.F., et al., Genetic variations in benzene metabolism and susceptibility to
multiple myeloma. Leuk Res, 2007. 31(6): p. 759-63.
54. Hosgood, H.D., 3rd, et al., Genetic variation in cell cycle and apoptosis related
genes and multiple myeloma risk. Leuk Res, 2009. 33(12): p. 1609-14.
55. Hosgood, H.D., 3rd, et al., Caspase polymorphisms and genetic susceptibility to
multiple myeloma. Hematol Oncol, 2008. 26(3): p. 148-51.
56. Brown, E.E., et al., Common variants in genes that mediate immunity and risk of
multiple myeloma. Int J Cancer, 2007. 120(12): p. 2715-22.
57. Spink, C.F., et al., Haplotypic structure across the I kappa B alpha gene
(NFKBIA) and association with multiple myeloma. Cancer Lett, 2007. 246(1-2):
p. 92-9.
58. Roddam, P.L., et al., Genetic variants of NHEJ DNA ligase IV can affect the risk
of developing multiple myeloma, a tumour characterised by aberrant class switch
recombination. J Med Genet, 2002. 39(12): p. 900-5.
59. Morgan, G.J., et al., Haplotypes in the tumour necrosis factor region and
myeloma. Br J Haematol, 2005. 129(3): p. 358-65.
60. Gold, L.S., et al., Associations of common variants in genes involved in
metabolism and response to exogenous chemicals with risk of multiple myeloma.
Cancer Epidemiol, 2009. 33(3-4): p. 276-80.
61. Kang, S.H., et al., Protective role of CYP1A1*2A in the development of multiple
myeloma. Acta Haematol, 2008. 119(1): p. 60-4.
62. Maggini, V., et al., Lack of association of NQO1 and GSTP1 polymorphisms with
multiple myeloma risk. Leuk Res, 2008. 32(6): p. 988-90.
63. Ortega, M., et al.,, GSTM1 and codon 72 P53 polymorphism in multiple myeloma.
Ann Hematol, 2007(86): p. 815-819.
64. Kim, H.N., et al., Polymorphisms involved in the folate metabolizing pathway and
risk of multiple myeloma. Am J Hematol, 2007. 82(9): p. 798-801.
65. Gonzalez-Fraile, M.I., et al., Methylenetetrahydrofolate reductase genotype does
not play a role in multiple myeloma pathogenesis. Br J Haematol, 2002. 117(4): p.
890-2.
30
66. Zheng, C., et al., Interleukin-10 gene promoter polymorphisms in multiple
myeloma. Int J Cancer, 2001. 95(3): p. 184-8.
67. Davies, F.E., et al., High-producer haplotypes of tumor necrosis factor alpha and
lymphotoxin alpha are associated with an increased risk of myeloma and have an
improved progression-free survival after treatment. J Clin Oncol, 2000. 18(15): p.
2843-51.
68. Zheng, C., et al., Cytotoxic T-lymphocyte antigen-4 microsatellite polymorphism
is associated with multiple myeloma. Br J Haematol, 2001. 112(1): p. 216-8.

31
Chapter 2: Genotypic Variation in the 8q24 Region and the Risk of
Multiple Myeloma

2.1. Introduction and Rationale
As described in the previous chapter, the 8q24 region has been identified as a risk
region for many solid tumors, including prostate cancer. The risk patterns of multiple
myeloma are similar to those of prostate cancer; African Americans have a 2-fold
increased risk in developing the disease and men are at a higher risk than women. This
region is also of interest because of the proximity of the oncogene MYC, which lies
~200kb telomeric to the 8q24 region. Translocations involving MYC are often observed
in MM, and have been associated with disease progression [1].
Given that the 8q24 region has exhibited strong associations with an increased
risk of prostate cancer [2], a disease with risk patterns similar to MM, and the role MYC
plays in the progression of MM, we hypothesized that genetic variation within the 8q24
region could also be associated with MM risk.
This manuscript examines the associations between 16 SNPs within the 8q24
region (previously reported to be associated with cancer risk) and MM risk in a pooled
analysis of two case-control and three nested case-control studies. The analyses were
performed in Whites and African Americans, separately. Although modest, the observed
associations suggest that genetic variation in the region may be associated with MM risk
in both Whites and African-Americans. These findings are consistent with a previous
32
European study (in Whites only), and if confirmed in other large populations, may imply
a role for the 8q24 region in the etiology of MM.

2.2 Abstract
Background: The 8q24 region 200 kb upstream of MYC has been associated with
prostate, breast, and colon cancer risk and was recently associated with multiple myeloma
(MM) risk in European whites. MYC plays a role in MM pathogenesis and progression.
Methods: We genotyped 16 SNPs in the 8q24 region previously associated with other
cancer risk in DNA samples from participants in two population-based and three nested
case-control studies of MM using TaqMan. A pooled analysis was performed to evaluate
the effect of the risk alleles on MM susceptibility among non-Hispanic whites (259 cases,
567 controls) and African Americans (114 cases, 186 controls) separately using
unconditional logistic regression adjusting for sex, age, and study site. Controls were
augmented with 1,137 white controls from Cancer Genetics Markers of Susceptibility
(CGEMS) data and 2,187 unaffected African Americans from the Multiethnic Cohort.
Results: With additional controls included, we observed modest associations between
two SNPs and MM risk among whites only (rs10086908: Odds Ratio (OR)=0.78, 95%
Confidence Interval (CI)=0.61-1.00; rs6983267: OR=0.81, 95%CI=0.67-0.99); the latter
similar in magnitude to that reported previously for MM. Among African Americans,
rs116041037 was associated with a 2-fold increased risk (OR=2.02, 95%CI=1.02-3.97).
Results were similar without additional controls, although CIs were wider.
33
Conclusions: In this hypothesis-generating study we found suggestive evidence of an
association between the 8q24 region and MM risk. Replication is needed to validate the
findings and explain discrepancies across race/ethnicity.
Impact: Genetic variation in the 8q24 region may be involved in MM etiology; further
studies are warranted

2.3 Introduction
Multiple myeloma (MM) is a rare and incurable B-cell malignancy consisting of
neoplastic plasma cells in the bone marrow, accounting for approximately 1% of all
cancers and 13% of hematologic cancers (www.cancer.org). The annual age-adjusted
incidence of MM is 5.7 cases/100,000 person-years [3], with age-specific incidence rates
increasing to ~35/100,00 person-years for those 80 years and older [4]. The 2- to 3-fold
higher risk among African Americans compared to whites, and a similar increased risk
among relatives of cases, suggests a heritable etiologic component [5]. Various candidate
single nucleotide polymorphisms (SNP) have been linked to MM risk, but most studies
were small and results have not been validated or are conflicting [6].
Although the function is unknown for most SNPs within the intergenic 8q24
region, rs6983267 is located within an enhancer shown to exert an effect on MYC through
long-range physical interaction [7-9]. This region 200 kb upstream of MYC is comprised
of distinct linkage disequilibrium (LD) blocks that have been associated with the risk of
solid tumors including prostate, breast, and colon cancer [2, 10-14]. In MM, MYC
translocations are late stage events, consistent with an increase in the proliferative
34
abilities of the neoplastic plasma cell and a decreased dependence on the stromal cell for
survival [15]. MYC rearrangements are rare in the precursor condition monoclonal
gammopathy of undetermined significance (MGUS) or smoldering MM, but occur in
15% of MM tumors, 45% of advanced MM tumors, and 90% of human myeloma cell
lines [15, 16]. These molecular events are associated with pathogenesis and progression,
and may have a role in susceptibility as well. In a relatively large study of MM (424
cases), Tewari et al. recently reported an association between MM risk and two SNPs in
the 8q24 region (rs16901979 and rs6983267) in a European population [17]. To further
investigate whether variation in the 8q24 region is involved in disease etiology, we
genotyped SNPs in participants from five study sites and conducted a pooled analysis to
examine the association of germline variation in the 8q24 region and the risk of MM.

2.4 Study Design and Methods
This study was approved by the institutional review boards at the University of
Southern California, Fred Hutchinson Cancer Research Center, Wayne State University,
Brigham and Women’s Hospital, and Harvard University, and signed informed consent
was obtained from all participants.
DNA samples from 485 MM cases and 938 controls were obtained from two
population-based case-control studies and three nested case-control studies. This pooled
analysis includes the subset of non-Hispanic whites, hereafter referred to as whites (259
cases and 567 population-based controls), and African Americans (114 cases and 186
35
population-based controls) with sufficient DNA for analysis (Table 2.1). A description
follows of each study site contributing samples:
MM cases diagnosed from 1999-2002 at 20-74 years of age were identified using
the Los Angeles, Seattle and Detroit Surveillance, Epidemiology, and End Results
(SEER) population-based cancer registries. These registries participated in a multi-center
study of non-Hodgkin lymphoma

and MM in collaboration with the National Cancer
Institute, and at each site, controls were identified by random-digit dialing (< 65 years of
age) or through Health Care Financing Administration rosters (65 years of age and older);
controls were frequency-matched to the non-Hodgkin lymphoma cases on age, race, and
sex [18]. The Los Angeles site is presented as a separate study from Seattle/Detroit [19].
The Multiethnic Cohort (MEC) is comprised of over 215,000 men and women
recruited from Hawaii and the Los Angeles area between 1993 and 1996. Participants are
between the ages of 45 and 75 at baseline, are primarily of Native Hawaiian, Japanese,
white, African American, or Latino ancestry, and completed a detailed questionnaire at
entry that assessed demographic information and lifestyle factors, including diet and
medical conditions [20]. Cases were ascertained by linking the MEC database to the
Hawaii and California SEER registries. DNA was extracted from blood or saliva samples
collected from 21 white and 51 African American MM cases; cases were matched with
two unaffected cohort participants with available DNA samples on sex, ethnicity, age at
cohort entry, and education (<=12 years, 13-16 years, and 17+ years).
Two additional nested case-control studies included in the pooled analysis are
from the Health Professionals Follow-up Study (HPFS) and the Nurses’ Health Study
36
(NHS) cohorts. The HPFS includes participants, identified from 18,018 United States
male health professionals, who returned an initial questionnaire in 1986 and who
provided blood samples between 1993 and 1994. DNA samples were available for 24
MM cases; these cases were matched to 48 unaffected controls on age at cohort entry,
sex, and race. The NHS cohort participants were identified from 32,826 nurses who
joined the cohort in 1976 and provided blood samples from 1989-1990 or buccal cell
samples from 2001-2004. DNA specimens were available for 58 cases (50 blood, 8
buccal); 116 controls (100 blood, 16 buccal) from the NHS were matched to MM cases
on age at cohort entry, race, sex, and DNA type. The two cohorts, described in detail
elsewhere [21, 22], were developed and are maintained at the Harvard School of Public
Health or Brigham and Women’s Hospital, respectively, and as in previous studies [23],
are combined for purposes of this analysis.

2.4.1 SNP Selection and Genotyping
Based on risk associations published before the start of this study [24, 25], 16
SNPs with pairwise r
2
< 0.4 (1000 Genomes June 2011 Data Release) were selected in
the 8q24 region (Table 2.2, Tables 2.5a and 2.5b). DNA was isolated from blood and
buccal samples using standard methods, and genotyping was performed on all 16 SNPs
using the fluorogenic 5’-nuclease assay (TaqMan). Probes were generated and labeled
with fluorogenic dyes (6-FAM or VIC) to determine each allele. The fluorescence profile
was measured using an ABI 7900HT Sequence Detection System and results were
analyzed using Sequence Detection Software (Applied Biosystems Inc.). Further
37
information regarding primers and probes is available upon request. The concordance
rate of the duplicate QC samples was 99.8% and the average SNP call rate was above
95%. Due to limited amounts of DNA in the HPFS and NHS cohorts, only seven of the
16 SNPs, which were originally reported to represent multiple risk regions on 8q24 [2],
were genotyped in those populations.

2.4.2 Statistical Analysis
Data from all studies were combined for analyses. We examined departure from
Hardy-Weinberg equilibrium (HWE) among controls, stratifying by race (whites and
African Americans).
To assess the effect of the alleles on MM risk, unconditional logistic regression
was used to calculate odds ratios (OR) and 95% confidence intervals (CI) assuming a log-
additive inheritance model and adjusting for age as a continuous variable, sex, and study
site. For each SNP, the allele previously associated with cancer risk was tested as the
variant allele. A likelihood ratio test for heterogeneity was performed across studies. We
did not adjust for multiple comparisons because these associations are hypothesis-
generating in the context of MM (note that an association between two of the SNPs we
tested and MM risk has subsequently been published since our study began).
To enhance statistical power, we performed a second analysis that included an
additional 2,187 African American controls from the MEC [26-28] and 1,137 white
controls from Cancer Genetics Markers of Susceptibility (CGEMS) data [29] combined
with the original control groups using unconditional logistic regression adjusting for sex.
38
In the MEC and CGEMS controls, rs13254738, rs6983561, rs7000448, and rs10090154
were imputed; rs13281615 and rs620861 were also imputed in the CGEMS controls. The
SNPs were imputed using MACH for MEC controls and using the IMPUTE2 software
application for CGEMS controls [30], referencing HapMap Phase 2 YRI and CEU
populations (www.hapmap.org) respectively. Prior to imputation for CGEMS controls,
SNP-level genotype concordance was checked between Illumina-genotyped HapMap
samples on the HumanHap550_v1 chip and the HapMap 2 database. SNPs having a
concordance <95% in any HapMap population or a call rate <90% were removed. In
addition, SNPs in CGEMS controls with a HWE p<10
-4
were removed. Prior to
imputation of MEC controls, only typed SNPs with a call rate > 95% and concordance
across known duplicate samples > 98% were retained for imputation. Statistical analyses
were performed using Stata release 11 (College Station, TX: StataCorp LP) and R version
2.12.1.

2.5 Results
All SNPs were in HWE among white and African American controls across
studies (p>0.05). Sex, race, and age at diagnosis are shown separately for cases and
controls by site (Table 2.1). The MEC contributes roughly half of the African American
participants; the mean age at diagnosis is older in this cohort than in the other
contributing studies. The greater number of females in the combined NHS and HPFS
study participants is consistent with the composition and size of the cohort populations.
The test for heterogeneity across studies was statistically significant for three SNPs,
39
rs13252298 (p=0.02), rs6983267 (p=0.004) and rs10090154 (p=0.04) in whites, but not
for any in African Americans (Tables 2.6a and 2.6b).
The allele frequencies of the 16 SNPs were consistent across control groups in
each study and similar to published estimates (Table 2.3) [14, 25, 31]. Six of the 16
SNPs showed similar allele frequencies among white and African American controls
(Table 2.3). The minor allele frequency (MAF) observed in the CGEMS and MEC
controls for a given SNP was similar to the corresponding MAF observed in the pooled
study population.
Several non-significant race-specific associations were observed in the initial
multicenter analysis (Table 2.4). Among whites, rs10086908 and rs6983267 were non-
significantly inversely associated with MM risk (OR=0.81, 95% CI=0.62-1.06 and
OR=0.83, 95% CI=0.67-1.03, respectively). These associations became slightly stronger
and more significant when cases were compared to the larger combined control group
(rs10086908: OR=0.78, 95% CI=0.61-1.00; rs6983267: OR=0.81, 95% CI=0.67-0.99)
(Table 2.4). We also observed a non-significant association between rs6983561 and MM
risk that was attenuated when additional controls were added.
The modest associations observed among whites were not evident among African
Americans (Table 2.4). However, we observed a statistically significant two-fold
increased risk for MM associated with the rare rs116041037 A-allele among African
Americans (OR=2.02, 95% CI=1.02-3.97, p=0.04) in the analysis supplemented with
additional controls. There were no associations between any other SNPs and MM in
African Americans.
40
2.6 Discussion
We conducted a large study of genetic risk factors for MM, and were the first to
examine 8q24 region genetic risk factors for MM in African Americans.
We observed a modest association for rs6983267 among whites that was similar
in magnitude to that previously reported in a study of European whites that included 424
cases and 429 controls, and auxiliary control data from 1,000 Irish blood donors [17].
Unlike studies of breast, colorectal and prostate cancer [2, 11, 14, 32], we did not find a
positive association between the putatively functional rs6983267 and MM risk. In fact,
we observed an inverse association among U.S. whites, comparable to that reported
among European whites by Tewari and colleagues [17]. A post-hoc random-effects
meta-analysis combining the estimates from our study and the study conducted by Tewari
produced a statistically significant estimate of decreased MM risk (OR=0.81, 95%
CI=0.71-0.91, p=0.001). We also found a modest inverse association with the variant
allele of rs10086908, which was not genotyped in the European study [17].
We examined rs6983561, a SNP in perfect LD with rs16901979 (r
2
=1.0, D’=1.0,
1000 Genomes June 2011 Data Release), shown to be associated with MM risk in the
European study. Unlike the previously published results for rs16901979, the risk
estimate for rs6983561 in the present study was not statistically significant and the
magnitude of the effect was attenuated when additional controls added. To our
knowledge, the European study is the only other evaluation of the 8q24 region and MM
susceptibility.
41
We found no associations between these SNPs and MM among African
Americans. Because the allele frequency of rs10086908 was similar among white and
African American controls, the borderline statistically significant finding seen in whites
may be due to chance. However, we did observe a positive association between the
variant allele of rs116041037 and MM risk exclusively among African Americans; this
SNP is rare (MAF =0.02) in African Americans and non-existent in whites. The two-fold
increased MM risk associated with this SNP is consistent with that observed for prostate
cancer [2, 33]. As these SNPs were discovered in a multiethnic population [2], we also
performed an analysis combining whites and African Americans; there were no
statistically significant findings, and the Breslow-Day test for heterogeneity for all SNPs
was not statistically significant.
Two 8q24 SNPs have been previously associated with other hematologic
malignancies. rs2456449 was linked to chronic lymphocytic leukemia (CLL) risk [34];
we observed no effect for MM. Another SNP in the 8q24 region, rs2019960, was
associated with classic Hodgkin lymphoma [35]. We did not genotype this SNP, nor was
it in LD with any that we assayed (r
2
>=0.80, 1000 Genomes June 2011 Data Release).
rs6983267 has been positively associated with both prostate and colorectal cancer
risk across different ethnic groups [2, 26, 36], in contrast to the inverse association with
MM risk observed here and by Tewari et al. [17]. Although rs6983267 is not directly
associated with differential MYC expression [37], it physically interacts with MYC in
normal and colon, breast, and prostate tumor tissue [38]. The relationship between
rs6983267 and MYC has not been examined in hematopoietic cells. It is possible that the
42
interaction of rs6983267 with MYC is tissue specific, or that the effect of this SNP is not
always mediated through MYC and instead may target other genes.
The study has several noteworthy strengths. In spite of the challenge of amassing
sufficient numbers, we conducted one of the largest genetic studies to date by pooling
samples from multiple institutions and adding additional controls. Furthermore, this is the
first study, to our knowledge, to examine 8q24 variation and risk of MM in African
Americans, a population with a well-documented excess risk of MM compared to whites
that is not understood.
This study also has several limitations. Statistical power to detect modest
associations was limited by the sample size for this rare cancer (incidence rate=
5.7/100,000 person-years and 5-year survival rate = 40%)
(http://seer.cancer.gov/statfacts/html/mulmy.html). We observed heterogeneity for three
SNPs where two would be expected to differ across the studies by chance. The observed
heterogeneity could be due to small sample sizes in the individual studies. Finally, we
genotyped 16 SNPs across a ~3,300 kb region and thus may not have captured all
possible risk alleles.
In conclusion, the present findings are supportive of a previously published
association of 8q24 region SNPs with MM susceptibility in whites, and suggestive of a
role for 8q24 genetic variation in MM etiology in African Americans as well. If
confirmed in other large populations, these findings may imply a role for alterations in
MYC or related signaling pathways in the etiology of MM, which is plausible given its
known role in MM pathogenesis. Further elucidation of the role of 8q24 variation and of
43
MYC in MM etiology is warranted in light of the paucity of knowledge on the etiology of
MM and the importance of developing strategies to prevent this incurable malignancy.
Acknowledgements

The authors would like to acknowledge the thoughtful input of Gerhard Coetzee,
the analytic support of Won Lee, the data management assistance of Catherine Suppan,
and the patients and other subjects who participated in this study. In addition, we would
like to thank the participants and staff of the NHS and HPFS for their valuable
contributions, as well as the following state cancer registries for their help: AL, AZ, AR,
CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY,
NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY.
Grant Support
This participating cohorts and work was supported by grants from the National
Institutes of Health 2P50CA100707-06 (KCA, NM, WC, DVC, KAR, CAH, DJVDB,
BMB, AJD, RKS, ML, LNK, AM); P30CA014089 (WC, DJDVB, CKE);
1R01CA134786-02 (WC, DVC, CAH, DJVDB, LNK, SKA); R37 CA54281 and P01
CA33619 (CAH, LNK, BEH); K05CA136967 (LB); R01 CA127435 (GAC, BMB); and K07
CA115687 (BMB). KCA and GAC are American Cancer Society Clinical Professors.
This project has been funded in whole or in part with Federal funds from the National
Cancer Institute SEER Population-based Registry Program, National Institutes of Health,
Department of Health and Human Services, under Contract No. N01-PC-35139, N01-PC-
67010-SOW 16 (WC, AM, LB), 5-P01CA033619-22 (LNK), N01-PC-67009 (AJD), and
N01-PC-65064 (RKS). The NHS was supported by P01 CA87969 and R01 CA49449,
44
and the HPFS was supported by P01 CA055075. The MEC was supported by National
Cancer Institute grants R37CA54281, R01CA63464, P01CA33619, U01CA136792, and
U01CA98758. The collection of incident multiple myeloma patients used in this
publication was supported by the California Department of Health Services as part of the
statewide cancer reporting program mandated by California Health and Safety Code
Section 103885. The ideas and opinions expressed herein are those of the authors, and no
endorsement by the State of California, Department of Health Services is intended or
should be inferred. This publication was made possible by grant number 1U58DP000807-
01 from the Centers for Disease Control and Prevention. Its contents are solely the
responsibility of the authors and do not necessarily represent the official views of the
federal government.

45

Table 2.1. Demographic Characteristics of White and African American Participants in the Studies Contributing to the
Pooled Analysis.
Los Angeles SEER Seattle/Detroit SEER MEC NHS and HPFS Total

Cases
n (%)
Controls
n (%)
Cases
n (%)
Controls
n (%)
Cases
n (%)
Controls
n (%)
Cases
n (%)
Controls
n (%)
Cases
n (%)
Controls
n (%)
Sex
Male 64 (60) 82 (57) 75 (60) 179 (56) 40 (56) 80 (56) 23 (33) 47 (32) 202 (54) 388 (52)
Female 43 (40) 62 (43) 49 (40) 141 (44) 32 (44) 64 (44) 47 (67) 98 (68) 171 (46) 365 (48)
Total 107 144 124 320 72 144 70 145 373 753
Race
White 71 (66) 93 (65) 98 (79) 288 (90) 21 (29) 42 (29) 69 (99) 144 (99) 259 (69) 567 (75)
AA 36 (34) 51 (35) 26 (21) 32 (10) 51 (71) 102 (71) 1 (1) 1 (1) 114 (31) 186 (25)
Total 107 144 124 320 72 144 70 145 373 753
Age
a

Mean age at
diagnosis (SD) 61 (10) 58 (12) 60 (9) 57 (12) 71(8) 71 (9) 66 (8) 66 (8)

64 (10)

61 (12)
Median age at
diagnosis 61 59 61 60 72 72 65 65

65

64
a
For controls, mean and median reference age was calculated

46

Table 2.2. Published Associations: 8q24 SNPs and Increased Cancer Risk.
SNP Position
a
Cancer Type
rs7008482 126336812 CRC
b [39]

rs12543663 127993841 Prostate
[24]

rs10086908 128081119 Prostate
[40]

rs1016343 128162479 Prostate
[24]

rs13252298 128164338 Prostate
[24]

rs13254738 128173525 Prostate
[2]

rs6983561 128176062 Prostate
[2]

rs116041037 128200991 Prostate
[2]

rs2456449 128262163 CLL
b [34]

rs620861 128404855 Prostate
[41]

rs13281615 128424800 Breast
[11]

rs6983267 128482487 Prostate, CRC, Ovary
[26, 42]

rs7000448 128510352 Prostate
[2]

rs10090154 128601319 Prostate
[2]

rs9642880 128787250 Bladder
[31]

rs10088218 129613131 Ovary
[43]

a
NCBI Build 36,
b
CRC – colorectal cancer; CLL – chronic lymphocytic leukemia,

47

Table 2.3. Allele Frequencies of Selected 8q24 SNPs among Controls.
SNP
Risk
Allele
Multicenter Study
White Controls
(N=567)
CGEMS Breast
Controls
(N=1137)
Multicenter Study
AA Controls
(N=186)
AA MEC
Controls
(N=2187)
rs7008482 G 0.32 0.31 0.79 0.80
rs12543663 C 0.31 0.30 0.14 0.15
rs10086908 T 0.70 0.69 0.73 0.75
rs1016343 T 0.19 0.22 0.19 0.22
rs13252298 A 0.69 0.72 0.92 0.92
rs13254738 C 0.31 0.34 0.61 0.58
rs6983561 C 0.03 0.03 0.43 0.41
rs116041037 A - - 0.02 0.02
rs2456449 G 0.35 0.36 0.22 0.20
rs620861 G 0.65 0.64 0.64 0.65
rs13281615 G 0.41 0.40 0.44 0.43
rs6983267 G 0.55 0.53 0.88 0.85
rs7000448 T 0.39 0.38 0.64 0.61
rs10090154 T 0.10 0.10 0.16 0.12
rs9642880 T 0.47 0.44 0.72 0.73
rs10088218 A 0.13 0.13 0.12 0.14

48

Table 2.4. The Association between 8q24 Genotypes and Risk of MM from a Pooled Analysis in Whites and African
Americans.

Multicenter Analysis –whites
b

ca/co=(259/567)
Combined with
CGEM Controls
c

ca/co=(259/1704)
Multicenter Analysis –AAs
b

ca/co=(114/186)
Combined with
MEC Controls
c

ca/co=(114/2373)
SNP RA
a
OR 95% CI P-value OR 95% CI P-value OR 95% CI P-value OR 95% CI P-value
rs7008482 G
0.90 (0.71-1.14) 0.37 0.93 (0.75-1.15) 0.51
1.02 (0.66-1.56) 0.93 1.06 (0.76-1.48) 0.75
rs12543663 C
0.87 (0.66-1.16) 0.34 0.94 (0.72-1.22) 0.64
1.08 (0.66-1.78) 0.75 0.99 (0.67-1.47) 0.96
rs10086908 T
0.81 (0.62-1.06) 0.13 0.78 (0.61-1.00) 0.05
1.18 (0.78-1.78) 0.44 1.08 (0.78-1.49) 0.64
rs1016343 T
1.08 (0.79-1.46) 0.64 1.00 (0.75-1.33) 0.99
1.05 (0.69-1.61) 0.81 0.94 (0.67-1.31) 0.72
rs13252298 A
0.99 (0.75-1.30) 0.94 0.98 (0.76-1.26) 0.88
1.00 (0.52-1.92) 0.99 1.15 (0.68-1.96) 0.60
rs13254738 C
1.12 (0.89-1.41) 0.34 1.07 (0.86-1.32) 0.55
0.91 (0.62-1.33) 0.62 1.11 (0.82-1.49) 0.51
rs6983561 C
1.33 (0.73-2.44) 0.35 1.29 (0.74-2.24) 0.37
0.80 (0.55-1.17) 0.25 0.87 (0.64-1.18) 0.37
rs116041037 A
- - - - - -
2.43 (0.85-6.96) 0.10 2.02 (1.02-3.97) 0.04
rs2456449 G
0.90 (0.69-1.17) 0.43 0.92 (0.72-1.18) 0.50
0.99 (0.63-1.57) 0.98 1.08 (0.76-1.55) 0.66
rs620861 G
0.98 (0.75-1.28) 0.88 1.01 (0.79-1.29) 0.92
1.12 (0.78-1.60) 0.53 1.09 (0.81-1.47) 0.58
rs13281615 G
0.95 (0.77-1.18) 0.65 1.00 (0.82-1.22) 0.99
0.70 (0.48-1.02) 0.06 0.82 (0.62-1.09) 0.17
rs6983267 G
0.83 (0.67-1.03) 0.09 0.81 (0.67-0.99) 0.04
0.90 (0.52-1.58) 0.72 1.05 (0.69-1.58) 0.83
rs7000448 T
0.94 (0.75-1.17) 0.59 0.95 (0.78-1.16) 0.63
1.05 (0.71-1.55) 0.80 1.2 (0.88-1.63) 0.26
rs10090154 T
0.98 (0.69-1.41) 0.92 1.02 (0.73-1.41) 0.93
0.94 (0.55-1.61) 0.82 1.34 (0.90-1.99) 0.15
rs9642880 T
0.92 (0.71-1.19) 0.52 0.98 (0.77-1.24) 0.84
0.88 (0.61-1.27) 0.48 0.81 (0.60-1.09) 0.16
rs10088218 A
0.99 (0.67-1.45) 0.95 0.95 (0.67-1.36) 0.79
1.43 (0.87-2.35) 0.16 1.23 (0.86-1.77) 0.26
a
RA=risk allele,
b
This analysis was adjusted for age (as a continuous variable), sex, and site. The NHS and HPFS were not included in the analysis of the following
SNPs: rs12543663, rs10086908, rs1016343, rs13252298, rs2456449, rs620861, rs9642880, rs10088218.
c
The combined analysis was adjusted for sex, and only 919
additional MEC controls were included in the analysis for rs116041037.

49

Table2.5a. Correlation between SNPs in Individuals of European Ancestry (1000 Genomes June 2011 data release).

rs10086908 rs10088218 rs10090154 rs1016343 rs12543663 rs13252298 rs13254738 rs13281615
rs10086908 1 0.066 0.007 0.019 0.004 0.002 0.003 0.002
rs10088218

1 0.004 0.033 0.013 0.033 0.025 0.003
rs10090154

1 0.000 0.037 0.001 0.007 0.000
rs1016343

1 0.024 0.139 0.258 0.001
rs12543663

1 0.000 0.011 0.001
rs13252298

1 0.159 0.000
rs13254738

1 0.000
rs13281615

1
rs2456449

rs620861

rs6983267

rs6983561

rs7000448

rs7008482

rs9642880

50

Table2.5a. Continued

rs2456449 rs620861 rs6983267 rs6983561 rs7000448 rs7008482 rs9642880
rs10086908 0.019 0.015 0.037 0.001 0.042 0.003 0.009
rs10088218 0.049 0.001 0.001 0.000 0.005 0.047 0.002
rs10090154 0.001 0.001 0.042 0.000 0.017 0.008 0.001
rs1016343 0.020 0.002 0.014 0.116 0.047 0.002 0.002
rs12543663 0.033 0.000 0.000 0.015 0.034 0.000 0.004
rs13252298 0.029 0.011 0.01 0.016 0.006 0.018 0.000
rs13254738 0.028 0.016 0.019 0.018 0.002 0.003 0.001
rs13281615 0.005 0.375 0.01 0.032 0.057 0.002 0.005
rs2456449 1 0.001 0.005 0.018 0.001 0.000 0.042
rs620861

1 0.033 0.003 0.076 0.002 0.004
rs6983267

1 0.006 0.197 0.002 0.010
rs6983561

1 0.025 0.000 0.000
rs7000448

1 0.000 0.002
rs7008482

1 0.004
rs9642880

1
51

Table 2.5b. Correlation between SNPs in Individuals of African Ancestry (1000 Genomes June 2011 data release).

rs10086908 rs10088218 rs10090154 rs1016343 rs116041037 rs12543663 rs13252298 rs13254738 rs13281615
rs10086908 1 0.006 0.000 0.016 0.007 0.299 0.002 0.003 0.001
rs10088218

1 0.001 0.006 0.012 0.001 0.011 0.007 0.075
rs10090154

1 0.001 0.049 0.007 0.003 0.001 0.000
rs1016343

1 0.006 0.001 0.003 0.000 0.014
rs116041037

1 0.004 0.000 0.011 0.007
rs12543663

1 0.001 0.000 0.002
rs13252298

1 0.004 0.010
rs13254738

1 0.000
rs13281615

1
rs2456449

rs620861

rs6983267

rs6983561

rs7000448

rs7008482

rs9642880

52

Table 2.5b. Continued

rs2456449 rs620861 rs6983267 rs6983561 rs7000448 rs7008482 rs9642880
rs10086908 0.007 0.034 0.007 0.025 0.016 0.001 0.044
rs10088218 0.000 0.003 0.000 0.000 0.003 0.005 0.010
rs10090154 0.015 0.009 0.008 0.017 0.004 0.001 0.026
rs1016343 0.055 0.005 0.006 0.021 0.058 0.013 0.014
rs116041037 0.008 0.003 0.024 0.023 0.004 0.005 0.008
rs12543663 0.007 0.028 0.004 0.000 0.005 0.005 0.000
rs13252298 0.001 0.006 0.000 0.014 0.034 0.001 0.042
rs13254738 0.054 0.007 0.011 0.221 0.007 0.001 0.002
rs13281615 0.032 0.073 0.002 0.012 0.005 0.030 0.014
rs2456449 1 0.003 0.051 0.078 0.032 0.002 0.008
rs620861

1 0.000 0.061 0.009 0.002 0.039
rs6983267

1 0.001 0.033 0.005 0.000
rs6983561

1 0.013 0.000 0.027
rs7000448

1 0.020 0.012
rs7008482

1 0.016
rs9642880

1
53

Table 2.6a. 8q24 Genotypes and Risk of Multiple Myeloma among Whites by Study Site.

Los Angeles
(71/93)
Seattle/Detroit SEER
(98/288)
Multiethnic Cohort
(21/42)
NHS and HPFS
c

(69/144)

P for
Heterogeneity
SNP RA
a
OR
b
95% CI p-value OR 95% CI p-value OR 95% CI p-value OR 95% CI p-value
rs7008482 G 1.12 (0.70-1.81) 0.63 0.79 (0.54-1.15) 0.22 0.92 (0.38-2.23) 0.86 0.87 (0.56-1.36) 0.55 0.74
rs12543663 C 0.68 (0.42-1.11) 0.12 1.03 (0.71-1.51) 0.86 0.71 (0.27-1.88) 0.49 - - - 0.42
rs10086908 T 1.11 (0.67-1.83) 0.69 0.75 (0.53-1.06) 0.11 0.54 (0.23-1.26) 0.15 - - - 0.28
rs1016343 T 1.14 (0.66-1.99) 0.63 1.10 (0.74-1.64) 0.64 0.89 (0.29-2.73) 0.84 - - - 0.84
rs13252298 A 1.16 (0.70-1.91) 0.57 1.11 (0.78-1.59) 0.57 0.26 (0.09-0.75) 0.01 - - - 0.02
rs13254738 C 1.19 (0.75-1.89) 0.47 1.14 (0.80-1.63) 0.47 1.31 (0.56-3.06) 0.53 1.00 (0.63-1.59) 0.99 0.95
rs6983561 C 1.42 (0.39-5.13) 0.59 0.91 (0.31-2.69) 0.86 0.34 (0.03-3.47) 0.37 2.96 (0.98-8.94) 0.05 0.24
rs2456449 G 0.85 (0.54-1.35) 0.50 0.96 (0.68-1.36) 0.82 0.69 (0.27-1.75) 0.43 - - - 0.79
rs620861 G 1.05 (0.65-1.70) 0.84 0.85 (0.60-1.21) 0.37 1.71 (0.74-3.96) 0.21 - - - 0.31
rs13281615 G 1.06 (0.67-1.68) 0.79 0.96 (0.70-1.32) 0.80 0.44 (0.17-1.12) 0.09 1.01 (0.66-1.55) 0.96 0.37
rs6983267 G 1.01 (0.63-1.62) 0.97 1.08 (0.77-1.51) 0.65 0.19 (0.06-0.59) 0.005 0.62 (0.41-0.93) 0.02 0.004
rs7000448 T 1.18 (0.77-1.81) 0.45 0.94 (0.67-1.33) 0.73 0.85 (0.36-2.02) 0.71 0.74 (0.46-1.17) 0.19 0.50
rs10090154 T 0.65 (0.29-1.47) 0.30 0.95 (0.55-1.66) 0.86 0.15 (0.02-1.32) 0.09 1.70 (0.91-3.19) 0.10 0.04
rs9642880 T 0.88 (0.56-1.39) 0.58 1.00 (0.72-1.39) 1.00 0.60 (0.23-1.57) 0.30 - - - 0.52
rs10088218 A 1.00 (0.49-2.02) 0.99 1.12 (0.69-1.81) 0.64 0.33 (0.07-1.43) 0.14 - - - 0.28
a
RA=risk allele,
b
Analyses adjusted for age (as a continuous variable), sex, and site
c
The NHS and HPFS were not included in the analysis of the following SNPs: rs12543663,
rs10086908, rs1016343, rs13252298, rs2456449, rs620861, rs9642880, rs10088218.

54

Table 2.6b. 8q24 Genotypes and Risk of Multiple Myeloma among African Americans by Study Site.

Los Angeles
(36/51)
Seattle/Detroit SEER
(26/32)
Multiethnic Cohort
(51/102)

P for
Heterogeneity
SNP RA
a
OR
b
95% CI p-value 95% CI p-value OR 95% CI p-value
rs7008482 G 0.83 (0.38-1.81) 0.65 0.33 (0.09-1.18) 0.09 1.43 (0.76-2.66) 0.26 0.26
rs12543663 C 1.53 (0.67-3.47) 0.31 1.19 (0.27-5.19) 0.82 0.82 (0.38-1.77) 0.62 0.49
rs10086908 T 0.86 (0.43-1.71) 0.66 1.34 (0.42-4.26) 0.62 1.48 (0.79-2.79) 0.22 0.47
rs1016343 T 0.95 (0.44-2.02) 0.88 3.45 (1.11-10.76) 0.03 0.80 (0.40-1.58) 0.52 0.19
rs13252298 A 0.55 (0.18-1.67) 0.29 2.00 (0.10-40.73) 0.65 1.80 (0.70-4.58) 0.22 0.39
rs13254738 C 0.90 (0.47-1.72) 0.74 1.38 (0.46-4.09) 0.56 0.79 (0.45-1.37) 0.40 0.82
rs6983561 C 0.64 (0.34-1.20) 0.16 0.70 (0.24-2.03) 0.51 0.95 (0.54-1.68) 0.85 0.65
rs116041037
c
A - - - 0.18 (0.01-4.76) 0.31 1.55 (0.33-7.30) 0.58 0.07
rs2456449 G 1.10 (0.56-2.16) 0.78 1.01 (0.31-3.22) 0.99 0.84 (0.38-1.86) 0.66 0.88
rs620861 G 0.81 (0.43-1.53) 0.52 1.55 (0.56-4.31) 0.40 1.18 (0.70-1.98) 0.54 0.40
rs13281615 G 0.68 (0.35-1.33) 0.26 0.90 (0.34-2.33) 0.82 0.64 (0.37-1.13) 0.13 0.60
rs6983267 G 1.46 (0.45-4.74) 0.53 0.41 (0.10-1.66) 0.21 0.89 (0.42-1.91) 0.77 0.53
rs7000448 T 0.70 (0.34-1.45) 0.33 0.61 (0.21-1.83) 0.38 1.38 (0.79-2.41) 0.25 0.25
rs10090154 T 1.04 (0.48-2.24) 0.92 0.37 (0.06-2.17) 0.27 1.25 (0.52-2.97) 0.62 0.49
rs9642880 T 1.09 (0.54-2.21) 0.81 0.90 (0.27-3.02) 0.87 0.78 (0.48-1.27) 0.32 0.72
rs10088218 A 2.25 (0.75-6.72) 0.15 0.98 (0.31-3.05) 0.97 1.20 (0.60-2.42) 0.61 0.70
a
RA=risk allele,
b
Analyses adjusted for age (as a continuous variable), sex, and site.
c
In the Los Angeles study, the estimate for rs116041037 could not be computed separately
due to unstable estimates, but was included in the pooled analysis presented in Table 3.

55
Chapter 2 References
1. Kuehl, W.M. and P.L. Bergsagel, Multiple myeloma: evolving genetic events and
host interactions. Nat Rev Cancer, 2002. 2(3): p. 175-87.
2. Haiman, C.A., et al., Multiple regions within 8q24 independently affect risk for
prostate cancer. Nat Genet, 2007. 39(5): p. 638-44.
3. Palumbo, A. and K. Anderson, Multiple myeloma. N Engl J Med, 2011. 364(11):
p. 1046-60.
4. Turesson, I., et al., Patterns of multiple myeloma during the past 5 decades: stable
incidence rates for all age groups in the population but rapidly changing age
distribution in the clinic. Mayo Clin Proc, 2010. 85(3): p. 225-30.
5. Alexander, D.D., et al., Multiple myeloma: a review of the epidemiologic
literature. Int J Cancer, 2007. 120 Suppl 12: p. 40-61.
6. Birmann BM, C.B., Muench K, Suppan CA, Cozen W., Epidemiology and
etiology of multiple myeloma., in Multiple Myeloma - A New Era of Treatment
Strategies, P.K.a.A. KC, Editor 2011, Bentham Science Publishers: Bentham
Books.
7. Jia, L., et al., Functional enhancers at the gene-poor 8q24 cancer-linked locus.
PLoS Genet, 2009. 5(8): p. e1000597.
8. Ahmadiyeh, N., et al., 8q24 prostate, breast, and colon cancer risk loci show
tissue-specific long-range interaction with MYC. Proc Natl Acad Sci U S A.
107(21): p. 9742-6.
9. Wright, J.B., S.J. Brown, and M.D. Cole, Upregulation of c-MYC in cis through a
large chromatin loop linked to a cancer risk-associated single-nucleotide
polymorphism in colorectal cancer cells. Mol Cell Biol, 2010. 30(6): p. 1411-20.
10. Freedman, M.L., et al., Admixture mapping identifies 8q24 as a prostate cancer
risk locus in African-American men. Proc Natl Acad Sci U S A, 2006. 103(38): p.
14068-73.
11. Easton, D.F., et al., Genome-wide association study identifies novel breast cancer
susceptibility loci. Nature, 2007. 447(7148): p. 1087-93.
12. Gudmundsson, J., et al., Genome-wide association study identifies a second
prostate cancer susceptibility variant at 8q24. Nat Genet, 2007. 39(5): p. 631-7.
56
13. Kiemeney, L.A., et al., Sequence variant on 8q24 confers susceptibility to urinary
bladder cancer. Nat Genet, 2008. 40(11): p. 1307-12.
14. Schumacher, F.R., et al., A common 8q24 variant in prostate and breast cancer
from a large nested case-control study. Cancer Res, 2007. 67(7): p. 2951-6.
15. Shou, Y., et al., Diverse karyotypic abnormalities of the c-myc locus associated
with c-myc dysregulation and tumor progression in multiple myeloma. Proc Natl
Acad Sci U S A, 2000. 97(1): p. 228-33.
16. Dib, A., et al., Characterization of MYC translocations in multiple myeloma cell
lines. J Natl Cancer Inst Monogr, 2008(39): p. 25-31.
17. Tewari, P., et al., Genetic variation at the 8q24 locus confers risk to multiple
myeloma. Br J Haematol.
18. Cozen, W., et al., Interleukin-6-related genotypes, body mass index, and risk of
multiple myeloma and plasmacytoma. Cancer Epidemiol Biomarkers Prev, 2006.
15(11): p. 2285-91.
19. Gold, L.S., et al., Associations of common variants in genes involved in
metabolism and response to exogenous chemicals with risk of multiple myeloma.
Cancer Epidemiol, 2009. 33(3-4): p. 276-80.
20. Kolonel, L.N., et al., A multiethnic cohort in Hawaii and Los Angeles: baseline
characteristics. Am J Epidemiol, 2000. 151(4): p. 346-57.
21. Colditz, G.A. and S.E. Hankinson, The Nurses' Health Study: lifestyle and health
among women. Nat Rev Cancer, 2005. 5(5): p. 388-96.
22. Giovannucci, E., et al., Risk factors for prostate cancer incidence and progression
in the health professionals follow-up study. Int J Cancer, 2007. 121(7): p. 1571-8.
23. Birmann, B.M., et al., Insulin-like growth factor-1- and interleukin-6-related gene
variation and risk of multiple myeloma. Cancer Epidemiol Biomarkers Prev,
2009. 18(1): p. 282-8.
24. Al Olama, A.A., et al., Multiple loci on 8q24 associated with prostate cancer
susceptibility. Nat Genet, 2009. 41(10): p. 1058-60.
25. Ghoussaini, M., et al., Multiple loci with different cancer specificities within the
8q24 gene desert. J Natl Cancer Inst, 2008. 100(13): p. 962-6.
26. Haiman, C.A., et al., A common genetic risk factor for colorectal and prostate
cancer. Nat Genet, 2007. 39(8): p. 954-6.
57
27. Haiman, C.A., et al., Characterizing genetic risk at known prostate cancer
susceptibility loci in African Americans. PLoS Genet, 2011. 7(5): p. e1001387.
28. Chen, F., et al., Fine-mapping of breast cancer susceptibility loci characterizes
genetic risk in African Americans. Hum Mol Genet, 2011. 20(22): p. 4491-503.
29. Mailman, M.D., et al., The NCBI dbGaP database of genotypes and phenotypes.
Nat Genet, 2007. 39(10): p. 1181-6.
30. Howie, B.N., P. Donnelly, and J. Marchini, A flexible and accurate genotype
imputation method for the next generation of genome-wide association studies.
PLoS Genet, 2009. 5(6): p. e1000529.
31. Cortessis, V.K., et al., Risk of urinary bladder cancer is associated with 8q24
variant rs9642880[T] in multiple racial/ethnic groups: results from the Los
Angeles-Shanghai case-control study. Cancer Epidemiol Biomarkers Prev.
19(12): p. 3150-6.
32. Zanke, B.W., et al., Genome-wide association scan identifies a colorectal cancer
susceptibility locus on chromosome 8q24. Nat Genet, 2007. 39(8): p. 989-94.
33. Hooker, S., et al., Replication of prostate cancer risk loci on 8q24, 11q13, 17q12,
19q33, and Xp11 in African Americans. Prostate. 70(3): p. 270-5.
34. Crowther-Swanepoel, D., et al., Common variants at 2q37.3, 8q24.21, 15q21.3
and 16q24.1 influence chronic lymphocytic leukemia risk. Nat Genet. 42(2): p.
132-6.
35. Enciso-Mora, V., et al., A genome-wide association study of Hodgkin's lymphoma
identifies new susceptibility loci at 2p16.1 (REL), 8q24.21 and 10p14 (GATA3).
Nat Genet. 42(12): p. 1126-30.
36. Tomlinson, I., et al., A genome-wide association scan of tag SNPs identifies a
susceptibility variant for colorectal cancer at 8q24.21. Nat Genet, 2007. 39(8): p.
984-8.
37. Pomerantz, M.M., et al., Evaluation of the 8q24 prostate cancer risk locus and
MYC expression. Cancer Res, 2009. 69(13): p. 5568-74.
38. Pomerantz, M.M., et al., The 8q24 cancer risk variant rs6983267 shows long-
range interaction with MYC in colorectal cancer. Nat Genet, 2009. 41(8): p. 882-
4.
58
39. Kupfer, S.S., et al., Novel single nucleotide polymorphism associations with
colorectal cancer on chromosome 8q24 in African and European Americans.
Carcinogenesis, 2009. 30(8): p. 1353-7.
40. Zheng, S.L., et al., Association between two unlinked loci at 8q24 and prostate
cancer risk among European Americans. J Natl Cancer Inst, 2007. 99(20): p.
1525-33.
41. Yeager, M., et al., Identification of a new prostate cancer susceptibility locus on
chromosome 8q24. Nat Genet, 2009. 41(10): p. 1055-7.
42. White, K.L., et al., Variation at 8q24 and 9p24 and risk of epithelial ovarian
cancer. Twin Res Hum Genet. 13(1): p. 43-56.
43. Goode, E.L., et al., A genome-wide association study identifies susceptibility loci
for ovarian cancer at 2q31 and 8q24. Nat Genet. 42(10): p. 874-9.

59
Chapter 3: Functional Polymorphisms in DNA Repair Pathways and
Multiple Myeloma Risk

3.1 Introduction and Rationale
DNA repair mechanisms are integral in B-cell differentiation and plasma cell
development [1]. As described in detail in the background chapter, errors in these
pathways may lead to the development of MM. We hypothesize that individual variation
in DNA repair function, represented by non-synonymous SNPs, may alter risk of MM.
We assess this hypothesis by examining functional polymorphisms within DNA
repair pathways and MM risk in a multiethnic case-control study in Los Angeles,
including relative and population-based controls. We try to replicate our most significant
associations in three separate replication sets. The first replication set includes 347 cases
and 758 controls from the Seattle/Detroit SEER case-control study, the MEC, the Nurses’
Health Study, and the Health Professionals Follow-up study. The second replication
includes 290 cases and 295 controls from a multiethnic GWAS conducted at the
University of California, San Francisco and the third replication set includes 305 cases
and 353 controls from a GWAS conducted at the Mayo clinic. We then combine data
from our most significant associations in a meta-analysis.

3.2 Abstract

Multiple myeloma (MM) is a hematologic cancer of post-germinal B-cells. DNA repair
plays an integral role in B-cell development and we hypothesize that genetic variation
within DNA repair pathways may play a role in MM risk. We evaluated the role of 42
60
non-synonymous single nucleotide polymorphisms in genes involved in DNA repair
pathways in MM risk in a multi-ethnic case-control study. The discovery set consisted of
148 cases, 114 matched relative controls, and 131 population-based controls. The
matched and unmatched controls were combined using a method which utilizes a binary
latent variable and takes into account the correlation between estimates. A replication set
consisted of three independent study sets: 342 cases and 751 controls from a multi-center
replication set, 290 cases and 205 controls from the University of California, San
Francisco, and 305 cases and 353 controls from the Mayo Clinic. We observed
significant associations in 6 single nucleotide polymorphisms (SNP) in the discovery set
(rs1801516, rs1650697, rs25489, rs2227999, rs2228528, and rs2228615); however the
associations were not statistically significant in any of the independent replications sets or
in the meta-analysis of all datasets. This study cannot rule out the involvement of genetic
variation in DNA repair pathways in MM risk, as we did not comprehensively capture the
variation of all pathways.

3.3 Introduction

Multiple myeloma (MM) is a rare (5.8/100,000 (http://seer.cancer.gov/statfacts/html/mulmy.html))
but incurable hematologic malignancy arising from a proliferation of monoclonal plasma
cells from post-germinal center B-cells [2]. It is responsible for approximately 2% of
cancer deaths and 20% of deaths from hematological malignancies [3]. A 2-3 fold
increased risk has been reported in families of MM cases, suggesting a genetic
component in the etiology of the disease [4].
61
DNA-repair mechanisms are integral to plasma cell differentiation and regulation,
and decreased DNA-repair capacity is hypothesized to be involved in uncontrolled clonal
proliferation of malignant MM plasma cells [2]. For example, DNA double-stranded
break repair is necessary for the development of antibody diversity and errors in this
process could lead to translocations of an oncogene adjacent to a promoter resulting in
uncontrolled cell growth [2]. There are five major DNA repair pathways: 1) direct
damage reversal, 2) base excision repair (BER), 3) nucleotide excision repair (NER), 4)
mismatch repair (MMR), and 5) double-stranded repair through homologous
recombination (HR) or non-homologous end joining (NHEJ) [5]. Deficiencies in DNA
repair capacity within these pathways have been associated with an increased risk of
breast and colorectal cancers [5-7], and polymorphisms associated with risk of other solid
tumor cancers such as bladder and prostate cancer, and certain hematologic malignancies,
including B-cell lymphoma, Non-Hodgkin lymphoma, and MM [8-12].
Few case-control studies have investigated the associations between
polymorphisms in DNA repair genes and MM risk, and results have been inconclusive
[13]. Hayden et al investigated genes involved in double-strand breakage (XRCC3,
XRCC4, and XRCC5) in 306 European cases, and reported an association between a
variant in XRCC4 (rs963248) and increased MM risk [9]. In a study consisting of 270
cases of European origin, Roddam et al observed that genetic variation within LIG4
resulted in a reduced risk of developing MM, suggesting involvement of the NHEJ
pathway and class switch recombination [14]. These findings have yet to be replicated in
independent studies.
62
While pathway-based analyses allow for interrogation of specific risk regions or
genes and are ideal for rare diseases such as MM, the contribution of polymorphisms
within DNA repair genes and MM risk remains unknown. This is due in part to the
difficulty in attaining a large number of cases necessary for adequate statistical power. To
evaluate the effects of heritable variation in DNA-repair pathways on MM risk, we
conducted a multi-ethnic population-based case-control study in Los Angeles, with
replication performed in three independent studies.

3.4 Methods

3.4.1 Participants
Los Angeles Set. The Los Angeles case-control study has been explained in detail
elsewhere [15]. Briefly, MM and plasmacytoma cases between the ages of 20-74 were
identified using rapid ascertainment from the Los Angeles Cancer Surveillance Program,
the population-based Surveillance, Epidemiology, and End Results (SEER) cancer
registry in Los Angeles County. Two groups of controls were selected. The population-
based controls were originally enrolled in a non-Hodgkin’s lymphoma study with the
National Cancer Institute. Cases were asked to list family members and one control was
selected for each case using a pre-determined hierarchical algorithm that prioritized
family members to obtain a second set of related controls. In total, DNA samples from
148 cases, 131 NCI controls, and 114 matched relative controls were available for
genotyping and analysis.
63
Multi-center Replication Set. The replication set consists of DNA samples from
347 cases and 758 controls from one population-based and three nested case-control
studies:
MM cases between the ages of 35-74 were identified using the Surveillance,
Epidemiology, and End Results (SEER) cancer registries in the Seattle and Detroit areas.
Similar to the Los Angeles study, the population-based controls were selected from a
previous non-Hodgkin’s lymphoma study [16], and represent the same geographical areas
from which the cases were recruited. In total, DNA samples for 138 myeloma cases and
340 population-based controls were obtained.
The Multiethnic Cohort is comprised of over 215,000 Native Hawaiian, Japanese,
White, African American, and Latino individuals who entered the cohort between 1993
and 1996 [17]. The MEC was cross-linked with the population-based Surveillance,
Epidemiology, and End Results (SEER) registries in California and Hawaii, and incident
MM cancer cases were identified. DNA samples from 127 myeloma cases and 254
controls (matched 2:1 on sex, ethnicity, age at cohort entry, and education (<=12 years,
13-16 years, and 17+ years)) were obtained.
The Nurses’ Health Study cohort (NHS) was developed in 1976 and was initially
comprised of 121,700 registered nurses ages 30 to 55 from 11 different states [18]. Blood
samples were later collected from 32,826 women between 1989 and 1990. DNA samples
from 58 myeloma cases and 116 controls (2:1 matched on age at cohort entry, race,
gender, and DNA type) were obtained.
64
The Health Professionals Follow-up study (HPFS) began in 1986 with 51,529 US
male health professionals; blood samples were later collected from 18,018 of these men
between 1993 and 1994 [19]. Samples from 24 cases and 48 controls from the HPFS (2:1
matched on age at cohort entry, race gender, and prevalent or incident case) were
obtained. Because of the similar study design and source populations, data from the NHS
and the HPFS were combined for analysis consistent with previous studies [20].
UCSF Set. Cases included patients with the diagnosis of Multiple Myeloma who
received care at UCSF between 1993 and 2010. Biospecimens included peripheral blood
samples after stem cell mobilization with granulocyte colony-stimulating factor (GCSF)
in preparation for autologous bone marrow transplantation and banked in liquid nitrogen
or peripheral blood that was collected as part of the research study from patients during a
follow-up appointment. Controls included patients who received care at the UCSF
General Medicine Clinic between 1995 and 2011. Samples from 300 cases and 300
controls were sent to Expression Analysis for genotyping. The cases were selected
based on having complete or near complete clinical records. We excluded Asian cases
since the number in this group was particularly small. Controls were selected based on
matching to the cases by gender and race/ethnicity.
Mayo Set. Incident MM cases were sampled from the Mayo Clinic regional
practice between 1998 and 2007, and within 9 months of the initial diagnosis. MM cases
from seven states (Minnesota, Iowa, Wisconsin, North Dakota, Michigan, South Dakota,
and Illinois) were included. Controls were sampled from the Mayo General Internal
Medicine clinics between 2002 and 2010. Controls were residents of Minnesota, Iowa or
65
Wisconsin, at least 20 years old, and had no prior history of MM, MGUS, lymphoma,
leukemia, or HIV infection [21]. Samples were available for 303 cases and 353 controls.

3.4.2 SNP Selection and Genotyping
Described elsewhere in detail [5], the initial SNP selection was based on a
characterization of linkage disequilibrium patterns in DNA repair genes with relationship
to breast cancer risk in multiethnic populations as follows: Briefly, characterization of
common variation in the coding and non-coding regions of 60 DNA repair related genes
was conducted by genotyping a high density of SNPs in the following multiethnic
samples: 20 CEPH trios (a subset of the 30 trios used in HapMap), European Americans
(n=70), African Americans (n=70), Native Hawaiians (n=70), Japanese Americans
(n=70) in the MEC, and Chinese from the Singapore Chinese Health Study (n=59) [5]. In
total, approximately 3000 SNPs were genotyped. From these, 1536 tag-SNPs were
selected using Tagger, a tagging approach which combines pair wise and multi-marker
approaches [22]. 253 non-synonymous SNPs were included: 105 coding SNPs were used
as tag-SNPs and another 148 coding variants with a minor allele frequency (MAF) ≥
1%were added after tagging was completed. These SNPs were selected from the five
ethnic populations (r
2
>=.80) across 60 DNA repair genes in the following five pathways:
base excision repair (BER), nucleotide excision repair (NER), double strand break repair
(DSB) via homologous recombination (HR) or non-homologous end-joining (NHEJ),
mismatch repair (MMR), and direct reversion repair.
66
Los Angeles Set. Genotyping was performed using the GoldenGate assay and
Illumina BeadArray technology in the University of Southern California’s Genomics
Center. Of the 1536 previously described SNPs, 53 SNPs with call rates <80% and 59
monomorphic SNPs were excluded. Of these 1424 SNPs, 42 non-synonymous SNPs
with a MAF >5% (in the controls) were selected for quality control and further analysis,
assuming that these SNPs best represent functional changes within DNA repair pathways
and are likely to have similar effects across racial/ethnic groups (Table 3.1) [21].
Multi-center Replication Set. 347 cases and 758 controls were genotyped for the
6 SNPs (of the 42 SNPs), rs1801516, rs1650697, rs25489, rs2227999, rs2228528, and
rs2228615 that were statistically significant in the Los Angeles discovery set (p<0.05)
using TaqMan. DNA samples from the discovery set (150 cases and 131 population-
based controls) were also genotyped to validate the Illumina panel results. Fifty-seven
additional NCI controls from the Los Angeles study not included in the discovery set
were included in the replication set, as DNA from additional controls became available.
UCSF Set. Cases and controls were genotyped using the Omni 5M (~4.2 million
SNPs). Three of the five significant SNPs were genotyped (rs25489, rs2227999,
rs2228528) and two were imputed using 1000 genomes data as the reference panel
(rs1801516, rs2228615). Of the 600 samples genotyped, we excluded 15 cases and 10
controls for poor genotyping quality.
Mayo Set. Cases and controls were genotyped using the Affymetrix 6.0 (~1.8
million SNPs). Three of the five significant SNPs were genotyped (rs25489, rs2227999,
67
rs2228528) and two were imputed using 1000 genomes data as the reference panel
(rs1801516, rs2228615).

3.4.3 Statistical Analysis
Los Angeles Set. Hardy-Weinberg equilibrium (HWE) was assessed in white and
African American controls in the 42 functional SNPs using a one-degree of freedom chi
squared test. Due to the small sample size, HWE was not assessed in other ethnicities. 28
cases and 7 relative controls with call rates <80% were excluded: 120 cases, 131 NCI
controls, and 107 relative controls were included in the analysis.
Assuming a log-additive risk, odds ratios (OR) and 95% confidence intervals
(CIs) were calculated using conditional logistic regression (cases and matched relative
controls) and unconditional logistic regression (cases and population-based controls).
Because not all cases had a matched relative, fewer cases were used in this analysis than
in the population-based case-control comparisons (n=87). All models were adjusted for
age (continuous), gender, and race (African American vs. non-African American),
because of the small numbers of other ethnicities.
To attain an estimate combining the matched and unmatched samples, a polytomous
conditional likelihood approach was used [23], which utilizes a binary latent variable and
takes into account the correlation between estimates when using the same case-set [23].
Multi-center Replication Set. After quality control procedures were employed,
genotyping data from 342 cases and 751 controls were available for the replication
analysis of the 6 SNPs. SNP call rates were above 95% in the genotyped data, and
68
concordance of duplicates was 100%. HWE was assessed in the White and African
American controls separately. Assuming a log-additive risk, ORs and 95% CIs were
calculated using unconditional logistic regression adjusting for age, race (five categories)
and gender. There was >98% concordance between the re-genotyped Los Angeles
samples and the initial assay.
UCSF Set. Genotype calls were made using Illumina BeadStudio Software
(Illumina, San Diego, CA). SNPs that had >10% missing rate were dropped for poor
quality. SNPs were excluded with p values <10e-6 for tests of HWE in Caucasians only.
Imputation was performed using IMPUTE2 with the 1000Genomes (KGP) data as the
reference dataset. Principal components analysis was used to infer genetic ancestry.
SNPs were analyzed for association with myeloma using logistic regression models and
adjusted for gender and for genetic ancestry using principal components 1-10. Analyses
were run using the program PLINK.
Mayo Set. SNPs with a call rate < 95% were excluded. Data was analyzed using
logistic regression adjusting for age, sex, and site. Imputation was performed using
Beagle with KGP (May 2011) as the reference set. SNPs with an r2 < 0.3 and MAF <0.01
were excluded from the analysis. Analyses were conducted using PLINK.
Meta-Analysis. We performed a random-effects meta-analysis to obtain combined
estimates from the Los Angeles, multi-center replication, the UCSF set and the Mayo set.

69
3.5 Results

All 42 SNPs were in Hardy-Weinberg equilibrium. Sex, race, and age by study
site are described in Table 3.2. Overall, the study populations are similar, except that the
mean age of diagnosis for cases in the MEC was higher than other participating studies
and the NHS and HPFS do not contribute multi-ethnic populations. Cases that were
excluded due to poor call rates were not different from those included on any
demographic variables presented in Table 3.2. The 57 additional genotyped NCI controls
are included in the Los Angeles NCI controls for purposes of demographic characteristic
description. Sex, race and age of the UCSF set and Mayo set are similar to the other
studies and are reported in Table 3.4.
Results for the Los Angeles discovery set and replication sets are presented in
Table 3.3. Of the 42 SNPs within DNA repair genes (Table 3.1), six SNPs (rs1801516,
rs1650697, rs25489, rs2227999, rs2228528, and rs2228615) were statistically
significantly associated with MM risk in the Los Angeles discovery set (P<.05) and were
genotyped in the multi-center replication set. One SNP (rs1650697) in the MSH3 gene
failed TaqMan genotyping. There was no statistically significant association with MM
risk and any of the five SNPs across the three independent replication sets. We observed
a possible inverse association with the A-allele of rs25489 in XRCC1 and MM risk in the
Los Angeles discovery set with similar inverse associations observed in the three
replication sets, however these results were not statistically significant. Results for the
Los Angeles discovery set and replication sets are presented in Table 3.3.
70
When all studies were combined using a random-effects meta-analysis we
observed an inverse association between the A-allele of rs25489 and MM risk when
studies were combined (OR=0.70, 95% CI=0.50, 0.96), however when a Bonferroni
correction (0.05/6=0.008) was applied no statistically significant associations remained
(Table 3.3).

3.6 Discussion
We conducted a hypothesis-driving case control study to assess genetic variation
within DNA repair genes and MM risk. We then replicated our most significant findings
in three independent study-sets and did not observe any statistically significant
associations with the risk of developing MM.
Although decreased DNA repair efficiency is hypothesized to create cumulative
mutations leading to solid-tumor carcinogenesis [8], we did not observe any association
between non-synonymous SNPs within DNA repair pathway genes and MM risk.
We were able to explore this hypothesis taking advantage of a relatively large (for MM)
discovery set and three large independent replication sets. We further increased our
power by using a novel analysis method developed at the University of Southern
California, which combines unmatched and matched case-control studies with the same
case-set [23].
In a high-risk, high-payoff study design, we examined non-synonymous SNPs
most likely to represent functional changes occurring within the gene, which permitted
use of a larger multi-ethnic population under the assumption that any observed
71
“functional change” would be the same across racial/ethnic populations [23]. In addition,
the use of a multi-ethnic population permits identification of a larger set of risk alleles
than studies limited to one racial/ethnic group [5]. However, by targeting only the 42
non-synonymous SNPs, we sacrificed broad coverage of these pathways. The original
Illumina panel included a large number of non-synonymous SNPs: 105 missense SNPs
used as tagSNPs and the manual addition of 148 missense SNPs with a MAF > 1% [24].
In the discovery set, 28 cases were removed due to poor genotyping call rates.
Although the demographic characteristics of these cases were similar to those included in
the analysis, it is possible that the excluded cases, who primarily provided buccal swab
rather than blood samples, may have had a more advanced stage of the disease, which
could limit the generalizability.
In conclusion, we did not observe any statistically significant findings after
replicating our discovery set results in three independent replication sets. We cannot rule
out the etiologic involvement of DNA repair pathways in MM, as these five SNPs do not
capture the entire variation of the genes in which they reside, nor do the genes represent
the complete pathways. Although the field of genetic epidemiology has embodied an
agnostic approach in genetic research, candidate gene studies are still important, as there
is the ability to thoroughly investigate genetic variation and disease within biologically
implicated pathways, thereby incorporating a-priori hypotheses while avoiding the
penalty of many false negatives in the newer agnostic approaches.

72

Table 3.1. Forty-two Functional SNPs
SNP Gene Pathway* SNP Gene Pathway
rs4994 ADRB3
-
rs1805404 PARP1 BER
rs1801516 ATM SIGN/RESP rs1805323 PMS2 MMR
rs2227928 ATR SIGN/RESP rs1059060 PMS2 MMR
rs2229032 ATR SIGN/RESP rs2640 PMS2 MMR
rs1045485 CASP8
-
rs1726801 POLD POL
rs2228528 CSB NER rs8305 POLI POL
rs2228527 CSB NER rs37370 PRLR
-
rs2228529 CSB NER rs3740955 RAG1 V(D)J Recombination
rs762679 DNA-PK DSB-NHEJ rs2227973 RAG1 V(D)J Recombination
rs7830743 DNA-PK DSB-NHEJ rs5030755 RPA1 NER
rs2239359 FANCA FA rs1042522 TP53 SIGN/RESP
rs7190823 FANCA FA rs2227999 XPC NER
rs9462088 FANCE FA rs2228001 XPC NER
rs2228615 ICAM5
-
rs1799793 XPD NER
rs2308327 MGMT REV rs2020955 XPF NER
rs12917 MGMT REV rs1800067 XPF NER
rs175080 MLH3 MMR rs17655 XPG NER
rs1650697 MSH3 MMR rs25489 XRCC1 BER
rs26279 MSH3 MMR rs2307177 XRCC1 BER
rs3219489 MUTYH
-
rs25487 XRCC1 BER
rs1805794 NBS1 DSB-HR rs3218536 XRCC2 DSB-HR

*SIGN/RESP, DNA damage signaling and response; NER, nucleotide excision repair; BER, base excision repair;
MMR, mismatch repair; FA, DNA cross link repair; DSB–NHEJ, double strand break repair - non-homologous end
joining; POL, DNA polymerase; DSB-HR, double strand break repair - homologous recombination.
73

Table 3.2. Demographic Characteristics of the Initial Los Angeles Study and the Multi-Center Replication Set
Los Angeles SEER Multiethnic Cohort Seattle/Detroit SEER NHS and HPFS

Cases
n(%)
Family controls
n(%)
NCI Controls
n (%)**
Cases
n (%)
Controls
n (%)
Cases
n (%)
Controls
n (%)
Cases
n (%)
Controls
n (%)
Sex
Male 90 (61) 54 (47) 111 (59) 65 (51) 130 (51) 82 (59) 191 (56) 24 (31) 48 (31)
Female 57 (39) 60 (53) 77 (41) 62 (49) 124 (49) 56 (41) 149 (44) 53 (69) 109 (69)

Race
White 72 (49) 57 (50) 94 (50) 21 (17) 42 (17) 101 (73) 288 (85) 69 (90) 144 (92)
African American 36 (25) 28 (25) 51 (27) 51 (40) 102 (40) 26 (19) 32 (9) 1 (1) 1 (.5)
Hispanic 16 (11) 12 (11) 34 (18) 34 (27) 68 (27) 2 (1) 8 (2) 0 (0) 1 (.5)
Other/unknown 22 (15) 15 (13) 9 (5) 21 (17) 42 (17) 9 (7) 12 (4) 7 (9) 11 (7)
Missing 1(.5) 2 (1) - - - - - - -

Age*
Mean age
at diagnosis (SD) 59 (10) 59 (12) 56 (13) 70 (10) 70 (10) 60 (10) 57 (12) 67 (8) 67 (8)
Median age at
diagnosis 61 59 59 72 72 60 60 66 66
*For controls, mean and median reference age was calculated. **The additional 57 NCI controls are included in this column for the demographic table

74

Table 3.3. Top DNA Repair SNP Estimates in the Discovery and Replication Sets, and the Meta-Analysis Results

Combined Los Angeles Set
Multi-center Replication Set
(339/751)
UCSF Replication Set
(118/87/130)
a,b
(290/295)
SNP Gene OR 95% CI P-Value OR 95% CI P-Value OR 95% CI P-Value
rs1801516 ATM 0.38 (0.20-0.73) 0.004 1.11 (0.83-1.49) 0.48 0.90 (0.63-1.28) 0.55
rs25489 XRCC1 0.38 (0.18-0.80) 0.01 0.74 (0.47-1.17) 0.2 0.65 (0.37-1.12) 0.12
rs2227999 XPC 0.31 (0.12-0.80) 0.02 1.38 (0.86-2.19) 0.18 0.63 (0.35-1.14) 0.13
rs2228528 CSB 1.50 (1.01-2.24) 0.04 1.00 (0.79-1.26) 0.99 0.92 (0.67-1.27) 0.61
rs2228615 ICAM5 0.69 (0.47-1.00) 0.05 0.97 (0.79-1.18) 0.73 0.82 (0.64-1.04) 0.10
a (ca/relative control/pop-based control) as the relative controls were analyzed using conditional logistic regression, only 87 cases were counted.
b (case/control)
c Combined Los Angeles set was adjusted for age, gender, and race (African-American vs non-African-American), Multi-center replication set was adjusted for
age, race (5 categories), and site, the UCSF replication set was adjusted for age, sex, and principal components 1-10, and the Mayo Replication set was adjusted
for age, sex, and principal components 1-10.

75

Table 3.3. Continued

Mayo Replication Set
Meta-Analysis
(305/353)
SNP Gene OR 95% CI P-Value OR 95% CI P-Value P-Het
rs1801516 ATM 0.68 (0.48-0.95) 0.02 0.76 (0.53-1.10) 0.15 0.01
rs25489 XRCC1 0.95 (0.58-1.54) 0.82 0.7 (0.50-0.96) 0.03 0.24
rs2227999 XPC 1.21 (0.78-1.86) 0.39 0.84 (0.49-1.44) 0.53 0.01
rs2228528 CSB 1.31 (0.96-1.77) 0.09 1.13 (0.92-1.39) 0.24 0.14
rs2228615 ICAM5 1.11 (0.88-1.39) 0.39 0.91 (0.76-1.09) 0.3 0.11
a (ca/relative control/pop-based control) as the relative controls were analyzed using conditional logistic regression, only 87 cases were counted.
b (case/control)
c Combined Los Angeles set was adjusted for age, gender, and race (African-American vs non-African-American), Multi-center replication set was adjusted for
age, race (5 categories), and site, the UCSF replication set was adjusted for age, sex, and principal components 1-10, and the Mayo Replication set was adjusted
for age, sex, and principal components 1-10.

76

Table 3.4. Demographic Characteristics for the UCSF and Mayo Replication Sets

UCSF
Replication Set
Mayo
Replication Set
Mean Age 55/61
61/62
Gender

Male 175/175
175/213
Female 115/120 130/140
Race

White 161/225 305/353
AA 29/31
0/0
Hispanic 38/39
0/0
Other/Unknown 62/0
0/0
Missing -
-
Total 290/295 305/353
77
Chapter 3 References

1. Shapiro-Shelef, M. and K. Calame, Regulation of plasma-cell development. Nat
Rev Immunol, 2005. 5(3): p. 230-42.
2. Palumbo, A. and K. Anderson, Multiple myeloma. N Engl J Med, 2011. 364(11):
p. 1046-60.
3. Kuehl, W.M. and P.L. Bergsagel, Multiple myeloma: evolving genetic events and
host interactions. Nat Rev Cancer, 2002. 2(3): p. 175-87.
4. Alexander, D.D., et al., Multiple myeloma: a review of the epidemiologic
literature. Int J Cancer, 2007. 120 Suppl 12: p. 40-61.
5. Haiman, C.A., et al., Comprehensive association testing of common genetic
variation in DNA repair pathway genes in relationship with breast cancer risk in
multiple populations. Hum Mol Genet, 2008. 17(6): p. 825-34.
6. Han, J., et al., Genetic variation in DNA repair pathway genes and
premenopausal breast cancer risk. Breast Cancer Res Treat, 2009. 115(3): p. 613-
22.
7. Peltomaki, P., Deficient DNA mismatch repair: a common etiologic factor for
colon cancer. Hum Mol Genet, 2001. 10(7): p. 735-40.
8. Goode, E.L., C.M. Ulrich, and J.D. Potter, Polymorphisms in DNA repair genes
and associations with cancer risk. Cancer Epidemiol Biomarkers Prev, 2002.
11(12): p. 1513-30.
9. Hayden, P.J., et al., Variation in DNA repair genes XRCC3, XRCC4, XRCC5 and
susceptibility to myeloma. Hum Mol Genet, 2007. 16(24): p. 3117-27.
10. Rudd, M.F., et al., Variants in the ATM-BRCA2-CHEK2 axis predispose to
chronic lymphocytic leukemia. Blood, 2006. 108(2): p. 638-44.
11. Shen, M., et al., Polymorphisms in DNA repair genes and risk of non-Hodgkin
lymphoma in a pooled analysis of three studies. Br J Haematol, 2010. 151(3): p.
239-44.
12. Tambini, C.E., et al., The importance of XRCC2 in RAD51-related DNA damage
repair. DNA Repair (Amst), 2010. 9(5): p. 517-25.
13. Vangsted, A., T.W. Klausen, and U. Vogel, Genetic variations in multiple
myeloma I: effect on risk of multiple myeloma. Eur J Haematol, 2012. 88(1): p. 8-
30.
78
14. Roddam, P.L., et al., Genetic variants of NHEJ DNA ligase IV can affect the risk
of developing multiple myeloma, a tumour characterised by aberrant class switch
recombination. J Med Genet, 2002. 39(12): p. 900-5.
15. Cozen, W., et al., Interleukin-6-related genotypes, body mass index, and risk of
multiple myeloma and plasmacytoma. Cancer Epidemiol Biomarkers Prev, 2006.
15(11): p. 2285-91.
16. De Roos, A.J., et al., Metabolic gene variants and risk of non-Hodgkin's
lymphoma. Cancer Epidemiol Biomarkers Prev, 2006. 15(9): p. 1647-53.
17. Kolonel, L.N., et al., A multiethnic cohort in Hawaii and Los Angeles: baseline
characteristics. Am J Epidemiol, 2000. 151(4): p. 346-57.
18. Colditz, G.A. and S.E. Hankinson, The Nurses' Health Study: lifestyle and health
among women. Nat Rev Cancer, 2005. 5(5): p. 388-96.
19. Giovannucci, E., et al., Risk factors for prostate cancer incidence and progression
in the health professionals follow-up study. Int J Cancer, 2007. 121(7): p. 1571-8.
20. Birmann, B.M., et al., Insulin-like growth factor-1- and interleukin-6-related gene
variation and risk of multiple myeloma. Cancer Epidemiol Biomarkers Prev,
2009. 18(1): p. 282-8.
21. Greenberg, A.J., et al., Single-nucleotide polymorphism rs1052501 associated
with monoclonal gammopathy of undetermined significance and multiple
myeloma. Leukemia, 2012.
22. de Bakker, P.I., et al., Efficiency and power in genetic association studies. Nat
Genet, 2005. 37(11): p. 1217-23.
23. Gebregziabher, M., et al., A polytomous conditional likelihood approach for
combining matched and unmatched case-control studies. Stat Med, 2010. 29(9):
p. 1004-13.
24. Edlund, C.K., et al., Snagger: a user-friendly program for incorporating
additional information for tagSNP selection. BMC Bioinformatics, 2008. 9: p.
174.

79
Chapter 4: Meta-Analysis of Genome-wide Association Studies of
Multiple Myeloma

4.1. Introduction and Rationale
The candidate-gene association approach is used to examine specific pathways
hypothesized to play a role in the etiology of multiple myeloma. By definition, this
method is only able to evaluate a restricted number of polymorphisms. Recently, a
genome-wide association study (GWAS) of 1,675 cases and 5,903 controls (described in
detail in chapter one) has identified novel risk loci associated with MM risk, including a
non-synonymous single nucleotide polymorphism (SNP) in exon 17 of ULK4, a gene
involved in the regulation of mTOR-mediated autophagy [1]. This study identified three
common variants that influence MM risk but it is possible that additional genetic risk
factors exist and were not discovered in this single study.
The primary specific aim of this study is to evaluate the contribution of common
genetic variation in whites to MM risk by performing a GWAS and then combining our
data with UCSF in a meta-analysis and the secondary specific aim is to replicate findings
from the published GWAS. This project is performed as part of the primary specific aim
of the ‘SPORE in Multiple Myeloma’ grant (USC PI: Dr. Cozen, Parent grant PI: Dr.
Anderson), in which the contribution of genetic polymorphisms in genes associated with
hypothesized etiologic pathways of MM were to be evaluated.

80
4.2 Abstract
Multiple myeloma (MM) is a hematologic malignancy involving an
invasive proliferation of monoclonal plasma cells. We performed a genome-wide meta-
analysis of 745 cases and 797 controls using the Illumina OmniExpress and the Illumina
Omni5 platform. Overall, the most significant association was observed for rs7443528
(OR=0.64, p-value=5.78 x 10
-7
), located ~350 kb upstream of PIK3R1 on chromosome
5q13, a gene that is involved in a Ras dependent kinase pathway that has been associated
with the proliferation and survival of MM cells. There was also a marginally significant
association with rs76580792, a SNP located in the intron region of HLA-DRB5 on
chromosome 6p21.3 (OR=1.55, p-value=2.27x10
-6
). We also replicated a previously
reported GWAS association (rs4487645, p-value=0.0009). Our results provide suggestive
evidence of heritable risk factors for MM; larger-powered studies are needed to confirm
these specific results.

4.3 Introduction
Multiple myeloma (MM), a clonal proliferation of malignant post germinal center
plasma cells in the bone marrow, accounts for 1% of all cancer deaths and 20% of all
hematological cancer deaths (acs.gov). To date, multiple myeloma remains incurable
with a five year survival rate of ~40% (SEER.gov), yet there are few known genetic or
environmental epidemiologic risk factors for this disease. There is a 2- to 3- fold higher
risk of disease in African Americans compared to whites and a similar increased risk in
relatives of MM cases [2], suggesting a heritable component to this disease. It is likely
81
that, similar to other cancers [3, 4], MM genetic susceptibility involves a polygenic
model of multiple low-risk variants.
Candidate-gene or single nucleotide polymorphism (SNP) association studies,
which assess genetic contribution to MM risk, have been small with inconsistent results
[5]. While this targeted approach is used to examine specific pathways hypothesized to
play a role in MM etiology, this method is only able to evaluate a restricted number of
polymorphisms. Recently, a genome-wide association study (GWAS) from the UK
based on 1675 cases and 5903 controls identified two genome-wide significant and one
promising (10
-7
) novel loci associated with MM risk, including a non-synonymous SNP,
rs1052501 in exon 17 of ULK4, a gene involved in the regulation of mTOR-mediated
autophagy, rs4487645 on chromosome 7, and rs6746082 on chromosome 2 [6]. Because
the findings have not yet been replicated by a separate study, other agnostic studies may
provide further information by validating these SNPs and identifying additional loci.
Therefore, we assembled a multi-center collaboration to conduct a GWAS of
MM in 515 white patients and 518 white controls in order to identify new risk variants.
To increase our power, we combined our results in a meta-analysis with those from a
GWAS conducted at UCSF with 290 cases and 295 controls.

4.4 Methods
This study was approved by the institutional review boards of the University of
Southern California, University of Washington, Wayne State University, the University
of British Colombia, the University of Alabama, the Victoria Cancer Council and the
82
University of Melbourne, and the University of California, San Francisco in accordance
with the Declaration of Helsinki. Informed consent was obtained from all participants in
this study.

4.4.1 Source of Subjects
USC/Dana Farber Cancer Institute SPORE Multi-center Study (referred to as USC):
DNA specimens from 515 cases and 518 controls were obtained from six
epidemiologic studies of multiple myeloma (Table 4.1). In all six studies, participants
providing samples were self-reported whites. Three of these six studies have been
previously described in detail in chapter two, which include the Los Angeles SEER
population-based case-control study (72 cases and 68 controls), the Seattle/Detroit SEER
population-based case-control study (101 cases and 255 controls), and the Multiethnic
cohort (MEC) nested case-control study (21 cases and 25 controls). The three additional
study sites are described as follows:
University of Alabama samples include hospital-based cases from the university
facility and local clinics, and population-based controls from the Birmingham
metropolitan catchment area for the hospitals (108 cases and 108 controls).
Cases contributed by the University of British Columbia (151 cases) were
ascertained through referrals by specialists who are part of the hematology network at the
British Colombia (BC) Cancer Agency. There are a limited number of specialists who see
~90% of MM patients in BC. Cases were frequency matched to Seattle NCI controls
from the Seattle/Detroit SEER population-based case-control study on gender and site.
83
Controls from Seattle were selected as they most accurately represent cases from
Vancouver.
Samples were obtained from a nested case-control study of 62 incident cases
matched on age and sex to 62 randomly sampled unaffected controls nested within the
Melbourne Collaborative Cohort Study, Australia [7].
UCSF:
Cases included patients with the diagnosis of MM who received care at UCSF
between 1993 and 2010. Biospecimens included peripheral blood samples after stem cell
mobilization with granulocyte colony-stimulating factor (GCSF) in preparation for
autologous bone marrow transplantation and banked in liquid nitrogen or peripheral
blood that was collected as part of the research study from patients during a follow-up
appointment. Initially, cases of all races with complete or nearly complete clinical
records were selected for this study, but Asian cases were later excluded as the number in
this group was particularly small. Controls from population-based and clinic-based case-
control studies of pancreas cancer conducted at UCSF from Jan 1995 through Dec 1999
(population-based study) [8], and from Jan 2006 to Aug 2011 (clinic-based study) [9],
were matched to cases on sex and race/ethnicity. DNA was extracted from blood
samples collected at the time of the study. Samples from 300 cases and 300 controls
were sent to Expression Analysis (San Francisco, CA) for genotyping.

84
4.4.2 Genotyping and quality control
USC:
Genotyping in cases and controls was completed on the Illumina OmniExpress
BeadChip, which resulted in 733,202 successfully genotyped SNPs. There were 1053
total samples, including 20 replicates, which were removed. The concordance rate for
blinded replicates was >99%, with a mean concordance of 99.94%. Of these 1,033 total
subjects, 75 were removed using the following exclusion criteria: self-reported and
genotyped sex differences (n=4), unintended replicates (n=4), low concordance cross
study replicates (n=1), samples with cryptic relatedness (n=3), MGUS samples (n=7), and
samples with call rates <95% (n=44). Individuals with an African or Asian component >
10% were also removed from the analysis (n=13). 41,534 SNPs were removed using the
following exclusion criteria: SNPs with a call rate <95% (n=41,436) and SNPs with >1
discordant genotype among sample replicates (n=98).
UCSF:
Of the 300 cases and 300 controls genotyped, two intended duplicate cases were
excluded. Six cases and one control were dropped due to high missing call rates (>10%).
Another two cases and four controls were dropped because of cryptic relatedness. Of the
4301332 SNPs on the Omni5 platform, 1038755 were monomorphic in our data and
therefore excluded from any analysis (MAF<0.005). An additional 1301 SNPs were
dropped due to poor call rate (>10%). After quality control measures, this study includes
290 cases and 295 controls from a multiethnic hospital-based case-control study (161 and
255 self-reported white cases and controls).
85
4.4.3 Statistical Analysis
USC:
Principal Component Analysis / STRUCTURE: Population stratification, defined as a
difference in allele frequencies between subpopulations due to systematic ancestry
differences, has the potential to bias genetic studies by causing spurious associations [10].
Using EIGENSTRAT, we performed a principal components analysis (PCA), which
calculates eigenvectors representing genetic variation within our study population [10].
Specifically, 1,073 ancestry informative markers (AIMs), or SNPs that have allele
frequency differences that vary across racial populations (HapMap 3), were used to
calculate these eigenvectors. In our study, population admixture was assessed using both
PCA and STRUCTURE (k=4), which clusters individuals to one or more sub-populations
based on whether their genotypes indicate admixture, and then groups similar populations
[11]. The top 10 principal components will be adjusted for in the analysis, and
eigenvectors 1-5 are plotted in Figure 4.1.
Imputation: SNP concordance was calculated using HapMap samples genotyped on the
OmniExpress and in the 1000 Genomes Project (KGP); those with concordance <95% in
any population were removed (36,724, ~5%). This concordance data was also used to
match alleles to the forward strand. SHAPEIT v1.r532 was used to pre-phase haplotypes
in order to decrease the computational time needed to impute the missing data and
imputation was performed using IMPUTE v2.2.2
(https://mathgen.stats.ox.ac.uk/impute/impute_v2.html) using KGP (March 2012 release) as
the reference set.
86
Q-Q Plot: A Q-Q plot is a graphical method in which two probability distributions can be
compared. When applied to genetic studies, the Q-Q plot estimates the inflation of test
statistics by plotting the ranked observed values for disease association against the
distribution expected under the null hypothesis that no true association exists [12]. The
inflation factor (λ) is calculated by dividing the mean of the test statistics by the mean of
the expected values from a chi-square distribution. Q-Q plots for USC, UCSF, and the
meta-analysis were plotted.
Statistical Analysis: Unconditional logistic regression adjusting for age, sex, and PCs 1-
10 (eigenvectors adjusting for population admixture) was used to assess the association of
each SNP with MM risk assuming a log-additive model. All analyses were conducted
using PLINK software and R version 2.15.1.
UCSF:
PCA: PCA was performed using an internally developed JAVA program which makes
use of JAMA (java linear algebra package, http://math.nist.gov/javanumerics/jama/) and
is available upon request. The program was checked against other PCA implementations
in Stata and R and achieved identical results. Approximately 20,000 of the SNPs which
passed QC were selected from across the genome and entered into the PCA program.
Imputation: SHAPEIT was used to phase haplotypes prior to imputation. Imputation was
done using IMPUTE2 using KGP (March 2012 release) as the reference dataset.
Statistical Analysis: Unconditional logistic regression models were used to test the
association between each SNP and myeloma susceptibility, assuming a log-additive
87
model. Each SNP association was adjusted for PC 1-10 and gender. All association
analyses were conducted using PLINK.

4.5 Results
Overall, the age and sex of all studies included in the USC study are similar,
although the cases and controls in the MEC are older than in other studies (Table 4.1).
The cases in the UCSF study are slightly younger than cases in the USC study (Table
4.2).
SNPs that deviated from Hardy-Weinberg equilibrium (HWE) in controls were
not removed from the analysis; however, no SNP presented in the results was out of
HWE (p<0.01). After performing quality control measures, analyses were conducted in
455 cases and 502 and then meta-analyzed with the UCSF study, consisting of 290 cases
and 295 controls.
From the Q-Q plot for the meta-analysis, the inflation factor λ was 1.024
(MAF>0.05), showing little evidence for inflation (Figure 4.2). Study-specific Q-Q plots
are provided in Figures 4.5a and 4.5b, and also show little evidence for inflation (λ=1.01
and 1.04 for the USC and UCSF Studies, respectively).
Overall, we did not observe any genome-wide significant associations between
SNPs and MM risk (p-value < 5 x 10
-8
; Figure 4.2). From the fixed-effects meta-
analysis, the most significant association was for SNP rs7443528 located ~350 kb
upstream of PIK3R1 on chromosome 5q13, although it did not reach genome-wide
significance (OR=0.64, p-value=5.78 x 10
-7
;

Figure 4.3). The T-allele of rs60248962 was
88
associated with a large increase in MM risk (OR=2.62, p-value=1.37 x 10
-6
). We also
observed an association with rs76580792, a SNP located in the intron region of HLA-
DRB5 on chromosome 6p21.3 (OR=1.55, p-value=2.27x10
-6
;

Figure 4.4). The top 25
significant associations are shown in Table 4.3.
We examined the three significant SNPs that were associated with MM risk in the
previously published GWAS [6]. Although the ORs were similar for all three SNPs, only
rs4487645 was statistically significant (p-value=0.0009, Table 4.4).

4.6 Discussion
We performed a meta-analysis and although we did not reach genome-wide
significance, there was suggestive evidence for novel risk loci in chromosomes 5, 6, and
10. We also replicated the effect of the C-allele of rs4487645, one of the three SNPs
reported in the only MM GWAS published to date. Although not statistically significant,
our ORs are very similar for the other two previously reported SNPs, showing a trend
toward replication (p-value< .22).
We observed an association with the SNP rs7443528, located upstream of
PIK3R1, a gene involved in a Ras dependent PI3K/Akt signaling pathway which directly
interacts with the Ras activated protein kinase (MAPK) pathway. Both pathways have
been associated with the proliferation and MM Cell survival and are currently being
targeted for potential therapeutic targets [13]. The second highest ranking SNP,
rs60248962, is located in the intron region of a long intergenic non-protein coding RNA
(LINC00200) on chromosome 10; however little is known about its function.
89
We also observed a borderline-significant association between a SNP
(rs76580792) located in the intron region of the HLA-DRB5 gene in the Class II human
leukocyte antigen (HLA) region on chromosome 6 and MM risk. The HLA region is one
of the most gene-dense regions of the chromosome and contains genes that encode
proteins involved in immune system functions. Specifically, the HLA class II region is
involved in the presentation of antigens by professional antigen presenting cells, which in
turn stimulate T-helper cells and antibody producing B-cells invoking an immune
response to foreign antigens [14, 15].
There is substantial evidence demonstrating that the HLA region is associated
with autoimmune disease and other B-cell neoplasms, such as nodular sclerosis Hodgkin
lymphoma [16, 17]. MM is a neoplasm of post-germinal center B-cells, and has recently
been associated with autoimmune disease [18]. Our data is the first to connect the HLA
region with increased MM risk, supporting the hypothesis that chronic inflammatory
response caused by an antigen or autoimmune condition could be associated with this
disease. However, there are limited studies on MM risk and the association with a
particular virus or autoimmune condition is unknown.
There are potential limitations to this study. Because MM is a rare disease and
patients are often too ill to participate in studies, it is difficult for studies to accrue
enough patients for adequate statistical power. Each study alone had low statistical
power; however, the ability to combine the data in a meta-analysis greatly enhanced our
power to detect novel loci. The UCSF Study case-control study consists of a multiethnic
population, and although the analysis was adjusted for race, there is still the potential for
90
residual confounding due to population stratification. Differential linkage disequilibrium
patterns are observed across race, which could potentially bias our results if a SNP is
tagging a “causal SNP” in one population but not in another. This could weaken the
observed effect and bias our results towards the null. Cases in the USC study were older
than the controls thus it could be argued that the possibility exists that young controls will
develop MM later in life, but given the rarity of the disease (~6/100,000) this is unlikely
to bias our results. A major strength of this study is the ability to meta-analyze our data
with the UCSF study, which greatly increases the statistical power to detect an effect.
In conclusion, we have replicated the association with one SNP from the only
published MM GWAS and we have identified novel risk loci that warrant further
investigation. Although our findings did not reach genome-wide significance, the loci
identified in this study are biologically relevant to the etiology of MM.

91

Table 4.1. Demographic Characteristics of the USC Multi-center Study

Los Angeles SEER Seattle/Detroit SEER Multiethnic Cohort Univ of Alabama Australia UBC

Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls Cases
Mean Age
*
61 (10) 59 (12) 59 (9) 58 (12) 72 (9) 69 (9) 64 (8) 66 (8) 69 (8) 61 (7) 64 (9)
Male 46 41 63 154 13 17 66 57 35 34 92
Female 26 27 38 101 8 8 42 51 27 28 59
Total 72 68 101 255 21 25 108 108 62 62 151
* mean age (SD)

92

Table 4.2. Demographic Characteristics of the UCSF Study

UCSF Study (ca/co)
Mean Age 55/61
Gender

Male 175/175
Female 115/120
Race

White 161/225
AA 29/31
Hispanic 38/39
Other/Unknown 62/0
Missing -
Total 290/295
93

Table 4.3. Top 25 SNP Associations, by Study and Meta-analyzed
SNP* Chr BP TA* USC-OR P-value TA* UCSF-OR P-value Allele1 Allele2 META-OR P-value HetPVal
rs6693301 1 34708075 T 1.98 5.98E-05 T 1.74 0.01404 T C 1.89 2.82E-06 0.6412
rs186678987 5 4811436 A 2.42 0.00174 A 2.74 0.00060 A G 2.57 3.55E-06 0.7654
rs114171913 5 67133532 A 1.55 0.00012 A 1.48 0.00722 A G 1.52 2.76E-06 0.8120
rs958499 5 67145399 T 0.66 0.00019 T 0.66 0.00710 T G 0.66 4.24E-06 0.9495
rs78289652 5 67145400 A 0.66 0.00019 A 0.66 0.00710 A G 0.66 4.24E-06 0.9495
rs72647226 5 67158606 A 0.66 0.00020 A 0.66 0.00445 A G 0.66 2.84E-06 0.9973
rs1907211 5 67159171 C 0.66 0.00020 C 0.66 0.00445 C G 0.66 2.84E-06 0.9973
rs1844702 5 67159407 T 0.66 0.00020 T 0.66 0.00445 T C 0.66 2.84E-06 0.9973
rs11956416 5 67160369 A 1.51 0.00020 A 1.51 0.00445 A G 1.51 2.84E-06 0.9973
rs138189656 5 67162098 T 0.66 0.00020 T 0.66 0.00445 T G 0.66 2.82E-06 0.9967
rs7443528 5 67162999 T 0.67 0.00036 T 0.59 0.00037 T G 0.64 5.78E-07 0.5058
rs138082873 5 67163326 T 0.66 0.00020 T 0.66 0.00445 T C 0.66 2.82E-06 0.9967
rs181922386 5 67163789 T 0.66 0.00018 T 0.65 0.00340 T C 0.66 1.95E-06 0.9620
rs146576146 5 67164024 A 1.51 0.00019 A 1.51 0.00445 A G 1.51 2.82E-06 0.9967
rs62368055 5 67165550 A 0.66 0.00020 A 0.66 0.00445 A G 0.66 2.81E-06 0.9960
rs12654037 5 67167721 A 0.66 0.00020 A 0.66 0.00445 A G 0.66 2.82E-06 0.9967
rs114220057 6 32455168 T 0.65 0.00026 T 0.64 0.00297 T C 0.65 2.48E-06 0.9654
rs76580792 6 32491459 A 1.43 0.00331 A 1.74 0.00012 A G 1.55 2.27E-06 0.3013
rs74662189 8 4017965 T 2.88 0.00033 T 4.22 0.00240 T C 3.20 3.26E-06 0.4916
rs58129199 10 1202129 T 2.91 0.00029 T 2.26 0.00241 T C 2.53 2.77E-06 0.5227
rs60157223 10 1203378 C 0.34 0.00029 C 0.44 0.00246 C G 0.40 2.86E-06 0.5228
rs72760976 10 1206102 A 0.34 0.00029 A 0.43 0.00210 A C 0.39 2.41E-06 0.5517
rs60248962 10 1206592 T 2.93 0.00027 T 2.38 0.00131 T C 2.62 1.37E-06 0.6095
rs4907654 13 1.13E+08 A 1.40 0.00179 A 1.60 0.00029 A G 1.48 2.42E-06 0.4261
rs4889931 17 77889963 A 1.81 0.00134 A 2.09 0.00079 A C 1.92 3.94E-06 0.6165
*These results are ordered by SNP and position, not statistical significance
*TA =Tested Allele
94

Table 4.4. Replication of Published SNPs Associated with MM Risk from a GWAS

Previously Published GWAS USC/UCSF Meta-Analysis
SNP Location RA
a
OR
b
P-value SNP RA OR P-value
rs4487645 7p15.3 C 1.38 3.33x10-15 rs4487645 C 1.32 0.0009
rs1052501 3p22.1 G 1.32 7.47x10-9 rs1052501 C 1.15 0.1521
rs6746082 2p23.3 A 1.29 1.22x10-7 rs6746082 A 1.12 0.2177
a – risk allele
b – Odds ratio, log additive model

95
Figure 4.1. Eigenvector Plots from Principal Components Analysis

Red: African populations
Purple: Asian populations
Green & Yellow: European populations
Pink: MEX population
Blue: Indian population
Black: SPORE Study population

96

Figure 4.2. Q-Q Plot Displaying the Observed versus the Distribution of the –
Log(P-Values) in the Meta-analysis (MAF >0.05).

97

Figure 4.3. LocusZoom Plot for Chromosome 5 - rs7443528

98

Figure 4.4. LocusZoom Plot for Chromosome 6 – rs76580792

99

Figure 4.5a. Q-Q plot for SPORE Study

Figure 4.5b. Q-Q plot for UCSF Study

100
Chapter 4 References
1. Broderick, P., et al., Common variation at 3p22.1 and 7p15.3 influences multiple
myeloma risk. Nat Genet, 2011. 44(1): p. 58-61.
2. Gebregziabher, M., et al., Risk patterns of multiple myeloma in Los Angeles
County, 1972-1999 (United States). Cancer Causes Control, 2006. 17(7): p. 931-8.
3. Witte, J.S. and T.J. Hoffmann, Polygenic modeling of genome-wide association
studies: an application to prostate and breast cancer. OMICS, 2011. 15(6): p.
393-8.
4. Ghoussaini, M. and P.D. Pharoah, Polygenic susceptibility to breast cancer:
current state-of-the-art. Future Oncol, 2009. 5(5): p. 689-701.
5. Vangsted, A., T.W. Klausen, and U. Vogel, Genetic variations in multiple
myeloma I: effect on risk of multiple myeloma. Eur J Haematol, 2012. 88(1): p. 8-
30.
6. Broderick, P., et al., Common variation at 3p22.1 and 7p15.3 influences multiple
myeloma risk. Nat Genet, 2012. 44(1): p. 58-61.
7. Giles, G.G. and D.R. English, The Melbourne Collaborative Cohort Study. IARC
Sci Publ, 2002. 156: p. 69-70.
8. Holly, E.A., C.A. Eberle, and P.M. Bracci, Prior history of allergies and
pancreatic cancer in the San Francisco Bay area. Am J Epidemiol, 2003. 158(5):
p. 432-41.
9. Duell, E.J., et al., A population-based, case-control study of polymorphisms in
carcinogen-metabolizing genes, smoking, and pancreatic adenocarcinoma risk. J
Natl Cancer Inst, 2002. 94(4): p. 297-306.
10. Price, A.L., et al., Principal components analysis corrects for stratification in
genome-wide association studies. Nat Genet, 2006. 38(8): p. 904-9.
11. Pritchard, J.K., M. Stephens, and P. Donnelly, Inference of population structure
using multilocus genotype data. Genetics, 2000. 155(2): p. 945-59.
12. Clayton, D.G., et al., Population structure, differential bias and genomic control
in a large-scale, case-control association study. Nat Genet, 2005. 37(11): p.
1243-6.
13. Steinbrunn, T., et al., Combined targeting of MEK/MAPK and PI3K/Akt
signalling in multiple myeloma. Br J Haematol, 2012.
101
14. Complete sequence and gene map of a human major histocompatibility complex.
The MHC sequencing consortium. Nature, 1999. 401(6756): p. 921-3.
15. de Bakker, P.I., et al., A high-resolution HLA and SNP haplotype map for disease
association studies in the extended human MHC. Nat Genet, 2006. 38(10): p.
1166-72.
16. Cozen, W., et al., A genome-wide meta-analysis of nodular sclerosing Hodgkin
lymphoma identifies risk loci at 6p21.32. Blood, 2012. 119(2): p. 469-75.
17. Huang, X., et al., Multiple HLA class I and II associations in classical Hodgkin
lymphoma and EBV status defined subgroups. Blood, 2011. 118(19): p. 5211-7.
18. Anderson, L.A., et al., Population-based study of autoimmune conditions and the
risk of specific lymphoid malignancies. Int J Cancer, 2009. 125(2): p. 398-405.

102
Chapter 5: Summary and Future Directions

This dissertation has evaluated genetic risk factors and MM risk utilizing both
candidate-pathway and agnostic approaches, with the objective of providing insight into
the etiology of MM. The 8q24 region has been associated with many different cancers,
including prostate cancer, which exhibits a risk profile similar to MM. The region is also
~200 kb centromeric to MYC, an oncogene that when translocated, is associated with
pathogenesis and progression of MM. In light of these findings, we hypothesized that
genetic variation within the 8q24 region could be associated with MM risk. The first
analysis of this dissertation examined genetic variation in 16 SNPs located within the
8q24 region and MM risk in a pooled analysis. We observed modest associations between
two SNPs and MM risk among whites only (rs10086908: Odds Ratio (OR)=0.78, 95%
Confidence Interval (CI)=0.61-1.00; rs6983267: OR=0.81, 95%CI=0.67-0.99). These
associations were not observed among African Americans; however, rs116041037 was
associated with a 2-fold increased risk of MM (OR=2.02, 95%CI=1.02-3.97). Our
findings are suggestive of an association between genetic variation in the 8q24 region and
MM risk in both whites and African Americans; however, replication in larger sample
sizes is critical to validate our findings and explain the discrepancies across
race/ethnicity.
The second analysis examined functional polymorphisms within DNA repair
pathways and MM risk in a multiethnic case-control pilot study, with significant findings
then analyzed in three independent study sets and combined in a meta-analysis. We
103
hypothesized that genetic variation within these pathways could be associated with MM
risk, as DNA repair mechanisms play an integral role in the formation of B-cells. There
were no statistically significant genetic associations with MM risk observed; however, we
did not comprehensively evaluate genetic variation within these pathways and a more
thorough investigation of these pathways is warranted.
The final analysis in this dissertation was a GWAS, with the primary aim to
identify novel common variants associated with MM risk. We combined our study of
455 cases and 502 controls with a UCSF study of 290 cases and 295 controls in a meta-
analysis. Although associations did not reach genome-wide significance, the results were
promising. We observed an association with two SNPs in the HLA region and MM risk,
with one SNP located in the intron region of the HLA-DRB5 gene and the other located
downstream of the gene. This is the first study to identify an association between the
HLA region, recently linked to other B-cell cancers, and MM risk. This unexpected but
biologically plausible finding is indicative of how the field of MM research can benefit
from the agnostic evaluation provided by a GWAS despite the rare frequency of the
disease. Overall, this research has contributed to the field of genetic epidemiology of
MM. The analysis revealed that five functional SNPs in ATM, XRCC1, XPC, CSB, and
ICAM5 (DNA repair genes) were not associated with MM. Additionally, we’ve shown
that a SNP previously associated with increased risk of prostate cancer may also be
linked to increased risk of MM in African Americans. Most importantly, our GWAS
study is the first to implicate the HLA region with MM risk. If this finding is replicated,
we will be able to move forward with hypotheses consisting of antigens and viruses
104
associated with chronic immune system stimulation and therefore increased risk of MM,
similar to other B-cell cancers.
As has been described in various chapters of this dissertation, MM is a
particularly challenging disease to study. It is a rare and lethal cancer with a relatively
short survival period, and there are challenges in accruing a large sample size. Due to the
many challenges of studying a rare cancer, relevant findings should not be based on p-
values alone, but on emerging patterns in the literature where true associations take
longer to become apparent. In order to better study this cancer, collaboration with
individual studies is necessary to advance this field as adequate statistical power to detect
and replicate genetic associations will only be possible when there is sufficient sample
size. Despite these challenges, the study of genetic epidemiology of MM should be
pursued, as there is still so much to be discovered regarding the etiology and
pathogenesis of MM.
USC is leading the formation of the North American MM Consortium, which will
consist of at least four institutions with GWAS data willing to collaborate and share
biologic specimens and data. With this resource, the field of MM research will be able to
move forward in discovering biological mechanisms involved in the etiology of this
disease. The GWAS data from the North American MM Consortium will also lead to a
more educated investigation of candidate pathways, where one can study the cumulative
variation of polymorphisms within genes in a given biological pathway. We will also be
able to clarify the previous associations of various SNPs and MM risk.
105
In the field of MM research, progress has been developing drugs with real impact
on survival based on tumor genomics; however, little progress has been made in
understanding the cause of MM. Functional epidemiology, or the integration of the
biological function with results from genetic epidemiology, should be utilized to follow
up associations with common low penetrance susceptibility loci identified in large
GWAS meta-analyses to facilitate early intervention or even prevention. However,
future directions of this functional work will rely on developing methods to demonstrate
the association of SNPs with regulatory elements, both in vitro and in vivo.
Ultimately, our goal is to improve the understanding of the mechanisms
underlying the disease process, which can lead to the identification of targeted areas for
risk stratification and personalized medicine, and eventually result in a decreased disease
burden in the population.
106
Comprehensive References

(1999). Complete sequence and gene map of a human major histocompatibility complex.
The MHC sequencing consortium. Nature 401(6756): 921-923.
Ahmadiyeh, N., Pomerantz, M. M., et al. 8q24 prostate, breast, and colon cancer risk loci
show tissue-specific long-range interaction with MYC. Proc Natl Acad Sci U S A
107(21): 9742-9746.
Al Olama, A. A., Kote-Jarai, Z., et al. (2009). Multiple loci on 8q24 associated with
prostate cancer susceptibility. Nat Genet 41(10): 1058-1060.
Alexander, D. D., Mink, P. J., et al. (2007). Multiple myeloma: a review of the
epidemiologic literature. Int J Cancer 120 Suppl 12: 40-61.
Anderson, K. C. and Carrasco, R. D. (2011). Pathogenesis of myeloma. Annu Rev Pathol
6: 249-274.
Anderson, L. A., Gadalla, S., et al. (2009). Population-based study of autoimmune
conditions and the risk of specific lymphoid malignancies. Int J Cancer 125(2):
398-405.
Baris, D., Brown, L. M., et al. (2000). Socioeconomic status and multiple myeloma
among US blacks and whites. Am J Public Health 90(8): 1277-1281.
Birmann, B. M., Tamimi, R. M., et al. (2009). Insulin-like growth factor-1- and
interleukin-6-related gene variation and risk of multiple myeloma. Cancer
Epidemiol Biomarkers Prev 18(1): 282-288.
Birmann BM, C. B., Muench K, Suppan CA, Cozen W. (2011). Epidemiology and
etiology of multiple myeloma. Multiple Myeloma - A New Era of Treatment
Strategies. P. K. a. A. KC. Bentham Books, Bentham Science Publishers.
Blattner, W. A., Blair, A., et al. (1981). Multiple myeloma in the United States,
1950--1975. Cancer 48(11): 2547-2554.
Bommert, K., Bargou, R. C., et al. (2006). Signalling and survival pathways in multiple
myeloma. Eur J Cancer 42(11): 1574-1580.
Broderick, P., Chubb, D., et al. (2011). Common variation at 3p22.1 and 7p15.3
influences multiple myeloma risk. Nat Genet 44(1): 58-61.

107
Broderick, P., Chubb, D., et al. (2012). Common variation at 3p22.1 and 7p15.3
influences multiple myeloma risk. Nat Genet 44(1): 58-61.
Brown, E. E., Lan, Q., et al. (2007). Common variants in genes that mediate immunity
and risk of multiple myeloma. Int J Cancer 120(12): 2715-2722.
Burma, S., Chen, B. P., et al. (2006). Role of non-homologous end joining (NHEJ) in
maintaining genomic integrity. DNA Repair (Amst) 5(9-10): 1042-1048.
Chahwan, R., Edelmann, W., et al. (2011). Mismatch-mediated error prone repair at the
immunoglobulin genes. Biomed Pharmacother 65(8): 529-536.
Chen, F., Chen, G. K., et al. (2011). Fine-mapping of breast cancer susceptibility loci
characterizes genetic risk in African Americans. Hum Mol Genet 20(22):
4491-4503.
Chng, W. J., Glebov, O., et al. (2007). Genetic events in the pathogenesis of multiple
myeloma. Best Pract Res Clin Haematol 20(4): 571-596.
Clayton, D. G., Walker, N. M., et al. (2005). Population structure, differential bias and
genomic control in a large-scale, case-control association study. Nat Genet
37(11): 1243-1246.
Colditz, G. A. and Hankinson, S. E. (2005). The Nurses' Health Study: lifestyle and
health among women. Nat Rev Cancer 5(5): 388-396.
Cortessis, V. K., Yuan, J. M., et al. Risk of urinary bladder cancer is associated with 8q24
variant rs9642880[T] in multiple racial/ethnic groups: results from the Los
Angeles-Shanghai case-control study. Cancer Epidemiol Biomarkers Prev 19(12):
3150-3156.
Cozen, W., Gebregziabher, M., et al. (2006). Interleukin-6-related genotypes, body mass
index, and risk of multiple myeloma and plasmacytoma. Cancer Epidemiol
Biomarkers Prev 15(11): 2285-2291.
Cozen, W., Li, D., et al. (2012). A genome-wide meta-analysis of nodular sclerosing
Hodgkin lymphoma identifies risk loci at 6p21.32. Blood 119(2): 469-475.
Crowther-Swanepoel, D., Broderick, P., et al. Common variants at 2q37.3, 8q24.21,
15q21.3 and 16q24.1 influence chronic lymphocytic leukemia risk. Nat Genet
42(2): 132-136.

108
Davies, F. E., Rollinson, S. J., et al. (2000). High-producer haplotypes of tumor necrosis
factor alpha and lymphotoxin alpha are associated with an increased risk of
myeloma and have an improved progression-free survival after treatment. J Clin
Oncol 18(15): 2843-2851.
de Bakker, P. I., McVean, G., et al. (2006). A high-resolution HLA and SNP haplotype
map for disease association studies in the extended human MHC. Nat Genet
38(10): 1166-1172.
de Bakker, P. I., Yelensky, R., et al. (2005). Efficiency and power in genetic association
studies. Nat Genet 37(11): 1217-1223.
De Roos, A. J., Gold, L. S., et al. (2006). Metabolic gene variants and risk of
non-Hodgkin's lymphoma. Cancer Epidemiol Biomarkers Prev 15(9): 1647-1653.
Dib, A., Gabrea, A., et al. (2008). Characterization of MYC translocations in multiple
myeloma cell lines. J Natl Cancer Inst Monogr(39): 25-31.
Duell, E. J., Holly, E. A., et al. (2002). A population-based, case-control study of
polymorphisms in carcinogen-metabolizing genes, smoking, and pancreatic
adenocarcinoma risk. J Natl Cancer Inst 94(4): 297-306.
Easton, D. F., Pooley, K. A., et al. (2007). Genome-wide association study identifies
novel breast cancer susceptibility loci. Nature 447(7148): 1087-1093.
Edlund, C. K., Lee, W. H., et al. (2008). Snagger: a user-friendly program for
incorporating additional information for tagSNP selection. BMC Bioinformatics 9:
174.
Enciso-Mora, V., Broderick, P., et al. A genome-wide association study of Hodgkin's
lymphoma identifies new susceptibility loci at 2p16.1 (REL), 8q24.21 and 10p14
(GATA3). Nat Genet 42(12): 1126-1130.
Fendly, B. M., Winget, M., et al. (1990). Characterization of murine monoclonal
antibodies reactive to either the human epidermal growth factor receptor or
HER2/neu gene product. Cancer Res 50(5): 1550-1558.
Ferlay, J., International Agency for Research on Cancer., et al. (1997). CI5VII electronic
database of Cancer incidence in five continents, vol. VII. IARC cancerBase no 2.
Lyon, France, International Agency for Research on Cancer,: 2 computer disks.
Fonseca, R., Bergsagel, P. L., et al. (2009). International Myeloma Working Group
molecular classification of multiple myeloma: spotlight review. Leukemia 23(12):
2210-2221.
109
Freedman, M. L., Haiman, C. A., et al. (2006). Admixture mapping identifies 8q24 as a
prostate cancer risk locus in African-American men. Proc Natl Acad Sci U S A
103(38): 14068-14073.
Gabrea, A., Leif Bergsagel, P., et al. (2006). Distinguishing primary and secondary
translocations in multiple myeloma. DNA Repair (Amst) 5(9-10): 1225-1233.
Gebregziabher, M., Bernstein, L., et al. (2006). Risk patterns of multiple myeloma in Los
Angeles County, 1972-1999 (United States). Cancer Causes Control 17(7):
931-938.
Gebregziabher, M., Guimaraes, P., et al. (2010). A polytomous conditional likelihood
approach for combining matched and unmatched case-control studies. Stat Med
29(9): 1004-1013.
Gellert, M. (2002). V(D)J recombination: RAG proteins, repair factors, and regulation.
Annu Rev Biochem 71: 101-132.
Ghoussaini, M. and Pharoah, P. D. (2009). Polygenic susceptibility to breast cancer:
current state-of-the-art. Future Oncol 5(5): 689-701.
Ghoussaini, M., Song, H., et al. (2008). Multiple loci with different cancer specificities
within the 8q24 gene desert. J Natl Cancer Inst 100(13): 962-966.
Giles, G. G. and English, D. R. (2002). The Melbourne Collaborative Cohort Study.
IARC Sci Publ 156: 69-70.
Giovannucci, E., Liu, Y., et al. (2007). Risk factors for prostate cancer incidence and
progression in the health professionals follow-up study. Int J Cancer 121(7):
1571-1578.
Gold, L. S., De Roos, A. J., et al. (2009). Associations of common variants in genes
involved in metabolism and response to exogenous chemicals with risk of
multiple myeloma. Cancer Epidemiol 33(3-4): 276-280.
Gonzalez-Fraile, M. I., Garcia-Sanz, R., et al. (2002). Methylenetetrahydrofolate
reductase genotype does not play a role in multiple myeloma pathogenesis. Br J
Haematol 117(4): 890-892.
Goode, E. L., Chenevix-Trench, G., et al. A genome-wide association study identifies
susceptibility loci for ovarian cancer at 2q31 and 8q24. Nat Genet 42(10):
874-879.

110
Goode, E. L., Ulrich, C. M., et al. (2002). Polymorphisms in DNA repair genes and
associations with cancer risk. Cancer Epidemiol Biomarkers Prev 11(12):
1513-1530.
Greenberg, A. J., Lee, A. M., et al. (2012). Single-nucleotide polymorphism rs1052501
associated with monoclonal gammopathy of undetermined significance and
multiple myeloma. Leukemia.
Gudmundsson, J., Sulem, P., et al. (2007). Genome-wide association study identifies a
second prostate cancer susceptibility variant at 8q24. Nat Genet 39(5): 631-637.
Haiman, C. A., Chen, G. K., et al. (2011). Characterizing genetic risk at known prostate
cancer susceptibility loci in African Americans. PLoS Genet 7(5): e1001387.
Haiman, C. A., Hsu, C., et al. (2008). Comprehensive association testing of common
genetic variation in DNA repair pathway genes in relationship with breast cancer
risk in multiple populations. Hum Mol Genet 17(6): 825-834.
Haiman, C. A., Le Marchand, L., et al. (2007). A common genetic risk factor for
colorectal and prostate cancer. Nat Genet 39(8): 954-956.
Haiman, C. A., Patterson, N., et al. (2007). Multiple regions within 8q24 independently
affect risk for prostate cancer. Nat Genet 39(5): 638-644.
Han, J., Haiman, C., et al. (2009). Genetic variation in DNA repair pathway genes and
premenopausal breast cancer risk. Breast Cancer Res Treat 115(3): 613-622.
Hayden, P. J., Tewari, P., et al. (2007). Variation in DNA repair genes XRCC3, XRCC4,
XRCC5 and susceptibility to myeloma. Hum Mol Genet 16(24): 3117-3127.
Hideshima, T., Mitsiades, C., et al. (2007). Understanding multiple myeloma
pathogenesis in the bone marrow to identify new therapeutic targets. Nat Rev
Cancer 7(8): 585-598.
Holly, E. A., Eberle, C. A., et al. (2003). Prior history of allergies and pancreatic cancer
in the San Francisco Bay area. Am J Epidemiol 158(5): 432-441.
Hooker, S., Hernandez, W., et al. Replication of prostate cancer risk loci on 8q24, 11q13,
17q12, 19q33, and Xp11 in African Americans. Prostate 70(3): 270-275.
Hosgood, H. D., 3rd, Baris, D., et al. (2009). Genetic variation in cell cycle and apoptosis
related genes and multiple myeloma risk. Leuk Res 33(12): 1609-1614.
Hosgood, H. D., 3rd, Baris, D., et al. (2008). Caspase polymorphisms and genetic
susceptibility to multiple myeloma. Hematol Oncol 26(3): 148-151.
111
Howie, B. N., Donnelly, P., et al. (2009). A flexible and accurate genotype imputation
method for the next generation of genome-wide association studies. PLoS Genet
5(6): e1000529.
Huang, X., Kushekhar, K., et al. (2011). Multiple HLA class I and II associations in
classical Hodgkin lymphoma and EBV status defined subgroups. Blood 118(19):
5211-5217.
Jia, L., Landan, G., et al. (2009). Functional enhancers at the gene-poor 8q24
cancer-linked locus. PLoS Genet 5(8): e1000597.
Kang, S. H., Kim, T. Y., et al. (2008). Protective role of CYP1A1*2A in the development
of multiple myeloma. Acta Haematol 119(1): 60-64.
Kerber, R. A. and O'Brien, E. (2005). A cohort study of cancer risk in relation to family
histories of cancer in the Utah population database. Cancer 103(9): 1906-1915.
Kiemeney, L. A., Thorlacius, S., et al. (2008). Sequence variant on 8q24 confers
susceptibility to urinary bladder cancer. Nat Genet 40(11): 1307-1312.
Kim, H. N., Kim, Y. K., et al. (2007). Polymorphisms involved in the folate metabolizing
pathway and risk of multiple myeloma. Am J Hematol 82(9): 798-801.
Klein, U. and Dalla-Favera, R. (2008). Germinal centres: role in B-cell physiology and
malignancy. Nat Rev Immunol 8(1): 22-33.
Koessel, S. L., Theis, M. K., et al. (1996). Socioeconomic status and the incidence of
multiple myeloma. Epidemiology 7(1): 4-8.
Kolonel, L. N., Henderson, B. E., et al. (2000). A multiethnic cohort in Hawaii and Los
Angeles: baseline characteristics. Am J Epidemiol 151(4): 346-357.
Kristinsson, S. Y., Bjorkholm, M., et al. (2009). Patterns of hematologic malignancies
and solid tumors among 37,838 first-degree relatives of 13,896 patients with
multiple myeloma in Sweden. Int J Cancer 125(9): 2147-2150.
Kristinsson, S. Y., Derolf, A. R., et al. (2009). Socioeconomic differences in patient
survival are increasing for acute myeloid leukemia and multiple myeloma in
sweden. J Clin Oncol 27(12): 2073-2080.
Kuehl, W. M. and Bergsagel, P. L. (2002). Multiple myeloma: evolving genetic events
and host interactions. Nat Rev Cancer 2(3): 175-187.
Kunkel, T. A. and Erie, D. A. (2005). DNA mismatch repair. Annu Rev Biochem 74:
681-710.
112
Kupfer, S. S., Torres, J. B., et al. (2009). Novel single nucleotide polymorphism
associations with colorectal cancer on chromosome 8q24 in African and European
Americans. Carcinogenesis 30(8): 1353-1357.
Kyle, R. A. and Rajkumar, S. V. (2007). Monoclonal gammopathy of undetermined
significance and smouldering multiple myeloma: emphasis on risk factors for
progression. Br J Haematol 139(5): 730-743.
Kyle, R. A. and Rajkumar, S. V. (2008). Multiple myeloma. Blood 111(6): 2962-2972.
Landgren, O., Kristinsson, S. Y., et al. (2009). Risk of plasma cell and
lymphoproliferative disorders among 14621 first-degree relatives of 4458 patients
with monoclonal gammopathy of undetermined significance in Sweden. Blood
114(4): 791-795.
Landgren, O., Kyle, R. A., et al. (2009). Monoclonal gammopathy of undetermined
significance (MGUS) consistently precedes multiple myeloma: a prospective
study. Blood 113(22): 5412-5417.
Landgren, O., Linet, M. S., et al. (2006). Familial characteristics of autoimmune and
hematologic disorders in 8,406 multiple myeloma patients: a population-based
case-control study. Int J Cancer 118(12): 3095-3098.
Landgren, O., Rajkumar, S. V., et al. (2010). Obesity is associated with an increased risk
of monoclonal gammopathy of undetermined significance among black and white
women. Blood 116(7): 1056-1059.
Landgren, O. and Weiss, B. M. (2009). Patterns of monoclonal gammopathy of
undetermined significance and multiple myeloma in various ethnic/racial groups:
support for genetic factors in pathogenesis. Leukemia 23(10): 1691-1697.
Larsson, S. C. and Wolk, A. (2007). Body mass index and risk of multiple myeloma: a
meta-analysis. Int J Cancer 121(11): 2512-2516.
Lichtman, M. A. (2010). Obesity and the risk for a hematological malignancy: leukemia,
lymphoma, or myeloma. Oncologist 15(10): 1083-1101.
Lima, C. S., Ortega, M. M., et al. (2008). Polymorphisms of methylenetetrahydrofolate
reductase (MTHFR), methionine synthase (MTR), methionine synthase reductase
(MTRR), and thymidylate synthase (TYMS) in multiple myeloma risk. Leuk Res
32(3): 401-405.
Lincz, L. F., Scorgie, F. E., et al. (2007). Genetic variations in benzene metabolism and
susceptibility to multiple myeloma. Leuk Res 31(6): 759-763.
113
Maggini, V., Buda, G., et al. (2008). Lack of association of NQO1 and GSTP1
polymorphisms with multiple myeloma risk. Leuk Res 32(6): 988-990.
Mailman, M. D., Feolo, M., et al. (2007). The NCBI dbGaP database of genotypes and
phenotypes. Nat Genet 39(10): 1181-1186.
McDuffie, H. H., Pahwa, P., et al. (2009). Clustering of cancer among families of cases
with Hodgkin Lymphoma (HL), Multiple Myeloma (MM), Non-Hodgkin's
Lymphoma (NHL), Soft Tissue Sarcoma (STS) and control subjects. BMC
Cancer 9: 70.
Morgan, G. J., Adamson, P. J., et al. (2005). Haplotypes in the tumour necrosis factor
region and myeloma. Br J Haematol 129(3): 358-365.
Ogmundsdottir, H. M., Haraldsdottirm, V., et al. (2005). Familiality of benign and
malignant paraproteinemias. A population-based cancer-registry study of multiple
myeloma families. Haematologica 90(1): 66-71.
Ortega, M., et al., (2007). GSTM1 and codon 72 P53 polymorphism in multiple
myeloma. Ann Hematol(86): 815-819.
Palumbo, A. and Anderson, K. (2011). Multiple myeloma. N Engl J Med 364(11):
1046-1060.
Peltomaki, P. (2001). Deficient DNA mismatch repair: a common etiologic factor for
colon cancer. Hum Mol Genet 10(7): 735-740.
Pomerantz, M. M., Ahmadiyeh, N., et al. (2009). The 8q24 cancer risk variant rs6983267
shows long-range interaction with MYC in colorectal cancer. Nat Genet 41(8):
882-884.
Pomerantz, M. M., Beckwith, C. A., et al. (2009). Evaluation of the 8q24 prostate cancer
risk locus and MYC expression. Cancer Res 69(13): 5568-5574.
Price, A. L., Patterson, N. J., et al. (2006). Principal components analysis corrects for
stratification in genome-wide association studies. Nat Genet 38(8): 904-909.
Pritchard, J. K., Stephens, M., et al. (2000). Inference of population structure using
multilocus genotype data. Genetics 155(2): 945-959.
Renshaw, C., Ketley, N., et al. Trends in the incidence and survival of multiple myeloma
in South East England 1985-2004. BMC Cancer 10: 74.
Roddam, P. L., Rollinson, S., et al. (2002). Genetic variants of NHEJ DNA ligase IV can
affect the risk of developing multiple myeloma, a tumour characterised by
aberrant class switch recombination. J Med Genet 39(12): 900-905.
114
Rudd, M. F., Sellick, G. S., et al. (2006). Variants in the ATM-BRCA2-CHEK2 axis
predispose to chronic lymphocytic leukemia. Blood 108(2): 638-644.
Schumacher, F. R., Feigelson, H. S., et al. (2007). A common 8q24 variant in prostate
and breast cancer from a large nested case-control study. Cancer Res 67(7):
2951-2956.
Shapiro-Shelef, M. and Calame, K. (2005). Regulation of plasma-cell development. Nat
Rev Immunol 5(3): 230-242.
Shen, M., Menashe, I., et al. (2010). Polymorphisms in DNA repair genes and risk of
non-Hodgkin lymphoma in a pooled analysis of three studies. Br J Haematol
151(3): 239-244.
Shou, Y., Martelli, M. L., et al. (2000). Diverse karyotypic abnormalities of the c-myc
locus associated with c-myc dysregulation and tumor progression in multiple
myeloma. Proc Natl Acad Sci U S A 97(1): 228-233.
Spink, C. F., Gray, L. C., et al. (2007). Haplotypic structure across the I kappa B alpha
gene (NFKBIA) and association with multiple myeloma. Cancer Lett 246(1-2):
92-99.
Steinbrunn, T., Stuhmer, T., et al. (2012). Combined targeting of MEK/MAPK and
PI3K/Akt signalling in multiple myeloma. Br J Haematol.
Tambini, C. E., Spink, K. G., et al. (2010). The importance of XRCC2 in RAD51-related
DNA damage repair. DNA Repair (Amst) 9(5): 517-525.
Tewari, P., Ryan, A. W., et al. Genetic variation at the 8q24 locus confers risk to multiple
myeloma. Br J Haematol.
Tibaldi, J. M., Lorber, D., et al. (1992). Postprandial hypoglycemia in islet beta cell
hyperplasia with adenomatosis of the pancreas. J Surg Oncol 50(1): 53-57.
Tomlinson, I., Webb, E., et al. (2007). A genome-wide association scan of tag SNPs
identifies a susceptibility variant for colorectal cancer at 8q24.21. Nat Genet
39(8): 984-988.
Tonon, G. (2007). Molecular pathogenesis of multiple myeloma. Hematol Oncol Clin
North Am 21(6): 985-1006, vii.
Turesson, I., Velez, R., et al. (2010). Patterns of multiple myeloma during the past 5
decades: stable incidence rates for all age groups in the population but rapidly
changing age distribution in the clinic. Mayo Clin Proc 85(3): 225-230.

115
Vachon, C. M., Kyle, R. A., et al. (2009). Increased risk of monoclonal gammopathy in
first-degree relatives of patients with multiple myeloma or monoclonal
gammopathy of undetermined significance. Blood 114(4): 785-790.
Vangsted, A., Klausen, T. W., et al. (2012). Genetic variations in multiple myeloma I:
effect on risk of multiple myeloma. Eur J Haematol 88(1): 8-30.
Velez, R., Beral, V., et al. (1982). Increasing trends of multiple myeloma mortality in
England and Wales; 1950-79: are the changes real? J Natl Cancer Inst 69(2):
387-392.
Wallin, A. and Larsson, S. C. (2011). Body mass index and risk of multiple myeloma: a
meta-analysis of prospective studies. Eur J Cancer 47(11): 1606-1615.
White, K. L., Sellers, T. A., et al. Variation at 8q24 and 9p24 and risk of epithelial
ovarian cancer. Twin Res Hum Genet 13(1): 43-56.
Witte, J. S. and Hoffmann, T. J. (2011). Polygenic modeling of genome-wide association
studies: an application to prostate and breast cancer. OMICS 15(6): 393-398.
Wright, J. B., Brown, S. J., et al. (2010). Upregulation of c-MYC in cis through a large
chromatin loop linked to a cancer risk-associated single-nucleotide polymorphism
in colorectal cancer cells. Mol Cell Biol 30(6): 1411-1420.
Yeager, M., Chatterjee, N., et al. (2009). Identification of a new prostate cancer
susceptibility locus on chromosome 8q24. Nat Genet 41(10): 1055-1057.
Zanke, B. W., Greenwood, C. M., et al. (2007). Genome-wide association scan identifies
a colorectal cancer susceptibility locus on chromosome 8q24. Nat Genet 39(8):
989-994.
Zheng, C., Huang, D., et al. (2001). Cytotoxic T-lymphocyte antigen-4 microsatellite
polymorphism is associated with multiple myeloma. Br J Haematol 112(1):
216-218.
Zheng, C., Huang, D., et al. (2001). Interleukin-10 gene promoter polymorphisms in
multiple myeloma. Int J Cancer 95(3): 184-188.
Zheng, S. L., Sun, J., et al. (2007). Association between two unlinked loci at 8q24 and
prostate cancer risk among European Americans. J Natl Cancer Inst 99(20):
1525-1533.
116
Appendices

Appendix A: Summary of Tables 1a-2b.

Results from a literature search conducted for MM and MGUS familial risk studies. The
key words for the literature search included familial risk, hereditary risk, family cases,
predisposition, MGUS, multiple myeloma, and lymphoma. Tables 1a and 1b. report the
studies that have been conducted for MM risk of first degree relative relatives of MM and
MGUS cases, including measures of effect, effect estimates, and a brief summary of the
conclusions. Tables 2a and 2b. report the studies that have been conducted for MGUS
risk of first degree relative relatives of MM and MGUS cases, including measures of
effect, effect estimates, and a brief summary of the conclusions.

Appendix B: Tables 1a, 1b, 2a, and 2b.

Table 1a. MM Risk of Relatives of MM Cases
MM Risk for First Degree Relatives of MM Cases
Author
Measure
of Effect
Effect
Estimates
Summary
Kristinsson et al.
RR
(95%CI)
2.1 (1.6-2.9)
First-degree relatives of MM cases had a 2-fold higher risk in
multiple myeloma
McDuffie et al.
OR
(95%CI)
1.38 (1.07-
1.78)
Overall, there was an increased risk of approximately 40% in
developing MM in any first-degree relative
Ogmundsdottir et
al.
RR
(95%CI)
Male - 1.64
(0.44-4.17)

Female -
3.23 (1.17-
7.01)
When MM risk was examined in males and females separately,
females seemed to be at a higher risk

Both - 2.33
(1.12-4.26)

Table 1b. MM Risk of Relatives of MGUS Cases
MM Risk for First Degree Relatives of MGUS Cases
Author
Measure of
Effect
Effect
Estimates
Summary
Landgren et al RR (95%CI) 2.9 (1.9-4.3)
First-degree relatives of MGUS cases had a 3-fold higher risk
in multiple myeloma

IgG/IgA
isotype
2.9 (1.7-4.9)
Suggestive higher risk for first-degree relatives when MGUS
isotype is IgG/IgA

IgM 1.9 (0.3-10)

117

Appendix B: Continued

Table 2a. MGUS Risk of Relatives of MM Cases
MGUS Risk for First Degree Relatives of MM Cases
Author
Measure of
Effect
Effect
Estimates
Summary
Kristinsson et al.
RR
(95%CI)
2.1 (1.5-3.1)
First-degree relatives of MM cases had an approximate 2-fold
higher risk in MGUS
Vachon et al.
RR
(95%CI)
2.0 (1.4-2.8)
First-degree relatives of MM cases had a 2-fold increased risk in
MGUS
Ogmundsdottir et
al.
RR
(95%CI)
null
No increase in risk of developing MGUS was noted in first-
degree relatives of MM cases in this study

Table 2b. MGUS Risk of Relatives of MGUS Cases
MGUS Risk for First Degree Relatives of MGUS Cases
Author
Measure of
Effect
Effect
Estimates
Summary
Vachon et al.
RR
(95%CI)
3.3 (2.1-4.8)
First-degree relatives of MGUS cases were three times as likely
to develop MGUS
Landgren et al.
RR
(95%CI)
2.8 (1.4-5.6)

IgG/IgA
isotype
4.0 (1.7-9.2)
Suggestive higher risk for first-degree relatives when MGUS
isotype is IgG/IgA

IgM 0 There were no first-degree relatives of IgM MGUS patients
118
Appendix C: Linkage Disequilibrium patterns in Whites and Africans in the 8q24 Region

Figure 1. Haploview LD plot showing linkage disequilibrium patterns in whites in the
8q24 region (CEU, 127.9-128.8Mb).

Figure 2. Haploview LD plot showing linkage disequilibrium in Africans in the 8q24
region (YRI, 127.9-128. 8Mb).

119
Appendix D: Graphical Depiction of Projects in this Dissertation

120

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Genetic studies of cancer in populations of African ancestry and Latinos

PDF

Genomic risk factors associated with Ewing Sarcoma susceptibility

PDF

The role of heritability and genetic variation in cancer and cancer survival

PDF

Environmental risk factors of Multiple Sclerosis: a twin study

PDF

Hormonal and genetic risk factors of endometrial cancer and trends in incidence and survival of adult acute lymphoblastic leukemia

PDF

Genetic and environmental risk factors for childhood cancer

PDF

Prostate cancer: genetic susceptibility and lifestyle risk factors

PDF

The role of inflammation in non-Hodgkin lymphoma etiology

PDF

Age related macular degeneration in Latinos: risk factors and impact on quality of life

PDF

Identifying genetic, environmental, and lifestyle determinants of ethnic variation in risk of pancreatic cancer

PDF

Meat intake, polymorphisms in the NER and MMR pathways and colorectal cancer risk

PDF

Genes and hormonal factors involved in the development or recurrence of breast cancer

PDF

Body size and the risk of prostate cancer in the multiethnic cohort

PDF

Utility of polygenic risk score with biomarkers and lifestyle factors in the multiethnic cohort study

PDF

Breast cancer in the multiethnic cohort study: Genetic (prolactin pathway genes) and environmental (hormone therapy) factors

PDF

Arm lymphedema in a multi-ethnic cohort of female breast cancer survivors

PDF

Factors that influence mammographic density: role of estrogen metabolism genes, biomarkers of inflammation, and lifestyle

PDF

The multiethnic nature of chronic disease: studies in the multiethnic cohort

PDF

Hormone therapy timing hypothesis and atherosclerosis

PDF

Common immune-related factors and risk of non-Hodgkin lymphomy

Asset Metadata

Creator Rand, Kristin Alyse (author)

Core Title Genetic risk factors in multiple myeloma

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Publication Date 11/26/2012

Defense Date 10/02/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Epidemiology,genetic risk factors,hematologic malignancy,multiple myeloma,OAI-PMH Harvest,single nucleotide polymorphism,SNP

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Cozen, Wendy (committee chair), Conti, David V. (committee member), Haiman, Christopher A. (committee member), McKean-Cowdin, Roberta (committee member), Taylor, Clive R. (committee member)

Creator Email kristin.rand@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-119332

Unique identifier UC11291014

Identifier usctheses-c3-119332 (legacy record id)

Legacy Identifier etd-RandKristi-1343.pdf

Dmrecord 119332

Document Type Dissertation

Rights Rand, Kristin Alyse

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA